Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

F. Protein Phosphatases

As we previously discussed (Section 19-2E), to prevent an

intracellular signaling pathway from being stuck in the

“on” position, its signals must be rapidly eliminated once

the message has been delivered. For proteins with phos-

phoTyr or phospho-Ser/Thr residues, this task is carried out

by a variety of protein phosphatases, of which ⬃500 are en-

coded by the human genome (around the same number as

species of protein kinases, which suggests that these fami-

lies have similar levels of complexity). The protein phos-

phatases, as we shall see, are not just simple housekeeping

enzymes but are signal transducers in their own right.Thus,

whereas kinases control the amplitude of a signaling re-

sponse, protein phosphatases control its rate and duration.

a. Protein Tyrosine Phosphatases Also Mediate

Signal Transduction

The enzymes that dephosphorylate Tyr residues, the pro-

tein tyrosine phosphatases (PTPs), which were discovered

by Nicholas Tonks, form a large family of diverse proteins

that are present in all eukaryotes (humans have 107 PTP

genes vs 90 PTK genes). Each PTP contains at least one

conserved ⬃280-residue phosphatase domain that has the

11-residue signature sequence (I/V)HCXAGXGR(S/T)G,

the so-called HCX

R motif, which contains the enzyme’s

catalytically essential Cys and Arg residues. The reaction

proceeds via the nucleophilic attack of the Cys thiolate

group on the P atom of the bound phosphoTyr to yield Tyr

and a cysteinyl–phosphate intermediate that is subse-

quently hydrolyzed. The Arg side chain participates in sub-

strate binding and stabilizes the cysteinyl–phosphate inter-

mediate.

The PTPs have been classified into three groups: (1) recep-

torlike PTPs, (2) intracellular PTPs, and (3) dual-specificity

PTPs, which can also dephosphorylate phospho-Ser/Thr

residues. The receptorlike PTPs are constructed much

like the RTKs (Fig. 19-25); that is, they have, from N- to

C-terminus, an ectodomain consisting of often multiple

repeating modules that occur in other proteins, a single

transmembrane helix, and a cytosolic domain consisting of

a catalytically active PTP domain that, in most cases, is fol-

lowed by a second PTP domain with little or no catalytic

activity. Nevertheless, these inactive PTP domains, which

are highly conserved, are important for the activity, speci-

ficity, and stability of the PTP as a whole.

Biochemical and structural analyses indicate that ligand-

induced dimerization of a receptorlike PTP reduces its

catalytic activity, probably by blocking its active sites. Intra-

cellular PTPs contain only one PTP domain, which is

flanked by regions containing motifs, such as SH2 domains,

that participate in protein–protein interactions. Structural

studies reveal that the active sites of receptorlike and intra-

cellular PTPs are too deep to bind phospho-Ser/Thr side

chains—as we also saw to be the case for both PTK and

SH2 domains (Sections 19-3Ac and 19-3Cb). However, the

active site pockets of dual-specificity PTPs are sufficiently

shallow to bind both phosphoTyr and phospho-Ser/Thr

residues.

b. SHP-2 Is Inactivated by Binding Its Unliganded

N-Terminal SH2 Domain

The cytoplasmic PTP SHP-2 (for SH2 domain-containing

phosphatase 2), which is expressed in all mammalian cells,

binds to PTKs that are activated by a variety of ligands,

including cytokines, growth factors, and hormones. The

591-residue SHP-2 consists of two tandem SH2 domains,

followed by a PTP domain and a 66-residue C-terminal tail

that contains Tyr-phosphorylation sites as well as a Pro-rich

segment that may bind SH3- or WW-containing proteins.

SHP-2’s PTP activity is increased ⬃10-fold on binding pep-

tides with a single phosphoTyr residue and ⬃100-fold with

those having two phosphoTyr residues and at much lower

peptide concentrations. SHP-2 binds to both growth factors

and certain cytokine receptors via its SH2 domains and,

when its C-terminal tail is phosphorylated, also functions

as an adaptor to recruit Grb2 so as to activate MAP kinase

pathways (Section 19-3D). Mutations in the gene encoding

SHP-2 are responsible for ⬃50% of the cases of Noonan

syndrome, a relatively common (1 in ⬃2000 live births) dis-

order principally characterized by cardiac abnormalities,

short stature, learning disabilities, and distinctive facial

features.

The X-ray structure of SHP-2 lacking its C-terminal tail

(Fig. 19-50), determined by Eck and Steven Shoelson, re-

veals that the N-terminal SH2 domain (N-SH2) interacts

extensively with the PTP domain. N-SH2 inhibits the PTP

by inserting its D¿E loop far into the PTP’s 9-Å-deep cat-

alytic cleft, where the loop interacts with the PTP’s cat-

alytic Arg and Cys residues and prevents the active site clo-

sure observed in the X-ray structure of a PTP in complex

with a phosphopeptide. In contrast, the more C-terminal

SH2 domain (C-SH2) does not have a significant interface

with either the N-SH2 or the PTP domains.

Section 19-3. Tyrosine Kinase–Based Signaling 721

Figure 19-50 X-ray structure of the protein tyrosine phosphatase

SHP-2. In this structure, its N-SH2 domain is gold with its D¿E

loop red, its C-SH2 domain is green, and its PTP domain is cyan,

with its 11-residue signature sequence, its CX

R motif, blue and

the side chain of its catalytically essential Cys residue shown in

ball-and-stick form with C green and S yellow. [Based on an

X-ray structure by Michael Eck and Steven Shoelson, Harvard

Medical School. PDBid 2SHP.]

