Iowa State University Capstones, Theses and Retrospective Theses and Dissertations Dissertations

1999 Substrate and inhibitor recognition by phosphorylase kinase and protein phosphatase-1 Cheryl Jean Suitt aB rtleson Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/rtd Part of the Biochemistry Commons

Recommended Citation Bartleson, Cheryl Jean Suitt, "Substrate and inhibitor recognition by phosphorylase kinase and protein phosphatase-1 " (1999). Retrospective Theses and Dissertations. 12549. https://lib.dr.iastate.edu/rtd/12549

This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. INFORMATION TO USERS

This manuscript has been rqntxiuced fix)m the microfihn master. UMI films the text directly fiom the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter &ce, while others may be from any ^e of computer printer.

The qualify of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper aligmnent can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Infinmation Company 300 North Zeeb Road, Ann Aibor MI 48106-1346 USA 313/761-4700 800/521-0600

Substrate and inhibitor recognition by

phosphorylase kinase and protein phosphatase-1

by

Cheryl Jean Svdtt Bartleson

A dissertation submitted to the graduate faculty in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Major: Biochemistry

Major Professor: Donald J. Graves

Iowa State University

Ames, Iowa

1999 DHX Number: 9924700

UMI Microform 9924700 Copyright 1999, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 u

Graduate College

Iowa State University

This is to certify that the Doctoral Dissertation of

Cheryl Jean Suitt Bartleson has met the dissertation requirements of Iowa State University

Signature was redacted for privacy. Major P^fessor

Signature was redacted for privacy. For the Major Program

Signature was redacted for privacy. For the Graduate College iii

TABLE OF CONTENTS

GENERAL INTRODUCTION 1 Phosphorylase kinase 1 General properties of phosphorylase kinase 1 Roles of subiinits 3 Catalytic properties 7 Substrate specificity 8 Ca2+/calmodnlin-dependent protein kinases/autoinhibition 10 Protein phosphatase-1 13 General properties 13 Roles of subxinits 15 Catalytic properties 16 Substrate specificity 17 Dissertation organization 19

ARGININE TO CITRULLINE REPLACEMENT IN SUBSTRATES OF PHOSPHORYLASE KINASE 21 Abstract 21 Introduction 22 Experimental procedvires 24 Results and discussion 26 Literature cited 30

SPECIFICITY DETERMINANTS OF PHK13 IMPORTANT FOR THE INHIBITION OF PHOSPHORYLASE KINASE 37 Abstract 37 Introduction 38 Experimental procedures 42 Results 45 Discussion 50 References 54 iv

EXPRESSION AND ANALYSIS OF PHK86, THE AUTOINHIBITORY CALMODULIN BINDING C-TERMINUS OF THE CATALYTIC SUB- UNIT OF PHOSPHORYLASE KINASE 75 Abstract 75 Introduction 76 Experimental procedures 78 Resxdts 82 Discussion 85 References 89

MUTANTS OF PHOSPHORYLASE-A ALTERED IN RECOGNITION BY PROTEIN PHOSPHATASE-1 105 Abstract 105 Introduction 106 Experimental procedures 109 Results 112 Discussion 116 References 120

GENERAL CONCLUSIONS 143

REFERENCES CITED 147

ACKNOWLEDGMENTS 155 1

GENERAL INTRODUCTION

Phosphorylase Kinase

General properties of phosphorylase kinase

Phosphorylase kinase plays a key regulatory role in glycogen metabohsm

by catalyzing the phosphorylation of glycogen phosphorylase, which converts

phosphorylase b, inactive, to phosphorylase a, active. Phosphorylase a, in turn,

facilitates the breakdown of glycogen to glucose-l-phosphate. Krebs and Fischer

first isolated phosphorylase kinase firom rabbit skeletal muscle (1). Although

phosphorylase kinase was the first protein kinase known to modulate another

enzyme's activity by phosphorylation (2), there are many unknowns regarding

the structure, function, and regulation of phosphorylase kinase. This incomplete

knowledge can be related to the complexity of phosphorylase kinase, in its

regulation and its structure. This enzyme is found in several tissues, including

smooth muscle, Hver, heart, brain (3-6). The rabbit skeletal muscle form is the

best characterized form.

Two major pathways exist for the regulation of glycogenolysis involving

phosphorylase kinase (7,8). The first pathway is hormonal stimidation in the

presence of epinephrine or glucagon. Rising levels of epinephrine result in

increased levels of cyclic AMP which in tvim initiates the activation of cycUc

AMP-dependent protein kinase (protein kinase A). Protein kinase A 2 phosphoiylates the a and P subunits of phosphorylase kinase yielding activated phosphorylase kinase. The second pathway involves neuronal and a-adrenergic hormonal control of glycogenolysis. These systems are coordinated by calcium, which is believed to serve as a mediator to link glycogenolysis with muscle contraction (9-12). Calcium binds to the 5 sub unit of phosphorylase kinase and stimulates calcium dependent activity. Additionally, exogenous calcium- calmodulin binds to calmodulin binding sites on the a and P subvmits resulting in activation.

Phosphorylase kinase is a very large and complex enzyme. The enzyme

is a hexadecamer with subunit composition of (a,3,7,8)4 and a molecular

weight of 1.3 x 10® kDa (8,13). The a, Mr 138 kDa, and p, Mr 125 kDa are

regulatory subunits that can be phosphorylated and can bind exogenous

calmodulin (14-17). Both events affect phosphorylase kinase activity (8,13).

Two isozymes of phosphorylase kinase have been isolated from white, fast-

twitch, and red, slow twitch, muscles (18,19). These isozymes differ in the size of the a subunit, a and a'. The difference in physiological function among

these isozymes is not clear. The y subimit, Mr 45 kDa, is the catalytic subunit (20-22). The 5 subunit, Mr 17 kDa, is almost identical to bovine brain calmodulin and confers calcium sensitivity to the enzyme (23,24).

Phosphorylase kinase is distinct from many other calmodulin-regxUated 3 systems, in that tight binding of the 5 subunit occurs in the absence of calcium (25).

Roles of subunits a and the regulatory subunits

The a and P subunits of phosphorylase kinase are regvilatory subunits.

Sequence comparisons reveal that a and P are homologous (15). Some sequence differences exist, the N-terminus of P contains an unique sequence, and the middle of a has two unique sequences, with respect to the P subimit

(15). The primary structures of a and P both contain a prenylation signal.

Prenylation of a and P has been verified and both proteins are farnesylated.

However, a and P £ire not additionally processed, like many other prenylated proteins (26).

Activation of holoenzyme residts fi:om phosphorylation of a and P by protein kinase A (8,13) or by Ca^^-dependent autophosphorylation (27,28).

Protein kinase A catalyzed activation of phosphorylase kinase residts firom phosphorylation of the p subunit. Once a minimum of 1 mole phosphate/

(aPyS) hexadecamer has been incorporated into the P subunit, phosphorylation of the a subunit can occxir (13). Phosphorylation of p is essential for activation and is independent of the phosphorylation level of a

(29). Phosphorylation of a does not activate phosphorylase kinase (13,29-31). 4

However, phosphorylation of a might be important in sustaining the activation level of phosphorylase kinase (8,29,32,33). If the metal ions Ca^^ and Mg2+ are present in addition to ATP, phosphorylase kinase can autophosphorylate and autoactivate. Multiple sites on the a and P subunits are phosphorylated (28,34,35). This autophosphorylation is an intramolecular phosphorylation (36,37). The similarity in the mechanism of activation, between autophosphorylation and phosphorylation by protein kinase A, is revealed in the fact that the same site in the (3 subunit is modified (38).

The a and P subunits contain calmodulin binding sites. Binding of

Ca^^-calmodulin to a and |3 (39) activates phosphorylase kinase (30,40

178,41). The binding of exogenous calmodulin, termed 5', to a and P is Ca^-^- dependent, as it can be inhibited by trifluoperazine, and calmodulin binding peptides melittin and mastoparan (42). Protein kinase A and calmodulin complement each other in the regvilation of phosphorylase kinase activity (43).

Calmodulin binding to P can block phosphorylation by protein kinase A.

Phosphorylation of P by protein kinase A can block calmodulin binding, a and

P have several proposed calmodulin binding sites based on comparison of their primary structures with other calmodulin binding sequences (14,15). 5

Phosphorylase kinase can be activated upon treatment with 1.8 M lithium bromide, which dissociates the holoenzjone into small active complexes (21). Renatured ct/p mixtures can inhibit the activity of renatured isolated y (44). Isolated or synthetic peptides of P, containing the proposed calmodulin binding site, can inhibit phosphorylase kinase (45,46). This inhibitory effect of ot/p and P inhibitory peptides is reheved by phosphorylation and calmodulin binding. Thus, a and P subunits may be involved in the negative regidation of the phosphotransferase activity of phosphorylase kinase by directly affecting the conformation, and/or activity of the Y subunit (44,46,47).

The catalytic ysubunit

Research in the early to mid I980's by Skuster et al. (20), Chan and

Graves (21,48,49), Kee and Graves (22,50) demonstrated that the y subunit is the catalytic subunit of phosphorylase kinase. Two active complexes, ay6 and y5, were isolated and purified after treating holoenzyme with lithium bromide (21). Both complexes exhibited activity similar to that of holoenzyme. Additionally, renatured HPLC purified y subunit contained phosphorylase kinase activity (22,50). Kinetic analysis of the renatured y subunit showed that the Km values for phosphorylase b and ATP were similar to the values obtained with holoenzyme (50). 6

The N-terminal region of y, residues 1-276, has homology to the

catalj^c domain of other protein kinases. Primary sequence alignment

results in 40% homology (51). The C-terminal region of y, residues 276-386,

contains two non-contiguous high affinity calmodulin binding domains.

PhKl3, residues 302-326, and PhK5, residues 342-366 (52,53). The full-

length y subvmit (1-386) has been expressed in E. coli and insect cells by our

laboratory and others (54-56). Unfortunately, purification of full-length y has

been difficult. As a result, most work has been done utilizing a truncated

form of y, residues 1-300, that can be expressed, purified, and renatured from

inclusion bodies (54). The crystal structure of a truncated form of y, residues

1-298, has been solved in the presence of Mn^^ and the non-hydrolyzable ATP

analog, |3-y-imidoadenosine 5' triphosphate (AMP-PNP) (57). The structure of

the catalytic domain of y is structurally similar to the crystal structure of

protein kinase A. No structural information is available on the C-terminal

region containing the calmodulin and autoinhibitory binding segments.

The 5 subunit

The 5 subvmit was not identified xmtil 1978, due primarily to its poor staining with Coomassie blue after gel electrophoresis (58). It is identical to bovine uterus calmodulin (17) and differs firom bovine brain calmodulin by two amino acids (13). Calmodulin is a calcium binding protein that mediates 7

control of a variety of enzjones in response to the physiological flux of calcium.

The 5 subunit acts as a built in sensor for calcium and is responsible for

conferring calcium sensitivity to phosphorylase kinase. Muscle phosphorylase

kinase requires calcium for activity. Holoenzjnne phosphorylase kinase

appears to bind calcium solely through the 5 subxanit (23,24). Tight binding of

the 5 subunit to phosphorylase kinase occurs in the absence of calcium,

making this interaction in phosphorylase kinase unique among calmodulin

regulated enzymes (16,25).

Catalytic properties

Nonactivated phosphorylase kinase exhibits little activity at

physiological pH, yet has considerable activity at higher pH (27). This pH

dependency changes upon activation by phosphorylation, calmodulin binding,

proteolysis, and dissociation into aY6 and yd complexes (59). The

nonactivated enzyme has a ratio of activity, pH 6.8 compared to pH 8.2, of

0.04-0.08, whereas the activated enzyme has a ratio of 0.3-1.0 (59). This

difference in pH dependency suggests that a functional group located in the

active site region important for catalysis may be affected by the state of activation. It has been proposed that a catalytic base helping pidl a proton from a reactive seryl residue of substrate during catalysis could explain the high tiunover rate of serine/threonine kinases (60,61). 8

Phosphorylase kinase requires Mg2+ to form active substrate with ATP,

for phosphotransferase activity. Excess over that required to form

MgATP2-, stimulates phosphorylase kinase activity (62). This stimulation

effect of free Mg2+ can be observed via a decrease in Km for phosphorylase b

and ATP, as well as, an increase in Vmax (63). Mn^-^, on the other hand, will

activate phosphorylase kinase at low concentrations and inhibit

phosphorylase kinase at high concentrations (50).

Substrate specificity

The phosphorylation of phosphorylase b by phosphorylase kinase is a

highly specific reaction. No other protein kinase is known to catalyze this

reaction. Phosphorylase kinase can phosphorylate other proteins, including

itself, troponin I, glycogen synthase, neuromodulin, and neurogranin (64).

However, only the phosphorylation of phosphorylase b has been shown to be

physiologically relevant. More recently, phosphorylase kinase was shown to exhibit dual-specificity by the demonstration of angiotensin II tyrosine phosphorylation. Mn2+ activates this tyrosine kinase activity of phosphorylase kinase (65). Phosphorylase kinase catalyzes the formation of phosphorylase a by phosphorylation of serine 14 in glycogen phosphorylase b

(66). Peptides of the phosphorylatable region of phosphorylase b are effective substrates of phosphorylase kinase (67). A phosphorylation study on a series 9

of sjnithetic peptide derivatives based on the phosphorylatable region,

^KHKQIS^'^VRGLis, revealed that 1) a basic amino acid is required N-

terminal (P-3), 2) a basic amino acid is preferred, but not required, C-

terminal (P+2) to the phosphorylatable serine, 3) threonine can not replace serine as an acceptor, 4) hydrophobic amino acids are preferred on either side of serine, at P-1 and P+1 (67-69).

An alternative approach to help define substrate interactions is to use mutants of the truncated y (1-300) with various peptide substrates. This approach was utilized via charge-reversal mutations to define the P-3 binding site in 7(1-300) as residue El10 (70). Crystallographic results with protein kinase A suggested that E127 of protein kinase A interacted with an arginine residue in the P-3 of its substrates (71-73). Mutation of this conserved glutamate residue, EllOK, in trxmcated 7(1-300) combined with the charge- reversal in peptide substrate, KREQISVRGL, led to the identification of an electrostatic interaction between these two residues. Recent results from X- ray crystallography of truncated 7(1-298) containing a peptide substrate, MC- peptide RQMSFRL, confirms this El10 interaction (74).

Additional important information on substrate specificity has been obtained through the use of peptide libraries. For truncated 7(1-300), the sequence, KRMMSFFL, was favored for optimal interactions (75). However, 10

no kinetic analysis was performed with this peptide substrate. A modified

version of this peptide, termed MC-peptide, was analyzed in the

crystallographic studies of trxmcated Y(1-298). The MC-peptide was found to

be a better substrate than the peptide derived firom the natural sequence of

phosphorylase b (74).

Ca2+/Calmodulin-dependent protein kinases/autoinhibition

The catalytic activity of many protein kinases appears to be regulated by an autoinhibitory mechanism, also known as intrasteric regulation (76,77).

Many autoinhibitory domains were discovered by the observation that many proteolyzed or trxmcated forms of these protein kinases are active. The autoinhibitory domain often, but not always resembles the consensus sequence of an enzyme's substrate. Protein kinase A was the first described case of intrasteric regulation by a pseudosubstrate domain located on the regulatory subxmit (78). Substantial evidence exists implicating the autoinhibitory mechanism in protein kinase G, protein kinase C, myosin light chain kinase, twitchin, calmodulin-dependent protein kinase I and phosphorylase kinase (79-82). Phosphorylase kinase is a member of the Ca2+

/calmodulin-dependent protein kinase family. Members of the Ca2+

/calmodulin-dependent protein kinase family, myosin light chain kinase, calmodulin-dependent protein kinases I, II, IV, and phosphorylase kinase 11

contain an autoinhibitory domain, which overlaps the calmodulin binding

regions (76). A possible mechanism for this intrasteric regxxlation is that the

autoinhibitory domain occupies the enzyme's active site. Binding of

calmodulin in the presence of calcium induces a conformational change

releasing the autoinhibitory domain from the active site, thus activating the

enzjone.

As previously mentioned, the crystsd structure of truncated 7(1-298)

has been solved (57). Knighton et al, modeled myosin hght chain kinase based

on protein kinase A complexed with the protein kinase inhibitor (PKE)

peptide. These results indicated that it was theoretically possible for the autoinhibitory domain to bind at the active site (83). A truncated form of

twitchin kinase, containing the kinase catalytic core and the sixty residue C-

terminal autoinhibitory domain was crystallized. Twitchin kinase is related to myosin Ught chain kinase. Although twitchin is not activated by calmodxilin, it is activated by the calciiun binding protein SlOO, which has homology to calmodulin. The crystal structure of twitchin kinase reveals that the C-termined autoinhibitory domain extends through the active site making extensive contacts with the active site region and passes over the activation loop (84). The more recent crystal structure of Ca^+Zcalmodulin-dependent protein kinase I illustrates another possibility. In this structiire, the 12

autoinhibitory calmodulin binding C-terminus makes multiple contacts with

the surface of the large lobe of the catalytic core, but then txims away from the

activation loop and travels along the small lobe (85). The use of small-angle

x-ray and neutron scattering resulted in the first structural view of a

truncated form of skeletal muscle myosin Ught chain kinase complexed with

calmodulin (86). Calmodulin was near the large lobe of the kinase cataljrtic core but away from the catalytic cleft, thus suggesting a major displacement of the autoinhibitory domain from the active site. Recently, this structural view was expanded via the elucidation of myosin light chain kinase/calmodulin complexed with AMP-PNP, and a peptide substrate derived from the regulatory light chain, pRLC. This structure illustrates the

"closed" active conformation of protein kinases. Myosin light chain kinase/calmodulin/substrate complex becomes more compact in overall shape.

The C-terminal autoinhibitory domain, boxmd to calmodulin, actually reorients and moves closer to the catalytic domain. New interactions are formed between calmodulin and the N-terminus of the kinase (87).

The regulatory C-terminal domains of the Ca^+Zcalmodulin-dependent protein kinase family may be involved in additional functions. The autoinhibitory domains of calmodulin-dependent protein kinase II and myosin

Hght chain kinase have been shown to be crucial for enzymatic stabilization 13

(88) or increased Ca^^/calmodulin-independent activity through autophosphorylation (89). Recent work with chimeric forms of smooth and skeletal myosin light chain kinases and Ca^+Zcalmodulin-dependent protein kinase II suggests that the regulatory C-terminal domains may affect substrate recognition and kinase activity (87).

