22.11.2014 Views

Principles of cell signaling - UT Southwestern

Principles of cell signaling - UT Southwestern

Principles of cell signaling - UT Southwestern

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Ser Thr<br />

IRAK-M<br />

TRIF/<br />

TICAM-1<br />

SOCS1<br />

/JAB<br />

ys63TRAF6<br />

0<br />

CAPK<br />

CAPK<br />

IRAK1<br />

Lys<br />

TBK-1<br />

P<br />

P<br />

TRAF2 TRAF1<br />

A20<br />

P Tyr STAT6<br />

K<br />

-2<br />

TNF<br />

TICAM-1<br />

P<br />

Ser<br />

IRAK1<br />

Thr<br />

Lys<br />

Ser<br />

IRAK1<br />

Thr Lys<br />

IKK<br />

Ser176<br />

Ser181 IKK<br />

IKK<br />

IKK<br />

P<br />

Ser176<br />

Ser181 IKK<br />

P<br />

IKK<br />

GM-CSFR<br />

GM-CSFR<br />

Uev1A<br />

Ubc13<br />

P Tyr STAT5<br />

PI3K<br />

MEKK<br />

Tyr STAT5<br />

GM-CSF<br />

TLR4<br />

PP2A<br />

TAB2<br />

P<br />

P<br />

P<br />

Thr184<br />

Ser192 TAK1 Thr187 P<br />

P P<br />

TBK-1<br />

P Ser TAB1<br />

Thr<br />

P<br />

P<br />

Tyr580<br />

P<br />

TRIF/<br />

TICAM-1<br />

TRAM<br />

MD-2<br />

P P<br />

Tyr STAT5<br />

Tyr<br />

Ser386<br />

Ser385<br />

IRF-3<br />

Ser Thr IRAK1<br />

Lys<br />

Ub<br />

P<br />

P<br />

P<br />

P Tyr STAT3<br />

Tyr STAT3<br />

Thr38 Ser473<br />

Akt/PKB<br />

P<br />

P<br />

Thr38 Ser473<br />

Akt/PKB<br />

Ser Thr IRAK1 Lys<br />

Ub<br />

PKA<br />

PKA<br />

TRAM<br />

ASK<br />

Lys21<br />

Lys21<br />

I B<br />

Ser32<br />

Lys22<br />

Ser36<br />

NF-<br />

B<br />

p65+p50<br />

Ser276<br />

Ser529<br />

P<br />

Lys21<br />

I B<br />

Ser32<br />

Lys22<br />

Ser36<br />

P<br />

PKA<br />

NF-<br />

B<br />

p65+p50<br />

Ser276<br />

Ser529<br />

P<br />

Tyr701<br />

TAB2<br />

Tyr701<br />

Ser727 STAT1<br />

IKK<br />

Thr184 Ser192 TAK1 Thr187<br />

P<br />

P<br />

Ser<br />

TAB2<br />

TAB1<br />

Ser32<br />

Ser36<br />

IKK<br />

Ser Thr IRAK1<br />

Lys<br />

Ub<br />

Thr184<br />

Ser192<br />

TAK1<br />

Thr187 P<br />

SEK1/MKK4<br />

P<br />

Tyr701<br />

Ser727<br />

STAT1<br />

P<br />

P<br />

Tyr STAT2<br />

IRF-9<br />

Thr<br />

Ser386<br />

P<br />

P<br />

CK II<br />

Tyr JAK1<br />

IL-10R<br />

Tyr<br />

Tyr JAK1<br />

Ser<br />

IL-10R<br />

TAB1<br />

NF- B<br />

p65+p50<br />

Ser276 Ser529<br />

IL-10R<br />

Tyr<br />

NF- B<br />

p65+p50<br />

Ser276 Ser529<br />

P<br />

P<br />

NF- B<br />

p65+p50<br />

Ser276 Ser529<br />

P<br />

NF-<br />

B<br />

PKA<br />

p65+p50<br />

Ser276<br />

Ser529<br />

IRF-9<br />

P<br />

P<br />

Ser<br />

Thr<br />

glycerol 3P<br />

TAB2<br />

TAB1<br />

P<br />

IL-10R<br />

Tyr<br />

Tyk2<br />

Tyr<br />

IL-10R<br />

Thr184<br />

Ser192<br />

TAK1<br />

Thr187<br />

P<br />

Ser<br />

TAB2<br />

TAB1<br />

Thr<br />

Ser Thr IRAK1<br />

Lys<br />

Ub<br />

Thr184<br />

Ser192<br />

TAK1<br />

Thr187<br />

Thr<br />

SEK2/MKK7<br />

SEK1/MKK4<br />

Tyk2 Tyr<br />

Tyr STAT5<br />

p50<br />

p60<br />

P<br />

Thr183 ERK1<br />

Tyr185ERK2<br />

P<br />

TG<br />

PKA<br />

malic enzyme<br />

P<br />

Tyr<br />

P<br />

Tyr542<br />

Tyr580 SHP-2<br />

Tyr580<br />

P<br />

lipid droplet<br />

SOCS3<br />

SHP-1<br />

Tyr JAK1<br />

Tyr759<br />

Tyr915<br />

IL-6R<br />

Tyr767<br />

Tyr905<br />

Tyr814<br />

Tyr JAK1<br />

Tyr JAK1<br />

Tyr STAT3 P Tyr STAT3<br />

Thr<br />

Thr<br />

Ub<br />

Lys21<br />

Ub<br />

IL-1ra<br />

malate<br />

Tyr580<br />

Tyr542 SHP-2<br />

P Tyr STAT5<br />

JNK<br />

JNK<br />

Tyr<br />

Tyr<br />

P<br />

NF- B<br />

p65+p50<br />

Ser276 Ser529<br />

Ser276 Ser529<br />

P<br />

P<br />

P<br />

P<br />

Ser32<br />

Ser36<br />

P<br />

P<br />

pyruvate<br />

carrier<br />

SEK2/MKK7<br />

PAFR<br />

TNF<br />

TG<br />

y<br />

Tyr<br />

P<br />

malate<br />

dehydrogenase<br />

I B<br />

GM-CSF<br />

IL-6<br />

p50<br />

IL-1<br />

oxaloacetate<br />

PKA<br />

NADH+H+<br />

Tyk2<br />

Tyr<br />

Tyr759<br />

Tyr915<br />

Tyr767<br />

IL-6R<br />

Tyr905<br />

Tyr814<br />

Tyk2<br />

Tyr<br />

Tyr759 Tyr915<br />

IL-6R<br />

Tyr767 Tyr905<br />

Tyr814<br />

Ub<br />

P<br />

Lys21<br />

I B<br />

Ser32<br />

Lys22<br />

Ser36<br />

Ub<br />

P<br />

PKA<br />

NF-<br />

B<br />

p65+p50<br />

Ser276<br />

Ser529<br />

IFN-<br />

IFN-<br />

IFNpyruvate<br />

pyruvate acetyl CoA<br />

A20<br />

P<br />