JWCL281_c19_671-743.qxd 3/16/10 7:17 PM Page 721

The phosphopeptide-binding sites on both SH2 do-

mains face away from the PTP domain and are therefore

fully exposed on the protein surface. However,the compar-

ison of the structure of N-SH2 complexed to a phospho-

peptide with that in the above autoinhibited form of SHP-

2 indicates that, in the autoinhibited form, N-SH2 adopts a

conformation in which it is unable to bind phosphoTyr. Ev-

idently, the conformations of N-SH2’s PTP-binding surface

and phosphopeptide binding site are allosterically linked

such that its binding of PTP and phosphopeptide are mutu-

ally exclusive. The C-SH2 domain does not participate in

PTP activation, although it almost certainly contributes

binding energy and specificity to the binding of a phospho-

peptide.

c. Bubonic Plague Virulence Requires a PTP

Bacteria lack PTKs and hence do not synthesize phos-

phoTyr residues. Nevertheless, PTPs are expressed by bac-

teria of the genus Yersinia, most notably Yersinia pestis, the

pathogen that causes bubonic plague (the flea-transmitted

“Black Death,” which, since the sixth century, has been re-

sponsible for an estimated ⬃200 million human deaths in-

cluding about one-third of the European population in the

years 1347–1350). The Y. pestis PTP, YopH, which is re-

quired for bacterial virulence, is far more active than other

known PTPs. Hence,when Yersinia injects YopH into a cell,

the cell’s phosphoTyr-containing proteins are catastrophi-

cally dephosphorylated. Although YopH is only ⬃15%

identical in sequence to mammalian PTPs, it contains all of

their invariant residues and their X-ray structures are

closely similar. This suggests that an ancestral Yersinia ac-

quired a PTP gene from a eukaryote. However, the discov-

ery of a dual-specificity protein phosphatase in a free-living

cyanobacterium raises the possibility that PTPs arose be-

fore the divergence of eukaryotes and prokaryotes.

d. Cells Contain Several Types of Protein

Ser/Thr Phosphatases

The protein Ser/Thr phosphatases were first character-

ized by Earl Sutherland (who also discovered the role of

cAMP as a second messenger; Section 18-3Eb) and by

Edmond Fischer and Edwin Krebs (who discovered the

role of protein phosphorylation in controlling glycogen

metabolism; Section 18-3C).The majority of these enzymes

are members of two protein families:the PPP family, which

consists of PP1, PP2A, and PP2B (PP for phosphoprotein

phosphatase); and the PPM family, which consists of

PP2C. The PPP and PPM families are unrelated to each

other or to the PTKs. We have already considered PP1 in

connection with the role of its catalytic subunit,PP1c,in de-

phosphorylating the proteins that regulate glycogen me-

tabolism as well as the roles of its targeting subunits, G

and G

, in binding PP1c to glycogen in muscle and liver

(Section 18-3Cg). Indeed, all PP1c’s are associated with

one or two regulatory (R) subunits that function to modu-

late the activity of their bound PP1c’s, target them to sub-

strates in specific subcellular locations, or modify their sub-

strate specificities. It is the large variety of these mostly

unrelated R subunits that permits the limited number

(1–8) of genetically distinct but closely similar (⬃90% se-

quence identity) PP1c’s in a eukaryotic cell to carry out

their diverse functions.

X-ray structures have shown that PPP catalytic centers

each contain an Fe

2

(or possibly an Fe

3

) ion and a Zn

2

(or possibly an Mn

2

) ion, whereas PPM catalytic centers

each contain two Mn

2

ions.These binuclear metal ion cen-

ters nucleophilically activate water molecules to dephos-

phorylate substrates in a single reaction step.

e. PP2A Is Structurally Variable and

Functionally Diverse

PP2A participates in a wide variety of regulatory

processes including those governing metabolism, DNA

replication, transcription, and development. It consists of

three different subunits:

1. An ⬃36 kD catalytic subunit (C), whose N-terminal

catalytic domain contains the ⬃280-residue catalytic core

common to all PPP family members. Its C-terminal re-

gulatory domain contains an activating binding site for

2

–calmodulin, an inactivating Tyr phosphorylation site

that is targeted by a variety of PTKs including the EGF and

insulin receptors, and a C-terminal autoinhibitory tail. The

C subunit is highly conserved from yeast to mammals.

2. An ⬃65 kD scaffold subunit (A; also called PR65),

with which the C subunit is tightly associated in the cell.

3. One of four dissimilar types of regulatory subunits

(B, B¿,B–, and B9–¿) that bind to both the A and C subunits

and, to a large extent, control the substrate specificity of

PP2A.

All of PP2A’s subunits have multiple isoforms and

splice variants that are expressed in a tissue-specific and

developmentally specific manner, thereby generating an

enormous panoply of enzymes that are targeted to differ-

ent phosphoproteins in distinct subcellular sites. This com-

plexity is a major cause of our limited understanding of

how PP2A carries out its diverse cellular functions, even

though it comprises between 0.3 and 1% of cellular pro-

teins and, together with PP1, accounts for 90% of the

Ser/Thr phosphatase activity in most cells.

The X-ray structure of PP2A’s A subunit (Fig. 19-51a),

determined by David Barford, reveals a remarkable sole-

noidal protein that consists of 15 imperfect tandem repeats

of a 39-residue sequence termed HEAT (because it occurs

in proteins named Huntingtin, EF3, A subunit of PP2A,

and TOR1). Successive HEAT repeats, which each consist

of two antiparallel helices joined by a short linker, stack on

one another with their corresponding helices nearly paral-

lel so as to form an ⬃100-Å-long right-handed superhelix

(helix of helices) with a hooklike shape.

The X-ray structure of a PP2A holoenzyme (complete

enzyme; Fig. 19-51b), determined independently by Yigong

Shi and Wenqing Xu, reveals, unexpectedly, that its regu-

latory subunit consists of 8 tandem HEAT-like repeats

arranged like those of the A subunit, despite their lack of

722 Chapter 19. Signal Transduction

JWCL281_c19_671-743.qxd 3/16/10 7:17 PM Page 722

sequence similarity. The C subunit binds to the A subunit’s

concave surface along a ridge of conserved hydrophobic

side chains spanning HEAT repeats 11 to 15. The regula-

tory subunit similarly interacts with the A subunit’s HEAT

repeats 2 to 8 and also binds to the C subunit via a ridge

spanning its own HEAT-like repeats 6 to 8. The highly

acidic, convex side of the regulatory subunit (lower part of

Fig. 19-51b) is thereby left unoccupied, which suggests that

it interacts with substrate proteins. The C subunit is struc-

turally similar to the catalytic subunits of PP1 and PP2B.