Protein Phosphatase-1

General properties

Protein phosphatase-1, PP-1, plays an equally important, complementary regulatory role to phosphorylase kinase in glycogen metaboUsm. Protein phosphatase-1 catalyzes the dephosphorylation of phosphorylase a. This dephosphorylation converts phosphorylase a, the active form, to phosphorylase b, an inactive form.

Such a phosphorylase a to phosphorylase b converting enzyme was first reported by Cori and Green in 1943. They proposed that the removal of AMP was responsible for the change fi"om phosphorylase a to phosphorylase b and thus named the enzyme PR enzyme, for prosthetic group removing enzyme

(90). Later PR enzyme was altered to mean phosphorylase rupturing enzyme with the discovery that the phosphorylase a to phosphorylase b conversion resulted in a halving of the molecular weight (91). In 1959 Krebs, Graves, and 14

Fischer identified PR enzjone as phosphorylase phosphatase, later identified

as protein phosphatase-1 (92).

Protein phosphatase-1 is one of four major serine/threonine phosphatases, types -1, -2A, 2B, and -2C. These protein phosphatases are distinguished via their substrate specificities and susceptibility to specific inhibitors (93). Protein phosphatase-1 can be inhibited by Inhibitor 1 and

Inhibitor 2 and will preferentially dephosphorylate the P subunit of phosphorylase kinase. In addition to its role in the regulation of glycogen metabohsm, protein phosphatase-1 plays an essential role in other cellular processes, such as cell division, muscle contraction, and protein synthesis

(94). Although protein phosphatase-1 can be found in many different tissues, like phosphorylase kinase, the most widely studied form is from rabbit skeletal muscle.

In muscle protein phosphatase-1, PP-IG, is a complex of the catal5^ic subunit, Mr 38kDa, with a glycogen binding subunit, Mr 124 kDa (94).

Protein phosphatase-IG activity and affinity for the glycogen binding subunit are regulated by phosphorylation. Phosphorylation of site 1 on the glycogen binding subunit, G, by insulin stimulated kinase results in activation of protein phosphatase-1 activity. Phosphorylation of site 2 by cAMP-dependent protein kinase, (PKA), causes the dissociation of protein phosphatase-1 15

catalytic subiinit, C, from the G subunit. The released protein phosphatase-1

catalytic subunit can translocate to the cytosol. This results in the inabilitj'

of protein phosphatase-1 catalytic subunit to dephosphorylate phosphorylase

a. Phosphorylated Inhibitor 1, phosphorylated by cAMP-dependent protein

kinase, can bind to protein phosphatase-1 catalytic subunit and inhibit its

activity.

Role of subunits

Diversity in the protein phosphatase family arises from the coupling of

the catalytic subunit with a variety of non-catalytic targeting or regulatory

subunits. The catalytic domains of protein phosphatases are highly

conserved, which suggests similar mechanisms and tertiary structxire (95).

Thus, structural and functional diversity results from the combination of catalytic subunits with their associated non-catalytic targeting or regulatory subunits to form heterodimers. In the cell protein phosphatase-1 exists as a complex of an isoform of the protein phosphatase catalytic subunit with a non-catalytic targeting or regulatory subunit. The non-catalj^ic targeting or regulatory subunit influences the activity, substrate specificity and the intracellular location of protein phosphatase-1 (94).

Protein phosphatases type-1, -2A, and -2B catalytic subunits share extensive sequence identity and are members of a single gene family (96). 16

The few invariant residues identified are dispersed across the primary

sequence, with three regions relatively abundant in invariant residues. The

crystal structures for recombinant protein phosphatase-1 catalytic subunit,

protein phosphatase 2B holoenzyme and complex with immunosuppressants

have been solved (85,97-99). The structxires illustrate that the catalytic

domains of these two phosphatases adopt an identical core structure. The

carboxyl termini of protein phosphatases are divergent. Evidence suggests

that these regions may be involved in regulation and substrate specificity

(95). The regulatory C-tenninus of protein phosphatase-1 catalytic subunit

would be positioned to interact with residues surrounding the active site and

restrict the access of substrate (97).

The targeting or regulatory subunit is the subvmit that directs the catalytic subunit to a specific cellular locus. These targeting or regvilatory subimits are believed to function by locating the protein phosphatase to a particular cellxalar loci in order to be in close proximity to substrate, or to sequester the protein phosphatase from substrates and ligands (100).

Catalj^ic properties

Protein phosphatases are metalloenzjones. Recombinant protein phosphatase-1 requires the addition of Mn2+ (101). The crystal structure of protein phosphatase-1 catal3^ic subunit indicates the presence of Fe2+ and 17

Mn2+. Catalysis by serine and threonine protein phosphatases has been

proposed to occur through a direct attack of an activated water molecule at

the phosphorous center of substrate, without transfer of phosphate to the

enzyme (97). This is consistent with kinetic data suggesting that serine and

threonine protein phosphatases do not form phosphoryl-enzyme

intermediates (97,102). The catalytic site is located at the base of a shallow

groove on the surface. Many of the invariant residues occur on intervening loops that converge at this site (97).

Substrate specificity

Protein phosphatase-1 dephosphorylates the P subunit of phosphorylase kinase 10-100 fold faster than the a subxinit of phosphorylase kinase (94). Phosphorylase a is the best and most commonly used substrate

(94). Protein phosphatase-1 does not dephosphorylate phosphopeptide substrates weU, even those derived from phosphorylase a (103-105). Failure to successfully dephosphorylate phosphopeptide substrates indicates that the determinants for recognition of substrate are not only located in the primary structure but also in the higher order of substrate. No peptide binding cleft is evident from the crystal structxire of protein phosphatase-1 (97). This is consistent with the preference of protein phosphatase-1 for intact protein substrates and suggests several modes of substrate recognition. The binding 18

of substrate is proposed to occur through interaction of N-tenninal basic

residues in substrate with acidic residues in the acidic groove, binding of

phosphoserine to the active site, and interaction of C-terminal residues to the

hydrophobic groove of protein phosphatase-1 (97).

Phosphopeptide studies have jdelded some important information on

protein phosphatase-1 specificity. A preference for basic residues located N-

terminal to the phosphoserine is a general determinant (106). Addtional

work indicates that a basic residue C-terminal to the phosphoserine might be

important. Work with phosphorylase phosphatase utilizing peptides derived

from phosphorylase a by Graves et al in 1960 was the first indication that a

C-terminal basic residue might be important for specificity (104).

Modification of the C-terminal basic residue, arginine 16 of phosphorylase by

nitrosylation, amidation, and more recently through substitution of arginine

16 to citruUine indicates that in fact, the positive charge present may be important for protein phosphatase-1 specificity (105,107). Kinetic analysis of the citruUine substituted phosphorylase peptides illustrates the preference for arginine 16 for enhanced turnover. Multiple interactions occiu- in phosphorylase a that are involved in stabilizing the N-terminal region (108).

Two dimensional NMR studies indicate an interaction of arginine 16 with the phosphoryl group of a peptide substrate (107). The structure of phosphorylase 19

a shows the phosphoserine interacting with arginine 69 and arginine 43'. All of these interactions may contribute to the overall structural specificity of protein phosphatase-1.

Mutants of protein phosphatase-1 catalytic subunit have provided some information towards understanding the catalysis and substrate specificity. Extensive mutagenesis of the acidic groove of protein phosphatase-1 resxdted in little effect on the kinetic parameters for phosphorylase a (109). These results may be misleading. The acidic groove contains multiple acidic residues and the approach of mutating only one amino acid at a time may have obscured any changes in interaction with phosphorylase a. Mutants of the arginine residue proposed to bind phosphosubstrate, R221Q and R221S, exhibit higher Km values for phosphorylase a and much lower Vmav values (109,110). Other mutants of protein phosphatase-1 have focused mainly on the invariant phosphatase residues (110,111). Results in these studies have been consistent with identification of the metal binding residues firom the crystal structure.

Dissertation organization

The remainder of this dissertation consists of four papers, that will be submitted for pubhcation. The first paper addresses the importance of arginine 16 of phosphorylase, in peptide substrates of phosphorylase kinase. 20 as a positive substrate specificity determinant. The second paper is a study of the specificity determinants of PhK13, a peptide inhibitor, that are important for the inhibition of phosphorylase kinase. The third paper is a study of the expression, purification, and initial characterization of PhK86. the C-terminal autoinhibitory domain of the catalytic subunit of phosphorylase kinase. The fourth paper siunmarizes work with mutants of phosphorylase a utilized to analyze the substrate specificity of protein phosphatase—1. The general conclusion follows the fourth paper. References cited in the general introduction and general conclusion follow the general conclusion. 21

ARGININE TO CITRULLINE REPLACEMENT IN

SUBSTRATES OF PHOSPHORYLASE KINASE

A paper to be submitted to Biochemical Biophysical Research

Communications

Cheryl Bartleson, Siquan Luo, Donald J. Graves, and Bruce L. Martin

Abstract

Synthetic peptides based on residues 9 to 18 of glycogen phosphorylase

were prepared containing citruUine at position 16, or at both positions 10 and

16. The peptides were compared as substrates for a recombinant, truncated

form of the catalytic subunit of phosphorylase kinase (residues 1-300). Both

the mono-citruUine and the di-citruUine peptides were phosphorylated less

effectively than the parent peptide; kcat/Km values were approximately 20%

the value with the parent peptide. Incorporation of the second citrtdline had

Little change in the effectiveness of the peptide as a substrate although the

kinetic parameters with the citruUine peptides did show differences. The change in phosphorylation did not seem to result from a change in peptide structure. Two dimensional NMR studies of di-citruUine peptide are reported and show no change in the solution structure of the peptide compared to parent peptide. Thus, the change in kinetic parameters with the modified peptides seemed an effect of arginine replacement and was likely a 22

consequence of the loss of charge inasmuch as the size of arginine and

citrulline are similar. The results also suggest that arginine-16 was more

important for phosphorylase kinase recognition than arginine-10. This result

was consistent with earlier studies using alanine to replace arginine in

sjmthetic peptides as substrates for the holoenzyme form of phosphorylase

kinase.

Introduction

The modification of proteins via phosphorylation and

dephosphorylation reactions is an ubiquitous mechanism for the regulation of

subsequent activity. The enzymes catalyzing these reactions, the protein

kinases and phosphatases, are broadly classified by target specificity into

serine and threonine specific, and tjnrosine specific enzymes. Protein kinases

can readily be characterized based on primary sequence determinants as

judged by the systematic use of synthetic peptides (1). In other studies, we

have examined the specificity of phosphorylase kinase. Using synthetic

peptides based on the phosphorylation site of glycogen phosphorylase, the role of residues surrounding the target site (serine-14 of the phosphorylase sequence) has been investigated (summarized in reference 2). Basic residues on both the amino terminal and carboxyl terminal sides of the target serine were foxmd to be important in recognition by phosphorylase kinase. This is 23

distinctive from most other serine protein kinases, which seem to require

basic residues only on the amino terminal side of the target residue. These

studies used synthetic peptides in which basic residues, lysine or arginine,

were often replaced by alanine residues for reactions catalyzed by the

holoenzyme form of phosphorylase kinase. Alanine differs not only in charge,

but also in size and the effects of this replacement may have other imderlying

causes.

Peptidylarginine deiminase was previously used as a tool to study the

role of arginine residues in glycogen phosphorylase function (3). Arginines

were converted to citrxiUine by a hydrolysis reaction res\ilting in the loss of the

positive charge with little change in the size of the side chain. The

replacement of arginine by citruUine did have effects on the properties of

phosphorylase with the arginine target having the greatest effect traced to

position-16 of phosphorylase. Not surprisingly, the ability of phosphorylase kinase to phosphorylate the protein was affected as well. Synthetic peptides containing citnilline were enzymatically generated using peptidylarginine deiminase, from a parent peptide of residues 9 to 18 (peptide P9-18) of the target sequence from phosphorylase: LysArgLysGlnlleSerValArgGlyLeu.

Both mono-citrulline and di-citruUine peptides were enzjnnatically produced.

Difficulties analyzing citrulline content provided imcertainty about the 24

composition of the mono-modified peptide, particularly whether modification

was position-10 or position-16 of the peptide or a mixture of both. The di-

citruUine peptide showed no arginine after amino acid analysis providing

confidence in the composition of this variant. Phosphorylation of these

peptides by phosphorylase kinase did show differences although a full kinetic

analysis was not done (3). In this work, citruUine-containing peptides have

been chemically prepared so the composition and sequence are definitively

known. The peptides were compared as substrates for the recombinant,

tr\mcated form of the catalytic subunit of phosphorylase kinase. Kinetic data

were collected and can be compared to peptides containing alanine residues to

provide information about the reqxiired properties of residues located near the target serine. Use of an isolated catalytic moiety of phosphorylase kinase removes any concern about the non-specific effects caused by the presence of the other subimits of the kinase.

Experimental procedures

Materials- Peptides used in these experiments were synthesized and purified to 95% purity by the Protein Facility at Iowa State University or

QCB, Inc. (Framingham, MA). [y-32p]ATP was the product of ICN Biomedicals

(Costa Mesa, CA). Other chemicals were from Aldrich Chemical (Milwakee,

WI) or Fisher Scientific (Pittsburgh, PA). 25

Proteins-The recombinant, tnmcated y subunit of phosphorylase

kinase was expressed, renatiired, and purified as described previously (4).

Protein concentrations were determined by the method of Bradford (5).

Peptide Phosphorylation- Kinase activity was measured as described

(4) in 60 mM Pipes, 60 mM Tris, pH 8.2; 10 mM MgCla, 5 mM DTT; and 1 mM [y-32p]-ATP at SCC. Reactions were initiated by the addition of enzyme and incubated at 30°C. After 3 minutes, 15 |il were spotted on phosphocellulose paper (P-81) and washed with 0.5% phosphoric acid. After 3 washes, the filter papers were washed with 95% ethanol, dried, placed in scintillation fluid, and radioactivity remaining on the papers counted. Initial velocities were with 0.33, 0.50, 0.66, 1.0 and 2.0 mM each peptide. Each substrate-velocity pair was determined at least in triplicate (n=3 for P9-18, and n=4 for P9-18-Cit and P9-18-Cit2). Rates were converted to units of nmol phosphate incorporated per minute per nmol enzyme.

Peptide Samples for NMR Spectroscopy- The di-citrulline peptide, P9-

18Cit2, was enzymatically synthesized as described (3) and purified by reverse phase high performance hquid chromatography. The chromatography was accompHshed on a Beckman 332 chromatography system on a 10 x 250 mm Vydac Cl8 column (Hesperia, CA). Elution was performed employing a gradient of acetonitrile in 0.1% trifluoroacetic acid over 25 minutes. Detection 26 was done at 230 nm with a Hitachi 100-40 UV-Vis variable wavelength detector. Purity was confirmed by thin-layer chromatography, high- performance hquid chromatography, and amino acid analysis on an Applied

Biosystems Model 420A amino acid analyzer and concentration estimated by amino acid analysis.

Two Dimensional NMR Spectroscopy- Two dimensional ^H-nuclear magnetic resonance spectra were collected on a Varian Unity-50 spectrometer. D2O (20%) was included for field frequency lock. Both NOESY and TOCSY two dimensional pulse sequences were used (6-8). For the

NOESY spectra, a mixing time of 300 milliseconds was used. For the TOCSY spectra, the mixing time was 30 milliseconds. Each spectrum was comprised of 32 or 64 FIDs of 4K data points. The sample concentration was 3.7 mM peptide and the pH was 5.7. The spectra were collected at 6°C. A spectral width of 5000 Hz was utilized with a pvilse width of 15.5 ms.

Results and discussion

Synthetic peptides were prepared with sequences and designations shown in Figure 1. Citnilline was used to replace arginine in position 16 in the mono-citrtdline peptide, P9-18Cit, and in both positions 10 and 16 in the di-citruUine peptide, P9-18Cit2. Enzymatically prepared versions of these modified peptides were previously reported to be substrates for 27

phosphorylase kinase (3), but no kinetic analyses were performed because of

low availability, and uncertain composition and sequence of the modified

peptides. Chemical synthesis abrogated these concerns.

There were differences between the citrulline containing peptides and

the parent peptide. Both modified peptides were phosphorylated, but more

slowly as shown in Figure 2. The replacement of arginine-16 with citrulline

caused a significant decrease in peptide phosphorylation and the decreased

phosphorylation was greater in the di-citrulline peptide. Kinetic

characterization was done and the parameters are collected in Table I. The

parameters showed that citrulline replacement primarily caused an effect on

the Km parameter of the peptides with a modest variation in the effect of

citrulline replacement on the kcat parameter. Each citrulline replacement

caused an increase in Km. Citrulline at position 16 only (peptide P9-18Cit) caused a 3-fold increase in Km- Change of arginine at position 10 also had an efiect on the Km parameter, but the effect was less than 2-fold compared to

P9-18Cit. In previous studies (2) replacement of arginine-10 alone with alanine had virtually no effect on the kinetic parameters. Replacement of arginine with lysine did cause changes in the kcat parameter with little difference in Km; the kcat term was reduced 100-fold. Replacement of arginine-

16 with alanine or glycine did affect both kinetic parameters detrimentally. 28

although the greatest effect was on kcat- No di-alanine containing peptide was

tested. Comparing these data, it seems that it was the loss of charge that

most influenced the Km for peptides with phosphorylase kinase. Only with the

loss of size as in the alanine peptides did the kcat values get drastically

influenced. CitruUine replacement did not cause similar declines in kcat. It

seems that charge is the critical requirement at position 16.

These data suggest that arginine replacement by citruUine had

consequences on peptide phosphorylation. A possible basis for charge

recognition is a change in the overall structiire of the citridline peptides in

solution. Two dimensional NMR was used to compare the solution structures

of peptides P9-18 and P9-18Cit2. Previously, we reported that NMR studies

of peptide 9-18 showed little secondary structure for this peptide at pH 5.7

(9). The same NMR studies were done on the di-citrulline peptide and only

this peptide was characterized as it was expected to show the most significant differences in structure. Peptide P9-18Cit2 showed little secondary structure. There were few cross peaks evident in the spectrum suggesting Uttle interaction between residues. The complete assignment of resonance for the di-citrulline peptide was done and resonances are shown in

Table II with assignments for the parent peptide included for comparison.

There were little differences between assignments consistent with similar 29

structures of the peptides. The notable difference was the e-nitrogen

hydrogens for the arginine or citruUine residues. In both P9-18 and P9-

18Cit2, the residues showed no difference between positions 10 and 16.