P<br />

Ser385<br />

Ser386<br />

P<br />

Ser386<br />

P<br />

NF- B<br />

p65+p50<br />

Ser276 Ser529<br />

Ser276 Ser529<br />

P<br />

P<br />

P<br />

P<br />

IL-6R<br />

gp130<br />

P<br />

P<br />

P<br />

Tyr759 Tyr915 Tyr759 Tyr915<br />

Tyr767 IL-6R IL-6R<br />

Tyr905<br />

Tyr767<br />

Tyr905<br />

P<br />

Tyr814<br />

P<br />

Tyr814<br />

P<br />

P<br />

P<br />

Tyr JAK1<br />

Tyk2<br />

Tyr<br />

P<br />

P<br />

SOCS3<br />

IL-1ra<br />

pyruvate<br />

carboxylase<br />

P<br />

Lys21<br />

I B<br />

Ser32<br />

Lys22<br />

Ser36<br />

P<br />

NF-<br />

B<br />

p65+p50<br />

Ser276<br />

Ser529<br />

IFN<br />

P<br />

P<br />

Ser385<br />

Ser386<br />

Ser386<br />

P<br />

P<br />

citrate<br />

liase<br />

PDH kinase<br />

PDH kinase<br />

P<br />

P<br />

pyruvate<br />

dehydrogenase<br />

dehydrogenase<br />

y<br />

Tyr JAK1<br />

IFN<br />

IFN<br />

Tyr440<br />

Tyr JAK1<br />

Tyk2 Tyr<br />

p38MAPK<br />

IFN-<br />

SCF<br />

R1<br />

IFN<br />

Tyr440<br />

R1<br />

Tyr701<br />

Ser727 STAT1<br />

P P<br />

Tyr STAT3<br />

Tyr<br />

P P<br />

Tyr Tyr STAT3<br />

P P<br />

Tyr Tyr STAT5<br />

TrCP<br />

UbcH5<br />

IFN<br />

acetyl CoA<br />

citrate<br />

Tyr1007<br />

R2<br />

Tyr<br />

P P<br />

Tyr Tyr<br />

STAT3<br />

SHP-1<br />

IFN<br />

R2<br />

Tyr<br />

JAK2<br />

Tyr1007<br />

PIAS3<br />

JAK2Tyr1007<br />

P<br />

P<br />

SOCS3<br />

P<br />

Tyr701<br />

Ser727 STAT1<br />

P<br />

P Tyr701 Tyr701<br />

P<br />

Ser727 STAT1<br />

Ser727<br />

P<br />

CoASH<br />

carnitine<br />

P<br />

P Tyr701<br />

Ser727 Tyr701 STAT1<br />

P<br />

P<br />

P Tyr701 Tyr701<br />

Ser727 STAT1<br />

P Ser727<br />

P<br />

carnitine<br />

acyl-CoA<br />

y<br />

P<br />

PIAS3<br />

P<br />

IL-4R<br />

Tyr<br />

IFN<br />

R1<br />

IFN<br />

R2<br />

Tyr440<br />

Tyr<br />

P<br />

P<br />

JAK2<br />

Tyr JAK1<br />

Tyr1007<br />

P<br />

SOCS1<br />

/JAB<br />

P<br />

P<br />

Thr38 Ser473<br />

Akt/PKB<br />

Thr38 Ser473<br />

Akt/PKB<br />

P<br />

IRF-9<br />

CoASH<br />

Tyr1007<br />

y<br />

P<br />

P<br />

acyl-CoA<br />

CoASH<br />

CPT I<br />

IL-4R<br />

Tyr580<br />

P<br />

P<br />

P<br />

P<br />

common<br />

Tyr<br />

chain<br />

JAK3Tyr<br />

SOCS1<br />

/JAB<br />

SOCS1<br />

/JAB<br />

SOCS1/JAB<br />

Tyr701<br />

Ser727<br />

STAT1<br />

Tyr STAT2<br />

IRF-9<br />

P<br />

Tyr STAT6<br />

IRF-2<br />

acylcarnitine<br />

acylcarnitine<br />

P<br />

J 3 y<br />

P<br />

IL-4R<br />

Tyr<br />

P Tyr STAT6<br />

Tyr<br />

TyrSTAT6<br />

MKP<br />

CACT<br />

CPT II<br />

P<br />

Tyr580<br />

Tyr542 SHP-2 SHP-1<br />

P<br />

SHIP<br />

PI3K<br />

PI3K<br />

PIAS1<br />

IRF-2<br />

IRF-2<br />

P P<br />

Tyr Tyr STAT6<br />

P P<br />

Tyr Tyr STAT6<br />

ATP<br />

ADP<br />

Tyr Fyn<br />

IRF-1<br />

y<br />

IL-4R<br />

Tyr<br />

Tyr JAK1<br />

P Tyr Fyn<br />

Tyr JAK1<br />

Tyr701<br />

Tyr701<br />

Ser727 STAT1<br />

Ser727<br />

P P Tyr701<br />

Ser727<br />

Tyr701<br />

STAT1<br />

P<br />

Ser727<br />

P<br />

PIAS1<br />

IRF-1<br />

IRF-1<br />

malonyl CoA<br />

fatty acid<br />

ATP<br />

synthetase<br />

IL-4<br />

JAK3Tyr<br />

IL-4<br />

y<br />

common<br />

Tyr<br />

chain<br />

JAK3Tyr<br />

Tyr<br />

IRS<br />

Ser312(307:R)<br />

p53<br />

NOSII/iNOS<br />

NOSII/iNOS<br />

common<br />

Tyr<br />

chain<br />

P<br />

SOCS3<br />

IRF-7<br />

Ser484<br />

Ser485 IRF-7<br />

Tyr<br />

Gab2<br />

P Tyr Gab2<br />

P<br />

Tyr<br />

IRS<br />

Ser312(307:R)<br />

P<br />

P Ser21 P<br />

Ser32 Ser374<br />

c-Fos<br />

Ser42 Ser113<br />

P Ser70 P<br />

P<br />

P<br />

P<br />

Ser63 Ser73<br />

c-Jun<br />

P<br />

P Ser21 P<br />

Ser32 Ser374<br />

c-Fos<br />

Ser42 Ser113<br />

P Ser70 P<br />

P<br />

P<br />

Tyr<br />

P<br />

P<br />

Ser63 Ser73<br />

c-Jun<br />

CPT1<br />

acyl CoA<br />

synthetase<br />

Tyr<br />

IRS<br />

Ser312(307:R)<br />

P<br />

P<br />

Ser484<br />

IRF-7<br />

Ser485<br />

P<br />

P<br />

Ser484<br />

IRF-7<br />

Ser485<br />

P<br />

calpain<br />

Tyr706<br />

Tyr544 Tyr974Tyr921<br />

M-CSFR<br />

Tyr559 Tyr807<br />

Tyr721<br />

P Tyr Gab2<br />

AP-1<br />

c-Fos+c-Jun<br />

Tyr580 Tyr542SHP-2<br />

P<br />

P<br />

Thr<br />

Tyr<br />

JNK<br />

Ser133<br />