Section 19-3. Tyrosine Kinase–Based Signaling 723

(a)

(b)

Figure 19-51 X-ray structure of protein phosphatase PP2A.

(a) The structure of an isolated scaffold (A) subunit. HEAT

repeats, which are drawn here in different colors, each consist of

two antiparallel helices joined by a short linker.These stack on

one another with their corresponding helices nearly parallel to

form an ⬃100-Å-long right-handed superhelix (helix of helices)

with a hooklike shape. Compare this structure to that of a

portion of human ankyrin (Fig. 12-39), which also forms a

right-handed solenoid, although consisting of ankyrin repeats.

[Courtesy of Bostjan Kobe, St. Vincent’s Institute of Medical

Research, Fitzroy,Victoria, Australia. X-ray structure by David

Barford, University of Oxford, U.K. PDBid 1B3U.] (b) The

structure of a PP2A heterotrimer viewed with the scaffold

subunit oriented approximately as in Part a. Here the scaffold

(A; 589 residues) and regulatory (B¿; 449 residues) subunits are

drawn in worm form, each colored in rainbow order from its

N-terminus (blue) to its C-terminus (red). In addition, the A

subunit is embedded in its transparent molecular surface. The

catalytic (C; 309 residues) subunit (magenta) is drawn in ribbon

form. Note the close structural resemblance of the A and B¿

subunits. [Based on an X-ray structure by Yigong Shi, Princeton

University. PDBid 2NPP.]

JWCL281_c19_671-743.qxd 3/16/10 7:17 PM Page 723

f. PP2B Is the Target of Immunosuppressant Drugs

PP2B, which is also known calcineurin (CaN), is unique

among protein Ser/Thr phosphatases in that it is activated

by Ca

2⫹

. CaN is a heterodimer composed of a catalytic A

subunit (CaNA) and a regulatory B subunit (CaNB).

CaNA contains an N-terminal catalytic domain followed

by a CaNB-binding domain, a calmodulin (CaM)-binding

domain, and a C-terminal autoinhibitory segment. CaNB,

which has 35% sequence identity with CaM, binds four

2⫹

ions via its four EF hand motifs (Section 18-3Ce).

CaN is activated by the binding of Ca

2⫹

to CaNB and

2⫹

–CaM to CaNA.

Calcineurin plays an essential role in the antigen in-

duced proliferation of T cells. As we discuss in Section

35-2D, the binding of an antigenic peptide to a T cell recep-

tor, a tyrosine kinase–associated receptor, initiates a com-

plex series of signaling events involving the Src-like PTKs

Lck and Fyn, a MAP kinase cascade,and the phosphoinosi-

tide cascade (Section 19-4), which, among other things,

releases Ca

2⫹

into the cytosol. The Ca

2⫹

, in turn, activates

CaN to dephosphorylate the transcription factor NFAT

(for nuclear factor of activated T cells). NFAT

in complex

with CaN is then translocated to the nucleus where, in

concert with other transcription factors, it induces some of

the early steps in T cell proliferation.

As we discussed Section 9-2B, the fungal products cy-

closporin A (CsA) and FK506 are highly effective im-

munosuppressants that are in clinical use for the preven-

tion of organ-transplant rejection and for the treatment of

autoimmune disorders (processes that are mediated by

T cells). CsA and FK506, respectively, bind to the peptidyl

prolyl cis–trans isomerases (rotamases) cyclophilin and

FK506 binding protein (FKBP12), which are therefore col-

lectively known as immunophilins. However, the observa-

tion that both CsA and FK506 (also known as tacrolimus)

are effective immunosuppressants at concentrations far be-

low those of the immunophilins suggests that it is the pres-

ence of the cyclophilin ⴢ CsA and FKBP12 ⴢ FK506 com-

plexes themselves, rather than the inhibition of their

rotamase activity, that interferes with T cell proliferation. It

is, in fact, the binding of either complex to CaN that pre-

vents it from dephosphorylating NFAT

and thereby sup-

presses T cell proliferation.

The X-ray structures of the bovine complex FKBP12 ⴢ

FK506–CaN, by Manuel Navia, and the corresponding hu-

man complex, by Ernest Villafranca, reveal how the

FKBP12 ⴢ FK506 complex binds to CaN (Fig. 19-52a). The

catalytic domain of CaNA, with its binuclear Fe

2⫹

–Zn

2⫹

center marking its active site, resembles those of other

protein Ser/Thr phosphatases of known structure. A

22-residue ␣ helix at the C-terminal end of this phosphatase

domain, which extends out from the phosphatase domain

by up to 40 Å, provides much of the CaNB-binding site.Be-

yond this helix, the C-terminal portion of CaNA, which

contains the CaM-binding site and the autoinhibitory seg-

ment, is not visible due to disorder. However, in the X-ray

structure of CaN alone, the autoinhibitory segment is seen

to bind in the CaNA active site so as to block the access of

substrate phosphoproteins. The structure of CaNB, which

724 Chapter 19. Signal Transduction

Figure 19-52 Calcineurin. (a) The X-ray structure of bovine

FKBP12 ⴢ FK506–CaN. The CaNA subunit is orange, the CaNB

subunit is blue, and the FKBP12 is cyan.The FK506 is drawn in

stick form with C green, N blue, and O red; the Fe

2⫹

and Zn

2⫹

ions in the CaNA active site are, respectively, represented by red

and green spheres; and the four Ca

2⫹

ions bound to CaNB are

represented by pale green spheres. [Based on an X-ray structure

by Manual Navia,Vertex Pharmaceuticals, Cambridge,

(b)

Massachusetts. PDBid 1TCO.] (b) The X-ray structure of human

cyclophilin ⴢ CsA–CaN. The CaN and its bound metal ions are

represented and oriented as in Part a, the cyclophilin is magenta,

and the CsA is drawn in stick form with C green, N blue, and O

red. Note the close resemblance between the two structures.

[Based on an X-ray structure by Hengming Ke, University of

North Carolina, Chapel Hill, North Carolina. PDBid 1M63.]