Collectively, these data support the conclusion that arginine-16 in

glycogen phosphorylase is involved in the recognition and effective

phosphorylation of phosphorylase by phosphorylase kinase. The positive

charge is not an absolute requirement consistent with earlier suggestions

(2,3), but phosphorylase kinase did prefer positive charge on the carboxyl side

of the target serine for most effective phosphorylation of substrate. Based on

comparisons to the cAMP dependent protein kinase, residues of the y-subunit

predicted to interact with the P+2 position are present in sub-domains II and

in (nomenclature of Hanks, et al., 10). There is an unique insert containing

three glutamic acid residues, E61-E62-V63-Q64-E65, in this part of the

protein. Recently, a role for positive charge in this position was further supported by the characterization of the crystal structure of a similar form of the catalytic subunit (residues 1-298) complexed with a heptapeptide (see

Figure 1) containing an arginine in the P+2 position as the penidtimate residue (11). The sequence of this heptapeptide is similar to preferred sequences identified by screening a synthetic peptide Library (12) yielding peptides with Phe, Arg, or Lys in the P+2 position. The resolved crystal 30

structure suggested that the arginine in the P+2 position was oriented toward

a glutamic acid residue, Glu-182, of phosphorylase kinase. Clearly, charge

interactions between the P+2 arginine and Glu-182, or another glutamate

residue, wotdd facilitate more effective interaction of the substrate with the enzyme. Conversion to citridline in position 16 may weaken this interaction resulting in poorer substrate turnover.

Literature cited.

1. Pinna, Lj\.., and Ruzzene, M. (1996) Biochim. Biophys. Acta 1314, 191-

225.

2. Graves, D.J. (1983) Meth. Enzymol. 99, 268-278.

3. Luo, S., Martin, B.L., and Graves, D.J. (1995) Arch. Biochem. Biophys. 318,

362-369.

4. Huang, C.-Y.F., Yuan, C.-J., Livanova, N.B., and Graves, D.J. (1993) Mol.

Cell. Biochem. 127/128, 7-18.

5. Bradford, M.M. (1976) Anal. Biochem. 72, 248-254.

6. Jenner, J., Meier, B.H., Bachmann, P., and Ernst, R.R. (1979) J. Chem.

Phys. 71, 4546-4553.

7. Braunschweiler, L., and Ernst, R.R. (1983) J. Magn. Reson. 53, 521-528.

8. Griesinger, C., Otting, G., Wuthrich, K, and Ernst, R.R. (1988) J. Am.

Chem. Soc. 110, 7870-7872. 31

9. Martin, B.L., Luo, S., Chen, M., Kintanar, A., and Graves, D.J. (1998) Arch.

Biochem. Biophys. In press.

10.Hanks, S.K., Quinn, A.M., and Hunter, T. (1988) Science 241, 42-52.

1I.Lowe, E.D., Noble, M.E.M., Skamnaki, V.T., Oikonomakos, N.G., Owen,

D.J., and Johnson, L.N. (1997) EMBO J. 16, 6646-6658.

12.Songyang, Z., Lu, KP., Kwon, Y.T., Tsai, L.-H., Filhol, O., Brickey, D-A.,

Soderling, T.R., Bartleson, C., Graves, D.J., DeMaggio, A.J., Hoekstra,

M.F., Blenis, J., Hvinter, T., and Cantley, L.C. (1996) Molec. Cell. Biol. 16,

6486-6493. 32

P9-18 Lys Arg Lys Gin He Ser Val Arg Gly Leu

P9-18Cit Lys Arg Lys Gin lie Ser Val Cit Gly Leu

P9-18Cit2 Lys Cit Lys Gin He Ser Val Cit Gly Leu

MC peptide Arg Gin Met Ser Phe Arg Leu

Figure 1. Sequences of the substrate peptides. The sequences of the

different peptides are shown illustrating the position of

citrulline incorporation. The heptapeptide variant used in

crystallographic studies is also shown. 33

500000 -m- P9-18 £ 450000 (M -t- P9-18Cit g 400000 -Tfc- P9-18Cit2 >§ 350000-|

I 300000 a 1250000 I 200000 (U CQ 150000H JS S* 100000 o ^ 50000

10 15 20 Time (min)

Figure 2. Phosphorylation of P9-18 peptides. Substrate peptides were phosphorylated as described in the text. The concentration of each peptide was 1.0 mM. Table I Kinetics of Peptide Phosphorylation

Substrate Peptide kcat (min ') Km (mM) kcnt/Kni

Recombinant Catalytic Subunit-Truncated Fragment of Residues 1-300 P9-18 3410+ 119 0.26 + 0.03 13170 (1.00)

P9-18Cit 2396 + 277 0.81 + 0.20 2951 (0.22)

P9-18Cit2 3293 + 157 1.44 + 0.12 2284 (0.17)

Holoenzyme Studies (Calculated from Reference 2)

P9-18 3771 + 351 0.9 + 0.2 4190 (1.00)

P9-18 10-Ala 3511 + 247 0.9 + 0.3 3901 (0.93)

P9-18 16-Ala 234 + 4 2.3 + 0.04 102 (0.02)

P9-18 16-Gly 156 + 2 1.0 + 0.12 156 (0.04) Table II Resonance Assignments for Di-Citrulline and Parent Peptides

P9-18Cit2

Residue g-NH g-CH &-CH X-CH 8-CH £-NH Other

Lys-9 3.89 1.82 1.42 2.97 Cit-10 8.42 4.25 1.79 1.51 3.08 6.39 1.70

Lys-11 8.60 4.21 1.74 1.41 2.97 Gin-12 8.55 4.34 1.95 2.32 2.03

110-13 8.44 4.16 1.84 1.17 0.84 0.88 1.45 (P-CHa) (Y-CHa) Ser-14 8.51 4.51 3.82 Val-15 8.38 4.13 2.09 0.91 Cit-16 8.42 4.29 1.79 1.51 3.08 6.39 1.70

Gly-17 8.44 3.90

Leu-18 7.80 4.19 1.56 1.56 Table II (continued)

P9-18

Residue a-NH a-CH fi-CH x-CH 8-CH £-NH Other

Lys-O Arg-10 8.43 4.31 1.79 1.63 3.19 7.22

Lys-ll 8.66 4.23 1.75 Gln-12 8.60 4.34 1.96 2.33 2.03

Ile-13 8.43 4.18 1.84 1.17 0.90 0.84 1.45 (P-CHs) (y-CHs) Ser-14 8.53 4.52 3.83 Val-15 8.39 4.13 2.08 Arg-16 8.48 4.29 1.85 1.63 3.16 7.22 1.75 Gly-17 8.45 3.91

Leu-18 7.86 4.19 1.57 1.57 0.89 37

SPECIFICITY DETERMINANTS OF PHK13 IMPORTANT

FOR THE INHIBITION OF PHOSPHORYLASE KINASE

A paper to be submitted to the Journal of Biological Chemistry

Cheryl Bartleson and Donald J. Graves

Abstract

The C-terminus of the catalytic y subunit of phosphorylase kinase contains two autoinhibitory calmodidin binding domains, designated PhK13 and PhK5. Peptides of these sequences inhibit trxmcated 7(1-300). Previous data show that the peptide PhKl3, residues 302-326, is a competitive inhibitor with respect to phosphorylase b, with a Ki of 1.8 ^iM (1). This result suggests that PhK13 could bind to the active site of trxmcated 7(1-300).

Variants of PhK13 were prepared to localize the determinants for interaction with the catalytic fragment 7(1-300). A dodecapeptide, PhKl3-l, containing residues 302-312, was foimd to be a competitive inhibitor with respect to phosphorylase b, with a Ki of 6.0 nM. PhKl3 has been proposed to function as a pseudosubstrate inhibitor, with C308 occupying the site that normally accomodates the phosphorylatable serine in phosphorylase b. To test this pseudosubstrate h5T)othesis, a PhK13-l variant changing C308 to serine was synthesized. Kinetic characterization of this peptide reveals it does not serve 38

as a substrate, but is a competitive inhibitor. Thus, additional PhK13-l

variants were designed based on previous knowledge of phosphorylase kinase

substrate determinants. Each variant was analyzed both for its ability to serve as a substrate, and as an inhibitor for trtmcated 7(1-300).

The inabihty to convert PhKl3-l into an efficient substrate resulted in questioning whether PhKl3-l could bind to the active site. Competitive inhibition is consistent with active site binding but does not prove active site binding. Additional analysis of the ATPase activity of trimcated 7(1-300) in the presence and absence of PhK13-l resulted in inhibition of the ATPase activity by PhKl3-l,

PhK13-l is a potent competitive inhibitor of phosphorylase kinase, with respect to phosphorylase b phosphorylation and ATPase activity.

Although PhKl3-l does not appear to function as a pseudosubstrate inhibitor, several specificity determinants employed in the recognition of phosphorylase b as substrate seem to be employed in the recognition of

PhK13-l as an inhibitor.

Introduction

The catalytic activity of many protein kinases appears to be regulated by an autoinhibitory mechanism, also known as intrasteric regulation (2,3). Many autoinhibitory domains were discovered by the 39 observation that numerous proteolyzed or truncated forms of these protein kinases resulted in activation. The autoinhibitory domain often, but not always, resembles the consensus sequence of an enzyme's substrate. Protein kinase A was the first described case of intrasteric regxilation by a pseudosubstrate domain located on the regulatory subunit (4-6). Substantial evidence exists impHcating this mechanism in protein kinase G, protein kinase C, myosin hght chain kinase, twitchin, calmodulin-dependent protein kinase I and phosphorylase kinase (2,3,7,8)-

Phosphorylase kinase is a member of the Ca^^ /calmodulin- dependent protein kinase family. Members of the Ca^-^ /calmodulin- dependent protein kinase family, myosin hght chain kinase, calmodulin- dependent protein kinases I, II, IV, and phosphorylase kinase contain an autoinhibitory domain, which overlaps the calmodvdin binding regions

(2,3). A possible mechanism for this intrasteric regulation is that the autoinhibitory domain occupies the enzyme's active site and binding of calmodvdin in the presence of calcixmi induces a conformational change releasing the autoinhibitory domain firom the active site, thus activating the enzyme. Alternatively, the autoinhibitory domain could bind outside of the active site and influence the active site indirectly. Binding of 40

calmodulin in the presence of calciimi to the autoinhibitory domain would result in a conformational change that is transmitted to the active site and results in activation of the enzjone.

Phosphorylase kinase catalyzes the conversion of phosphorylase b, inactive, to phosphorylase a, active, by phosphorylating serine residue

14. This reaction is highly specific. Phosphorylase kinase is the only protein kinase known to catalyze this reaction in vivo. Holoenzyme phosphorylase kinase has a molecular mass of 1300 kDa. It is comprised of a hexadecamer, containing four of each of the four subunits, (aPY5)4.

The a and P subvmits are regulatory. The 5 subimit is equivalent to bovine brain calmodulin and confers calcium sensitivity to the enzyme.

The Y subunit is the catalytic subimit.

The N-terminal 300 amino acids of y encompass the catalytic domain of phosphorylase kinase. The C-terminal 86 amino acids contain two non-contiguous high affinity calmodvdin binding segments, PhK13, residues 302-326, and PhK5, residues 342-366 (9) (Figure 1). The catalytic

Y subxinit is unique in that the 5 subunit remains tightly bound in the absence of calciimi and the autoinhibitory C-terminal domain is comprised of two calmodulin binding domains. Dasgupta et al. determined the Ki values for PhK13 and PhK5 utilizing holoenzyme phosphorylase kinase to 41

be 300 nM and 20 fiM, respectively (10). These values fall into the range of

Ki values obtained with other protein kinase autoinhibitory domains (2,3).

The peptides PhK13 and PhK5 differentisdly inhibit truncated y (1-300).

Previous data show that PhKlB is a competitive inhibitor with respect to

phosphorylase b, with a Ki of 1.8 |iM (1), suggesting that PhK13 may act as

a pseudosubstrate with cysteine 308 occupying the site corresponding to the

serine of the phosphorylatable substrate. Figure 2 illustrates the

ahgnment of PhKl3 with 9-18, the peptide substrate derived from the

phosphorylatable sequence of phosphorylase b.

Kinetic studies utilizing two truncated 7(1-300) mutants, EllOK

and E153R, with charge reversals of 9-18 peptide, demonstrated that

these mutants recognize the P"^ Gysine) and P-2 (glutamine)

positions/residues of substrate, respectively (1). The crystallization of triincated 7(1-300) with a peptide substrate confirms this assignment

(11). PhK13 contains a lysine at P-3 and a valine at P-2. Analysis of PhK13 binding to the truncated 7(1-300) mutants EllOK and E153R resulted in increasing Ki's for PhKl3, compared to wild-type truncated 7(1-300). The

K, increased 14-fold with EllOK and 8-fold with E153R (1). These resialts are consistent with PhK13 binding to the active site of truncated 7(1-300). 42

The aim of this work was to test the hypothesis that PhK13

functions as a pseudosubstrate inhibitor. Conversion of PhK13 from an

inhibitor to a substrate would be indicative of active site binding. In lieu

of structural information, this would support the model, in which the C-

terminus of phosphorylase kinase interacts directly with the active site, as

is observed with twitchin kinase. In addition the specificity determinants of PhK13 important for the inhibition of phosphorylase kinase were examined in an effort to further our understanding of this molecular interaction.

Experimental procedures

Materials- [y-^^PJ-ATP was purchased from ICN Biomedicals (Costa

Mesa, CA). All other reagents were commercially available from Fisher

(Pittsburgh, PA), Sigma (St. Louis, MO) or Boehringer Mannheim

(IndianapoHs, IN). Phosphorylase b was ptirified as described (12).

PhKlS, "/residues 302-326, PhKl3 analogs, residues 302-314, 9-18 peptide, residues corresponding to the phosphorylatable region of phosphorylase b, were synthesized at the Iowa State University Protein

Facility. All peptides were purified by reverse-phase HPLC utihzing a

C18 column. 43

Peptide Concentration Z)eiermimziions-Peptide concentrations were

determined by amino acid analysis at the Iowa State University Protein

Facility and by DTNB titration for the cysteine containing peptides (13).

Expression, Purification, and Renaturation of Truncated ySubunit-

Tramcated Y(1-300) was expressed, ptirified, and renatured as previously

described (14). Recombinant tnmcated Y(1-300) was purified to homogeneity

as judged by SDS-PAGE (data not shown). Protein concentrations were

determined by Bradford assay with commercially prepared reagents from

BioRad (Hercules, CA) (15).

Activity Assay-The protein kinase activity of tnmcated Y(1-300) was

determined by incorporation into phosphorylase b as previously described

with minor modification (16,17). The standard activity assay contained 50

mM PIPES, 50 mM Tris, pH 8.2, 10 mM MgCh, 10 mg/ml phosphorylase b, 5

mM DTT, ImM [Y-32P]-ATP, and 0.5 |ig/ml tnmcated Y(1-300). The reaction is performed at 30°C and initiated by the addition of enzyme. After incubation 15 |j1 of reaction mixture was spotted on ET-31 Whatman filter paper, and washed with trichloroacetic acid (18). The radioactivity incorporated was determined by liquid scintillation counting.

Substrate Analyses-AIL peptides were first evaluated as substrates for tnmcated 7(1-300) using the standard activity assay, but with non- 44

radioactive ATP, and peptide concentrations of 1-2 mM. These samples were

analyzed by MALDI-TOF-MS using a Finnigan Lasermat 2000, with equal

molar ratios of ACH matrix to peptide, in the Iowa State University Protein

Facility. The second evaluation of peptides as substrates for truncated y(l-

300) utilized the standard activity assay with [y-^zpj-ATP.

Inhibition Studies-ln the inhibition assay (Figure 3), the standard

activity assay was used except, for the addition of PhK13-l and the final concentration of phosphorylase b was 4 (iM. In the kinetic analyses (Table II), the concentration ranges of phosphorylase b varied from 4 |iM-90 inhibitor peptide concentrations varied from IjiM-lmM and the concentration of ATP was fixed at ImM.

Kinetic data were analyzed with either Enzfitter (Elsevier Science

Publishers) or Enzyme Kinetics (Trinity Software). The inhibition data were fitted to equations for competitive or non-competitive inhibition. The inhibition constants were determined from the linear secondary plots (data not shown). All analyses were repeated a minimum of three times in triplicate.

ATPase Assa^'s-ATPase assays were performed at 30°C as described

(19). Final concentrations of reagents in the assay were as follows; 50 mM

Hepes, pH 8.2, 0.1 mM DTT, 20 mM MgCla, 1 mM ATP. Reactions were 45

initiated by the addition of truncated 7(1-300) to a final concentration of 10

M-g/ml. PhKl3-l concentrations varied from 0.1 nM —500 jiM. At the

specified time 20 ^il aliquots were removed from the assay mixture and

pipetted into vials containing 20 (il of a mixtiure of 2.8 N sulfuric acid, 4.3%

silicotxmgstic acid, plus 100 ^il of an organic mixture containing benzene and

isobutanol, 1:1. After mixing the solution, 20 (J1 of 10% ammonium

molybdate was added and the sample vortexed for 10s. An aliquot of 20 was removed from the upper organic phase and spotted on ET31 Whatman filter paper. Liquid scintillation counting was used to determine the amount of [32p].p| released.

Results

IdentificatioTL of Inhibitory Section of PhKl3- A variant of PhKl3 was prepared to localize the determinants for interaction with trimcated Y(1-

300). A dodecamer of the N-terminus of PhKl3, PhKl3-l,

GiCFKVICLTVXA, containing cysteine 308 inhibits tnmcated Y(1-300)

(Figure 3). A peptide of the C-terminus of PhK13, LASVRIYYQYRRVKP, does not inhibit (10), suggesting that the major interaction in the active site occurs through the segment containing cysteine. PhK13-l was a competitive inhibitor of truncated 7(1-300), with respect to phosphorylase b, and has a Ki of 6 jiM (Table II). 46

Pseudosubstrate Analysis of PhKlS-Maxime is the most common amino

acid occupjdng the phosphate acceptor site in autoinhibitory and/or

pseudosubstrate domains (3). In many autoinhibitory protein kinases,

substitution of this alanine residue to serine resulted in conversion of the

inhibitor to a substrate. If cysteine 308 occupies the phosphate acceptor site, does the equivalent serine substituted peptide then serve as a substrate?

We made the PhK13-l variant, PhK13-l(C308S), changing C308 to serine.