CREB<br />

Tyr580<br />

nucleus<br />

acetyl CoA<br />

carboxylase<br />

NADPH<br />

Grb2<br />

acetyl CoA carboxylase<br />

xanthine<br />

oxidase<br />

Tyr771<br />

PLC<br />

Tyr783 Tyr1254<br />

PP2B<br />

P<br />

Thr183 ERK1<br />

Tyr185<br />

ERK2<br />

P<br />

P<br />

P<br />

P<br />

P<br />

Ser63 Ser73<br />

Ser63 Ser73<br />

c-Jun<br />

P Tyr Gab2<br />

P<br />

Tyr580 Tyr542SHP-2<br />

P<br />

P<br />

P<br />

Ser369 Thr577<br />

RSK<br />

Ser386 Ser227<br />

P<br />

xanthine<br />

P<br />

Ser<br />

SOS<br />

Thr<br />

Ser63 Ser73<br />

c-Jun<br />

c-jun<br />

P<br />

Ser133<br />

CREB<br />

fatty acid<br />

synthetase<br />

Tyr706<br />

P<br />

P<br />

fatty acid synthetase<br />

NADPH<br />

oxidase<br />

PP2A<br />

c-fos<br />

P<br />

RasGAP<br />

GTP<br />

GDP<br />

Ras<br />

Grb2<br />

Ser21<br />

Ser32 Ser374<br />

c-Fos<br />

Ser42 Ser113<br />

Ser70<br />

Grb2<br />

PI3K<br />

PP2B<br />

Ser<br />

SOS<br />

Thr<br />

acyl CoA synthetase<br />

Fc RIa<br />

chain<br />

?<br />

Ser<br />

SOS<br />

Thr<br />

proteasome<br />

MKP<br />

P<br />

Tyr771<br />

Tyr783<br />

P<br />

P<br />

Ser383<br />

Elk-1<br />

Ser389<br />

P<br />

Site-2 protease<br />

PLC<br />

Pi<br />

Ser383<br />

Ser389 Elk-1<br />

Thr183<br />

Thr183 ERK1<br />

Tyr185ERK2<br />

Tyr185<br />

. O2 -<br />

RXR<br />

RasGAP<br />

SOD<br />

chain<br />

Tyr<br />

Tyr<br />

Tyr1254<br />

P<br />

chain<br />

Tyr<br />

P<br />

Tyr518<br />

GDP<br />

SREBP1c<br />

/bHLH<br />

SREBP1c<br />

/bHLH<br />

GTP<br />

Ras<br />

P<br />

P<br />

Thr183<br />

Thr183 ERK1<br />

Tyr185ERK2<br />

Tyr185<br />

P<br />

P<br />

MPO<br />

LOOH<br />

Golgi<br />

LXR<br />

LXR<br />

Tyr341 Ser4<br />

Raf<br />

Ser338 Ser62<br />

Tyr341 Ser4<br />

Raf<br />

Ser338 Ser62<br />

LXR<br />

27-hydroxyChol<br />

SREBP1c<br />

/bHLH<br />

Tyr518<br />

Tyr518<br />

Syk<br />

Tyr519<br />

Src<br />

P<br />

Ser<br />

Grb2<br />

SOS<br />

Thr<br />

P<br />

P<br />

Ser369<br />

P<br />

P<br />

Ser369<br />

Ser386<br />

P<br />

LXR<br />

SCAP<br />

HOCl<br />

P<br />

P<br />

Thr577<br />

P<br />

R<br />

9<br />

r<br />

Site<br />

39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 589<br />

14<br />

<strong>Principles</strong> <strong>of</strong> <strong>cell</strong><br />

<strong>signaling</strong><br />

Melanie H. Cobb and Elliott M. Ross<br />

The University <strong>of</strong> Texas <strong>Southwestern</strong> Medical Center at Dallas<br />

This image represents about 10% <strong>of</strong> the map <strong>of</strong> the known <strong>signaling</strong> interactions and<br />

reactions in the mouse macrophage. Preparing such a map in a computable format is<br />

the first step in analyzing a large <strong>signaling</strong> network. This map was prepared by the group<br />

led by Hiroaki Kitano at the Systems Biology Institute, Tokyo, using their CellDesigner<br />

program. Map courtesy <strong>of</strong> Kanae Oda, Yukiko Matsuoka, and Hiroaki Kitano (The Systems<br />

Biology Institute).<br />

Ub Lys63 TRAF6<br />

Lys63TRAF6<br />

Ub<br />

Lys63TRAF6<br />

Ub<br />

P<br />

Ser385<br />

IRF-3<br />

Ub Lys63 TRAF6<br />

Lys22 I B P<br />

P<br />

Tyr542 SHP-2<br />

Lys22 I B IL-10<br />

Ser385<br />

IRF-3<br />

Lys22 I B PKA<br />

Ser385<br />

IRF-3<br />

P Ser727<br />

TyrJAK1<br />

P<br />

Tyr697<br />

Tyr706<br />

P Tyr Gab2<br />

Tyr542 SHP-2<br />

Thr JNK<br />

Tyr542 SHP-2<br />

Syk<br />

Tyr519<br />

P<br />

Tyr519 Syk<br />

Ser386 R<br />

RSK Ser227<br />

14.1<br />

14.2<br />

14.3<br />

14.4<br />

14.5<br />

14.6<br />

14.7<br />

14.8<br />

14.9<br />

14.10<br />

14.11<br />

14.12<br />

14.13<br />

14.14<br />

14.15<br />

14.16<br />

14.17<br />

14.18<br />

14.19<br />

14.20<br />

CHAPTER O<strong>UT</strong>LINE<br />

Introduction<br />

Cellular <strong>signaling</strong> is primarily chemical<br />

Receptors sense diverse stimuli but initiate a limited<br />

repertoire <strong>of</strong> <strong>cell</strong>ular signals<br />

Receptors are catalysts and amplifiers<br />

Ligand binding changes receptor conformation<br />

Signals are sorted and integrated in <strong>signaling</strong> pathways<br />

and networks<br />

Cellular <strong>signaling</strong> pathways can be thought <strong>of</strong> as<br />

