Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

The activities of both glycogen phosphorylase and glyco-

gen synthase are controlled by bicyclic cascades. Let us now

examine the enzymatic interconversions involved in these bi-

cyclic cascades. We shall specifically focus on the covalent

modifications of glycogen phosphorylase and glycogen syn-

thase, the structural effects of these covalent modifications,

and how these structural changes affect the interactions of

their allosteric effectors.We shall then consider the cyclic cas-

cades as a whole, studying the various modification enzymes

involved and their “ultimate” allosteric effectors. Finally, we

shall see how the various cyclic cascades of glycogen metab-

olism function in different physiological situations.

C. Glycogen Phosphorylase Bicyclic Cascade

In 1938, Carl and Gerty Cori found that glycogen phospho-

rylase exists in two forms, the b form that requires AMP for

activity, and the a form that is active without AMP. It nev-

ertheless took 20 years for the development of the protein

chemistry techniques through which Edwin Krebs and

Edmond Fischer demonstrated, in 1959, that phosphory-

lases a and b correspond to forms of the protein in which a

specific residue, Ser 14, is enzymatically phosphorylated or

dephosphorylated, respectively.

a. Glycogen Phosphorylase: The Cascade’s

Target Enzyme

The activity of glycogen phosphorylase is allosterically

controlled, as we saw, through AMP activation and ATP,

G6P, and glucose inhibition (Section 18-3A). Superim-

posed on this allosteric control is control by enzymatic in-

terconversion through a bicyclic cascade involving the ac-

tions of three enzymes (Figs. 18-13 and 18-14, left):

1. Phosphorylase kinase, which specifically phosphory-

lates Ser 14 of glycogen phosphorylase b (Fig. 18-13,

enzyme F

2. Protein kinase A, which phosphorylates and thereby

activates phosphorylase kinase (Fig. 18-13, enzyme F

3. Phosphoprotein phosphatase-1, which dephosphory-

lates and thereby deactivates both glycogen phosphorylase a

and phosphorylase kinase (Fig. 18-13, enzymes R

and R

In an interconvertible enzyme system, the “modified” form

of the enzyme bears the prefix m and the “original” (unmod-

ified) form bears the prefix o, whereas the enzyme’s most ac-

tive and least active forms are identified by the suffixes a and

b, respectively. In this case, o-phosphorylase b (unmodified,

least active) is the form under allosteric control by AMP,

ATP, and G6P (Fig. 18-10, left). Phosphorylation to yield m-

phosphorylase a (modified, most active) all but removes the

effects of these allosteric modulators. In terms of the sym-

metry model of allosterism (Section 10-4B), the phosphory-

lation of Ser 14 shifts the enzyme’s T (inactive) 34 R (active)

equilibrium in favor of the R state (Fig. 18-10, right). Indeed,

phosphorylase a’s Ser 14-phosphoryl group is analogous to

an allosteric activator: It forms ion pairs with two Arg side

chains on the opposite subunit, thereby knitting the subunits

together in much the same way as does AMP when it binds

tightly to a site between the subunits (Fig. 18-11b).

Section 18-3. Control of Glycogen Metabolism 651

Figure 18-13 A bicyclic enzyme cascade. See the legend of

Fig. 18-12 for symbol definitions. In a bicyclic cascade, one of

the modifying enzymes (F

) is also subject to covalent



ADP

ATP

ADP

ATP



• e





• e

modification. It is active in the modified state (F

) and inactive

in the unmodified state (F

JWCL281_c18_638-670.qxd 2/26/10 2:24 PM Page 651

In the resting cell, the concentrations of ATP and G6P

are high enough to inhibit phosphorylase b. The level of

phosphorylase activity is therefore largely determined by the

fraction of the enzyme present as phosphorylase a. The

steady-state fraction of phosphorylated enzyme (E

) de-

pends on the relative activities of phosphorylase kinase

), protein kinase A (F

), and phosphoprotein phos-

phatase-1 (R

and R

).This interrelationship is remarkably

elaborate for glycogen phosphorylase. Let us consider the

actions of these enzymes.

b. Protein Kinase A: A Crucial Regulatory Link

Phosphorylase kinase, which converts phosphorylase b

to phosphorylase a, is itself subject to covalent modification

652 Chapter 18. Glycogen Metabolism

Protein kinase

A (PKA)

(inactive)

m-Phosphoprotein

phosphatase

inhibitor-1 a

Phosphoprotein

phosphatase-1

(inactive)

(α β γ δ)

o-Phosphorylase

kinase b

o-Glycogen

phosphorylase b

m-Glycogen

phosphorylase a

ATP ADP

m-Glycogen

synthase b

o-Glycogen

synthase a

ATP ADP

4cAMP

ATP ADP

m-Phosphoprotein

phosphatase

inhibitor-1 a

o-Phosphoprotein

phosphatase

inhibitor-1 b

DEPHOSPHORYLATION

SYSTEM

PHOSPHORYLATION

SYSTEM

+ R

(cAMP)