Based on our knowledge of phosphorylase kinase substrate specificity, this

PhK13-l(C308S) peptide was expected to function as a substrate. This vEiriant was analyzed as a substrate using 9-18 peptide as a comparison, but we were unable to detect any phosphorylation in this peptide using MALDI-

TOF MS (Figure 4). Analysis of PhK13-l(C308S) as an inhibitor demonstrated that this variant is a competitive inhibitor of truncated 7(1-

300) with respect to phosphorylase b. The change of C308S resulted in the K, increasing 30-fold, from 6 pM to 175 |iM (Table II).

Inhibitor Specificity Determinants-To leam more about the contribution of other residues in the sequence of PhK13-l, additional PhKl3-l variants were designed based on prior knowledge of phosphorylase kinase substrate determinants. Residues previously shown to affect specificity of phosphorylase kinase for phosphorylase were systematically substituted 47

(20). Table I lists the PhK13-l variants designed. PhK13-l(F304E/C308S) adds a basic residue in the P-* position, since phosphorylase b contains an arginine at P"*. The optimal sequence derived from peptide library work places a phenylalanine at P '*, implying that phenylalanine can be accommodated in this position (21). PhKl3-l(V306Q/C308S) replaces valine with glutanune at P-2. The recent crystal structure of truncated y(1-

298) complexed with a modified version of the optimal library sequence indicates that valine should fit in place of glutamine (11). In the library sequence, a methionine residue was selected in this position (21). Arginine in the P^^ position has been shown to be a positive determinant for phosphorylase kinase recognition and this substitution resulted in peptide

PhK13-l(C308S/T310R). Each of these peptides was analyzed for its ability to serve as a substrate of truncated 7(1-300). As evident from the apparent lack of detectable phosphorylation by MA LP I analysis or 32-P incorporation, none of the PhK13-l variants listed in Table I were able to function as substrates for trxmcated Y(1-300) (data not shown). Every PhKl3-l variant was a competitive inhibitor of trimcated 7(1-300) with respect to phosphorylase b as substrate. The resulting changes in inhibition constants varied widely, depending on the substituted amino acid residue. Table II summarizes these results. 48

Inhibition Analysis of PhKl3-l/RG Variants-T^o additional PhKl3-l variants were designed (Table HI). The glycine residue at found in phosphorylase b, neurogranin, and neuromodulin was incorporated into

PhK13-l (22). The first glycine variant PhK13-l(C308S/T310RAr311G) tested up to ImM, no longer inhibited truncated 7(1-300) (data not shown).

This peptide variant was phosphorylated as analyzed by MALDI-TOF MS, using 9-18 peptide as a positive control (Figure 5). Phosphorylation of this peptide proceeded at a much reduced rate, approximately 100 fold less, when compared to 9-18 peptide (Figure 6). Additional kinetic analysis was not performed. The resulting conversion of this peptide fi:om inhibitor to substrate by the incorporation of glycine could change the orientation of the serine. If incorporation of glycine does change the orientation of serine, then the corresponding peptide with cysteine would also change in orientation. This potential change in orientation could weaken the interaction with truncated

7(1-300) and we may no longer observe inhibition with the peptide PhKl3-l

(TSIORA'^SIIG). Initial results with peptide PhKl3-l (T310IW311G) show no inhibition of truncated 7(1-300) with peptide concentrations up to 400

(Figure 7).

ATPase Aciiuify-Substitution of serine for cysteine resulted in much poorer inhibitor. The distance between the 7 phosphate of ATP and the 49 acceptor residue might be changed in the presence of the larger cysteine residue. Although there is no evidence for cysteine phosphorylation by a protein kinase catalyzed reaction, if the cysteine residue in PhK13-l is in closer proximity to the y phosphoryl group of ATP, it may affect the ATPase activity of trxmcated y(1-300). Phosphorylation could occur, but because a cysteine phosphate would be very labile, it could break down and re-form, making the product extremely difficult to detect. One way to test this h3T>othesis is to examine the ATPase activity of truncated 7(1-300) in the presence and absence of PhK13-l. Protein kinases have a small intrinsic hydrolytic activity towards ATP. They are able to utilize the hydroxyl group of water as a phosphate acceptor at a very slow rate. ATPase activity is monitored by detecting the release of Pi. Thus, if truncated y(1-300) was attempting to utilize the cysteine as a phosphate acceptor, we should detect an increase in the amounts of Pi released in the presence of PhK13-l when compared to the ATPase activity of trxincated y(1-300) alone. Alternatively, the peptide PhKl3-l could block access of water into the active site, thus preventing ATPase activity. Initial results on the analysis of the ATPase activity reveal that PhK13-l inhibits the ATPase activity of truncated 7(1-

300) (Figure 8). This result indicates that PhK13-l may bind tightly, filling 50

the active site, and either exclude water from entering or disrupt the y

phosphate orientation so that ATPase activity is diminished.

Discussion

A dodecamer of the N-terminus of PhK13, PhK13-l,

GKFKVICLTVLA, containing cysteine 308, was a competitive inhibitor

with a Ki of 6 |iM. The low Ki of PhK13-l, compared to the Ki of 1.8 jiM for

PhK13, indicates that most the residues important for the inhibition of

truncated 7(1-300) are located in the N-terminal half of PhK13.

Conversion of PhKl3-l to a potential substrate for truncated 7(1-300)

by changing C308 to serine was unsuccessful. If upon PhK13-l(C308S)

binding, the seryl group could not reach the terminal 7 phosphoryl group of

ATP, this peptide would not serve as a substrate, but it could act as a competitive inhibitor. Indeed, PhK13-l(C308S) competitively inhibits

truncated 7(1-300). The Kiof PhK13-l(C308S) increases 30 fold over the Ki of the parent peptide, PhK13-l, from 6 [iM to 175 ^iM.

Further analysis of the specificity determinants in PhK13-l identified

other potentially important residues, in addition to cysteine 308, that are

involved in the inhibition of trtincated 7(1-300) by PhK13-l. The increase in

K. from 175 [iM in PhK13-l(C308S) to 802 ^lM in PhK13-l(F304R/C308S)

indicates that phenylalanine in the P-4 position is an important specificity 51

determinant for inhibition. This phenylalanine is eqmvalent to the

conserved hydrophobic anchoring residue found in many calcium, calmodulin-

dependent protein kinases. This hydrophobic anchoring residue is important

for calmodulin binding as well as inhibition. The Km for 9-18 peptide,

containing arginine in the equivalent position, is ImM. Substitution of

arginine for phenylalanine raised the Ki close to the mM range. Inclusion of

the arginine at P+2 decreased the Ki down to an 8-fold difference, when

compared to PhK13-l, emphasizing the preference phosphorylase kinase has

for arginine at

The inability of the glycine variant, PhK13-l(C308S/RG), to inhibit

truncated Y(1-300) is consistent with previous results utilizing a cysteine

substituted 9-18 peptide substrate, KRKQICVRGL, which does not inhibit

truncated 7(1-300), and does not serve as a substrate (23). Previous work

utilizing a cysteine substituted Kemptide, LRRACLG, showed that this

peptide was unable to function as a substrate for protein kinase A (24).

Although this peptide could serve as an inhibitor of protein kinase A with respect to Kemptide, the Ki of 1.9 mM is very high compared to the micromolar Km of Kemptide (24). This suggests that the valine-leucine pair found in PhKl3-l, and several of the variants, is causing steric interference and changes the position of the substituted serine residue. The substitution 52 of glycine for valine could relieve this interference, enabling phosphorylation of the peptide substrate.

PhK13-l is a much better inhibitor than PhK13-l(C308S). What characteristics of cysteine make it a better inhibitor than serine? Many protein kinase pseudosubstrates contain alanine in the phosphoacceptor position. Alanine has a smaller hydrophobic side chain than cysteine, although the size of an alanine chain side would be closer to that of serine.

The larger cysteine may change the distance between the y phosphate of

ATP with respect to the acceptor residue. This could result in the SH- of cysteine causing steric interference with the ypliosphate of ATP.

Alternatively, PhK13-l binding to truncated 7(1-300) could fill the active site, excluding water. Elimination of water in the active site would prevent

ATPase activity. The inhibition of ATPase activity in the presence of

PhKl3-l supports PhKl3-l binding to the active site of trtmcated y(1-300).

Competitive kinetics obtained for the PhKl3-l variants is consistent with these peptides binding at the active site. Previous data firom this lab utilizing charge reversal mutants of truncated y(1-300) with PhK13 supports PhK13 binding at the active site. The mutants EllOK and E153R exhibited higher Ki values with PhK13 than trxmcated 7(1-300), and PhK13 remained a competitive inhibitor. However, the lack of phosphorylation 53

observed for the majority of PhK13-l variants is troublesome. An

alternative possibility is the binding of PhK13-l at another site on

trimcated 7(1-300), that is distinct from the active site. Binding of PhK13-

1 at this site coxild alter the conformation or orientation of the N- and C-

terminal lobes of the catalytic kinase core, thus altering the active site. A

change in conformation around the active site could result in altering metal ion binding, ATP binding, the orientation of the catalytic base, and/or the binding of substrate. Structural evidence exists that illustrates each possibility. The crystal structvire of twitchin kinase reveals the C-terminal autoinhibitory domain extends through the active site, making extensive contacts with the active site, and passes over the activation loop (25). The more recent crystal structiure of /calmodulin-dependent protein kinase

I shows the C-terminal autoinhibitory domain making multiple contacts with the surface of the large lobe of the catalytic core, but then txims away from the activation loop and travels along the small lobe (26). Recent work from small x-ray and neutron scattering resulted in a structiire for myosin light chain kinase complexed with calmodulin. Calmodulin was near the catalytic core, but away from the active site, suggesting displacement of the automhibitory domain. The additional structxire has been expanded to include the substrates AMP-PNP and a peptide substrate derived from the 54

regulatory light chain (27). Here the overall complex becomes more

compact, i.e. "closed", jdelding the active conformation of a protein kinase.

It is not currently possible to definitely distinguish which of these modes of

binding will ultimately be utilized by the autoinhibitory C-terminus of

phosphorylase kinase. Thus, the exact binding of PhKl3-l awaits

structural determination.

References

1. Huang, C.-Y. F., Yuan, C.-J., Blumenthal, D. K., and Graves, D. J. (1995)

JBC 270, 7183-88.

2. Soderhng, T. R. (1990) JBC 265,1823-26.

3. Kemp, B. E., and Pearson, R. B. (1991) BBCA 1094, 67-16.

4. Corbin, J. D., Sugden, P. H., West, L., Flockhart, D. A., Lincoln, T. M.,

and McCarthy, D. (1978) JBC 253, 3997-03.

5. Takio, K., Smith, S. B., Krebs, E. G., Walsh, K. A., and Titani, K. (1982)

Proc. Natl. Acad. Sci. U.SA. 79, 2544-48.

6. Titani, K., Sasagawa, T., Ericsson, L. H., Kimiar, S., Smith, S. B., Krebs,

E. G., and Walsh, K. A. (1984) Biochemistry 23, 4193-99.

7. Lei, J., Tang, X., Chambers, T. C., Pohl, J., and Benian, G. M. (1994)

JBC 269, 21078-85. 55

8. Heierhorst. J., Probst, W. C., Vilim, F. S., Buku, A., and Weiss, K. R.

(1994) JBC 269, 21086-93.

9. Dasgupta, M., Honeycutt, T., and Blumenthal, D. K. (1989) JBC 264,

17156-63.

10. Dasgupta, M., and Blumenthal, D. K. (1995) JBC 270, 22283-89.

11. Lowe, E. D., Noble, M. E. M., Skamnaki, V. T., Oikonomakos, N. G.,

Owen, D. J., and Johnson, L. N. (1997) EMBOJ16, 6646-52.

12. Krebs, E. G., Kent, A. B., Graves, D. J., and Fischer, E. H. (1958)

Proceedings of the International Symposium on Enzyme Chemistry,

Tokyo and Kyoto 19571, 41-43.

13. EUman, G. L. (1959) Arch. Biochemistry and Biophysics 82, 70-77.

14. Huang, C.-Y. F., Yuan, C.-J., Livanova, N. B., and Graves, D. J. (1993)

Molecular and Cellular Biochemistry 127/28, 7-18.

15. Bradford, M. M. (1976) Analytical Biochemistry 72, 248-254.

16. Kee, S. M., and Graves, D. J. (1986) JBC 261,4732-37.

17. Kee, S. M., and Graves, D. J. (1987) JBC 262, 9448-53.

18. Reimann, E. M., Walsh, D., and Krebs, E. G. (1971) JBC 246, 1986-95.

19. Pollard, T. D. (1982) Methods in Enzy 85,123-30.

20. Graves, D. J. (1983) Methods in Enzymology, Academic Press, New

York. 21. Songyang, Z., Lu, K. P., Kwon, Y. T., Tsai, L. H., Filhol, O., Cochet, C.,

Brickey, D. A., Soderling, T. R., Bartleson, C., Graves, D. J., DeMaggio,

A. J., Hoekstra, M. F., Blenis, J., Hxmter, T., and Csintley, L. C. (1996)

Molecular and Cellular Biology 16(11), 6486-6493.

22. Paudel, H. KL, Zwiers, H., and Wang, J. H. (1993) JBC 268, 6207-13.

23. Yuan, C.-J. (1993) in Biochemistry and Biophysics, pp. 151, Iowa State

University, Ames.

24. Mendelow, M., Prorok, M., Salemo, A., and Lawrence, D. S. (1993) JBC

268, 12289-96.

25. Hu, S.-H., Parker, M. W., Lei, J.-Y., Wilce, M. C. J., Benian. G., and

Kemp, B. E. (1994) Nature 369, 581-84.

26. Goldberg, J., Nairn, A. C., and Knriyan, J. (1996) Cell 84, 875-887.

27. Krueger, J. K., Zhi, G., Stull, J. T., and Trewhella, J. (1998) Biochemistry

37, 13997-04. Figure 1. Schematic of the catalytic subunit y of phosphorylase kijiase.

The N-terminal region of 7, residues 1-276, has homology to the catalytic domains of other

protein kinases. The C-terminal region, residues 276-386, contains two non-contiguous high

affinity calmodulin binding domains, PhK13, resides 302-326, and PhK5, residues 342-366. 1 300 386

I III I I III iy catalytic core calmodulin binding Figure 2. Sequence alignment of PhK13 with the phosphorylatable region ofphosphorylase_b. (:) represents close chemical similarity (.) represents distant chemical similarity ( I ) represnts amino acid identity phosphorylase b EllO

9KRKQISVRGLA19 I I I ; : .11 303G KFKVICLTVLA313SVRI YYQYRRVKP

Y 1-300 g

(:) represents close chemical similarity

(.) represents distant chemical similarity

( 1 ) represents amino acid identity 61

100 % Activity 90-

80- 70-

30-.

20-

10-

0 5 10 20 30 35 40 45 50

[PliKl3-l], ^iM

Figiire 3. Inhibition of truncated "^l-SOO) by PhKl3-l. % Activity of trxmcated Y(1-300) verus [PhK13-l] varied from 0-50 |xM. Phosphorylase was used as substrate. Assay described in Experimental procedures. TABLE I Sequence Comparisons of Peptide Substrates and Inhibitors of y (1-300)

9-18 K-R-K-Q-I-S-V-R-G-L

PhK13-l G-K-F-K-V-I-C-L-T-V-L-A

PhK13-l(C308S) G-K-F-K-V-I-S-L-T-V-L-A 0> PhK13-l(F304R/C308S) G-K-R-K-V-I-S-L-T-V-L-A

PhK13-l(V306Q/C308S) G-K-F-K-^-I-S-L-T-V-L-A

PhK13-l(C308S/T310R) G-K-F-K-V-I-S-L-R-V-L-A Figure 4. MALDl Spectra of 9-18 compared to PhK13- IS,

MALDI spectra of 9-18 peptide and reaction product compared with the

PhK13-IS peptide and reaction product. Reactions conditions described

under Experimental procedures. Panel A is the 9-18 spectra, showing the

predicted weight of 1180, and panel B is the reaction spectra, with the

expected product peak at 1259. Panel C is the PhK13-lS peptide spectra.

Panel D is the PhKl3-lS reaction spectra. Panel A Panel B 100

90

70

60

40

1100 12iOO 1300 ^ Panel C Panel D

27.0

26.5 kO (ba> m 26.0 cs CO

25.5

25.0

24.5

24.0

23.5

23.0

22.5

22.0

21.5

21.0

20.5

20.0

19.5

19.0

18.5

18.0 1100 1200 1300 1400 1200 1300 1400 1500 1800

Mass (m/z) Mass (m/z)

Figure 4 (continued) Table II Ki values and kinetic mechanisms for the inhibition of truncated y (1-300)

Y (1-300) form PhK 13 variant Mode of inhibition Ki (jlM)

wild-type GKFKVICLTVLA Competitive 6+1 wild-type GKFKVISLTVLA Competitive 175 + 3 wild-type GKRKVISLTVLA Competitive 802+ 1,5 wild-type GKFKVISLTVLA Competitive 80 + 5 wild-type GKFKVISLRTVLA Competitive 44 + 4 TABLE III Sequences of PhK13 Glycine Variants

9-18 KRKQISVRGL

PhKl 3-1 GKFKVICLTVLA a PhKlS-KCSOSS/TSlORWailG) GKFKVISLRGLA

PhK13-l(T310RW31lG) GKFKVICLRGLA Figure 5. MALDI Spectra of PhKl3-l S/RG.

MALDI spectra of the PhK13-l S/RG peptide and reaction product.

Reaction conditions described under Experimental procedures. Panel A

is the spectra of PhK13-lS/RG peptide and panel B is the spectra of the

reaction product indicating that no phosphorylated product is observed. A Panel B

65 85 80 CO(O 60 n 75 I I 55 70

50 65 60 45 55

CO 40- 50

35- 45

00 40 30- 35 25 30

20 25

20 15 15 10 1250 1300 1350 1400 10 12^ ISOO 1350 1400 1

Mass (m/z) Mass (m/z) 70

Panel A

800000 • 9-18 700000-

600000-

500000-

I 400000-

300000-

200000-

100000-

0 5 10 15 20 25 30 time, minutes

Figure 6A. [32P] incorporation into peptide 9-18. Panel A is the reaction with 9-18. Assay conditions are described in Experimental procedures. The concentration of 9-18 peptide was ImM. 71

Panel B

8000 • PhK13-l(S/RG) 7000-

6000-

5000-

4000-

3000-

2000-

1000-

10 15 20 time, minutes

Figvire 6B. [32P] incorporation into peptide PhK13-l(S/RG). Panel B is the reaction with PhK13-l(S/RG). Assay conditions are described in Experimental procedures. The concentration of PhK13-l S/RG was 2mM. 72

% Activity

> 80-

0 50 100 150 200 250 300 350 400 [PhK13-l/RG], \iM

Figure 7. Inhibition of truncated y(l-300) by PhKl3-l/RG. % Activity of trimcated 7(1-300) in the presence of peptide PhK13-l/RG. Assay conditions were as described under experimented, procedures. The amounts of PhK13-l/RG are as indicated. Figure 8. ATPase inhibition by PhK13-l.