biochemical logic circuits<br />

Scaffolds increase <strong>signaling</strong> efficiency and enhance<br />

spatial organization <strong>of</strong> <strong>signaling</strong><br />

Independent, modular domains specify protein-protein<br />

interactions<br />

Cellular <strong>signaling</strong> is remarkably adaptive<br />

Signaling proteins are frequently expressed as multiple<br />

species<br />

Activating and deactivating reactions are separate and<br />

independently controlled<br />

Cellular <strong>signaling</strong> uses both allostery and covalent<br />

modification<br />

Second messengers provide readily diffusible pathways<br />

for information transfer<br />

Ca2+ <strong>signaling</strong> serves diverse purposes in all eukaryotic<br />

<strong>cell</strong>s<br />

Lipids and lipid-derived compounds are <strong>signaling</strong><br />

molecules<br />

PI 3-kinase regulates both <strong>cell</strong> shape and the activation<br />

<strong>of</strong> essential growth and metabolic functions<br />

Signaling through ion channel receptors is very fast<br />

Nuclear receptors regulate transcription<br />

G protein <strong>signaling</strong> modules are widely used and highly<br />

adaptable<br />

14.21<br />

14.22<br />

14.23<br />

14.24<br />

14.25<br />

14.26<br />

14.27<br />

14.28<br />

14.29<br />

14.30<br />

14.31<br />

14.32<br />

14.33<br />

14.34<br />

14.35<br />

14.36<br />

Tyr JAK1 GM-CSFR<br />

Tyr<br />

JAK2<br />

Tyr1007<br />

Heterotrimeric G proteins regulate a wide variety <strong>of</strong><br />

effectors<br />

NADP + Cl -<br />

P Ser727 STAT1 e<br />

O 2 H 2 O<br />

Heterotrimeric G proteins are controlled P<br />

pyruvate<br />

by a regulatory<br />

-<br />

2<br />

F<br />

hypoxanthine<br />

P Tyr STAT2<br />

NAD +<br />

Fe 3+<br />

H + H +<br />

e<br />

GTPase - Tyr542 SHP-2<br />

cycle<br />

O 2<br />

e -<br />

Small, monomeric GTP-binding proteins are multiuse<br />

switches<br />

Protein phosphorylation/dephosphorylation is a major<br />

regulatory mechanism in the <strong>cell</strong><br />

Two-component protein phosphorylation systems are<br />

<strong>signaling</strong> relays<br />

Pharmacological inhibitors <strong>of</strong> protein kinases may be<br />

used to understand and treat disease<br />

Phosphoprotein phosphatases reverse the actions <strong>of</strong><br />

kinases and are independently regulated<br />

Covalent modification by ubiquitin and ubiquitin-like<br />

proteins is another way <strong>of</strong> regulating protein function<br />

The Wnt pathway regulates <strong>cell</strong> fate during development<br />

and other processes in the adult<br />

Diverse <strong>signaling</strong> mechanisms are regulated by protein<br />

tyrosine kinases<br />

Src family protein kinases cooperate with receptor<br />

protein tyrosine kinases<br />

MAPKs are central to many <strong>signaling</strong> pathways<br />

Cyclin-dependent protein kinases control the <strong>cell</strong> cycle<br />

Diverse receptors recruit protein tyrosine kinases to the<br />

plasma membrane<br />

What’s next?<br />

Summary<br />

References<br />

589


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 590<br />

(GPCR)<br />

G protein<br />

coupled<br />

receptor<br />

Heterotrimeric<br />

G protein<br />

14.1<br />

Introduction<br />

All <strong>cell</strong>s, from prokaryotes through plants and<br />

animals, sense and react to stimuli in their environments<br />

with stereotyped responses that allow<br />

them to survive, adapt, and function in<br />

ways appropriate to the needs <strong>of</strong> the organism.<br />

These responses are not simply direct physical<br />

or metabolic consequences <strong>of</strong> changes in the<br />

local environment. Rather, <strong>cell</strong>s express arrays<br />

<strong>of</strong> sensing proteins, or receptors, that recognize<br />

specific extra<strong>cell</strong>ular stimuli. In response to<br />

these stimuli, receptors regulate the activities<br />

<strong>of</strong> diverse intra<strong>cell</strong>ular regulatory proteins that<br />

in turn initiate appropriate responses by the<br />

<strong>cell</strong>. The process <strong>of</strong> sensing external stimuli and<br />

conveying the inherent information to intra<strong>cell</strong>ular<br />

targets is referred to as <strong>cell</strong>ular signal<br />

transduction.<br />

Cells respond to all sorts <strong>of</strong> stimuli. Microbes<br />

respond to nutrients, toxins, heat, light, and<br />

chemical signals secreted by other microbes.<br />

Cells in multi<strong>cell</strong>ular organisms express receptors<br />

specific for hormones, neurotransmitters,<br />

autocrine and paracrine agents (hormonelike<br />

compounds from the secreting <strong>cell</strong> or <strong>cell</strong>s<br />

Overview <strong>of</strong> major receptor types in a <strong>cell</strong><br />

Receptor<br />

protein<br />

kinase<br />

Ion<br />

channel<br />

Transcription<br />

factor<br />

Twocomponent<br />

complex<br />

(<br />

Sensor<br />

Histidine<br />

kinase<br />

Response<br />

regulator<br />

NUCLEUS<br />

Transmembrane<br />

scaffold<br />

(<br />

E1<br />

E2<br />

E1<br />

E2<br />

Guanylyl<br />

cyclase<br />

FIGURE 14.1 Receptors form a rather small number <strong>of</strong> families that share common<br />

mechanisms <strong>of</strong> action and overall similar structures.<br />

nearby), odors, molecules that regulate growth<br />

or differentiation, and proteins on the outside<br />

<strong>of</strong> adjacent <strong>cell</strong>s. A mammalian <strong>cell</strong> typically<br />

expresses about fifty distinct receptors that sense<br />

different inputs, and, overall, mammals express<br />

several thousand receptors.<br />

Despite the diversity <strong>of</strong> <strong>cell</strong>ular lifestyles<br />

and the enormous number <strong>of</strong> substances sensed<br />

by different <strong>cell</strong>s, the general classes <strong>of</strong> proteins<br />