(α β γ δ)

m-Phosphorylase

kinase a

Other

kinases

Protein kinase

A (PKA)

(active)

Phosphoprotein

phosphatase-1

(active)

Figure 18-14 Schematic diagram of the major enzymatic

modification/demodification systems involved in the control of

glycogen metabolism in muscle. Modification (phosphorylation)

systems are shaded in yellow, demodification (dephosphorylation)

systems are shaded in lavender, active enzymes/inhibitors are

shaded in green, and inactive enzymes/inhibitors are shaded in

orange. Dashed yellow and purple arrows indicate facilitation of

a modification and demodification reaction. Note that glycogen

phosphorylase activity is controlled by a bicyclic enzyme cascade

(left) and glycogen synthase activity is controlled by both a

bicyclic and a monocyclic enzyme cascade (right). By convention,

the modified form of the enzyme bears the prefix m and the

“original” (unmodified) form bears the prefix o. The most active

and least active forms of the enzymes are identified by the

suffixes a and b, respectively. Further control of phosphoprotein

phosphatase-1 covalent modification is diagrammed in Fig. 18-22.

See Guided Exploration 15: Control of Glycogen Breakdown, and

the Animated Figures

JWCL281_c18_638-670.qxd 2/26/10 2:24 PM Page 652

(Fig. 18-14). For phosphorylase kinase to be fully active,

2

must be present (see below) and the protein must be

phosphorylated.

In both the glycogen phosphorylase and glycogen syn-

thase cascades, the primary intracellular signal, e

,is

adenosine-3ⴕ,5ⴕ-cyclic monophosphate (3ⴕ,5ⴕ-cyclic

AMP or cAMP). The cAMP concentration in a cell is a

function of the ratio of its rate of synthesis from ATP by

adenylate cyclase (AC; also called adenylyl cyclase)

and its rate of breakdown to AMP by a enzymes known

as cAMP-phosphodiesterases (cAMP-PDEs; Section

19-2E):

HO OH

–

PPO

–

ATP

3ⴕ,5ⴕ-Cyclic AMP

(cAMP)

–

adenylate cyclase

phosphodiesterase

HO OH

–

AMP

AC is, in turn, activated by certain hormones (Sections

18-3Ea and 19-2D).

cAMP is absolutely required for the activity of protein

kinase A [PKA; also called cAMP-dependent protein

kinase (cAPK)], an enzyme that phosphorylates specific

Ser and/or Thr residues of numerous cellular proteins,

including phosphorylase kinase and glycogen synthase.

These proteins all contain PKA’s consensus recognition

sequence, Arg-Arg-X-Ser/Thr-Y, where Ser/Thr is the

phosphorylation site, X is any small residue, and Y is a large

hydrophobic residue. In the absence of cAMP, PKA is an

inactive heterotetramer consisting of two regulatory (R)

and two catalytic (C) subunits, R

. The cAMP binds to

the regulatory subunits so as to cause the dissociation of

active catalytic monomers (Fig. 18-14; top). The intracellu-

lar concentration of cAMP therefore determines the fraction

of PKA in its active form and thus the rate at which it phos-

phorylates its substrates. In fact, in all known eukaryotic

cases, the physiological effects of cAMP are exerted

through the activation of specific protein kinases.

The X-ray structure of the 350-residue C subunit of

mouse PKA in complex with Mg

2

–ATP and a 20-residue

inhibitor peptide was determined by Susan Taylor and

Janusz Sowadski (Fig. 18-15), and that of a similar complex

of the porcine heart enzyme was determined by Robert

Huber.The C subunit, as are other kinases of known struc-

ture (e.g., Figs. 17-5 and 17-15), is bilobal. It has an N-

terminal domain that consists of a 5-stranded  sheet and

an  helix, and a larger C-terminal domain that is mainly 

helical. A deep cleft between the lobes is occupied by the

2

–ATP and the segment of the inhibitor peptide that

includes the above 5-residue consensus sequence.This cleft

therefore contains PKA’s catalytic site, with the small do-

main contributing the nucleotide-binding site and the large

subunit supplying the substrate-binding and catalytic

residues.

The C subunit of PKA must be phosphorylated at Thr

197 for activity. Thr 197 is part of the so-called activation

loop (comprising residues 184–208), which is located at the

“mouth” of the cleft between PKA’s N- and C-terminal do-

mains. The phosphoryl group at Thr 197 knits together the

various components of PKA into its active conformation

via extensive interactions with the protein. Most notably,

the phosphoryl group interacts with Arg 165, a conserved

residue that is adjacent to Asp 166, the catalytic base that

activates the substrate protein’s target Ser/Thr hydroxyl

group for phosphorylation, thereby properly orienting

PKA’s active site residues.

Protein kinases play key roles in the signaling pathways

by which many hormones, growth factors, neurotransmit-

ters, and toxins affect the functions of their target cells

(Chapter 19), as well as in controlling metabolic pathways.