% ATPase activity in the presence of increasing amounts of PhK13-l.

Assay conditions described in experimental procedures. The concentrations

of PhK13-l utilized are as indicated. I % ATPase Activity

<1

[PhK13-l], mM 75

EXPRESSION AND ANALYSIS OF PHK86, THE

AUTOINHIBITORY CALMODULIN BINDING C-

TERMINUS OF THE CATALYTIC SUBUNIT OF

PHOSPHORYLASE KINASE

A paper to be submitted to the Joximal of Biological Chemistry

Cheryl Bartleson and Donald J. Graves

Abstract

The N-terminus, residues 1-296, of the catalytic y subxmit of phosphorylase kinase has homology to the catalytic domains of other protein kinases. The C-terminus, residues 297-386, contains two non-contiguous high affinity, autoinhibitory calmodulin-binding regions, designated PhK13 and

PhK5. Peptides of these sequences differentially inhibit truncated 7(1-300).

PhK13, residues 302-326, is a competitive inhibitor and PhK5, residues 342-

366, is a non-competitive inhibitor. Although the crystal structure of the truncated 7(1-298) has been solved, no structural information is available on how the C-terminus binds to tnmcated 7(1-298). Previous attempts to purify full-length 7 in amounts suitable for structure/function studies have been imsuccessful due to proteolysis at the C-terminus. We have expressed the C- terminus of the catalytic 7 subunit, referred to as PhK86. PhK86 76

encompasses both the PhK13 and PhK5 regions. Expression of PhK86 as a

histidine-tagged fusion protein in E. coli resulted in formation of insoluble

inclusion bodies. A purification and renatiu-ation protocol has been developed

that will lead to successful production of amounts of PhK86 suitable for

structxire/function analyses. Renatured PhK86 appears functional. PhK86

binds calmodulin and truncated 7(1-300). Initial results with eluted PhK86

indicate that PhK86 inhibits tnmcated 7(1-300) with respect to

phosphorylase b.

Introduction

Phosphorylase kinase plays a key regulatory role in glycogen

metabohsm by catalyzing the phosphorylation of glycogen phosphorylase

kinase, which converts phosphorylase b, inactive, to phosphorylase a, active.

The enzyme is a hexadecamer with a subunit composition of (a,p,7,8)4, and a

molecular weight of 1300 kDa. The a, Mr, 138 kDa, and p, Mr 125 kDa, are

regidatory subunits that can be phosphorylated by protein kinase A and nan

bind exogenous calmodulin, 6'. The 7 subunit, Mr 45 kDa, is the catalj^ic sub unit. The 5 subunit, Mr 17 kDa, is almost identical to bovine brain calmodulin and confers calcium sensitivity to the enzyme. Phosphorylase kinase is distinct from many other calmodulin-regulated systems, in that tight binding of the 5 subimit occxirs in the absence of calcium (1). 77

The catalytic activity of many calcixim-ca Imodulin regulated protein

kinases appears to be regulated by an autoinhibitory mechanism (2-4). Many

autoinhibitory domains were discovered by the observation that proteolyzed

or truncated forms of these protein kinases resulted in activation. The

autoinhibitory domain is often located in the C-terminus near or overlapping

the calmodulin binding region. The proposed mechanism for this intrasteric

regulation is that the autoinhibitory domain occupies or blocks access to the

enzyme's active site. The binding of calmodulin, in the presence of calcium, induces a conformational change that releases the autoinhibitory domain, activating the enzyme.

The C-terminus, PhK86, of the cataljrtic y subimit of phosphorylase kinase contains two non-contiguous autoinhibitory calmodulin binding domains (5). Although, the crystal structure of trxmcated Y(1-298) has been solved, no structural information is available on how the C-terminus, PhK86, interacts with the catalytic core and the other subimits of phosphorylase kinase to regulate activity (6).

The full-length Y subrmit (1-386) has been expressed in E. coli and insect cells by ovir laboratory and others (7-9). Unfortimately, due to proteolysis during purification, attempts to produce svifficient amoimts of full- length Y suitable for structural studies has been vmsuccessful. As a result. 78

most work has been done utilizing a tnmcated form of y, residues 1-300, that

can be expressed, purified, and renatxired firom inclusion bodies (7). In the absence of full-length y, an alternative is to express the regulatory domain,

PhK86. PhK86 must be generated in quantities suitable for structiire/function studies in order to determine how it interacts with the kinase core of phosphorylase kinase, in addition to interactions that would be present with the other subunits in holoenzyme.

Experimental procedures

Materials-A}l chemicals were purchased firom Boehringer Mannheim

(Indianapolis ,IN), Sigma (St. Louis, MO), or Fisher (Pittsburgh, PA).

Restriction enyzmes and T4 DNA ligase were firom New England Biolabs

(Beverly, MA). The pET15b plasmid and E. coli BL21(DE3) were purchased firom Novagen (Madison, WI). Isopropyl-P-D-thiogalactop5nranoside was purchased firom 5 Prime-3 Prime (Boulder, CO). The immobilized metal affinity chromatography, IMAC, utilized were either, the Ni2+-based His-

Resin firom Novagen (Madison, WI), or the Co^-^ based TALON resin firom

Clontech (Palo Alto, CA). The nitrocellulose membrane used for Western blotting was Hybond ECL firom Amersham Life Sciences (Arlington Heights,

IL). Anti-calmodulin was firom Upstate Biotech (Lake Placid, NY). The monoclonal y antibody, mAb88, was a generous gift firom Dr. Gerald M. 79

Carlson. [y-^zpj-ATP was purchased from ICN Biomedicals (Costa Mesa,

CA). Phosphorylase b and tnmcated 7(1-300) were purified as previously

described (7,10). Calmodulin was purified from bovine brain as described

(11).

DNA manipu/aiion-Plasmid pGPK4 (12) was digested with the restriction enzymes EcoRI and HindlTT to produce a 1.2 kb fragment that contained the coding region of the fiiU-length cataljrtic y subunit. The 1,2 kb y fragment was subcloned into pUCllS. The C-terminal residues 297-386 of the catalytic subunit, y, were isolated from pUC118-y via a Ndel-Hindlll restriction digest and subsequently gel purified. The new 0.6 kb PhK86 fragment was hgated to Ndel- and HindHI digested pETlSb expression vector with T4 DNA Hgase. Ligation reactions were carried out at 4°C overnight.

The nucleotide sequence of the coding region in the resulting plasmid, pETl5b-PhK86, was confirmed by sequencing in the Iowa State University

DNA Facility. The expression plasmid, pET15b-PhK86 was transformed into E. coli strain BL21(DE3).

Expression of PhK86-Fov expression of PhK86, E. coli BL21(DE3) containing pETl5b-PhK86 were cultured into 5 ml of LB broth containing

100 M-g/ml ampicillin and grown overnight shaking at 37°C. Overnight cultiires were subsequently diluted 1:100 in 500 ml of LB plus ampicillin and 80

grown at either 22°C or 37 °C with constant shaking. Cells were induced at

an optical density (OD)6oo of either 0.6 or 0.8 with isopropyl-P-D-

thiogalactopyranoside (IPTG) at a final concentration of 0.4 mM. The

induction of PhK86 expression proceeded for 3 hours at 37°C, or 12 hours at

22°C. Cells were harvested by centrifugation at 5,000 x g for 10 minutes.

Cell pellets were firozen overnight at -20°C and used immediately.

Purification of PhK86 Inclusion Bodies-A31 buffers utilized throughout the purification procedure contained protease inhibitors at the following final concentrations: 0.7 ng/ml pepstatin A, 0.5 mM phenylmethylsulfonyl fluoride

(PMSF), 0.5 ^ig/ml leupeptin, 0.2 mg/ml benzamidine, and 0.5 p-g/ml aprotinin. The cell pellets were resuspended in cold deionized water and lysed by passage through a prechilled French pressure cell two to three times at 5000 lb/inch psi. The resulting cell lysate was centrifuged at 12,000 x g for

15 minutes at 4°C. The collected pellet was suspended in 20 mM Tris-HCl, pH 8.0. The suspension was incubated 1 hour at room temperatiire while shaking. Sodium deoxycholate was added to a final concentration of 1 mg/ml and shaking continued at room temperature for an additional hour. The

PhK86 inclusion bodies were collected at 12,000 x g for 15 minutes at 4°C and suspended in 50 mM Tris-HCl pH 8.0. The final PhK86 inclusion bodies were dissolved in denaturing binding buffer: 5 mM imidazole, 0.5 M NaCl, 20 81

mM Tris-HCl, pH 8.0, 0.1% NP-40 and 6 M guanidine hydrochloride or 8 M

urea, and stored at 4°C.

Purification and Renaturation of His-tagged PhK86-Tb.e dissolved

PhK86 inclusion bodies were filtered through a 0.45 pjn filter prior to IMAC

purification. PhK86 inclusion bodies were loaded onto the IMAC resin and

then washed with denaturing wash biaffer: 60 mM imidazole, 0.5 M NaCl, 20

mM Tris-HCl, pH 8.0, 0.1% NP-40, and 6 M guanidine hydrochloride or 8 M

urea. PhK86 was then refolded by running a continuous gradient from

denatiiring wash buffer, containing 6 M guanidine hydrochloride or 8 M urea

to nondenaturing wash buffer, containing no guanidine hydrochloride or urea.

Elution of PhK86 utilized 1 M imidazole, 0.5 M NaCl, and 20 mM Tris-HCl,

pH 8.0. Fractions collected were analyzed by SDS-PAGE for the presence of

PhK86. Fractions containing PhK86 were pooled and dialyzed into 50 mM

Tris-HCl, pH 7.8, 50 mM Hepes, 50 mM NaCl, with 5 mM DTT being added after two dialysis buffer changes. Protein concentration of eluted PhK86 was determined using commericially available Bradford reagents from BioRad

(Herculues, CA).

ImmunobloUingSamples were amalyzed by SDS-polyacrylamide gel electrophoresis, followed by semi-dry blotting at 36 mA for 30 minutes to nitrocellulose membrane. Immtinoblotting anaylsis utilized anti-calmodidin 82

or anti-y mAbSS. Proteins were detected according to the ECL protocol of

Amersham.

PhK86 inhibition analysis-The inhibition of truncated 7(1-300) by

PhK86 was determined by incorporation into phosphorylase b as

previously described with minor modification (13,14). The standard assay

contained 50 mM Pipes, 50 mM Tris, pH 8.2, 10 mM MgCla, 4 |iM

phosphorylase b, 5 mM DTT, ImM [y-^zpj-ATP, and 0.5 ng/ml trxmcated y(1-

300). The reaction is performed at 30°C and initiated by the addition of

enzjrme. At specified time points 15 pi of reaction mixture was spotted on

ET-31 Whatman filter paper, and washed with trichloroacetic acid (15). The

radioactivity incorporated was determined by liquid scintillation coimting.

Results

Expression Analysis of PhK86-la the pETl5b His-tag expression system, a series of six consecutive histidine residues is placed N-terminal to

PhK86. PhK86 csm be separated firom the histidine residues, if necessary, via a thrombin cleavage site. The histidine tag sequence enables rapid one-step purification by binding divalent cations, in particular Ni2+ and Co^^, immobilized on metal chelation resin. Expression in pET15b is under control of the T7 lac promoter and is induced with isopropyl-^-D thiogalactopyranoside. 83

Expression of PhK86 was analyzed in cells grown at 37°C (Figure 1).

Cells were induced with 0.4mM IPTG at two different optical densities, either

(OD)6OO of 0.6 or 0.8. High level expression of eukaryotic proteins in E. coli

often residts in the formation of insoluble protein aggregates referred to as

inclusion bodies. PhK86 can be seen in the cell peUet when induced at 37°C.

Reduction in growth temperatxire can result in soluble expression. Therefore, we additionally examined PhK86 expression at 22°C (Figure 2). PhK86 can be foimd in the insoluble inclusion body pellet regardless of temperature or ceU density upon induction.

Renaturation of PhK86-lnclusion bodies can offer high jdeld and relatively easy purification for proteins of interest. PhK86 inclusion bodies were purified as described in experimental procedures. After dissolving the

PhK86 inclusion bodies in denaturing buffer, and subjecting the sample to immobilized metal affinity chromatography, PhK86 was eluted in denaturing buffer. Initial trials on refolding PhK86 utilized a procedure previously successful in our laboratory for refolding truncated 7(1-300). This method involves dilution of PhK86 to 10 fig/ml, followed by overnight incubation at

4°C. Unfortunately, all detectable PhK86 precipitated during this procedure

(data not shown). The next attempted procedure involved eluting PhK86 in denatiiring buffer and subjecting PhK86 to the step-wise removal of guanidine 84 or xirea via dialysis. Each dialysis buffer change reduced the amount of guanidine or virea present. PhK86 precipitated in this procedure after removal from IM guanidine or urea (data not shown). The third attempt to renature PhK86 utilized the metal ion affinity resin to immobilize PhK86. It has been suggested that immobilizing the protein may limit protein-protein interaction and help limit aggregation (16). PhK86 was bound to the metal affinity resin in denaturing buffer. Next a gradient was nm from denaturing buffer to non-denaturing buffer in 80 column volumes. A sample of renatxired immobilized-PhK86 can be seen in Figure 3.

Functional Analysis of PhK86-We believed that PhK86 had been successfrdly renatured. However, PhK86 lacks a detectable enzjonatic activity to test whether refolding was successful. Thus, we turn to an alternative assessment of functionality. Samples of PhK86 bound to the metal affinity resin were incubated with calmodulin in the presence and absence of calcium and EGTA, as a suspension (Figure 4). No detectable calmodulin boxmd to the metal affijiity resin itself. Immobilized-PhK86 bound calmodulin. A second test for functionality including the binding of truncated Y(1-300) to immobilized-PhK86 (Figure 5). Truncated Y(1-300) did bind to the metal affinity resin under the current conditions used. The signal produced in the immobilized-PhK86 +truncated 7(1-300) lane is 85

approximately two fold stronger, indicating that PhK86 did bind tnincated

y(l-300).

Elution of PhKSGStrixctvxal studies of PhK86 will need large amounts

of refolded, soluble protein. Figure 6 illustrates the protein elution profile for

PhK86 from the metal affinity resin as detected by Bradford. This profile is

from 4 ml PhK86 inclusion body sample at 2 mg/ml nm over a 1 ml metal affinity column. All fractions collected were analyzed by SDS-PAGE for the presence of PhK86 (Figure 7). Fractions containing PhK86 were pooled and dialyzed to remove the imidazole. The concentration of PhK86 obtained from this small column was 6 jiM. We are confident that higher yields of PhK86 can be obtained from scaling this procedure up to a larger sample volxime.

PhK86 Inhibition of truncated y(l-300)-V\iKSQ contains both PhKl3 and PhK5, two inhibitory regions of the catalj^ic y subunit of phosphorylase kinase. Eluted PhK86 was analyzed as an inhibitor of truncated y(1-300), with respect to phosphorylase b as substrate. Although only a relatively small concentration of PhK86 was available for testing, PhK86 inhibited trimcated 7(1-300) (Figure 8).

Discussion

The protein kinase activity of many members of the calcium, calmodulin-dependent protein kinase family, of which phosphorylase kinase 86

is a member, may be regulated by intrasteric regulation. The C-terminus,

PhK86, of the catalytic y subxanit of phosphorylase kinase comprises two non­

contiguous autoinhibitory calmodulin binding domains, PhK13 and PhK5.

Structurally, it is unknown how the C-terminus of the y catalytic subunit

interacts with its kinase core, or the other subimits, a, P and 5, found in

holoenzyme, to regulate activity. Attempts to produce full-length y(l-386) for structural studies have been unsuccessful. Thus, an alternative approach was

to express PhK86.

Two possible modes of binding are possible for PhK86. PhK86 could

bind directly to the active site blocking access of substrate, phosphorylase b.

Data supporting this possibility comes from peptide studies utilizing PhK13, a segment of PhK86. PhKlS and a shorter version, PhK13-l are competitive inhibitors of truncated y(l-300), with respect to phosphorylase b (Bartleson, unpublished data)(17). Indirect evidence comes from the differential inhibition of PhK5. PhK5 is a non-competitive inhibitor of truncated y(l-300), yet, PhK5 competitively inhibits holoenzjnne phosphorylase kinase, with respect to phosphorylase b (17,18). Precedence for this mode of binding is illustrated by the crystal structmre of twitchin kinase (19). The C-terminal autoinhibitory domain of twitchin traverses the active site. The second possibility is that PhK86 binds outside of the active site but that the binding 87

of PhK86 influences the binding of substrate in the active site. The crystal

structure of calmodulin-dependent protein kinase I indicates that this may be

the mode of inhibition for this kinase (20). Recent small angle X-ray and

neutron scattering studies indicate that this may also be the mode of

inhibition for myosin light chain kinase (21,22). The C-terminal domains of calmodulin-dependent protein kinase I and myosin light chain kinase do not

traverse the active site but form multiple contacts with the kinase catalytic

core. In order to test these models of interaction in phosphorylase kinase, an

unproteolyzed, intact regulatory domain, PhK86, must be generated.

We have successfully expressed PhK86 as a histidine tagged fusion protein in E. coli. Soluble expression of PhK86 was not observed under conditions tested. PhK86 is expressed as an inclusion body. Inclusion bodies can provide an abundant, relatively pxire source of protein. A refolding protocol has been developed that involves refolding PhK86 while immobilized on the metal affinity resin, in an effort to minimize protein-protein interactions that can result in aggregation. Renatxiration of PhK86 employs the use of a continuous gradient that reduces the presence of denatiirant.

One caveat of working with renatured proteins can be in testing if the refolded product has been properly renatured. Often, enzymatic activity can be utilized as a reasonable indicator. PhK86 lacks a testable enzymatic 88 activity. This led us to utilize an alternative method to ascertain whether

PhK86 was properly renatured. Functionality of PhK86 was examined through the ability of calmodulin to bind to renatured, immobilized PhK86.