and mechanisms involved in signal transduction<br />

are conserved throughout living <strong>cell</strong>s, as<br />

shown in FIGURE 14.1.<br />

• G protein-coupled receptors,<br />

composed <strong>of</strong> seven membrane-spanning<br />

helices, promote activation <strong>of</strong> heterotrimeric<br />

GTP-binding proteins called<br />

G proteins, which associate with the inner<br />

face <strong>of</strong> the plasma membrane and<br />

convey signals to multiple intra<strong>cell</strong>ular<br />

proteins.<br />

• Receptor protein kinases are <strong>of</strong>ten<br />

dimers <strong>of</strong> single membrane-spanning<br />

proteins that phosphorylate their intra<strong>cell</strong>ular<br />

substrates and, thus, change<br />

the shape and function <strong>of</strong> the target proteins.<br />

These protein kinases frequently<br />

contain protein interaction domains that<br />

organize complexes <strong>of</strong> <strong>signaling</strong> proteins<br />

on the inner surface <strong>of</strong> the plasma<br />

membrane.<br />

• Phosphoprotein phosphatases reverse<br />

the effect <strong>of</strong> protein kinases by removing<br />

the phosphoryl groups added<br />

by protein kinases.<br />

• Other single membrane-spanning enzymes,<br />

such as guanylyl cyclase, have<br />

an overall architecture similar to the receptor<br />

protein kinases but different enzymatic<br />

activities. Guanylyl cyclase<br />

catalyzes the conversion <strong>of</strong> GTP to 3′:5′-<br />

cyclic GMP, which is used to propagate<br />

the signal.<br />

• Ion channel receptors, although diverse<br />

in detailed structure, are usually<br />

oligomers <strong>of</strong> subunits that each contain<br />

several membrane-spanning segments.<br />

The subunits change their conformations<br />

and relative orientations to permit<br />

ion flux through a central pore.<br />

• Two-component systems may either<br />

be membrane spanning or cytosolic. The<br />

number <strong>of</strong> their subunits is also variable,<br />

but each two-component system<br />

contains a histidine kinase domain or<br />

subunit that is regulated by a <strong>signaling</strong><br />

molecule and a response regulator that<br />

590 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 591<br />

contains a phosphorylatable aspartate<br />

(Asp) residue.<br />

• Some receptors are transmembrane<br />

scaffolds that change either the conformation<br />

or oligomerization <strong>of</strong> their<br />

intra<strong>cell</strong>ular scaffold domains in response<br />

to extra<strong>cell</strong>ular <strong>signaling</strong> molecules,<br />

or ligands, and, thus, recruit<br />

interacting regulatory proteins to a common<br />

site on the membrane.<br />

• Nuclear receptors are transcription<br />

factors, <strong>of</strong>ten heterodimers, that may<br />

reside in the cytoplasm until activated<br />

by agonists or may be permanently located<br />

in the nucleus.<br />

The biochemical processes <strong>of</strong> signal transduction<br />

are strikingly similar among <strong>cell</strong>s.<br />

Bacteria, fungi, plants, and animals use similar<br />

proteins and multiprotein modules to detect<br />

and process signals. For example, evolutionarily<br />

conserved heterotrimeric G proteins and G<br />

protein-coupled receptors are found in plants,<br />

fungi, and animals. Similarly, 3′:5′ cyclic AMP<br />

(cAMP) is an intra<strong>cell</strong>ular <strong>signaling</strong> molecule<br />

in bacteria, fungi, and animals; and Ca2+ serves<br />

a similar role in all eukaryotes. Protein kinases<br />

and phosphoprotein phosphatases are used to<br />

regulate enzymes in all <strong>cell</strong>s.<br />

Although the basic biochemical components<br />

and processes <strong>of</strong> signal transduction are conserved<br />

and reused, they are <strong>of</strong>ten used in wildly<br />

divergent patterns and for many different physiological<br />

purposes. For example, cAMP is synthesized<br />

by distantly related enzymes in bacteria,<br />

fungi, and animals, and acts on different proteins<br />

in each organism; it is a pheromone in<br />

some slime molds.<br />

Cells <strong>of</strong>ten use the same series <strong>of</strong> <strong>signaling</strong><br />

proteins to regulate a given process, such as<br />

transcription, ion transport, locomotion, and<br />

metabolism. Such <strong>signaling</strong> pathways are assembled<br />

into <strong>signaling</strong> networks to allow the<br />

<strong>cell</strong> to coordinate its responses to multiple inputs<br />

with its ongoing functions. It is now possible<br />

to discern conserved reaction sequences<br />

in and between pathways in <strong>signaling</strong> networks<br />

that are analogous to devices within the circuits<br />

<strong>of</strong> analog computers: amplifiers, logic gates,<br />

feedback and feed-forward controls, and memory.<br />

This chapter discusses the principles and<br />

strategies <strong>of</strong> <strong>cell</strong>ular <strong>signaling</strong> first and then discusses<br />

the conserved biochemical components<br />

and reactions <strong>of</strong> <strong>signaling</strong> pathways and how<br />

these principles are applied.<br />

14.2<br />

Cellular <strong>signaling</strong> is<br />

primarily chemical<br />

Key concepts<br />

• Cells can detect both chemical and physical<br />

signals.<br />

• Physical signals are generally converted to<br />

chemical signals at the level <strong>of</strong> the receptor.<br />

Most signals sensed by <strong>cell</strong>s are chemical, and,<br />

when physical signals are sensed, they are generally<br />

detected as chemical changes at the level<br />

<strong>of</strong> the receptor. For example, the visual photoreceptor<br />

rhodopsin is composed <strong>of</strong> the protein<br />

opsin, which binds to a second component,<br />

the colored vitamin A derivative cis-retinal (the<br />

chromophore). When cis-retinal absorbs a<br />

photon, it photoisomerizes to trans-retinal,<br />

which is an activating ligand <strong>of</strong> the opsin protein.<br />

(For more on rhodopsin <strong>signaling</strong> see 14.20<br />

G protein <strong>signaling</strong> modules are widely used and<br />

highly adaptable). Similarly, plants sense red and<br />

blue light using the photosensory proteins phytochrome<br />

and cryptochrome, which detect photons<br />

that are absorbed by their tetrapyrrole or<br />

flavin chromophores. Cryptochrome homologs<br />

are also expressed in animals, where they probably<br />

mediate adjustment <strong>of</strong> the diurnal cycle.<br />

A few receptors do respond directly to physical<br />

inputs. Pressure-sensing channels, which exist<br />

in one form or another in all organisms,<br />

mediate responses to pressure or shear by changing<br />

their ionic conductance. In mammals, hearing<br />

is mediated indirectly by a mechanically<br />

operated channel in the hair <strong>cell</strong> <strong>of</strong> the inner ear.<br />

The extra<strong>cell</strong>ular domain <strong>of</strong> a protein called cadherin<br />

is pulled in response to acoustic vibration,<br />

generating the force that opens the channel.<br />

Cells sense mechanical strain through a<br />

number <strong>of</strong> <strong>cell</strong> surface proteins, including integrins.<br />

Integrins provide signals to <strong>cell</strong>s based on<br />

their attachment to other <strong>cell</strong>s and to molecular<br />

complexes in the external milieu.<br />

One major group <strong>of</strong> physically responsive<br />

receptors is made up <strong>of</strong> channels that sense electric<br />

fields. Another interesting group are<br />

heat/pain-sensing ion channels; several <strong>of</strong> these<br />

heat-sensitive ion channels also respond to<br />

chemical compounds, such as capsaicin, the<br />

“hot” lipid irritant in hot peppers.<br />

Whether a signal is physical or chemical, the<br />

receptor initiates the reactions that change the<br />

behavior <strong>of</strong> the <strong>cell</strong>. We will discuss how these<br />

effects are generated in the rest <strong>of</strong> the chapter.<br />

14.2 Cellular <strong>signaling</strong> is primarily chemical 591


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 592<br />

Ligand A<br />

14.3<br />

Receptors sense diverse<br />

stimuli but initiate a<br />

limited repertoire <strong>of</strong><br />

<strong>cell</strong>ular signals<br />

Key concepts<br />

• Receptors contain a ligand-binding domain and an<br />

effector domain.<br />

• Receptor modularity allows a wide variety <strong>of</strong><br />

signals to use a limited number <strong>of</strong> regulatory<br />

mechanisms.<br />

• Cells may express different receptors for the same<br />

ligand.<br />

• The same ligand may have different effects on the<br />

<strong>cell</strong> depending on the effector domain <strong>of</strong> its<br />

receptor.<br />

Receptors mediate responses to amazingly diverse<br />

extra<strong>cell</strong>ular messenger molecules; hence,<br />

the <strong>cell</strong> must express a large number <strong>of</strong> receptor<br />