Indeed, 518 human proteins constituting ⬃2.3% of human

genes are predicted to be protein kinases, which accounts

for the observation that ⬃30% of the proteins in mam-

mallian cells are phosphorylated.The 1000 different pro-

tein kinases that have been sequenced share a conserved

catalytic core corresponding to residues 40 to 280 of PKA’s

Section 18-3. Control of Glycogen Metabolism 653

JWCL281_c18_638-670.qxd 2/26/10 2:24 PM Page 653

C subunit. In addition to phosphorylating other proteins,

many protein kinases are themselves phosphoproteins

whose activities are controlled by phosphorylation,often at

their activation loops. However, since PKA is normally

fully phosphorylated at Thr 197 and resistant to dephos-

phorylation, it is unclear whether its activity is regulated in

vivo by phosphorylation/dephosphorylation.

c. PKA’s R Subunit Competitively Inhibits

Its C Subunit

The R subunit of PKA has a well-defined domain struc-

ture that was first characterized by limited proteolysis. It

consists of, from N- to C-terminus, a dimerization domain,

an autoinhibitor segment, and two tandem homologous

cAMP-binding domains,A and B. In the R

complex, the

autoinhibitor segment, which resembles the C subunit’s

substrate peptide, binds in the C subunit’s active site (as

does the inhibitory peptide in Fig. 18-15) so as to block sub-

strate binding. Thus, the R subunit is a competitive

inhibitor of PKA’s substrate proteins.

Each R subunit cooperatively binds two cAMPs. When

the B domain lacks bound cAMP, it masks the A domain so

as to prevent it from binding cAMP. However, the binding

of cAMP to the B domain triggers a conformational

change that permits the A domain to bind cAMP, which, in

turn, releases the C subunits from the complex (see below).

Taylor determined the X-ray structure of the R subunit

lacking its N-terminal 91 residues and in complex with two

cAMPs (Fig. 18-16a). This truncated protein is unable to

dimerize but, in the absence of cAMP, forms a tight inactive

complex with the C subunit, and on binding cAMP releases

active C subunits as do intact R

dimers.As previously pre-

dicted by sequence alignments, the A and B domains are

structurally similar to each other and to the prokaryotic

cAMP-binding transcriptional regulator named catabolite

gene activator protein (CAP; Section 31-3Cb). The autoin-

hibitory segment, which in the free R subunit is extremely

sensitive to proteolysis, has its first 21 residues disordered

in the X-ray structure.

The X-ray structure of the truncated R subunit in com-

plex with the C subunit binding AMPPNP (Fig. 18-16b),

also determined by Taylor, reveals that the R subunit has

undergone a massive conformational reorganization rela-

tive to its cAMP-binding structure.The most striking such

change is that the 25-residue, bent 2-helix segment linking

domains A and B (purple in Fig. 18-16a) has coalesced to

form a single straight helix such that the ␤ sandwich in the

R subunit’s B domain has rotated by ⬃180° relative to

that in the A domain and separated from it. As a conse-

quence of this ⬃60-Å screwlike shift of the B domain, the

elongated globular structure of the cAMP complex

changes to a dumbbell-like shape in which the interface

between the A and B domains in the cAMP complex is re-

placed by extensive interactions with the large domain of

the C subunit. In addition, the helical regions of the A and

B domains undergo extensive conformational changes

that eliminate their cAMP binding sites by separating

their phosphate-binding pockets from their adenine-bind-

ing pockets. All of this positions the R subunit’s autoin-

hibitor segment in the C subunit’s active site cleft, thereby

inactivating it.

654 Chapter 18. Glycogen Metabolism

Figure 18-15 X-ray structure of the catalytic (C) subunit of

mouse protein kinase A (PKA). The protein is in complex with

ATP and a 20-residue peptide segment of a naturally occurring

protein kinase inhibitor.The N-terminal domain is pink and its

C-terminal domain is cyan with its activation loop blue.The

polypeptide inhibitor is orange and its pseudo–target sequence,

Arg-Arg-Asn-Ala-Ile, is purple, with the Ala replacing the Ser to

be phosphorylated green (note that the enzyme’s true target

sequence is Arg-Arg-X-Ser/Thr-Y, where X is a small residue,Y

is a large hydrophobic residue, and Ser/Thr, which is replaced by

Ala in the polypeptide inhibitor, is the residue that the enzyme

phosphorylates).The ATP and the phosphoryl group of

phosphoThr 197 are shown in space-filling form and the side

chains of the catalytically essential Arg 165,Asp 166, and Thr 197

are shown in stick form, all colored according to atom type

(C green, N blue, O red, and P orange). Note that the inhibitor’s

pseudo–target sequence is in close proximity to the ATP’s ␥

phosphate group, the group that the enzyme transfers. [Based on

an X-ray structure by Susan Taylor and Janusz Sowadski,

University of California at San Diego. PDBid 1ATP.]