Western analysis with anti-calmodiilin demonstrates calmodulin binding to

PhK86. Calmodvdin boimd PhK86 in the absence of added calcium. However, the binding of calmodulin to PhK86 was enhanced upon the addition of calcixmi. In the sample lacking PhK86, we were xmable to detect calmodulin, indicating that the PhK86 was necessary for the interaction with calmodulin.

A second test of functionality involved checking the ability of truncated Y(1-

300) to bind to renatured, immobilized PhK86. Western analysis with anti-y indicates truncated 7(1-300) binding to PhK86. Non-specific binding of truncated y(1-300) to the IMAC resin was observed imder the conditions utilized. Experimental conditions will be adjxisted in an effort to reduce the interaction of tnmcated y(1-300) with the IMAC resin.

Initial results with eluted, soluble PhK86 indicates that PhK86 inhibits truncated 7(1-300). Work is in progress to scale up the expression and increase the 3deld of PhK86, in order to generate amoimts suf&cient for structiaral studies. 89

References

1. Yuan, C.-J., and Graves, D. J. (1989) Archives of Biochemistry and

Biophysics 274, 317-26.

2. Kemp, B. E., and Pearson, R. B. (1991) BBCA 1094, 67-76.

3. Soderling, T. R. (1990) JBC 265, 1823-26.

4. Crivici, A., and Ikura, M. (1995) Annual Rev. Biophys Biomol Struct 24,

85-116.

5. Dasgupta, M., Honeycutt, T., and Blumenthal, D. K. (1989) JBC 264,

17156-63.

6. Owen, D. J., Noble, M. E. M., Garman, E. F., Papageorgiou, A. C., and

Johnson, L. N. (1995) Structure 3, 467-82.

7. Huang, C.-Y. F., Yuan, C.-J., Livanova, N. B., and Graves, D. J. (1993)

Molecular and Cellular Biochemistry 127/28, 7-18.

8. Cox, S., and Johnson, L. N. (1992) Protein Engineering 5(8), 811-819.

9. Lanciotti, R. A., and Bender, P. K. (1994) Biochem J. 299, 183-189.

10. Krebs, E. G., Kent, A. B., Graves, D. J., and Fischer, E. H. (1958)

Proceedings of the International Symposium on Emyme Chemistry,

Tokya and Kyoto 19571, 41-43.

11. Sharma, R. K., Taylor, W. A., and Wang, J. H. (1983) Methods n

Enzymol 102, 210-219. 90

12. Chen, L.-R., Yuan, C.-J., Somasekhar, G., Wejksnora, P., Peterson, J. E.,

Myers, A. M., Graves, L., Cohen, P. T. W., da Cmze Silva, E. F., and

Graves, D. J. (1989) Biochem. Biophys. Res. Commun. 161, 764-753.

13. Kee, S. M., and Graves, D. J. (1986) JBC 261, 4732-37.

14. Kee, S. M., and Graves, D. J. (1987) JBC 262, 9448-53.

15. Reimann, E. M., Walsh, D., and Krebs, E. G. (1971) JBC 246, 1986-95.

16. Stempfer, G., and HoU-Neugebauer, B. (1996) Nature Biotechnol. 14,

329-334.

17. Huang, C.-Y. F., Yuan, C.-J., Blvimenthal, D. K., and Graves, D. J. (1995)

JBC 270, 7183-88.

18. Dasgupta, M., and Bliunenthal, D. K. (1995) JBC 270, 22283-89.

19. Hu, S.-H., Parker, M. W., Lei, J.-Y., Wilce, M. C. J., Benian, G., and

Kemp, B. E. (1994) Nature 369, 581-84.

20. Goldberg, J., Nairn, A. C., and Kuriyan, J. (1996) Cell 84, 875-887.

21. Krueger, J. K, Olah, G. A., Rokop, S. E., Zhi, G., Sttdl, J. T., and

Trewhella, J. (1997) Biochemistry 36(6017-6023).

22. Krueger, J. K, Zhi, G., Stidl, J. T., and Trewhella, J. (1998) Biochemistry

S7, 13997-04. Figure 1. SDS-PAGE analysis of PhK86 eypressian @37° C.

A) total protein control, pETlSb;

B-D) total protein PhK86;

E) supernatant PhK86;

F-G) inclusion body PI1K86;

H) final PhK86 inclusion body solublized in guanidine. 92

MW ABODE FGHMW Figure 2. SDS-PAGE analysis of PhK86 expression @ 22 °C.

Panel A)

A) OD 0.8 pETlSb supernatant

B) OD 0.8 pET15b-PhK86 supernatant

C) OD 0.6 pET 15b-PhK86 supernatant

MW) Gibco prestained molecular weight markers

D) OD 0.8 pETb supernatant + EMAC resin

E-F) OD 0.6 pET 15b-PhK86 supernatant + IMAC

Panel B)

A) OD 0.8 pETlSb pellet + EMAC resin

B) OD 0.6 pET15b-resin

C) OD 0.8 pET15b-PhK86 pellet + IMAC-resin

MW) Gibco prestained molecxilar weight markers 94

A.

ABCMW D E F 95

B C MW

15 kDa

Figure 2 (continued) 96

Figure 3. SDS-PAGE Analysis of Refolded PhK86. A) refolded PhK86, 14.9 kDa from the metal affinity colimm MW) Gibco prestained molecvilar weight markers Figure 4: Western analysis of calmodulin binding to PhK86.

A) PhK86 + calmodulin + calcixmi;

B) PI1K86 + calmodulin;

C) PhK86 + calmodulin -1- 0.2 mM calcium+ 2 mM EGTA;

D) PhK86 +calmodulin + 2mM EGTA;

E) PhK86

F) Calmodulin 98 Figure 5: Western analysis of truncated y 1-300 binding to PhK86.

A) IMAC-resin+PhK86;

B-C) IMAC-resin + truncated y 1-300;

D-E) IMAC-resin+PhK86 + truncated y 1-300. Anti-y antibody

was a generous gift from Dr. Gerald Carlson. Note: under the

experimental conditions used, non-specific binding of truncated

y(l-300) to the IMAC-resin was observed. 100 101

0.8 A 595 nm 0.7-

0.6-

0.5-

0.3-

0.2-

0.1-

0 2 4 6 8 10 12 14 16 18 20 Fraction number

Figvire 6. Elution profile of PhK86 from IMAC chromatography. Protein concentration in individual fractions was determined by Bradford assay described in experimental procedures. Figxire 7: SDS-PAGE Analysis of Eluted PhK86.

A) Fraction number 4

B) Fraction number 6

C) Fraction number 8

D) Fraction number 10

E) Fraction number 12

MW) Gibco prestained molecular weight markers

F) Pooled, dialyzed PhK86 sample 103

A B C D E MW F 104

35000 • control 30000- • lxPhK86 25000- • 2xPhK86 20000-

15000-

10000-

5000-

1 min. 2 min. 5 min.

Figure 8. Initial time course of truncated y(1-300) inhibition by PhK86. Ix PhK86 is equivalent to 1 |iM PhK86 and 2x PhK86 is equivalent to 2 |iM PI1K86. Substrate was phosphorylase h. Assay conditions as described in experimental procedures with n=3. 105

MUTANTS OF PHOSPHORYLASE-a ALTERED IN

RECOGNITION BY PROTEIN PHOSPHATASE !

A paper to be submitted to the Journal of Biological Chemistry

Cheryl Bartleson, Alyssa C. Biom, and Donald J. Graves

Abstract

Little information is known about the substrate specificity of protein

phosphatase-1. The success obtained studying protein kinase substrate

specificity with peptide substrates has not extended to protein phosphatases.

Protein phosphatases are believed to recognize higher order structure in

substrates in addition to the primary sequence surrovmding the

phosphoserine or threonine. Peptide studies have revealed a preference for

basic residues N-terminal to the phosphoserine. Studies suggest that

arginine 16 in phosphorylase a may be a positive determinant.

In order to develop our knowledge of specificity determinants for protein

phosphatase-1, mutants of phosphorylase b have been converted to their corresponding phosphorylase a and examined for their efficacy as substrates for protein phosphoatase-1. Mutants focused on the N-terminal primary sequence surrounding the phosphoserine. R16A, R16E, and I13G phosphorylase b mutants were generated and converted to phosphorylase a

using phosphorylase kinase. 106

Protein phosphatase-1 prefers the positive charge on arginine 16.

R16A-a exhibits a similar Km but reduced Vmax, and R16E-a has an increased

Km and a decreased Vmax, when compared to phosphorylase a. I13G-a has a

similar Km but an increased Vmax. Surprisingly, I13G-a has no phosphorylase

a activity in the absence of AMP. Thus, I13G-a is the first reported mutant of

phosphorylase that is not activated by phosphorylation. Additional resvdts

with this mutant reveals that I13G-a may actually be more phosphorylase b

like in its properties.

Introduction

Protein phosphatase-1 is one of four major serine/threonine

phosphatases identified by substrate specificity and susceptibility to specific

inhibitors (1). These phosphatases are designated as phosphatases type-1, -

2A, -2B, also known as calcineurin, and -2C. Protein phosphatase-1, PP-1, is considered a broad specificity phosphatase. However, the best known

physiological protein substrate for protein phosphatase-1, is phosphorylase a

(2). Little information is available on the specificity of protein phosphatase-

1. The wealth of knowledge of protein kinase specificity derived firom peptide studies has not extended to the protein phosphatases. This could be due to the fact that phosphopeptides are not very good substrates for many phosphatases (3). Protein phosphatases seem to be less dependent on 107

primary structure and more dependent on aspects of secondary structure for

recognition of substrate.

Some information has been gained &om phosphopeptide studies. The

preference for basic residues N-terminal to the phosphoserine was found to be a general determinant for protein phosphatase-1 (3). An additional potential

preference for a basic residue C-terminal to the phosphoserine might exist.

Peptides derived from phosphorylase a lacking the residue arginine 16 fail to function as protein phosphatase-1 substrates (4). Modification of arginine 16 in peptide substrates by nitrosylation, and amidation, and more recently through substitution to citniUine indicates that a positive charge may be important for substrate recognition (5,6). Analysis of the citruUine substituted peptides substrates indicated that arginine 16 was important for enhanced tmrnover (6). Many interactions occur in the stabilization of the N- terminal region of phosphorylase a (7). NMR studies show an interaction of arginine 16 with the phosphoserine in peptide substrate (6). The crystal structure of phosphorylase a shows the phosphoserine complexed with arginine 69 and arginine 43', an arginine present in the neighboring phosphorylase (7). All of these interactions may contribute to the overall specificity of protein phosphatase-1. 108

Limited knowledge has been gained &om mutagenesis experiments

involving protein phosphatase-1. The acidic groove, proposed to bind the N-

terminal basic residues found in many protein phosphatase-1 substrates, was extensively mutated. No major changes in the kinetic parameters for

phosphorylase a were observed. The systematic mutagenesis involved changing one amino acid at a time, and this might have led to some compensation in the mutants by the surrounding acidic residues still available (8).

Phosphorylase mutants were constructed to examine the specificity of protein phosphatase-1 using a physiologically relevant protein substrate. The selected mutants change the primary structure around the phosphorylatable site in the N-terminus of phosphorylase b. The basic residue foimd at the P+2 position of phosphorylase was changed to R16A and R16E respectively. This arginine residue is a positive specificity determinant for phosphorylase kinase, and is a potential positive determinant for protein phosphatase-1. A hydrophobic residue in position P-1 was changed to glycine, I13G. Isoleucine is involved in the formation of the 3io helical segment found in phosphorylase a. Incorporation of glycine into peptide substrates of phosphorylase kinase had a large detrimental effect on phosphorylation (9). Substituting glycine may disrupt the helical structure and make this region more conformationally 109 flexible. All mutant phosphorylase a's were compared to wild-type-a with respect to substrate recognition by protein phosphatase-1.

Experimental procedures

Materials. All chemicals were commercially available from Boehringer

Mannheim (Indianapolis, IN), Fisher (Pittsburgh, PA), or Sigma (St. Louis,

MO). The trxmcated cataljrtic y subunit of phosphorylase kinase was renatured as described (10). Recombinant protein phosphatase-1, (PP-1), was purchased from New England Biolabs (Beverly, MA). Alkedine phosphatase,

(AP), was available from Sigma. Calcineurin, (CaN), was a generous gift of

Dr. Bruce L. Martin. Protein phosphatase-2A, (PP2A), was purchased from

Upstate Biotechnology (Lake Placid, NY). [y-32P]ATP was from ICN

Biomedicals (Costa Mesa, CA).

Preparation of Wild-type and Mutant Phosphorylase b. Wild-type phosphorylase was purified as described (11). Oligonucleotides used in site- directed mutagenesis were synthesized in the Iowa State University DNA

Facility. Phosphorylase mutant R16A was generated by single-stranded mutagenesis using the Kxinkel method (12). Phosphorylase mutants R16E, and I13G was generated by double stranded mutagenesis using the Qmck 110

Change Site-Directed Mutagenesis Kit from Stratagene (LaJoUa, CA).

Recombinant phosphorylases were expressed and p\iri&ed as described (13).

Preparation of Wild-type and Mutant Phosphorylase a. Wild-t3rpe and

the individual mutant phosphorylase b's were converted to phosphorylase a as

described by Krebs and Fischer, with some modification (14). Tnincated 7(1-

300) was used with 1 mM [y-^zpjATP, 10 mM MgCla, 5 mM dithiothreitol, 50 mM Pipes, and 50 mM Tris, pH 8.2. At a specified time, sample aliquots were removed firom the reactions and spotted on ET-31 Whatman filter paper.

Samples were washed and counted to determine the amoimt of radioactivity incorporated into phosphorylase a (15). Phosphorylase a activity was measured in the presence and absence of AMP as described (16). 32p. phosphorylase a's were dialyzed into 50 mM imidazole, pH7.2, 5 mM dithiothreitol and stored at A°C.

Protein Concentration Determinations. Protein concentrations of wild- type and mutant phosphorylases were determined spectrophotometrically at

280 nm (17).

Circular dichroism. The circular dichroism spectra of wild-type phosphorylase b and a, and I13G b and a was determined in the Iowa State

University Protein Facility using a Jasco J-710 spectropolarimeter (Jasco Ill

Corporation, Tokyo, Japan). Samples were diluted in 10 mM phosphate buffer, pH 6.8. Phosphorylase was diluted to 0.1 mg/ml for the analysis.

The wavelength scan was rxm from 200 to 260 nm utilizing a 0.1 cm cell.

Analytical ultracentrifugation. Equihbrium velocity centrifugation of wild-tjrpe phosphorylase b and a, and I13G b and a was performed in the Iowa

State University Protein Facility using a model XL-A Maxima Optima

Beckman centrifuge. Data was analyzed with Beckman XL-A software.

Samples were in 40 mM P-glycerol phosphate, pH 6.8, 100 mM KCl.

Assays of protein phosphatase activities. PP-1 assays were performed as previously described (18). The final assay mixture of 0.1 ml, contained 0.2 mg/ml p2-P]phosphorylase a, 50 mM imidazole, pH 7.2, 2mM dithiothreitol,

0.5 mg/ml bovine serum albumin, 5 mM caffeine, 0.2 mM MnCla. Alkaline phosphatase assays were performed at 30°C. The 0.1 ml reactions contained

25 mM Tris, pH 8.0, ImM MgCh, 0.2 mg/ml [32-P]phosphorylase a, 2mM dithiothreitol, 0.5 mg/ml bovine serxmi albimiin, 5 mM caffeine. The reaction conditions used for protein phosphatase-2A were the same as for protein phosphatase-1. Assays with calcineurin contained 25 mM Mops, pH 7.0, 1 mM MnCl2, 10 jig/ml calmodulin, 0.1 mM CaCb, 2 mM dithiothreitol, 0.5 mg/ml bovine senun albumin, 5 mM caffeine, and 0.2 mg/ml

[32p]phosphorylase a. Reactions were started with the addition of the 112

specified protein phosphatase. Samples were incubated at 30®C for 5 to 20

minutes. Reactions were terminated by the addition of 10 |il of 100%

trichloroacetic acid. Samples were microcentrifuged for 10 minutes to

precipitate protein. 50 |il of supernatant was spotted into 10 ml of

scintillation cocktail and coimted to determine the pmol P^P] released.

Kinetic data were analyzed with Enzyme Kinetics (Trinity Software).

Phosphorylase-a concentrations varied from 0.4 up to 18 jiM depending

on the mutant phosphorylase a utilized.

Results

Phosphorylase a Activities in the presence and absence of AMP. Wild-type

phosphorylase a and R16A-a, R16E-a, and I13G-a activities were assayed in the presence and absence of AMP. After phosphorylation, the activity of phosphorylase a in the absence of AMP should be equivalent to the activity in the presence of AMP. Do the mutant phosphorylase a's exhibit similar activities compared to wild-type-a? As shown in Figure 1, wild-ts^pe-a, R16A- a, and R16E-a all exhibit similar activities +/- AMP. Interestingly, I13G-a has no detectable activity in the absence of AMP.

CD spectra of wild-type-a compared to I13G-a. The lack of activity observed in

I13G-a after phosphorylation, led us to examine the secondary structure of this phosphorylase compared to wild-type-a. The N-terminal 113

phosphorylatable region in phosphorylase ^ residues 10-18,appears to be in a

random conformation (7). Upon phosphorylation, the N-terminus of

phosphorylase a moves, rotates up, and interacts at the subunit interface,

with a 3io helix forming aroxmd the phosphorylatable serine (7). Figure 2,

panel A illustrates the CD spectra of wild-type-b, compared with I13G-b. The

secondary structures of these two proteins are almost identical. The

secondary structures of wild-type-a and I13G-a are compared in Figure 2,

panel B, respectively. The secondary structures of the phosphorylated wild-

type-a and I13G-a are again very similar, indicating that a change in secondary structure has occurred in I13G-a upon phosphorylation. However,

this secondary change did not result in activation of phosphorylase.

Analytical ultracentrifugation. The quaternary structures of wild-type- b, and-a were compared to I13G-b and -a using equilibrium velocity centrifugation. Wild-type-b exists as a dimer and wild-type-a forms a tetramer. The results are illustrated in Figure 3, panels A and B. Figure 4, panel A shows that I13G-b exists as a dimer. Panel B of Figure 4 demonstrates that I13G-a does not form a tetramer, but remains a dimer.