varieties, each able to bind its extra<strong>cell</strong>ular<br />

ligand. In addition, each receptor must be able<br />

to initiate a <strong>cell</strong>ular response. Receptors, thus,<br />

contain two functional domains: a ligandbinding<br />

domain and an effector domain,<br />

which may or may not correspond to definable<br />

structural domains within the protein.<br />

The separation <strong>of</strong> ligand-binding and effector<br />

functions allows receptors for diverse ligands<br />

to produce a limited number <strong>of</strong> evolutionarily<br />

conserved intra<strong>cell</strong>ular signals through the action<br />

<strong>of</strong> a few effector domains. In fact, there are<br />

Receptors have a ligand-binding domain and an effector domain<br />

Output<br />

1<br />

Output<br />

2<br />

Ligand A<br />

Output<br />

1<br />

Ligand B<br />

Output<br />

1<br />

CHIMERIC<br />

RECEPTOR<br />

Ligand C<br />

LBD1 LBD1 LBD1 LBD2 LBD3<br />

ED1 ED2<br />

ED1 ED1<br />

Output<br />

2<br />

ED2<br />

FIGURE 14.2 Receptors can be thought <strong>of</strong> as composed <strong>of</strong> two functional domains,<br />

a ligand-binding domain (LBD) and an effector domain (ED). The twodomain<br />

property implies that two receptors that respond to different ligands<br />

(middle) could initiate the same function by activating similar effector domains,<br />

or that a <strong>cell</strong> could express two receptor is<strong>of</strong>orms (left) that respond to<br />

the same ligand with distinct <strong>cell</strong>ular effects mediated by different effector domains.<br />

It also implies that one can create an artificial chimeric receptor with<br />

novel properties.<br />

only a limited number <strong>of</strong> receptor families, which<br />

are related by their conserved structures and <strong>signaling</strong><br />

functions (see Figure 14.1).<br />

There are several useful correlates to the<br />

two-domain nature <strong>of</strong> receptors. For example,<br />

a <strong>cell</strong> can control its responsiveness to an extra<strong>cell</strong>ular<br />

signal by regulating the synthesis or<br />

degradation <strong>of</strong> a receptor or by regulating the<br />

receptor’s activity (see 14.10 Cellular <strong>signaling</strong> is<br />

remarkably adaptive).<br />

In addition, the nature <strong>of</strong> a response is generally<br />

determined by the receptor and its effector<br />

domain rather than any physicochemical<br />

property <strong>of</strong> the ligand. FIGURE 14.2 illustrates the<br />

concept that a ligand may bind to more than<br />

one kind <strong>of</strong> receptor and elicit more than one<br />

type <strong>of</strong> response, or several different ligands<br />

may all act identically by binding to functionally<br />

similar receptors. For example, the neurotransmitter<br />

acetylcholine binds to two classes<br />

<strong>of</strong> receptors. Members <strong>of</strong> one class are ion channels;<br />

members <strong>of</strong> the other regulate G proteins.<br />

Similarly, steroid hormones bind both to nuclear<br />

receptors, which bind chromatin and regulate<br />

transcription, and to other receptors in<br />

the plasma membrane.<br />

Conversely, when multiple ligands bind to<br />

receptors <strong>of</strong> the same biochemical class, they<br />

generate similar intra<strong>cell</strong>ular responses. For example,<br />

it is not uncommon for a <strong>cell</strong> to express<br />

several distinct receptors that stimulate production<br />

<strong>of</strong> the intra<strong>cell</strong>ular <strong>signaling</strong> molecule cAMP.<br />

The effect <strong>of</strong> the receptor on the <strong>cell</strong> will also be<br />

determined significantly by the biology <strong>of</strong> the<br />

<strong>cell</strong> and its state at any given time.<br />

Ligand binding and effector domains may<br />

evolve independently in response to varied selective<br />

pressures. For example, mammalian and<br />

invertebrate rhodopsins transduce their signal<br />

through different effector G proteins (G t<br />

and<br />

G q<br />

, respectively). Another example is calmodulin,<br />

a small calcium-binding regulatory protein<br />

in animals, which in plants appears as a<br />

distinct domain in larger proteins.<br />

The receptor’s two-domain nature allows<br />

the <strong>cell</strong> to regulate the binding <strong>of</strong> ligand and<br />

the effect <strong>of</strong> ligand independently. Covalent<br />

modification or allosteric regulation can alter<br />

ligand-binding affinity, the ability <strong>of</strong> the ligand-bound<br />

receptor to generate its signal or<br />

both. We will discuss these concepts further in<br />

14.13 Cellular <strong>signaling</strong> uses both allostery and covalent<br />

modification.<br />

Receptors can be classified either according<br />

to the ligands they bind or the way in which<br />

they signal. Signal output, which is character-<br />

592 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 593<br />

istic <strong>of</strong> the effector domain, usually correlates<br />

best with overall structure and sequence conservation.<br />

(Receptor families grouped by their<br />

functions are the organizational basis <strong>of</strong> the second<br />

half <strong>of</strong> this chapter.) However, classifying<br />

receptors pharmacologically, according to their<br />

specificity for ligands, is particularly useful for<br />

understanding the organization <strong>of</strong> endocrine<br />

and neuronal systems and for categorizing the<br />

multiple physiological responses to drugs.<br />

Expression <strong>of</strong> a receptor that is not normally<br />

expressed in a <strong>cell</strong> is <strong>of</strong>ten sufficient to<br />

confer responsiveness to that receptor’s ligand.<br />

This responsiveness <strong>of</strong>ten occurs because the<br />

<strong>cell</strong> expresses the other components necessary<br />

for propagating the intra<strong>cell</strong>ular signal from the<br />

receptor. The precise nature <strong>of</strong> the response will<br />

reflect the biology <strong>of</strong> the <strong>cell</strong>. Experimentally,<br />