(a)

JWCL281_c19_671-743.qxd 6/11/10 6:58 AM Page 724

has four bound Ca

2⫹

ions, resembles that of Ca

2⫹

–CaM in

complex with a helical target peptide (Fig. 18-19) except

that CaNB’s two globular domains are on the same side of

its bound peptide rather than on opposite sides as are those

of Ca

2⫹

–CaM. CaNB thereby forms a continuous groove in

which the CaNA helix binds.

FKBP12 ⴢ FK506, a mixed inhibitor of CaN (Section

14-3C), binds to it so as to contact both CaNA and CaNB,

with the portion of FK506 that extends out from its

FKBP12-binding site forming a significant part of this in-

terface. The structures of FKBP12 and CaN in this com-

plex closely resemble those in the X-ray structures of

these proteins alone. It therefore appears that FK506 pro-

vides a critical component of this contact. Nevertheless, no

part of the FKBP12 ⴢ FK506 complex is within 10 Å of

CaN’s phosphatase site (although the CaN autoinhibitory

segment has been displaced). This accounts for the obser-

vations that FKBP12 ⴢ FK506 strongly inhibits CaN from

dephosphorylating a 20-residue phosphopeptide but actu-

ally increases the rate at which CaN dephosphorylates the

much smaller p-nitrophenylphosphate by a factor of 3.

The X-ray structure of human cyclophilin ⴢ CsA–CaN

(Fig. 19-52b), independently determined by Hengming Ke

and Harrison, is strikingly similar to that of FKBP12 ⴢ

FK506–CaN despite the fact that there is no structural re-

semblance of cyclophilin ⴢ CsA to FKBP12 ⴢ FK506. Cy-

clophilin binds to essentially the same region of CaNA as

does FKBP12 with CsA forming an essential component of

this contact, thereby limiting access to the CaN active site

in the same way as does FKBP12 ⴢ FK506. However, the

characteristics of rotamases that apparently uniquely suit

them for their roles in CaN inhibition remain enigmatic.

4 THE PHOSPHOINOSITIDE CASCADE

Extracellular signals often cause a transient rise in the cytoso-

lic [Ca

2⫹

], which, in turn, activates a great variety of enzymes

through the intermediacy of calmodulin and its homologs.

An increase in cytosolic [Ca

2⫹

] triggers such diverse cellular

processes as glycogenolysis (Section 18-3Ce) and muscle

contraction (Section 35-3C).What is the source of this Ca

2⫹

and how does it enter the cytosol? In certain types of cells,

neurons (nerve cells; Fig. 1-10d), for example, the Ca

2⫹

orig-

inates in the extracellular fluid. However, the observation

that the absence of extracellular Ca

2⫹

does not inhibit

certain Ca

2⫹

-mediated processes led to the discovery that,

in these cases, cytosolic Ca

2⫹

is obtained from intracellular

reservoirs, mostly the endoplasmic reticulum (and its equiv-

alent in muscle, the sarcoplasmic reticulum). Extracellular

stimuli leading to Ca

2⫹

release must therefore be mediated

by an intracellular signal.

The first clue as to the nature of this signal came from

observations that the intracellular mobilization of Ca

2⫹

and the turnover of phosphatidylinositol-4,5-bisphosphate

(PIP

or PtdIns-4,5-P

; Fig. 19-53), which occurs mainly in

the plasma membrane as a minor (⬍1%) component of its

cytosolic leaflet, are strongly correlated. This information

led Robert Michell to propose, in 1975, that PIP

hydrolysis

is somehow associated with Ca

2⫹

release.

A. Ca

2ⴙ

, Inositol Trisphosphate, and Diacylglycerol

Are Second Messengers

Investigations, notably by Mabel and Lowell Hokin,

Michael Berridge, and Michell, eventually revealed that

PIP

is part of an important second messenger system, the

phosphoinositide cascade, that mediates the transmission

of numerous hormonal signals including those of vaso-

pressin, CRF, TRF (Section 19-1H), acetylcholine (a neuro-

transmitter;Section 20-5Cb), epinephrine (with ␣

-adrenergic

receptors; Section 19-1F), EGF, and PDGF. Remarkably,

this system yields up to three separate types of second

messengers through the following sequence of events

(Fig. 19-54):

1–3. The ligand–receptor interactions described below

activate a phosphoinositide-specific phospholipase C

(PLC; Section 19-4B) to hydrolyze PIP

to inositol-1,4,5-

trisphosphate (IP

or Ins-1,4,5-P

) and sn-1,2-diacylglycerol

(DAG or DG)

Section 19-4. The Phosphoinositide Cascade 725

Figure 19-53 Molecular formula of the phosphatidylinositides.

The head group of these glycerophospholipids is myo-inositol

that may be phosphorylated at its 3-, 4-, and/or 5-positions. R

predominantly the hydrocarbon tail of stearic acid (an 18:0 fatty

acid;Table 12-1) and R

is predominantly the hydrocarbon tail of

arachidonic acid (a 20:4 fatty acid).

⫺

OPO

⫺

OPO

⫺

HHO

OH H

⫺

PO H

OPO

⫺

OPO

⫺

HHO

PLC

⫹

DAG

PIP

PCH

OCR

CHCR

⫺

HHO

OY H

Y ⫽ H or PO

⫺

JWCL281_c19_671-743.qxd 6/30/10 1:49 PM Page 725

[the stereospecific numbering (sn) system is described in

the legend of Fig. 12-3]. PLCs catalyze the hydrolysis of the

bond linking a glycerophospholipid to its phosphoryl group

as indicated in Fig. 19-55 (which also shows the actions of

other types of phospholipases). Note that this reaction oc-

curs at the interface between the aqueous phase and the

membrane in such a way that both PIP

and its amphi-

pathic hydrolysis product DAG remain associated with the

membrane during the catalytic reaction.

4. The water-soluble IP

, acting as a second messenger,

diffuses through the cytoplasm to the ER, from which it

stimulates the release of Ca

2

into the cytoplasm by bind-

ing to and thereby opening an ER-bound transmembrane

2

-specific ion channel known as the IP

receptor (ion

channels are discussed in Chapter 20).