The calculated molecvdar weight for I13G-b was 215,430 Da, and I13G-a was

197,050 Da. 114

AMP activation curves of wild-type-a and I13G-a. Phosphoiylase b can

be activated allosterically by AMP. Phosphorylase b cooperatively binds

AMP, while phosphorylase a does not (19). H3G-a has no activity after

phosphorylation. Thus, it is possible that I13G-a might be similar to wild-

type-b. The AMP binding cvirves of wild-type-a and I13G-a are shown in figure

5. Figure 6 shows I13G-b and I13G-a. The AMP binding of I13G-a appears to

be more like wild-type-b.

Effect of Glucose-1-phosphate, GlP, on the activity ofI13G-a. The classic

model of cooperativity is based on the existence of multiple conformational

states, ranging from a T state, low affinity, to an R state, with high affinity for

substrate. Wild-type-b has a Km of 50 mM for glucose-1-phosphate, whereas

wild-type-a has a Km of 3.8 mM (20). Figure 7 illustrates the activities of

wild-type-b and I13G-b compared to wild-type-a and I13G-a in the presence of increasing amounts of glucose-1-phosphate, in the absence of AMP. Both wild-type-b and I13G-b exhibit increased activities with increased glucose-1- phosphate. No activity was detected in any of the 113G-a samples. I13G-a appears to more T state than R state.

Dephosphorylation of wild-type and mutant phosphorylase a's with PP-1.

Dephosphorylation of wild-type-a was compared to dephosphorylation of the mutants R16A-^ R16E-a, and I13G-a using protein phosphatase-1. The 115

residts of the time courses can be seen in Figure 8. Protein phosphatase-1

dephosphorylated all of the phosphorylase a's to varying extents. R16E-a was

dephosphorylated at the lowest rate. R16A-a was dephosphorylated at a rate approximately one-half that of wild-type phosphorylase a, while I13G-a was dephosphorylated at a faster initial rate.

Comparison of dephosphorylation of wild-type-a and mutant phosphorylases with different protein phosphatases. Figure 9 illustrates the differences in dephosphorylation for wild-type-a, R16A-a, R16E-a, and I13G-a with protein phosphatase-1, protein phosphatase 2-A, alkaline phosphatase, and calcineurin. Consistent with the time course of dephosphorylation presented in Figure 8, protein phosphatase-1 dephosphorylated all of the phosphorylase a's to different extents. Rl6E-a was the least preferred substrate for protein phosphatase-1. R16A-a was dephosphorylated at a rate about 50% compared to wild-type-a. I13G-a was actually a better substrate than wild-type-a. Protein phosphatse -2A dephosphorylatd wild-type-a and

R16A-a equally, and preferred R16E-a the least. I13G-a was again preferred.

Although I13G-a was much preferred by protein phosphatase 2-A, compared to wild-type-a, 63% versus 21%. Calcineurin followed the same substrate preference pattern of protein phosphatase 2-A, albeit to a lesser extent for

I13G-a. No significant dephosphorylation of wild-type-a or any mutant, 116

R16A-a, Rl6E-a or I13G-a, was detected using alkaline phosphatase. More

units of the other protein phosphatases, protein phosphatase-2A , calcineurin,

and alkaline phosphatase were used than units of protein phosphatase-l.

Kinetic analyses of wild-type-a and mutant-q^s with PP-1. Table I

contains a summary of the kinetic analyses for dephosphorylation of wild-

type-a compared with R16A-a, R16E-a, and I13G-a using protein

phosphatase-1. All experiments were run with using equivalent phosphate to

determine the concentration of phosphorylase a mutants that were not 100%

phosphorylated. The Km for I13G-a was similar to the Km for wild-type-a,

although the Vmax was 2 fold greater for I13G-a than that for wild-type-a.

R16E-^exhibited approximately a 3 fold increase in Km, and a 6.5 fold

decrease in Vmax. R16A-a had a similar Km to wild-type a, but had a 3.3 fold lower Vmax.

Discussion

We were interested in addressing the structural basis for recognition of phosphorylase a by specific phosphatases, protein phosphatase-1 in particular. What makes the dephosphorylation reaction much more efiective with protein substrate compared to phosphopeptide substrate? Why do some phosphatase, alkaline phosphatase for example, dephosphorylate phospho-9- 117

18 peptide, but do not dephosphorylate phosphorylase a? Mutants of

phosphorylase a were made to address this question of specificity.

The phosphorylase a mutants, R16A-a and R16E-a were active in the

presence of AMP, as well as in the absence of AMP after phosphorylation.

Both mutants exhibited activities very similar to wild-tjrpe-a. Although I13G-

a shows activity similar to wild-type-a in the presence of AMP, we were

unable to detect any phosphorylase a activity in the absence of AMP after

phosphorylation.

The lack of I13G-a activity after phosphorylation led to some further

characterization of I13G-a compared to wild-t3rpe-b £md wild-type-a. In the

crystal structure of phosphorylase b, the N-terminus, containing the

phosphorylatable serine, forms a random structure (7). After

phosphorylation, the structure of the N-terminus changes to a 3io helix, and

rotates up to the subunit interface (7). Although we can not detect such

conformational changes, the absence of activity in I13G-a suggests that a

change in conformation has occurred. The CD spectra indicate that the overall

secondary structure of I13G-b and I13G-a do not differ significantly from wild-

type-b and wild-tj^pe-a. The conformational change occurring upon phosphorylation of I13G-a, is not the same change that occurs in wild-type-a.

This is evident from the lack of I13G-a activity after phosphorylation. 118

The quaternary structure of phosphorylase b is a dimer. Wild-type-a forms a tetramer. The molecular weight of the phosphorylase monomer is

97,500 Da, the dimer is 195,000 Da, and the tetramer is 390,000 Da. I13G-b has an average molecular weight of 215,430 Da, and I13G-a has an average molecular weight of 197,050 Da. The results &om equilibrium velocity centrifiigation indicate that I13G-a is a dimer. From this data it appears that I13G-a is more like wild-type-b.

Additional support for I13G-a being similar to wild-type-b comes from the AMP activation curves. The apparent cooperativity that I13G-a exhibits is consistent with I13G-a behaving more like wild-tjrpe-b. The T state, low affinity, of phosphorylase b has a high Km for glucose-1-phosphate, while the R state, high affinity, of phosphorylase a has a much reduced Km for glucose-1- phosphate. When the activities of wild-type-b, and wild-type-a, are compared to the activities of I13G-b and I13G-a in the presence of increasing amounts of glucose-1-phosphate, but lacking AMP, I13G-a exhibits no activity. This resxilt is consistent with I13G-a being similar to the T state, like phosphorylase b.

The similarity in the Km for I13G-a, compared to wild-type-a indicates that protein phosphatase-1 recognizes similar features common to both of these conformations of phosphorylase a. The increased Vmax might translate 119

to a more accessible phosphoserine residue. Increased accessibility of the

phosphoserine residue could explain why I13G-a was dephosphorylated to a

greater extent by protein phosphatase-2A and calcineurin. Alternatively, the

substitution of isoleucine with glycine into the N-terminus of phosphorylase b

could disrupt the ability of the N-terminus to form a 3io helix and dock at the subxmit interface. This potential alternative conformation may be more favorable for dephosphorylation.

Substitution of Rl6 with citruUine in peptide substrate derivatives of

9-18, a peptide of the phosphorylatable region of phosphorylase b, resulted in decreased dephosphorylation of this peptide by protein phosphatase-1 (6). It was suggested that R16 either enhances substrate turnover for protein phosphatase-1, or alternatively, that protein phosphatase-1 recognizes the complex formed from the interaction of arginine with the target phosphoserine. This interaction could be with arginine 16. Alternatively, this interaction might occur with residues arginine 69 and arginine 43' in phosphoiylase a.

Our results confirm the importance of arginine 16 for phosphorylase a recognition by protein phosphatase-1. Both of the R16 phosphorylase-a mutants, R16A-a and R16E-a were dephosphorylated by protein phosphatase-1 at a lesser rate. The Km and Vma* for R16E change 120

significantly. Rl6A-a has a similar Km compared to wild-type-a, but the Vmax

decreases. It appears that protein phosphatase-1 does prefer the positive

charge of arginine 16 for optimal turnover, but it will dephosphorylate the

R16A-a and R16E-a mutants. Alkaline phosphatase was able to

dephosphorylate the citrulline substituted phosphopeptides of phosphorylase

a (6), yet we were unable to detect any significant dephosphorylation of

phosphorylase a or any of the mutants with alkaline phosphatase. These

results again emphasize protein phosphatase specificity for higher order structure in addition to the primary structiire. Further work involving

mutagenesis of phosphorylase a, as weU as other physiologically relevant protein phosphatase substrates is warranted. The combination of available crystal structures of phosphorylase a with the crystal structure of protein phosphatase-1 should help guide the planning of future experiments.

Additional experiments with phosphoprotein substrates of protein phosphatases will help broaden our understanding of protein phosphatase substrate specificity.

References

1. Ingebritsen, T. S., and Cohen, P. (1983) Science 221, 331-338.

2. BoUen, M., and Stalmans, W. (1992) Critical Reviews in Biochemistry

and Molecular Biology 27(3), 227-281. 121

3. Pinna, L. A., and Donella-Deana, A. (1994) Biochimica et Biophysica

Acta 1222, 415-431.

4. Graves, D. J., Fischer, E. H., and Krebs, E. G. (1960) JSC 235, 805-809.

5. Graves, D. J., Martensen, T. M., Tu, J.-I., and Tessmer, G. M. (1973) in

Third International Symposium held in Seattle June 5-8, pp. 53-61,

Springer-Verlag.

6. Martin, B. L., Luo, S., Kintanar, A., Chen, M., and Graves, D. J. (1998)

Archives of Biochemistry and Biophysics 359, In Press.

7. Barford, D., Hu, S.-H., and Johnson, L. N. (1991) JMB 218, 233-60.

8. Zhang, L., and Lee, E. Y. C. (1997) Biochemistry 36, 8209-8214.

9. Graves, D. J. (1983) Methods in Enzymology, Academic Press, New

York.

10. Huang, C.-Y. F., Yuan, C.-J., Livanova, N. B., and Graves, D. J. (1993)

Molecular and Cellular Biochemistry 127/28, 7-18.

11. Krebs, E. G., Kent, A. B., Graves, D. J., and Fischer, E. H. (1958)

Proceedings of the International Symposium on Enzyme Chemistry,

Tokya and Kyoto 19571, 41-43.

12. Kunkel, T. A. (1985) PNAS 82, 488-92.

13. Browner, M. F., Rasor, P., Tugendereich, S., and Fletterick, R. J. (1991)

Protein Engineering 4, 351-57. 122

14. Krebs, E. G., and Fischer, E. H. (1962) Methods in Emymo. 5, 373-376.

15. Reimann, E. M., Walsh, D., and Krebs, E. G. (1971) JBC 246, 1986-95.

16. niingworth, B., and Cori, C. F. (1953) Biochem. Prep. 3, 1-9.

17. Kastenschmidt, L. L., Kastenschmidt, J., and Helmreich, E. (1968)

Biochemistry 17, 3590-3608

18. KiUHea, S. D., Aylward, J. H., Mellgren, R. L., and Lee, E. Y. C. (1978)

Arch. Biochem. Biophysics 191, 638-646.

19. Johnson, L. N., and Barford, D. (1990) JBC 265(5), 2409-2412.

20. Uhing, R. J., Lentz, S. R., and Graves, D. J. (1981) Biochemistry 20,

2537-2544. 123

AMP

-AMP

R16A R16E I13G

Figure 1. Phosphorylaseji activites in the presence and absence of AMP. Assay conditions are described under experimental procedures. The concentration of AMP is ImM. Figure 2. Circular dichroism spectra of wild-type phosphorylase and I13G phosphorylase. Panel A: Spectra of wild-type-b compared to I13G-b. Panel B: Spectra of wild-type-a compared to H3G-a. 125

Panel A

40

20

0 -

-20 190 200 220 240 26C VVaveiengthtnml 0.1 mg/ml phos b " 1 mg/ml I13G b 126

Panel B

40

20

CD

-20 ISO 200 220 240 260 Waveienath[nml

0.1 mg/ml phos a 0.1 mg/ml I13G a

Figure 2 continued Figure 3. Analytical ultracentrifugation results for wild-type-^ and wild- type-q.. Experimental procedures are decribed in the methods. Panel A is wild-type-b. Panel B is wild-type-a. 128

Panel A

0.04 <0 cP T5 0.02- 0.00- €to— 'co ^ o o -0.02-1 o 0^

0.6 -

6.0 Radius (cm) 129

Panel B

0) 0.06^

S 0.024 •o o.ooq W -0.02H 2 -0.044

1.0

0.8 -

8 0 6

CO

O 0.4 ^ (0 < 0.2 -

0.0 -

Radius (cm)

Figure 3 continued Figure 4. Analytical ultracentrifugatioTt results for I13G-b and IlSG-q.. Experimental procedures are decribed in the methods. Panel A is I13G-b. Panel B is I13G-a. Absorbance Residuals I poop pop b io O) 132

Panel B

«2 0.02- oo

0.00- od 0) -0.02-

0.8 -

0.6 -

n O 0.4- tf)

0.2 -

0.0 -T T1 1 T1 r- 6.4 6.5 6.6 Radius (cm)

Figure 4 continued Figure 5. AMP activation curves of u)ild-type-b and wild-type-a.

Assay conditions described in Experimental procedures.

Wild-type-b and wild-type-a in the presence of increasing

amounts of AMP. 0.35 phos. b

^ phos. a 0.3-

0.25-

0.2- o CO 03 0.15- 1^

0.1-

0.05-

0-)—f~~i—I—I—I—I—I—I—r T 1 1 1 1 1 1 1 1 - ' ' I ' 0 50 100 150 250 300 350

lAMP], iiM Figure 6. AMP activation curves of I13G-b and I13G-q.

Assay conditions described in Experimental procedures.

I13G-b and I13G-a in the presence of increasing

amounts of AMP. 0.5 I13G-b

0.45: I13G-a

0.4:

0.35:

0.3-

§ 0.25- C£) < 0.2H 0)CO

0.15-

0.1-

0.05-

0 I I I 1 I I I I ' I I I I r I "I" I' I I I ' ' 0 50 100 150 300 350

[AMP], nM Figure 7. Effect of glucose-l-phosphate on wild-type and I13G

phosphorylases.

Panel A shows wild-type-b and I13G-b in the presence of

increasing glucose-l-phosphate concentrations. Panel B

shows wild-type-a and I13G-a in the presence of increasing

glucose-l-phosphate concentrations. All assay were in the

absence of AMP. Assay described in experimental

procedures. A 660 nm o o o o o o b b CO o> I , , . , I , j_i

wt-b, 16 mM GlP

wt-b, 50 mM GlP

wt-b, 100 mM GlP

wt-b, 200 mM GlP CO 00 I13G-b, 16mMGlP

I13G-b, 50 mM GlP

I13G-b, 100 mM GlP

I13G-b, 200 mMGlP 139

A 660 nm

0-18-

0.16- 0-14- 6 0.12-

< 0.08-

0.06- 0.04- 0.02-

Figure 7 (continued) 140

140 • wt 120- • R16A • R16E 100- • I13G 80-

« 60-

Q. 40-

20-

0123456789 time, minutes

Figxire 8. Dephosphorylation of wild-type-Qj R16A-Q, R16E-Si ond I13G-Q, with protein phosphatase-1. Assay conditions are described in experimental procedures. 141

160 n wt 140- • R16A 120- • • R16E • I13G

PP-1 PP-2A AP CaN

Figure 9. Comparison of dephosphorylation of wild-type and mutant phosphorylases with different protein phosphatases. Assay conditions used for each protein phosphatase are described in experimental procedures. 142

Table I Kinetic Constants for Phosphorylase a Mutants with Protein

Phosphatase-1

Km (jiM) Vmax (nmol P/min/mg)

wild-type-a 0.94 + 0.05 5124 + 600

R16A-a 1.46 + 0.3 1533 + 200

R16E-a 3.4 + 0.3 778 + 150

I13G-a 1.0 + 0.14 13,208 + 1,500 143

GENERAL CONCLUSIONS

Many crucial cellular processes are regulated by conformational

changes that result &om reversible phosphorylation of specific serine,

threonine, or tyrosine residues. Control by phosphorylation is achieved by

protein kinases and protein phosphatases that catalyze the opposing

phosphorylation and dephosphorylation reactions, respectively. Protein

phosphorylation and dephosphorylation were discovered in 1955 with glycogen

phosphorylase (112,113).

Phosphorylase kinase is a key enzyme involved in control of glycogen metaboUsm. Phosphorylase kinase catalyzes the phosphorylation of glycogen phosphorylase converting phosphorylase fi*om its inactive b form, to the active a form. Glycogen phosphorylase is the only recognized substrate of phosphorylase kinase in vivo. One aim of this research was to understand the specificity of phosphorylase kinase for phosphorylase. Peptide studies have jdelded a wealth of information about the specificity determinants for many protein kinases. Peptides of the phosphorylatable region of phosphorylase b are effective substrates of phosphorylsise kinase (67). It was previously estabhshed that phosphorylase kinase prefers basic residues N-terminal to the phosphorylatable serine residue 14 (67-69). It has been suggested that a basic residue C-terminal to the phosphorylatable serine might be an 144

additional specificity determinant for phosphorylase kinase recognition of

phosphorylase. Peptide studies utilizing arginine to citruUine substitutions

for the C-terminal arginine, R16Cit, and for both the C-terminal and N-

terminal arginines, R16Cit and RlOCit, were synthesized. Phosphorylation

of the citndline peptides was poor. These resxilts suggest that arginine 16 is

important for the recognition and effective phosphorylation of phosphorylase

by phosphorylase kinase. The fact that citruUine is similar in size to arginine

implies that charge is the critical requirement at position 16. Work in

progress in the lab on arginine 16 mutants of phosphorylase will help expand

our imderstanding of phosphorylase kinase recognition of phosphorylase.

Phosphorylase kinase is a member of the calcium-calmodulin regulated

protein kinase family. The catal3rtic activity of these protein kinases appears

to be regulated intrasterically by autoinhibitory domains in their C-termini

(76). The C-terminus of phosphorylase kinase, PhK86, contains two non­

contiguous autoinhibitory regions, PhKl3 and PhK5. Peptides of these

regions are competitive inhibitors of phosphorylase kinase (70,114). This suggested that PhK13 could bind to the active site of phosphorylase kinase and might function as a pseudosubstrate inhibitor. In this work, variants of

PhK13 were prepared to localize the determinants for interaction with phosphorylase kinase. All of the residues necessary for inhibition of 145

phosphorylase kinase are located in the N-terminal half of PhK13, peptide

PhK13-l. Substitution of residue cysteine 308 with serine indicates that this

cysteine residue is important for the inhibition of phosphorylase kinase.