responsiveness to a compound can be induced<br />

by introducing the cDNA that encodes the receptor.<br />

For example, mammalian receptors may<br />

be expressed in yeast, such that the yeast respond<br />

visibly to receptor ligands, thus providing<br />

a way to screen for new chemicals (drugs)<br />

that activate the receptor.<br />

Finally, it is possible to create chimeric receptors<br />

by fusing the ligand-binding domain<br />

from one receptor with the effector domain<br />

from a different receptor (Figure 14.2). Such<br />

chimeras can mediate novel responses to the<br />

ligand. With genetic modification <strong>of</strong> the ligandbinding<br />

domain, receptors can be reengineered<br />

to respond to novel ligands. Thus, scientists can<br />

manipulate <strong>cell</strong> functions with nonbiological<br />

compounds.<br />

14.4<br />

Receptors are catalysts<br />

and amplifiers<br />

Key concepts<br />

• Receptors act by increasing the rates <strong>of</strong> key<br />

regulatory reactions.<br />

• Receptors act as molecular amplifiers.<br />

Receptors act to accelerate intra<strong>cell</strong>ular functions<br />

and are, thus, functionally analogous to enzymes<br />

or other catalysts. Some receptors,<br />

including the protein kinases, protein phosphatases,<br />

and guanylate cyclases, are themselves<br />

enzymes and thus classical biochemical catalysts.<br />

More generally, however, receptors use<br />

the relatively small energy <strong>of</strong> ligand binding to<br />

accelerate reactions that are driven by alternative<br />

energy sources. For example, receptors that<br />

are ion channels catalyze the movement <strong>of</strong> ions<br />

across membranes, a process driven by the electrochemical<br />

potential developed by distinct ion<br />

pumps. G protein-coupled receptors and other<br />

guanine nucleotide exchange factors catalyze<br />

the exchange <strong>of</strong> GDP for GTP on the G protein,<br />

an energetically favored process dictated by the<br />

<strong>cell</strong>’s nucleotide energy balance. Transcription<br />

factors accelerate the formation <strong>of</strong> the transcriptional<br />

initiation complex, but transcription itself<br />

is energetically driven by multiple steps <strong>of</strong><br />

ATP and dNTP hydrolysis.<br />

As catalysts, receptors enhance the rates <strong>of</strong><br />

reactions. Most <strong>signaling</strong> involves kinetic rather<br />

than thermodynamic regulation; that is, <strong>signaling</strong><br />

events change reaction rates rather than<br />

their equilibria (see the next section). Thus, <strong>signaling</strong><br />

is similar to metabolic regulation, in<br />

which specific reactions are chosen according to<br />

their rates, with thermodynamic driving forces<br />

playing only a supportive role.<br />

In all <strong>signaling</strong> reactions, receptors use their<br />

catalytic activities to function as molecular amplifiers.<br />

Directly or indirectly, a receptor generates<br />

a chemical signal that is huge, both<br />

energetically and with respect to the number<br />

<strong>of</strong> molecules recruited by a single receptor.<br />

Molecular amplification is a hallmark <strong>of</strong> receptors<br />

and many other steps in <strong>cell</strong>ular <strong>signaling</strong><br />

pathways.<br />

14.5<br />

Ligand binding changes<br />

receptor conformation<br />

Key concepts<br />

• Receptors can exist in active or inactive<br />

conformations.<br />

• Ligand binding drives the receptor toward the<br />

active conformation.<br />

A central mechanistic question in receptor function<br />

is how the binding <strong>of</strong> a <strong>signaling</strong> molecule<br />

to the ligand-binding domain increases the activity<br />

<strong>of</strong> the effector domain. The key to this<br />

question is that receptors can exist in multiple<br />

molecular conformations, some active for <strong>signaling</strong><br />

and others inactive. Ligands shift the<br />

conformational equilibrium among these conformations.<br />

The structural changes that occur<br />

during the receptor’s inactive-active isomerization<br />

and how ligand binding drives these<br />

changes are exciting areas <strong>of</strong> biophysical research.<br />

However, the basic concept can be described<br />

simply in terms <strong>of</strong> coupling the<br />

conformational isomerizations <strong>of</strong> the ligandbinding<br />

and effector domains.<br />

14.5 Ligand binding changes receptor conformation 593


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 594<br />

How do ligands activate (or not activate) a<br />

receptor? Most <strong>of</strong> the basic regulatory activities<br />

<strong>of</strong> receptors can be described by a simple scheme<br />

that considers the receptors as having two interconvertible<br />

conformations, inactive (R) and<br />

active (R*). R and R* are in equilibrium, which<br />

is described by the equilibrium constant J.<br />

Because unliganded receptors are usually<br />

minimally active, J1), then ligand binding<br />

will shift the conformation to the R* state to an<br />

equivalent extent (i.e., J*/J>>1). The relative<br />

activation by a saturating concentration <strong>of</strong> ligand,<br />

J*/J, will exactly equal the ligand’s relative<br />

selectivity for the active receptor conformation,<br />

K*/K. This argument is generally valid for the regulation<br />

<strong>of</strong> a protein’s activity by any regulatory<br />

ligand.<br />

This model explains many properties <strong>of</strong> receptors<br />

and their ligands both simply and quantitatively.<br />

• First, J must be greater than zero for the<br />

equilibrium to exist. Thus, even unliganded<br />

receptor has some activity.<br />

Overexpressed receptors frequently display<br />

their intrinsic low activity.<br />

• Because physiological receptors are<br />

nearly inactive in the absence <strong>of</strong> ligand,<br />

J must be much less than 1 and is probably<br />

less than 0.01; most receptors are<br />

less than 1% active without agonist.<br />

• Ligands can vary in their selectivities<br />

between R and R*. Their abilities to activate<br />

will also vary. Some ligands, referred<br />

to as agonists, can drive formation<br />

<strong>of</strong> appreciable R*. Others, known as partial<br />

agonists, will promote submaximal<br />

activation. Chemical manipulation<br />

<strong>of</strong> a ligand’s structure will <strong>of</strong>ten alter its<br />

activity as an agonist. These relationships<br />

are depicted graphically in FIGURE<br />

14.3.<br />

• A ligand that binds equally well to both<br />

the R and R* states will not cause activation.<br />

However, such a ligand may still<br />

occupy the binding site and thereby<br />

competitively inhibit binding <strong>of</strong> an activating<br />

ligand. Such competitive inhibitors,<br />

referred to as antagonists, are<br />

frequently used as drugs to block unwanted<br />

activation <strong>of</strong> a receptor in various<br />

disease states.<br />

• A ligand that binds preferentially to R<br />

594 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 595<br />

relative to R* will further shift the conformational<br />

equilibrium to the inactive<br />

state and cause net inhibition. Such ligands<br />

are called inverse agonists.<br />

Because J is already low, effects <strong>of</strong> inverse<br />

agonists may only be noticeable<br />

if a receptor is overexpressed or if the<br />

receptor is mutated to increase its intrinsic<br />

activity (i.e., the mutation increases<br />

J).<br />

• The extent to which an agonist stimulates<br />

a receptor is unrelated to its affinity.<br />

Both agonists and antagonists may<br />

bind with either high or low affinity.<br />

Affinity does determine the receptor’s<br />

sensitivity—that is, how low a concentration<br />

<strong>of</strong> ligand can the receptor detect.<br />

Affinities <strong>of</strong> receptors for natural regulatory<br />

ligands vary enormously, with<br />

physiologic K d<br />

values ranging from<br />


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 596<br />

RECEPTORS<br />

TRANSDUCERS<br />

EFFECTORS<br />

Convergent and divergent <strong>signaling</strong> pathways<br />

Linear,<br />

parallel<br />

Convergent Divergent Multiply<br />

branched<br />

FIGURE 14.4 Signaling pathways use convergent and divergent branching to coordinate<br />

information flow. The diagrams at top show how even a simple, threelevel<br />

<strong>signaling</strong> network can sort information. Convergence or divergence can<br />

take place at multiple points along a <strong>signaling</strong> pathway. As an example <strong>of</strong> complexity,<br />

the lower portion <strong>of</strong> the figure shows a small segment (~10%) <strong>of</strong> the G<br />

protein-mediated <strong>signaling</strong> network in a mouse macrophage <strong>cell</strong> line. It omits<br />

several interpathway regulatory mechanisms and completely ignores inputs from<br />

non-G protein-coupled receptors. Pathway map courtesy <strong>of</strong> Lily Jiang, University<br />