726 Chapter 19. Signal Transduction

Protein kinase

C (PKC)

inositol

polyphosphate

5-phosphatase

Protein

(inactive)

Protein

(inactive)

Protein

(active)

External

signal

3 3 6

qα

GDP

GTP GDP

qα

GTP

ATP

ADP

– CaMCa

-gated Ca

transport channel

ATP

GDP

+ P

PIP

DAGDAG PS

Cellular

response

Plasma membrane

Cytosol

Endoplasmic reticulum

membrane

kinase

Phospho-

lipase C

(PLC)

Figure 19-54 Role of PIP

in intracellular signaling. (1) The

binding of a ligand to a cell-surface receptor, R, activates a

phosphoinositide-specific phospholipase C through the

intermediacy of what is shown here as (2) a G protein (G

;Fig.

19-17) but in many cases is an RTK, an NRTK, or, possibly, Ca

2

Phospholipase C catalyzes the hydrolysis of PIP

to IP

and DAG

(3). The water-soluble IP

stimulates the release of Ca

2

sequestered in the endoplasmic reticulum (4), which in turn

Figure 19-55 A phospholipase is named according to the bond

that it cleaves on a glycerophospholipid. X is a phosphoinositol

group for this discussion.

activates numerous cellular processes through the intermediacy

of calmodulin and its homologs (5). The nonpolar DAG remains

associated with the inner leaflet of the membrane, where it

activates protein kinase C to phosphorylate and thereby

modulate the activities of a number of cellular proteins (6). This

latter activation process also requires the presence of the

membrane lipid phosphatidylserine (PS) and Ca

2

. See the

Animated Figures

PCH

OCR

CHCR



phospholipase A

phospholipase C phospholipase D

JWCL281_c19_671-743.qxd 3/16/10 7:17 PM Page 726

5. The Ca

2

, in turn, stimulates a variety of cellular

processes, mainly through the intermediacy of calmodulin

and its homologs.

6. The amphipathic DAG is constrained to remain in

the inner leaflet of the plasma membrane, where it never-

theless also acts as a second messenger by activating pro-

tein kinase C (PKC; Section 19-4C) in the presence of Ca

2

and phosphatidylserine (PS; which is located exclusively on

the cytosolic face of the plasma membrane).This membrane-

bound enzyme (actually a family of enzymes; Section 19-

4C), in turn, Ser/Thr-phosphorylates and thereby modu-

lates the activities of several different proteins including

glycogen synthase (Sections 18-3D). DAG, which predomi-

nantly has a stearoyl group at its 1-position and an arachi-

donoyl group at its 2-position, is further degraded in some

cells by cytosolic phospholipase A

(cPLA

) to yield

arachidonate, the major substrate for the biosynthesis of

prostaglandins, prostacyclins, thromboxanes, leukotrienes,

and lipoxins. These paracrine hormones, as we discuss in

Section 25-7, mediate or modulate a wide variety of physi-

ological functions.

B. The Phospholipases C

The phosphoinositide-specific PLCs in mammals are clas-

sified in six families comprising 13 isozymes according to

their sequences and modes of regulation: (1–4), (1–2),

(1, 3, 4), ε, (1–2), and  (Fig. 19-56; the isozyme origi-

nally named PLC- is actually a proteolytic fragment of

PLC-1), several of which have splice variants.All of these

PLCs require the presence of Ca

2

ion for enzymatic

activity. The PLC- isozymes (⬃760 residues) consist of,

from N- to C-terminus, an ⬃120-residue pleckstrin homol-

ogy (PH) domain (Section 19-3Ce); an ⬃130-residue EF

hand domain that contains four EF hand motifs (Fig.

18-18); two conserved regions known as X and Y that

together form the ⬃250-residue PLC catalytic domain and

which are separated by an ⬃60-residue linker; and an

⬃120-residue C2 domain, a domain that in many cases

binds Ca

2

and which occurs in ⬃650 human proteins that

mainly participate in signal transduction and membrane

interactions. The PLC- isozymes (⬃1200 residues) have

an additional ⬃420-residue C-terminal tail that has been

implicated in both membrane association and regulation

by G proteins (see below). In contrast, the PLC-

isozymes (⬃1270 residues) contain an ⬃420-residue insert

between X and Y that consists of an additional PH domain

that is split by two SH2 domains that are implicated in

binding to activated PTKs (see below) as well as an SH3

domain. PLC-ε (⬃2300 residues), which is activated by

several Ras-related G-proteins, has an N-terminal RasGEF

domain and two C-terminal Ras-binding (RA) domains.

PLC- (⬃1000 residues), which occurs only in nerve

tissue, contains a C-terminal Ser/Pro (S/P)-rich region of

unknown function. PLC- (⬃650 residues), the smallest of

the PLCs and the only one that lacks a PH

domain, occurs only in sperm. The observation that the

PLCs in plants and lower eukaryotes such as yeast are

-like suggests that the various PLC isozymes in mammals

evolved from a primordial PLC-.

a. G

q

 GTP, G



, and Rac  GTP Activate PLC-s by

Bringing Them Into Contact with the Membrane

The PLC-s are hormonally regulated by certain G

protein-coupled receptors (e.g., those for histamine, vaso-

pressin, TSH, thromboxane A

, and angiotensin II) via

their associated heterotrimeric G proteins, as indicated in

Fig. 19-54. In particular,they are activated through their in-

teractions with the  subunits of the G

subfamily (Section

19-2C) in complex with GTP. G

q

 GTPS activates the

PLC- isoforms in order of potency 1 3 2, with the

position of 4 in this hierarchy indeterminate because it is

inhibited by GTPS. Moreover, an important aspect of the

regulation of PLC-s by G

q

 GTP is that the PLC-s

function as GAPs to increase the GTPase activity of G

q

50-fold, thereby limiting the activating function of G

q

The PLC- isoforms are independently activated by G



complexes, which may be supplied by the dissociation of

heterotrimeric G proteins other than G

q



. Moreover,

their order of efficacy with G



differs from that with

q

 GTP: 3 2 1, with 4 insensitive to the pres-

ence of G



. Although the concentration of G



required

for maximal activation of the PLC-s is much greater than

that of G

q

 GTP, their final extents of activation are simi-

lar. The sites on PLC-2 that interact with G



are its PH

domain and a 10-residue segment near the N-terminus of

its Y region. The region of G



that interacts with PLC-s

overlaps the region through which it binds G



subunits,

thereby explaining why a G



cannot simultaneously bind a

PLC- and a G



 GDP.