PhKl3-l is a potent competitive inhibitor of phosphorylase kinase with respect to phosphorylase b phosphorylation and ATPase activity. Although

PhK13-l does not appear to function as a pseudosubstrate, several specificity determinants employed in the recognition of phosphorylase b as substrate seem to be utilized in the recognition of PhKl3-l as an inhibitor.

The crystal structure of a truncated form, residues 1-298, of the catalytic sub unit of phosphorylase kinase has been solved (57). No structural information is available on the C-terminus, PhK86, residues 297-386, of the catalytic subunit of phosphorylase kinase. Expression of PhK86 as a histidine-tagged fusion protein in E. coli resulted in formation of insoluble inclusion bodies. A piudfication and renatxiration protocol was developed that will lead to successful production of amounts of PhK86 suitable for future structure/function analyses.

Another key enzyme involve in control of glycogen metabolism is protein phosphatase-1. Protein phosphatase-1 catalyzes the dephosphorylation of phosphorylase a, active, to yield phosphorylase b, inactive. Phosphorylase a is the best and most commonly used substrate for 146

protein phosphatase-1 (94). Much less is known about the specificity of

protein phosphatase-1. The success obtained studjdng protein kinase

substrate specificity with peptide substrates has not extended to protein

phosphatases. Protein phosphatases seem to recognize higher order structure

in substrates in addition to the primary sequence surrounding the

phosphoserine/threonine residue. Basic residues N-terminal to the

phosphoserine seem to be a general determinant (106). Other studies

suggested that arginine 16 might be important for protein phosphatase-1

recognition of phosphorylase a (104,105,107). Mutations focused around the

phosphoserine residue of phosphorylase were studied with protein

phosphatase-1. Recognition of the mutant, I13G was altered. No detectable

activity was observed in this mutant after phosphorylation. Characterization of I13G resulted in the conclusion that this mutant may be more phosphorylase b like in its properties. Results with the arginine 16 mutants of phosphorylase support a preference for basic residue C-terminal to the phosphoserine for effective recognition and dephosphorylation of phosphorylase a by protein phosphatase-1. Additional mutants of phosphorylase a will help broaden our understanding of protein phosphatase-

1 specificity. 147

REFERENCES CITED

1. Krebs, E. G., and Fischer, E. H. (1956) BBA 20,150-157.

2. Krebs, E. G., Kent, A. B., Graves, D. J., and Fischer, E. H. (1958) Proceedings of the International Symposium on Enzyme Chemistry, Tokyo and Kyoto 19571, 41-43.

3. Nanun, D. H. (1971) J. Pharmacol Exp Ther 178, 299.

4. Daegelen-Proux, D., Pierres, M., Alexandre, Y., and Dreyfus, J.-C. (1976) BBA 452, 398.

5. Vandenheede, J. R., Keppens, S., and DeWulf, H. (1976) Febs Letters (61), 213.

6. Drummond, G. I., and Bellward, G. (1970) J. Neurochem 17, 475.

7. Sutherland, E. W. (1950) Recent Progttorm Res 5,441-463.

8. Hayakawa, T., Perkins, J.P., Walsh, DA., and Krebs, E.G. (1973) Biochemistry 12, 567-573.

9. Ozawa, E., Hosoi, K-, and Ebashi, S. (1967) JBC (Tokyo) 61, 531-533.

10. Brostrom, C. O., Hunkeler, F. L., and Krebs, E. G. (1971) JBC 246, 1961-1967.

11. Meyer, W. L., Fischer, E. H., and Krebs, E. G. (1964) Biochemistry 3, 1033-1039.

12. StuU, J. T., and Mayer, S. E. (1971) JBC 246, 5716-5723.

13. Cohen, P. (1973) Eur. J. Biochem. 34, 1-14. 148

14. Zander, N. F., Meyer, H. E., Hoffmann-Posorske, E., Crabb, J. W., L. M. G. Heilmeyer, J., and Eilimann, M. W. (1988) PNAS 85, 2929-2933.

15. Kilimann, M. W., Zander, N. F., Kuhn, C. C., Crabb, J. W., Meyer, H. E., and Jr, L. G. K (1988) PNAS 85, 9381-85.

16. Reimann, E. M., Titani, K., Ericsson, L. H., Wade, R. D., Fischer, E. H., and Walsh, K. A. (1984) Biochemistry 23, 4185-92.

17. Grand, R. J., Shenolikar, S., and Cohen, P. T. (1981) Eur. J. Biochem. 113, 359-67.

18. Biurchell, A., Cohen, P. T. W., and Cohen, P. (1976) Febs Letters 67,17- 22.

19. Jennissen, H. P., and L. M. G. Heilmeyer, J. (1974) Febs Letters 42, 77- 80.

20. Skuster, J. R., Chan, K. J. J., and Graves, D. J. (1980) JBC 255(5), 2203-2210.

21. Chan, K F. J., and Graves, D. J. (1982) JBC 257, 5539-5547.

22. Kee, S. M., and Graves, D. J. (1986) JBC 261, 4732-37.

23. HeOmeyer, L. M. G., Jr, Groschel-Stewart, U., Jahnke, U., Kilmann, M. W., Kohse, K. P., and Varsanyi, M. (1980) Adv. Enzyme. Regul. 18,121- 144.

24. Kohse, K. P., and L. M. G. Heilmeyer, J. (1981) Eur. J. Biochem. 117, 507-513.

25. Yuan, C.-J., and Graves, D. J. (1989) Archives of Biochemistry and Biophysics 274, 317-26.

26. Heilmeyer, L. M. G., Jr, Serwe, M., Weber, C., Metzger, J., Ho£&nann- Posorske, E., and Meyer, H. E. (1992) PNAS 89, 9554-58.

27. Krebs, E. G., Love, D. S., Bratvold, G. E., Trayser, K A., and Fischer, E. H. (1964) Biochemistry 3, 1022-1033. 149

28. Walsh, D. A., Perkins, J. P., Brostrom, C. O., Ho, E. S., and Krebs, E. G. (1971) JBC 246, 1968-76.

29. Ramachandran, C., Goris, J., Waelkens, E., Merlevede, W., and Walsh, D. A. (1987) JBC 262, 3210-3218.

30. Cohen, P. (1980) Eur. J. Biochem. 11, 563-74.

31. Ganapathi, M. K., and Lee, E. Y. C. (1984) Arch Biochem. Biophys. 233, 19-31.

32. Carlson, G. M., Bechtel, P. J., and Graves, D. J. (1979) Adv. Enzymol. Relat. Areas Mol Bio. 50, 41.

33. Singh, T. J., and Wang, J. H. (1977) JBC 252, 625-632.

34. Wang, J. H., StuU, J. T., Huang, T.-S., and Krebs, E. G. (1976) JBC 251, 4521-4527.

35. Carlson, G. M., and Graves, D. J. (1976) JBC 251(23), 7480-7486.

36. Fitzgerald, T. J., Trempe, M. R., and Carlson, G. M. (1987) JBC 262, 11239-11246.

37. King, M. M., Fitzgerald, T. J., and Carlson, G. M. (1983) JBC 258, 9925-9930.

38. Hallenbeck, P. C., and Walsh, D. A. (1983) JBC 258(22), 13493-13501.

39. Picton, C., Klee, C. B., and Cohen, P. (1980) Eur. J. Biochem. Ill, 553- 561.

40. DePaoli-Roach, A. A., Roach, P. J., and Lamer, J. (1979) JBC 254, 4212-4219.

41. Shenolikar, S., Cohen, P. T. W., Cohen, P., Nairn, A. C., and Perry, S. V. (1979) Eur. J. Biochem. 100, 329-337. 150

42. Paudel, H. K., Xu, Y.-H., Jarrett, H. W., and Carlson, G. M. (1993) Biochemistry 32,11865-72.

43. Cox, O. E., and Edstrom, R. D. (1982) JBC 257,12728-12733.

44. Paudel, H. K., and Carlson, G. M. (1987) JBC 262, 11912-15.

45. Newsholme, P., Angelos, K L., and Walsh, D. A. (1992) JBC 267, 810-8.

46. Sanchez, V. E., and Carlson, G. M. (1993) JBC 268,1889-95.

47. Newsholme, P., and Walsh, D. A. (1992) Biochemisty Journal 283, 845- 8.

48. Chan, K. F. J., Hurst, M. O., and Graves, D. J. (1982) JBC 257(7), 3655- 3659.

49. Chan, K. F. J., and Graves, D. J. (1982) JBC 257, 5948-5955.

50. Kee, S. M., and Graves, D. J. (1987) JBC 262, 9448-53.

51. Hanks, S. K., Qtiinn, A. M., and Hunter, T. (1988) Science 241, 42-52.

52. Dasgupta, M., Honeycutt, T., and Blimienthal, D. K. (1989) JBC 261, 4732-37.

53. James, P., Cohen, P., and Carafoli, E. (1991) JBC 266, 7087-7091.

54. Huang, C.-Y. F., Yuan, C.-J., Livanova, N. B., and Graves, D. J. (1993) Molecular and Cellular Biochemistry 127/28, 7-18.

55. Cox, S., and Johnson, L. N. (1992) Protein Engineering 5(8), 811-819.

56. Lanciotti, R. A., and Bender, P. K. (1994) Biochem J. 299, 183-189.

57. Owen, D. J., Noble, M. E. M., Garman, E. F., Papageorgiou, A. C., and Johnson, L. N. (1995) Structure 3, 467-82.

58. Cohen, P., Burchell, A., Foulkes, J. G., Cohen, P. T. W., Vanaman, T. C., and Nairn, A. C. (1978) FEBS Letters 92(2), 287-293. 151

59. Chan, K.-F. J., and Graves, D. (1984) in Calcuim and Cell Function Vol. V, pp. 1-31.

60. Yoon, M. Y., and Cook, P. F. (1987) Biochemistry 26, 4118-4123.

61. Taylor, S. S., bubis, J., Toner-Webb, J., Saraswat, L. D., First, E. A., Beuchler, J. A., Knighton, D. R., and Sowadski, J. (1988) Faseb J 2, 2677.

62. Kim, G., and Graves, D. J. (1973) Biochemistry 12, 2090-2095.

63. Singh, T. J., Akatsnka, A., and Huang, K-P. (1982) Arch. Biochem. Biophy. 218, 360-368.

64. Pickett-Gies, C. A., and Walsh, D. A. (1986) The Enzymes, 17A.

65. Yuan, C. J., Haung, C.-Y. F., and Graves, D. J. (1993) JBC 268,17683- 86.

66. Nolan, C., Novoa, W. B., Krebs, E. G., and Fischer, E. H. (1964) Biochemistry 3, 542-551.

67. Tessmer, G. W., Skuster, J. R., Tabatabai, L. B., and Graves, D. J. (1977) JBC 252, 5666-71.

68. Tabatabai, L. B., and Graves, D. J. (1978) JBC 253, 2196-202.

69. Viriya, J., and Graves, D. J. (1979) BBRC 87, 17-24.

70. Huang, C.-Y. F., Yuan, C.-J., Blumenthal, D. K, and Graves, D. J. (1995) JBC 270, 7183-88.

71. Gibbs, C. S., and Zoller, M. J. (1991) Biochemistry 30, 5329-5334.

72. Knighton, D. R., Zheng, J., Eyck, L. F. T., Xuong, H. H., Taylor, S. S., and Sowadski, J. M. (1991b) Science 253, 414-420.

73. Knighton, D. R., Zheng, J., Eyck, L. F. T., Ashford, V. A., Xuong, N. H., Taylor, S. S., and Sowadski, J. M. (1991a) Science 253, 407-414. 152

74. Lowe, E. D., Noble, M. E. M., Skanmaki, V. T., Oikonomakos, N. G., Owen, D. J., and Johnson, L. N. (1997) EMBOJ16, 6646-52.

75. Songyang, Z., Lu, K. P., Kwon, Y. T., Tsai, L. H., Filhol, O., Cochet, C., Brickey, D. A., Soderling, T. R., Bartleson, C., Graves, D. J., DeMaggio, A. J., Hoekstra, M. F., Blenis, J., Hunter, T., and Cantley, L. C. (1996) Molecular and Cellular Biology 16(11), 6486-6493.

76. Kemp, B. E., and Pearson, R. B. (1991) BBA 1094, 67-76.

77. Soderling, T. R. (1990) JBC 265,1823-26.

78. Corbin, J. D., Sugden, P. H., West, L., Flockhart, D. A., Lincoln, T. M., and McCarthy, D. (1978) JBC 253, 3997-03.

79. Takio, K., Smith, S. B., Krebs, E. G., Walsh, K. A., and Titani, K. (1982) Proc. Natl. Acad. Sci. U.S.A. 79,2544-48.

80. Titani, K., Sasagawa, T., Ericsson, L. H., Kumar, S., Smith, S. B., Krebs, E. G., and Walsh, K. A. (1984) Biochemistry 23, 4193-99.

81. Heierhorst, J., Probst, W. C., Vilim, F. S., Buku, A., and Weiss, K. R. (1994) JBC 269, 21086-93.

82. Lei, J., Tang, X., Chambers, T. C., Pohl, J., and Benian, G. M. (1994) JBC 269, 21078-85.

83. Knighton, D. R., Pearson, R. B., Sowadski, J. M., Means, A. R., Eyck, L. F. T., Taylor, S. S., and Kemp, B. E. (1992) Science 258,130-34.

84. Hu, S.-H., Parker, M. W., Lei, J.-Y., Wilce, M. C. J., Benian, G., and Kemp, B. E. (1994) Nature 369, 581-84.

85. (joldberg, J., Huang, H.-B., Kwon, Y.-g., Greengurd, P., Nairn, A C., and Kuriyan, J. (1996) Nature 376, 745-753.

86. Krueger, J. K, Olah, G. A, Rokop, S. E., Zhi, G., StuU, J. T., and Trewhella, J. (1997) Biochemistry 36(6017-6023). 153

87. Krueger, J. K-, Zhi, G., StuU, J. T., and Trewhella, J. (1998) Biochemistry 37, 13997-04.

88. Faux, M. C., Pearson, R. B., Mitchehill, K. L, and Kemp, B. E. (1993) JBC 268, 12484-91.

89. Ishida, A., and Fujisawa, H. (1995) JBC 270, 2163-70.

90. Cori, G. T., and Green, A. A. (1943) JBC 151, 31.

91. Keller, P. J., and Cori, G. T. (1953) Biochim. Biophys. Acta. 12, 253.

92. Krebs, E. G., Graves, D. J., and Fischer, E. H. (1959) JBC 234, 2867.

93. Ingebritsen, T. S., and Cohen, P. (1983) Science 221, 331-338.

94. BoUen, M., and Stalmans, W. (1992) Critical Reviews in Biochemistry and Molecular Biology 27(3), 227-281.

95. Barford, D. (1996) TIBS 21.

96. Barton, G. J. (1993) Protein Engineering 6, Zl-AQ.

97. Egloff, M.-P., Cohen, P. T. W., Reinemer, P., and Barford, D. (1995) JMB 254, 942-959.

98. Kissinger, C. R., Parge, H. E., Knightong, D. R., Lewis, C. T., Pelletier, L. A., Fempczyk, A., Kalish, V. J., Tucker, D., Showalter, R. E., Mooman, E. W., Gastinel, L. N., Habuka, N., Chen, X., Maldonado, F., Barker, J. E., Bacquet, R., and VUlafrance, J. E. (1995) Nature 378, 641-644.

99. Griffith, J. P., Kim, J. L., Kim, E. E., Sintchak, M. D., Thomson, J. A., Fitzgibbon, M. J., Fleming, M. A., Caron, P. R., Hsiao, K., and Navia, M. A. (1995) Cell 82, 507-522.

100. Hubbard, M. J., and Cohen, P. (1993) TIBS 18, 172-177.

101. Zhang, Z. J., Bai, G., Deans-Zirattu, S., Browner, M. F., and Lee, E. Y. C. (1992) JBC 267, 1484-1490. 154

102. Martin, B. L., and Graves, D. J. (1994) Biochim. Biophys. Acta. 1206, 136-142.

103. McNaU, S. J., and Fischer, E. H. (1988) JBC 263, 1893-1897.

104. Graves, D. J., Fischer, E. H., and Krebs, E. G. (1960) JBC 235,805-809.

105. Graves, D. J., Martensen, T. M., Tu, J.-L, and Tessmer, G. M. (1973) in Third International Symposium held in Seattle June 5-8, pp. 53-61, Springer-Verlag.

106. Pinna, L. A., and Donella-Deana, A. (1994) Biochimica et Biophysica Acta 1222, 415-431.

107. Martin, B. L., Luo, S., Eintanar, A., Chen, M., and Graves, D. J. (1998) Archives of Biochemistry and Biophysics 359, In Press.

108. Barford, D., Hu, S.-H., and Johnson, L. N. (1991) JMB 218, 233-60.

109. Zhang, L., and Lee, E. Y. C. (1997) Biochemistry 36, 8209-8214.

110. Huang, H.-B., Horiuchi, A., Goldberg, J., Greengard, P., and Nairn, A. C. (1997) Proceedings of the National Academy of Science USA 94, 3530- 3535.

111. Zhang, J., Zhang, Z. J., Brew, K., and Lee, E. Y. C. (1996) Biochemistry 35, 6276-6282.

112. Fischer, E. H., and Krebs, E. G. (1955) JBC 216, 121-132.

113. Sutherland, E. W., and Wosilait, W. D. (1955) Nature 175, 169-170.

114. Dasgupta, M., and Blumenthal, D. K. (1995) JBC 270, 22283-89. 155

ACKNOWLEDGMENTS

I would like to express my deepest gratitude to my major professor, Dr.

Donald Graves, for his encouragement and gmdance throughout the course of my graduate career. I look forward to oxir future interactions. Additional thanks goes to Dr. Bruce L. Martin for his wonderful conversations, and to

Dr. Louise N. Johnson for her insights into crystallography. A big thanks to my labmates in the Graves' lab, with special thanks to Alyssa C. Biorn for her assistance and friendship. I thank my husband, John Bartleson for his loving support and patience. Thank you to my family for teaching me to beheve that I could accomphsh anything I set out to do. IMAGE EVALUATION TEST TARGET (QA-3)

150mm

IIVHGE, Inc 1653 East Main Street Rochester, NY 14609 USA Ptione: 716/482-0300 Fax: 716/288-5989