See

Interactive Exercise 10 and Kinemage Exercise 15-1

JWCL281_c18_638-670.qxd 10/19/10 7:29 AM Page 654

d. Phosphorylase Kinase: Coordination of Enzyme

Activation with [Ca

2ⴙ

]

Phosphorylase kinase (PhK) is activated by Ca

2⫹

con-

centrations as low as 10

⫺7

M as well as by covalent modifica-

tion. This 1300-kD enzyme consists of four nonidentical

subunits that form the active oligomer (␣␤␥␦)

. The iso-

lated ␥ subunit is capable of full catalytic activity (ability to

convert phosphorylase b to phosphorylase a), whereas the

␣, ␤, and ␦ subunits are inhibitors of the catalytic reaction.

The ␦ subunit, which is known as calmodulin (CaM),

confers Ca

2⫹

sensitivity on the complex. When Ca

2⫹

binds

to calmodulin’s four Ca

2⫹

-binding sites, this ubiquitous

eukaryotic regulatory protein undergoes an extensive

conformational change (see below) that activates phospho-

rylase kinase. Glycogen phosphorylase therefore becomes

phosphorylated and the rate of glycogen breakdown

increases. The physiological significance of this Ca

2⫹

activa-

tion process is that nerve impulses trigger muscle con-

traction through the release of Ca

2⫹

from intracellular

reservoirs (Section 35-3C). This transient increase in cytoso-

lic [Ca

2⫹

] induces both muscle contraction and the increase

in glycogen breakdown that supplies glycolysis, which in

turn, generates the ATP required for muscle contraction.

e. Calmodulin: A Ca

2ⴙ

-Activated Switch

Calmodulin is a ubiquitous eukaryotic Ca

2⫹

-binding

protein that participates in numerous cellular regulatory

processes. In some of these, CaM functions as a monomeric

protein, whereas in others (e.g., PhK) it is a subunit of a

larger protein.The X-ray structure of this highly conserved

Section 18-3. Control of Glycogen Metabolism 655

Figure 18-16 X-ray structures of the regulatory (R) subunit of

bovine protein kinase A (PKA). (a) The R subunit lacking its

N-terminal 91 residues (which form its dimerization domain) in

complex with cAMP. The N-terminal region, which includes its

autoinhibitor segment, is red, domain A is green, domain B is

orange, and the 2-helix segment linking domains A and B is

purple. The cAMPs, which are drawn in space-filling form

colored according to atom type (C cyan, N blue, O red, and P

orange), each bind at the opening of an 8-stranded ␤ sandwich

with their phosphate group abutting the N-terminal end of a

short helical segment. (b) The truncated R subunit in complex

with the C subunit that is binding AMPPNP. The R subunit is

colored as in Part a and the C subunit is colored as in Fig. 18-15.

The C subunit is rotated 180° about the vertical axis relative to

Fig. 18-15 and the ␤ sandwich of domain A is oriented similarly

to that in Part a. Note the dramatic conformational differences

between the R subunit in Parts a and b, which in large part is

caused by the coalescence and straightening of the two helices

linking domains A and B in the cAMP complex.Also note that

the now ordered N-terminal portion of the R subunit’s

autoinhibitor segment occupies the C subunit’s active site cleft.

[Based on X-ray structures by Susan Taylor, University of

California at San Diego. PDBids 1RGS and 2QCS.]

(a) (b)

JWCL281_c18_638-670.qxd 6/30/10 11:59 AM Page 655

148-residue protein, determined by Charles Bugg, has a cu-

rious dumbbell-like shape in which CaM’s two globular do-

mains are connected by a seven-turn  helix (Fig. 18-17).

CaM has two high-affinity Ca

2

-binding sites on each of its

globular domains, both of which are formed by nearly su-

perimposable helix–loop–helix motifs known as EF hands

(Fig. 18-18) that also occur in numerous other Ca

2

-sensing

proteins of known structure. The Ca

2

ion in each of these

sites is octahedrally coordinated by oxygen atoms from the

backbone and side chains of the loop as well as from a

protein-associated water molecule.

The binding of Ca

2

to either domain of CaM induces a

conformational change in that domain, which exposes an

otherwise buried Met-rich hydrophobic patch. This patch,

in turn, binds with high affinity to the CaM-binding domain

of the phosphorylase kinase  subunit, as well as to the

CaM-binding domains of numerous other Ca

2

-regulated

proteins (many of which interact with CaM that is free in

solution), and in doing so modulates the activities of these

proteins. These CaM-binding domains have little mutual

sequence similarity but are all basic amphiphilic  helices.

In fact, ⬃20-residue segments of these helices, as well as

synthetic amphiphilic helices composed of only Leu, Lys,

and Trp residues, are bound by Ca

2

–CaM as tightly as the

target proteins themselves.

Despite uncomplexed CaM’s extended appearance in

its X-ray structure (Fig. 18-17), a variety of studies indicate

that both of its globular domains can simultaneously bind

to a single target helix. Evidently, CaM’s central  helix

serves as a flexible linker rather than as a rigid spacer, a

property that probably further increases the range of tar-

get sequences to which CaM can bind. This idea is con-

firmed by the NMR structure (Fig. 18-19) of (Ca

2

)

–CaM

in complex with its 26-residue CaM-binding target

polypeptide of skeletal muscle myosin light chain kinase

(MLCK; a homolog of the PKA C subunit, which phospho-

rylates and thereby activates the light chains of the muscle

protein myosin; Section 35-3Da), which was determined by

Marius Clore, Angela Gronenborn, and Ad Bax. Indeed,

the extended conformation of CaM’s central helix in Fig.