The Rho family small G protein named Rac1, in complex

with GTP, activates PLC-2,-3,and -2.The X-ray structure

of a complex of PLC-2 lacking its C-terminal tail (but

still catalytically active in vitro) with Rac1  GTPS was

Section 19-4. The Phosphoinositide Cascade 727

Figure 19-56 Domain organization of the six classes of

phosphoinositide-specific PLCs.

SH2 SH2 SH3

RasGEF X

PLC-δ (1,3,4)

PLC-β (1–4)

PLC-γ (1–2)

PLC-ε

PH EF-hand C2

PLC-η (1–2)

PLC-ζ

JWCL281_c19_671-743.qxd 3/16/10 7:17 PM Page 727

determined by John Sondek. It reveals (Fig. 19-57) that PLC-

␤2’s remaining four domains are linked by extended

polypeptide segments that form tight interactions with these

domains. EF hand motifs 1 and 2 form a lobe that is closely

superimposable on the N- and C-terminal lobes of calmod-

ulin (Figs. 18-17 and 18-19),as do EF hands 3 and 4.Although

EF hands are usually associated with Ca

2⫹

ions, the residues

necessary to do so are poorly conserved in the PLCs. Thus,

although mutations to this domain are generally deleterious

to a PLC’s catalytic activity, its function is unclear.

The X and Y regions of PLC-␤2’s catalytic domain form

a TIM (or ␣/␤) barrel (Section 8-3B). The X region con-

tributes a typical ␤␣␤␣␤␣␤ motif to form nearly half the

barrel with the Y region contributing its remaining portion.

The visible parts of the partially disordered ⬃70-residue

peptide linking X and Y occlude the enzyme’s active site,

which as is the case for all TIM barrel enzymes, is located in

the mouth of the barrel at the C-terminal ends of its

␤ strands. Deletion of this apparently autoinhibitory X–Y

linker increases the enzyme’s phospholipase activity by

⬃10-fold, although it can still be further activated by vari-

ous G proteins. Although the X–Y linkers of the several

PLCs share little sequence identity, all of them contain

clusters of negatively charged residues. This suggests that

when a PLC approaches the negatively charged cytosolic

face of a membrane, its X–Y linker is electrostatically re-

pelled from the PLC’s active site, thereby permitting these

enzymes access to their membrane-bound PIP

substrates.

All PLCs require a bound Ca

2⫹

ion for catalytic activity.

The constellation of active site residues in the PLC-␤2

structure closely resemble that in the X-ray structure of

PLC-␦1 in complex with IP

and Ca

2⫹

ion. In this latter

structure, a 6-coordinate Ca

2⫹

is bound at the bottom of

the active site, with one of its ligands contributed by the

2-hydroxyl group of IP

and the remainder formed by highly

conserved Asp,Glu,and Asn side chains.The catalytic reac-

tion is therefore postulated to occur via a mechanism anal-

ogous to that of the RNase A–catalyzed hydrolysis of RNA

(Fig. 15-3) in which PIP

’s 2-hydroxyl group nucleophili-

cally attacks the neighboring 1-phosphate group to form

DAG and a cyclic phosphodiester intermediate that is sub-

sequently hydrolyzed to yield IP

. The Ca

2⫹

ion (rather

than a His side chain as in RNase A) is properly positioned

to promote the deprotonation of the 2-hydroxyl group so

as to enhance its nucleophilicity and to subsequently help

stabilize the developing negative charge on the pentava-

lent phosphorus in the catalytic reaction’s transition state

(Fig. 16-6b). This explains why 2-deoxy-PIP

is not hy-

drolyzed by mammalian phosphinositide-specific PLCs.

Rac1 ⴢ GTP␥S binds to the PH domain of PLC-␤2 exclu-

sively via Rac1’s switch regions (Fig. 19-57), which explains

why Rac1 ⴢ GDP does not bind to PLC-␤2. However, compar-

ison of the X-ray structure of PLC-␤2 alone with that of its

Rac1 ⴢ GTP␥S complex indicates that the binding of

Rac1 ⴢ GTP␥S does not alter the conformation of PLC-␤2.

How, then, does Rac1 ⴢ GTP␥S activate PLC-␤2? Most PH

domains bind membrane-associated phosphoinositides (see

below) but those of PLC-␤2 and -␤3 lack the residues neces-

sary to do so. Instead, Rac1’s C-terminal geranygeranyl mem-

brane anchor (Section 12-3Ba) functions to localize PLC-␤2

and probably PLC-␤3 to the cytosolic face of the plasma mem-

brane (Fig. 19-57b), where they can efficiently hydrolyze phos-

phinositides. Similarly, the membrane anchored G

q␣

ⴢ GTP

and G

␤␥

serve to localize PLC-␤2 to the membrane.

The C2 domain consists of a sandwich of two 4-stranded

antiparallel ␤ sheets with variable interstrand loops. The

X-ray structure of PLC-␦1 complexed with the calcium

analog lanthanum revealed that the C2 domain contains

three Ca

2⫹

ions sites in close proximity to one another.All

of these metal ions lie in a crevice at one end of the ␤ sand-

wich, where they are exposed on the surface of the enzyme.