18-17 is probably an artifact arising from crystal packing

forces, considering that this helix’s central two turns con-

tact no other portion of the protein and hence are maxi-

mally solvent-exposed (almost all other known  helices

are at least partially buried in a protein). Moreover, a

polypeptide with the sequence of this helix assumes a ran-

dom coil conformation in aqueous solution. Nevertheless,

the flexible linker is essential to the function of CaM: In the

presence of Ca

2

, CaM’s individual domains (obtained by

656 Chapter 18. Glycogen Metabolism

Figure 18-17 X-ray structure of rat testis calmodulin. This

monomeric 148-residue protein, which is colored in rainbow

order from N-terminus (blue) to C-terminus (red), contains two

remarkably similar globular domains separated by a seven-turn 

helix.The two Ca

2

ions bound to each domain are represented

by cyan spheres. The side chains liganding the Ca

2

ions are

drawn in stick form colored according to atom type (C green, N

blue, and O red). [Based on an X-ray structure by Charles Bugg,

University of Alabama at Birmingham. PDBid 3CLN.]

See

Kinemage Exercise 16-1

Figure 18-18 EF hand. The Ca

2

-binding sites in many

proteins that function to sense the level of Ca

2

are formed by

helix–loop–helix motifs named EF hands. [After Kretsinger,

R.H., Annu. Rev. Biochem. 45, 241 (1976).]

See Kinemage

Exercise 16-1

EF hand

E helix

F helix

JWCL281_c18_638-670.qxd 2/26/10 2:24 PM Page 656

tryptic cleavage), when in high concentration, are able to

bind their target proteins but fail to even marginally acti-

vate them unless present in several hundred-fold excess.

How does Ca

2

–CaM activate its target protein ki-

nases? MLCK contains a C-terminal segment whose se-

quence resembles that of MLCK’s target polypeptide on

the light chain of myosin but lacks a phosphorylation site.

A model of MLCK, based on the X-ray structure of the

30% identical C subunit of PKA, strongly suggests that this

autoinhibitor peptide inactivates MLCK by binding in its

active site. Indeed, the excision of MLCK’s autoinhibitor

peptide by limited proteolysis permanently activates this

enzyme. MLCK’s CaM-binding segment overlaps this au-

toinhibitor peptide. Evidently, the binding of Ca

2

–CaM to

this peptide segment extracts the autoinhibitor from MLCK’s

active site, thereby activating this enzyme (Fig. 18-20).

2

–CaM’s other target proteins, including the phos-

phorylase kinase  subunit, are presumably activated in the

Section 18-3. Control of Glycogen Metabolism 657

(a)

(b)

(inactive)

(active)

regulatory domain

Catalytic domain

Cellular response

– CaM

CaM-binding

protein kinase

Substrate

protein

Active site

CaM binding

site

Figure 18-20 Schematic diagram of the Ca

2

–CaM-dependent activation of protein kinases.

Autoinhibited kinases have an N- or C-terminal “pseudosubstrate” sequence (red ) that binds at

or near the enzyme’s active site (brown) so as to inhibit its function. This autoinhibitory segment

is in close proximity with or overlaps a Ca

2

–CaM-binding sequence. Consequently, Ca

2

–CaM

(green) binds to this sequence so as to extract it from the enzyme’s active site, thereby activating

the enzyme to phosphorylate other proteins (purple). [After Crivici, A. and Ikura, M., Annu. Rev.

Biophys. Biomol. Struct. 24, 88 (1995).]

Figure 18-19 NMR structure of (Ca

2ⴙ

)

–CaM from

Drosophila melanogaster in complex with its 26-residue target

polypeptide from rabbit skeletal muscle myosin light chain

kinase (MLCK). The N terminal domain of CaM is blue, its

C-terminal domain is red, the target polypeptide is green, and the

2

ions are represented by cyan spheres. (a) A view of the

complex in which the N-terminus of the target polypeptide is on

the right, and (b) the perpendicular view as seen from the right

side of Part a. In both views, the pseudo-2-fold axis relating the

N- and C-terminal domains of CaM is approximately vertical.

Note how the middle segment of the long central helix in

uncomplexed CaM (Fig. 18-17) has unwound and bent (bottom

loop in Part b) such that CaM forms a globular protein that

largely encloses the helical target polypeptide within a

hydrophobic tunnel in a manner resembling two hands holding a

rope (the target polypeptide assumes the random coil

conformation in solution). However, the conformations of CaM’s

two globular domains are essentially unchanged by the

complexation. Evidently, CaM’s bound Ca

2

ions serve to

organize and stabilize the target binding conformations of its

globular domains. [Based on an NMR structure by Marius Clore,

Angela Gronenborn, and Ad Bax, National Institutes of Health.

PDBid 2BBM.]

See Kinemage Exercise 16-2

JWCL281_c18_638-670.qxd 2/26/10 2:24 PM Page 657

same way.The X-ray structures of two homologous protein

kinases support this so-called intrasteric mechanism, those

of calmodulin-dependent protein kinase I (CaMKI) and

twitchin kinase. Although the details of binding of the au-

toinhibitory sequence differ for each of these protein ki-

nases, the general mode of autoinhibition and activation by

2⫹

–CaM is the same.

PKA’s R subunit, as we have seen, contains a similar au-

toinhibitory sequence adjacent to its two tandem cAMP-

binding domains. In this case, however, the autoinhibitory

peptide is allosterically ejected from the C subunit’s active

site by the binding of cAMP to the R subunit (which lacks

a Ca

2⫹

–CaM binding site).

f. Phosphorylase Kinase’s ␥ Subunit Is Controlled

by Multiple Autoinhibitors

Phosphorylase kinase’s 386-residue ␥ subunit consists of

an N-terminal kinase domain, which is 36% identical in

sequence to the C subunit of PKA, and a C-terminal regula-

tory domain, which contains a CaM-binding peptide and an

overlapping autoinhibitor segment. Evidently, Ca

2⫹

–CaM

relieves this inhibition, as is diagrammed in Fig. 18-20.This

explains why the N-terminal 298-residue segment of the

PhK ␥ subunit, termed PhK␥

(t for truncated), displays

catalytic activity comparable to that of fully activated PhK

but is unaffected by Ca

2⫹

or phosphorylation signals.

The X-ray structure of PhK␥

in complex with ATP and

a heptapeptide related to the natural substrate was deter-

mined by Johnson (Fig. 18-21). It reveals, as expected, that

PhK␥

structurally resembles PKA (Fig. 18-15) as well as

other protein kinases of known structure including CaMKI

and twitchin kinase. Comparisons of these various struc-

tures shed light on how the catalytic activity of PhK is

regulated. Numerous protein kinases, including PKA, are

activated by the phosphorylation of Ser, Thr, and/or

Tyr residues in their activation loops, which, as we saw in

Fig. 18-15, interacts with a conserved Arg residue that

thereby correctly positions the adjacent catalytically im-

portant Asp residue. However, the PhK ␥ subunit is not

subject to phosphorylation. Rather, its activation loop

residue that might otherwise be phosphorylated is Glu 182,

whose negative charge mimics the presence of a phosphate

group by interacting with Arg 148 so as to correctly posi-

tion Asp 149 (Fig. 18-21).Thus, the PhK catalytic site main-

tains an active conformation but, in the absence of Ca

2⫹

,is

inactivated by the binding of its C-terminal autoinhibitor

segment.

Sites on both the ␣ and ␤ subunits of PhK are subject to

phosphorylation by PKA (Fig. 18-14).This activates PhK at

much lower Ca

2⫹

concentrations than otherwise, and full

enzyme activity is obtained in the presence of Ca

2⫹

only

when both these subunits are phosphorylated. The ␤ sub-

unit does, in fact, have an autoinhibitor sequence, suggest-

ing that phosphorylation changes its conformation so as to

make it unavailable for inhibiting the ␥ subunit’s active

site.This would explain the synergistic effect of phosphory-

lation and Ca

2⫹

on the activity of PhK: Ca

2⫹

–CaM se-

questers the ␥ subunit’s autoinhibitory segment, whereas

phosphorylation of the ␤ subunit removes yet another au-

toinhibitor.The way in which the phosphorylation of the ␣

subunit modulates the activity of PhK is, as yet, unknown.

g. Phosphoprotein Phosphatase-1

The steady-state phosphorylation levels of most en-

zymes involved in cyclic cascades are maintained by the

opposition of kinase-catalyzed phosphorylations and the

hydrolytic dephosphorylations catalyzed by phosphopro-

tein phosphatases. The phosphatase involved in the cyclic

cascades controlling glycogen metabolism is phosphoprotein

phosphatase-1. This enzyme, as is indicated in Fig. 18-14,

hydrolyzes the phosphoryl groups from m-glycogen phos-

phorylase a, both the ␣ and ␤ subunits of phosphorylase

kinase, and two other proteins involved in glycogen meta-

bolism, as discussed below.

The catalytic subunit of phosphoprotein phosphatase-1

(PP1), which is designated PP1c, hydrolyzes phosphoryl

groups on Ser/Thr residues via a single step mechanism.

658 Chapter 18. Glycogen Metabolism

Figure 18-21 X-ray structure of rabbit muscle PhK␥

complex with ATP and a heptapeptide (RQMSFRL). This

heptapeptide is related in sequence to the enzyme’s natural

substrate (KQISVRG).The protein is shown in the “standard”

protein kinase orientation with its N terminal domain pink, its

C-terminal domain cyan, and its activation loop blue. The

heptapeptide is orange, with its residue to be phosphorylated

(Ser) green.The ATP is shown in space-filling form and the side

chains of the catalytically essential Arg 148,Asp 149, and Glu 182

are shown in stick form, all colored according to atom type

(C green, N blue, O red, and P orange). Note the structural simi-

larities and differences between this protein and the homologous

C subunit of PKA (Fig. 18-15). [After an X-ray structure by

Louise Johnson, Oxford University, Oxford, U.K. PDBid 2PHK.]