It therefore seems likely that, in vivo, they associate with

anionic head groups such as that of phosphatidylserine on

the surface of a membrane. Since the extensive interface

between the C2 and catalytic domain appears to be rigid,

this interaction probably helps bind the catalytic domain to

the membrane such that it can productively interact with

728 Chapter 19. Signal Transduction

(a)

(b)

Figure 19-57 X-ray structure of PLC-␤2 lacking its

C-terminal tail in complex with Rac1 ⴢ GTP␥S. The pleckstrin

homology (PH) domain of PCL-␤2 is blue, its EF hand (EF)

domain is orange, its TIM (␣/␤) barrel is red, and it’s C2 domain

is yellow. The Rac1 is green with its switch regions (Sw1 and

Sw2) red and its bound GTP␥S drawn in stick form in cyan. (a)

View toward the cytosolic surface of the membrane. (b) View

rotated 90° about the horizontal axis relative to Part a.The

geranylgeranyl group at the C-terminus of Rac1 and the PIP

substrate (black sticks) are modeled as sites of membrane

attachment. Note that the three loops of the TIM barrel, which

form a hydrophobic ridge, are inserted into the membrane.

[Courtesy of John Sondek, University of North Carolina School

of Medicine. PDBid 2FJU.]

JWCL281_c19_671-743.qxd 6/4/10 10:55 AM Page 728

PIP

molecules. This association appears to be supple-

mented through the interactions of a hydrophobic ridge

comprised of three loops from one side of the active site

opening that is postulated to penetrate into the mem-

brane’s nonpolar region during catalysis (Fig. 19-57b). This

would explain how the enzyme can catalyze the hydrolysis

of PIP

to DAG and IP

while the former two compounds

remain associated with the membrane.

The C-terminal tail that is unique to PLC-␤ isoforms

(Fig. 19-56) has been implicated, via studies of truncated

enzymes, in binding to G

q␣

ⴢ GTP. The X-ray structure of

this ⬃420-residue segment reveals that it consists almost

entirely of three long helices which form a coiled coil that

dimerizes along its long axis. It contains a large number of

basic residues that are clustered on one face of the dimer

and whose mutation results in reduced responses to

q␣

ⴢ GTP. This positively charged surface is likely to inter-

act with acidic phospholipids so as to recruit its attached

PLC-␤ to the membrane. It therefore appears that the acti-

vation of PLC-bs is almost entirely the result of their close

association with the negatively charged membrane.

b. The Pleckstrin Homology Domain Tethers

PLC-␦1 to the Membrane

The X-ray structure of PLC-␦1’s N-terminal PH domain

in complex with IP

, determined by Joseph Schlessinger

and Sigler,reveals that the IP

binds to a positively charged

surface of the protein (Fig. 19-58). This is consistent with

the PH domain’s proposed role as a membrane anchor, as

are the observations that this PH domain binds PIP

with

much greater affinity (K

⫽ 1.7 ␮M) than does PLC-␦1’s

catalytic domain (K

⬎ 0.1 mM). Since the peptide seg-

ment linking the PH domain to the rest of the enzyme is

probably flexible, it appears that the PH domain in PLC-␦1

functions to tether the enzyme to the membrane. This ac-

counts for kinetic measurements indicating that the en-

zyme catalyzes multiple cycles of PIP

hydrolysis without

releasing the membrane.

Despite the foregoing, the way in which the PLC-␦

isozymes are regulated is poorly understood. The higher

sensitivities of the PLC-␦ isozymes to Ca

2⫹

ion compared

to those of the other PLCs suggests that the PLC-␦

isozymes are regulated by changes in intracellular [Ca

2⫹

Thus, the activation of PLC-␦ isozymes may occur second-

arily to the receptor-mediated activation of other PLC

isozymes through their induction of the opening of Ca

2⫹

channels (Fig. 19-54).

c. Protein Tyrosine Kinases Recruit PLC-␥

Isozymes to the Membrane

The PLC-␥ isozymes in a wide variety of cells are acti-

vated by certain protein growth factors including EGF,

PDGF, FGF, and NGF. These growth factors cause their

corresponding receptors, which are RTKs (Section 19-3A), to

autophosphorylate at particular Tyr residues. Some of

these phosphoTyr sites are specifically bound,as mutational

studies indicate, by the more N-terminal SH2 (N-SH2) do-

main on PLC-␥1 (Figs. 19-56 and 19-27a) but not its C-SH2

domain. The N-SH2 domain binds peptides containing a

phosphoTyr residue followed by at least 5 predominantly

hydrophobic residues, in contrast to the SH2 domain of

Src, which preferentially binds pYEEI containing peptides

(Section 19-3Cb).

The activated receptors for all four of the above growth

factors phosphorylate PLC-␥1 at the same three Tyr

residues, 771, 783 (located between the C-SH2 and SH3 do-

mains), and 1254 (located in the C-terminal tail). In fact,

mutating Tyr 783 to Phe completely blocks the activation of

PLC-␥1 by PDGF, although this mutant PLC-␥1 still asso-

ciates with the PDGF receptor. Conversely, mutating

certain RTK autophosphorylation sites (e.g., the PDGF

receptor’s Tyr 1021) disrupts their binding of PLC-␥1 and

hence its activation, even though these mutant receptors

catalyze detectable levels of growth factor–dependent Tyr

phosphorylation on PLC-␥1. Evidently, growth factor–

induced activation of PLC-␥1 requires both the activating

Tyr phosphorylation of PLC-␥1 and its association with the

growth factor receptor, the latter presumably bringing

PLC-␥1 into contact with its PIP

substrate in the cytosolic

leaflet of the plasma membrane. The PLC-␥ isozymes may

also be activated by NRTKs, such as members of the Src

and JAK families (all of which are membrane associated),

that have been activated by tyrosine kinase–associated re-

ceptors (Section 19-3E). The function of the PLC-␥ SH3

domain is unclear.

Section 19-4. The Phosphoinositide Cascade 729

Figure 19-58 X-ray structure of the pleckstrin homology domain

of PLC-␦1 in complex with PIP

. The peptide chain is shown in

ribbon form colored in rainbow order from its N-terminus (blue)

to its C-terminus (red).The PIP

is drawn in stick form with

C green, O red, and P orange.The PH domain consists largely of

the ␤ barrel/ sandwich of 7 antiparallel strands and C-terminal

␣ helix common to the numerous PH domains of known structure.

[Based on an X-ray structure by Joseph Schlessinger, New York

University Medical Center, and Paul Sigler,Yale University.

PDBid 1MAI.]