JWCL281_c18_638-670.qxd 6/3/10 1:49 PM Page 658

The X-ray structure of PP1c indicates that it contains a bin-

uclear metal ion center (both metals are Mn

2

in the re-

combinant enzyme) which, it is proposed, activates a water

molecule (promotes its ionization to OH



, Section 15-1Cb)

for nucleophilic attack on the phosphoryl group.

PP1c binds to glycogen through the intermediacy of

regulatory proteins in both muscle and liver. In muscle,

PP1c is only active when it is bound to glycogen through

this glycogen-binding G

subunit. The activity of PP1c

and its affinity for the G

subunit are regulated by phos-

phorylation of the G

subunit at two separate sites (Fig.

18-22). Phosphorylation of site 1 by insulin-stimulated

protein kinase activates phosphoprotein phosphatase-1,

whereas phosphorylation of site 2 by PKA (which can

also phosphorylate site 1) causes the enzyme to be

released into the cytoplasm, where it cannot dephos-

phorylate the glycogen-bound enzymes of glycogen

metabolism.

In the cytosol, PP1c is also inhibited by its binding

to the protein phosphoprotein phosphatase inhibitor 1

(inhibitor-1). This latter protein provides yet another ex-

ample of control by enzymatic interconversion: It too is

modified by PKA and demodified by PP1c (Fig. 18-14,

bottom), although, in this case, a Thr, not a Ser, is phospho-

rylated/dephosphorylated. The protein is a functional in-

hibitor only when it is phosphorylated. The concentration

of cAMP therefore controls the fraction of an enzyme in its

phosphorylated form, not only by increasing the rate at

which it is phosphorylated, but also by decreasing the rate at

which it is dephosphorylated. In the case of glycogen phos-

phorylase, an increase in [cAMP] results not only in an

increase in this enzyme’s rate of activation, but also in a de-

crease in its rate of deactivation.

The activity of phosphoprotein phosphatase-1 in liver is

also controlled by its binding to glycogen through the inter-

mediacy of a glycogen-binding subunit, here named G

When bound to G

, PP1c is activated toward dephosphory-

lation of the glycogen-bound enzymes of glycogen metabo-

lism. However, G

is not subject to control via phosphory-

lation as is G

in the muscle. Rather, the binding of

m-phosphorylase a to G

strongly inhibits the activity of

PP1c by an allosteric mechanism.

Among the major conformational changes that glycogen

phosphorylase undergoes in converting from the T to the R

state is the movement of the Ser 14-phosphoryl group from

the surface of the T-state (inactive) enzyme to a position

buried a few angstroms beneath the protein’s surface at the

dimer interface in the R-state (active) enzyme (Figs. 18-11b).

Both the R and T forms of phosphorylase a strongly bind the

ⴢ PP1c complex, but only in the T-state enzyme is the Ser

14-phosphoryl group accessible for hydrolysis by PP1c. Con-

sequently, under the conditions that phosphorylase a con-

verts to the T state (Section 18-3G), PP1c hydrolyzes its now

exposed Ser 14-phosphoryl group.This converts m-phospho-

rylase a to o-phosphorylase b, which has only a low affinity

for binding the G

ⴢ PP1c complex and hence does not in-

hibit PP1c. One effect of phosphorylase a demodification,

therefore, is to relieve the inhibition of PP1c and thus allow

it to excise the phosphoryl groups of other susceptible phos-

phoproteins. Since phosphorylase a has a high affinity for

the G

ⴢ PP1c complex and is in ⬃10-fold greater concentra-

tion, relief of PP1c inhibition only occurs when more than

Section 18-3. Control of Glycogen Metabolism 659

Glycogen

Decreased phosphorylation

leading to increased

glycogen synthesis

Increased phosphorylation

leading to increased

glycogen breakdown

subunit PP1c

PP1c

Glycogen G

subunit

Protein kinase A

(PKA)

Epinephrine

Insulin-stimulated

protein kinase

ATP

ADP

(less active)

(more active)

(inactive)

Glycogen G

subunit

Insulin

Figure 18-22 The antagonistic effects of insulin and

epinephrine on glycogen metabolism in muscle. This occurs

through their effects on the phosphoprotein phosphatase-1

catalytic subunit, PP1c, via its glycogen-bound G

subunit. Green

discs and dashed arrows indicate activation.

JWCL281_c18_638-670.qxd 2/26/10 2:24 PM Page 659

⬃90% of the glycogen phosphorylase is in the o-phosphory-

lase b form. Glycogen synthase is among the proteins that

are dephosphorylated by the G

ⴢ PP1c complex when it is

no longer inhibited by phosphorylase. However, in contrast

to phosphorylase, dephosphorylation activates glycogen

synthase. This enzyme is involved in its own bicyclic cascade

whose properties we shall now examine.

D. Glycogen Synthase Bicyclic Cascade

Like glycogen phosphorylase, glycogen synthase exists in

two enzymatically interconvertible forms:

1. The modified (m; phosphorylated) form that is inac-

tive under physiological conditions (the b form).

2. The original (o; dephosphorylated) form that is

active (the a form).

m-Glycogen synthase b is under allosteric control; it is

strongly inhibited by physiological concentrations of ATP,

ADP, and P

and hence the modified enzyme is almost

totally inactive in vivo. The activity of the unmodified en-

zyme is essentially independent of these effectors, so a

cell’s glycogen synthase activity varies with the fraction of

the enzyme in its unmodified form.

The mechanistic details of the interconversion of modi-

fied and unmodified forms of glycogen synthase are partic-

ularly complex and are therefore not as well understood as

those of glycogen phosphorylase. It has been clearly estab-

lished that the fraction of unmodified glycogen synthase is,

in part, controlled by a bicyclic cascade involving phospho-

rylase kinase (PhK) and phosphoprotein phosphatase-1,

enzymes that are also involved in the glycogen phosphory-

lase bicyclic cascade (Fig. 18-14, right). This demodification

process is facilitated by G6P, whose binding to m-glycogen

synthase b induces it to undergo a conformational change

that exposes its phosphoryl groups to the surface of the

protein, thereby making them available for dephosphory-

lation by phosphoprotein phosphatase-1.

Glycogen synthase is phosphorylated at several sites.

Several protein kinases are known to at least partially de-

activate human muscle glycogen synthase by phosphory-

lating this homotetramer at 1 or more of 9 Ser residues in

the N- and C-terminal segments on its 737-residue sub-

units. These enzymes include PhK, PKA (so glycogen syn-

thase deactivation may also be considered to occur via a

monocyclic cascade), CaMKI (which is activated by the

presence of Ca

2⫹

), protein kinase C (PKC; which responds

to the extracellular presence of certain hormones via a

mechanism described in Sections 18-3G and 19-4Cb),

AMP-dependent protein kinase (AMPK; which responds

to ATP availability and hence acts as a fuel gauge; Sections

25-5a and 27-1), glycogen synthase kinase-3 [GSK3; which

is inhibited by insulin (Sections 18-3Ea and 18-3F), whose

presence therefore results in the dephosphorylation and

hence activation of glycogen synthase], and casein kinases

1 and 2 (which participate in a variety of cellular control

processes).Why glycogen synthase deactivation is so elab-

orately controlled compared to its activation or the activa-

tion/deactivation of glycogen phosphorylase is unclear, al-

though, whatever the reasons, it closely monitors the or-

ganism’s metabolic state.

E. Integration of Glycogen Metabolism

Control Mechanisms

Whether there is net synthesis or degradation of glycogen

and at what rate depends on the relative balance of the ac-

tive forms of glycogen synthase and glycogen phosphory-

lase. This, in turn, largely depends on the rates of the phos-

phorylation and dephosphorylation reactions of the two

bicyclic cascades.These cascades, one controlling the rate of

glycogen breakdown and the other controlling the rate of

glycogen synthesis, are intimately related. They are linked

by protein kinase A and phosphorylase kinase, which,

through phosphorylation, activate glycogen phosphorylase

as they inactivate glycogen synthase (Fig. 18-14). The cas-

cades are also linked by phosphoprotein phosphatase-1,

which in liver is inhibited by phosphorylase a and therefore

is unable to activate (dephosphorylate) glycogen synthase

unless it first inactivates (also by dephosphorylation) phos-

phorylase a.

a. Hormones Are Important Regulators

of Glycogen Metabolism

Glycogen metabolism is largely regulated by the pep-

tide hormone insulin (Fig. 7-2) acting in opposition to

glucagon, another peptide hormone,

⫹

N - His - Ser - Glu- Gly - Thr - Phe - Thr - Ser - Asp- Tyr - 10

Ser - Lys - Tyr - Leu - Asp- Ser - Arg - Arg - Ala - Gln- 20

Asp- Phe- Val - Gln - Trp - Leu - Met - Asn - Thr - COO

⫺

Glucagon

together with the adrenal hormones epinephrine (adrena-

line) and norepinephrine (noradrenaline):

Hormonal stimulation of cells at their plasma mem-

branes occurs through the mediation of transmembrane

proteins called receptors. Different cell types have different

complements of receptors and thus respond to different sets

of hormones. For example, both muscle and liver cells have

abundant insulin and adrenergic receptors (receptors re-

sponsive to epinephrine and norepinephrine), whereas

glucagon receptors are more prevalent in liver than in ske-

latal muscle.

X = CH

Epinephrine

X = H Norepinephrine

660 Chapter 18. Glycogen Metabolism

JWCL281_c18_638-670.qxd 6/30/10 12:00 PM Page 660