JWCL281_c19_671-743.qxd 6/11/10 6:58 AM Page 729

d. PLC-d Is Activated by Ras  GTP

The presence of RasGEF and Ras-binding (RA) do-

mains on PLC-

ε suggests that PLC-ε is activated by

Ras  GTP. This is, in fact, the case, as indicated by the ob-

servation that PLC-

ε binds Ras  GTP with high affinity

but does not bind Ras  GDP. Since Ras is membrane-

anchored, this interaction brings PLC-

ε into proximity with

the membrane, much as we saw with the Rac1  GTP acti-

vation of PLC-2 (Section 19-4Ba). Although the growth

factor–induced activation of Ras is terminated by the hy-

drolysis of its bound GTP to GDP, the resulting Ras  GDP

may be rapidly converted to Ras  GTP by the RasGEF do-

main of PLC-

ε, thereby prolonging the receptor-mediated

activation of PLC-

ε. PLC-ε may also be activated by G

12

C. The Protein Kinases C

Protein kinase C (PKC), a member of the AGC family of

protein kinases (AGC for PKA,PKG, and PKC), is the

Ser/Thr protein kinase that transduces the numerous signals

mediated by the release of DAG (Fig. 19-54). In mammals, it

comprises a family of ten ⬃700-residue monomeric isozymes

classified in three subfamilies: the “conventional” PKCs (,

I, II, and , of which I and II are splice variants of the

same gene), the “novel” PKCs (,

ε, , and ), and the “atypi-

cal” PKCs [ and ι(human)/(mouse)]. The conventional

PKCs, which are activated by both DAG and Ca

2

, each con-

sist of an N-terminal autoinhibitory pseudosubstrate (which

resembles the enzyme’s target peptide but with the Ser/Thr

phosphorylation site replaced by Ala) followed by four con-

served domains, C1 through C4 (C1 for conserved region 1 of

PKC, etc.). The DAG-binding C1 domain, which occurs in

50 other proteins including Raf (in which it does not bind

DAG), consists of two tandemly repeated ⬃50-residue Cys-

rich motifs, C1A and C1B. However, only C1B binds DAG.

C2, which often binds Ca

2

ion, is also a component of PLC

(Fig. 19-57) as well as numerous other signaling proteins. C3

and C4 form the N- and C-terminal lobes of the protein ki-

nase, which is similar in sequence and structure to the cat-

alytic subunit of PKA (Fig. 18-15). The protein kinase is

maintained in its inactive state through its binding of the

pseudosubstrate (as with MLCK; Section 18-3Ce).The novel

PKCs, which are activated by DAG but not by Ca

2

, resem-

ble the conventional PKCs except that their C2 domains do

not bind Ca

2

.The atypical PKCs, which are unresponsive to

both DAG and Ca

2

,have only one Cys-rich motif in their C1

domains and lack C2 domains.

In addition to their different regulatory properties, the

various PKCs are localized to different subcellular com-

partments (e.g., plasma membrane, nuclear membrane, en-

doplasmic reticulum, Golgi apparatus, mitochondria) in

ways that vary with the cell type and external stimuli. This

specificity is provided both by targeting sequences on the

PKC as well as by scaffolding proteins that localize individ-

ual PKCs to specific membrane microdomains in close

proximity to their substrate and regulatory proteins. For

example, the members of a family of membrane-associated

proteins known as RACKs (for receptors of activated

C-kinases) each anchor a specific activated PKC to a par-

ticular subcellular location.

a. The C1 and C2 Domains Anchor PKC to the

Plasma Membrane

Phorbol esters such as 12-O-myristoylphorbol-13-acetate

(which occurs in croton seed oil and was used as a drastic

purgative in folk medicine) are potent activators of protein

kinase C; they structurally resemble DAG but bind to PKC

with ⬃250-fold greater affinity. Consequently, phorbol

esters are the most effective known tumor promoters,

substances that are not in themselves carcinogenic but

increase the potency of known carcinogens. They do so by

inhibiting the apoptosis (programmed cell death; Section

34-4E) and stimulating the proliferation of precancerous

cells (cells with only some of the mutations that make them

cancerous but which nevertheless reduce the cell’s sensitiv-

ity to apoptotic signals and increase it for proliferative

signals) as well as stimulating metabolic processes that

generate carcinogenic agents (e.g., free radicals). This in-

creases the probability that such a cell will complete its ma-

lignant transformation. Not surprisingly, therefore, altered

PKC activity is associated with various types of cancers.

The X-ray structure of the C1B motif of PKC in com-

plex with 12-O-myristoylphorbol-13-acetate, determined by

Hurley, reveals that this 50-residue motif is largely knit to-

gether by two Zn

2

ions, each of which is tetrahedrally lig-

anded by one His and three Cys side chains (Fig. 19-59).The

phorbol ester binds in a narrow groove between two loops

that consist mainly of nonpolar residues. Since phorbol es-

ters are also nonpolar,the entire top third of the complex, as

shown in Fig. 19-59, forms a highly conserved hydrophobic

surface. Very few soluble proteins have such a large fraction

of their surface formed by a continuous nonpolar region.

Moreover,the middle third of the protein surface,that below

the nonpolar region, forms a positively charged belt about

the protein.This suggests that, in vivo, the hydrophobic por-

tion of the complex is inserted into the nonpolar region of its

associated membrane such that the motif’s positively

charged belt interacts with the membrane’s negatively

charged head groups. This hypothesis is supported by NMR

measurements indicating that residues on the ligand-binding

portion of C1B interact with lipid.The fatty acyl group that

is esterified to phorbol’s 12-position in effective tumor

promoters presumably extends into the membrane so as to

help anchor the C1 domain to the membrane.

The comparison of this structure with that of C1B alone

indicates that C1B does not undergo significant structural

change on binding phorbol ester. Evidently, phorbol esters,

and presumably DAG, activate PKC by anchoring it to the

membrane rather than by an allosteric mechanism.The C2

O C

12-O-Myristoylphorbol-13-acetate

730 Chapter 19. Signal Transduction

JWCL281_c19_671-743.qxd 3/16/10 7:17 PM Page 730