<strong>of</strong> Texas <strong>Southwestern</strong> Medical Center.<br />

maps. Signaling networks are also spatially complex.<br />

They may include components in various<br />

sub<strong>cell</strong>ular locations, with initial receptors and<br />

associated proteins in the plasma membrane, but<br />

with downstream proteins in the cytoplasm or intra<strong>cell</strong>ular<br />

organelles. Such complexity is necessary<br />

to allow the <strong>cell</strong>s to integrate and sort<br />

incoming signals and to regulate multiple intra<strong>cell</strong>ular<br />

functions simultaneously.<br />

The complexity and adaptability <strong>of</strong> <strong>signaling</strong><br />

networks, like the one shown in the lower<br />

half <strong>of</strong> Figure 14.4, make their dynamics at the<br />

whole-<strong>cell</strong> level difficult or impossible to grasp<br />

intuitively. Signaling networks resemble large<br />

analog computers, and investigators are increasingly<br />

depending on computational tools to understand<br />

<strong>cell</strong>ular information flow and its<br />

regulation. First, many <strong>signaling</strong> interactions<br />

that include only two or three proteins exert<br />

functions analogous to traditional computational<br />

logic circuits (see the next section). The<br />

theory and experience with such circuits in electronics<br />

facilitate understanding biological <strong>signaling</strong><br />

functions as well.<br />

The enormous complexity <strong>of</strong> <strong>cell</strong>ular <strong>signaling</strong><br />

networks can be simplified by considering<br />

them to be composed <strong>of</strong> interacting <strong>signaling</strong><br />

modules, i.e., groups <strong>of</strong> proteins that process signals<br />

in well-understood ways. A <strong>cell</strong>ular <strong>signaling</strong><br />

module is analogous to an integrated circuit<br />

in an electronic instrument that performs a<br />

known function, but whose exact components<br />

could be changed for similar use in another device.<br />

The concept <strong>of</strong> modular construction facilitates<br />

both qualitative and quantitative<br />

understanding <strong>of</strong> <strong>signaling</strong> networks. We will refer<br />

to many standard <strong>signaling</strong> modules later in<br />

the chapter. Examples include monomeric and<br />

heterotrimeric G protein modules, MAPK cascades,<br />

tyrosine (Tyr) kinase receptors and their<br />

binding proteins, and Ca2+ release/uptake modules.<br />

In each case, despite the numerous phylogenetic,<br />

developmental, and physiologic<br />

variations, understanding the basic function <strong>of</strong><br />

that class <strong>of</strong> module conveys understanding <strong>of</strong> all<br />

its incarnations. Last, the evolutionary importance<br />

<strong>of</strong> modules is significant; once the architecture<br />

<strong>of</strong> a module is established it can be reused.<br />

For larger-scale networks, multiplexed,<br />

high-throughput measurements on living <strong>cell</strong>s<br />

have been combined with powerful kinetic modeling<br />

strategies to allow an increasingly accurate<br />

quantitative depiction <strong>of</strong> information flow<br />

within <strong>signaling</strong> modules or entire networks.<br />

Such models, with sound and experimentally<br />

based parameter sets, can describe <strong>signaling</strong><br />

processes in systems too complex for intuitive<br />

or ad hoc analysis. They are also vital as tests <strong>of</strong><br />

understanding because they can predict experimental<br />

results in ways that can be used to test<br />

the validity <strong>of</strong> the model. Well-grounded models<br />

can then be used (cautiously) to suggest the<br />

mechanisms <strong>of</strong> systems for which data sets remain<br />

unattainable. At even greater levels <strong>of</strong><br />

complexity, the theories and tools <strong>of</strong> computer<br />

science are increasingly giving useful systemslevel<br />

analyses <strong>of</strong> signal flow in <strong>cell</strong>s. Using computational<br />

tools to analyze large arrays <strong>of</strong><br />

quantitative data allows us to understand <strong>cell</strong>ular<br />

information flow and its regulation.<br />

596 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 597<br />

Developing quantitative models <strong>of</strong> <strong>signaling</strong><br />

networks is a frontier in <strong>signaling</strong> biology. These<br />

models both help describe network function<br />

and pinpoint experiments to clarify mechanism.<br />

14.7<br />

Cellular <strong>signaling</strong><br />

pathways can be thought<br />

<strong>of</strong> as biochemical logic<br />

circuits<br />

Key concepts<br />

• Signaling networks are composed <strong>of</strong> groups <strong>of</strong><br />

biochemical reactions that function as<br />

mathematical logic functions to integrate<br />

information.<br />

• Combinations <strong>of</strong> such logic functions combine as<br />

<strong>signaling</strong> networks to process information at more<br />

complex levels.<br />

As introduced in the preceding section, processes<br />

that <strong>signaling</strong> pathways use to integrate and direct<br />

information to <strong>cell</strong>ular targets are strikingly analogous<br />

to the mathematical logic functions that are<br />

used to design the individual circuits <strong>of</strong> electronic<br />

computers. Indeed, there are biological equivalents<br />

<strong>of</strong> essentially all <strong>of</strong> the functional components<br />

that computer scientists and engineers<br />

consider in the design <strong>of</strong> computers and electronic<br />

control devices. To understand <strong>signaling</strong> pathways,<br />

it is, therefore, useful to consider groups <strong>of</strong><br />

reactions within a pathway as constituting logic circuits<br />

<strong>of</strong> the sort used in electronic computing, as<br />

illustrated in FIGURE 14.5. The simplest example is<br />

when two stimulatory pathways converge. If sufficient<br />

input from either is adequate to elicit the<br />

response, the convergence would constitute an<br />

“OR” function. If neither input is sufficient by itself<br />

but the combination <strong>of</strong> the two elicits the response,<br />

then the converging pathways would<br />

create “AND” functions. AND circuits are also referred<br />

to as coincidence detectors—a response<br />

is elicited only when two stimulating pathways<br />

are activated simultaneously.<br />

AND functions can result from the combination<br />

<strong>of</strong> two similar but quantitatively inadequate<br />

inputs. Alternatively, two mechanistically<br />

different inputs might both be required to elicit<br />

a response. An example <strong>of</strong> the latter would be<br />

a target protein that is allosterically activated<br />

only when phosphorylated, or that is activated<br />

by phosphorylation but is only functional when<br />

recruited to a specific sub<strong>cell</strong>ular location.<br />

The opposite <strong>of</strong> an AND circuit is a NOT<br />

function, where one pathway blocks the stim-<br />

Logical (Boolean)<br />

A<br />

B<br />

A + B<br />

A<br />

B<br />

A + B<br />

A<br />

B<br />

A + B<br />

A OR B<br />

A AND B<br />

A NOT B<br />

Response<br />

Response<br />

Response<br />

Response<br />

Response<br />

Simple logic circuits<br />

Response<br />

Quantitative (Analog)<br />

A + fixed [B]<br />

Response<br />

Response<br />

Additive<br />

ulatory effect <strong>of</strong> another. Simple logic gates are<br />

observed at many locations in <strong>cell</strong>ular <strong>signaling</strong><br />

pathways.<br />

We can also think about convergent <strong>signaling</strong><br />

in quantitative rather than Boolean terms<br />

by considering the additivity <strong>of</strong> inputs to a distinct<br />

process (see Figure 14.5, right). The OR<br />

function referred to above can be considered to<br />

be the additive positive inputs <strong>of</strong> two pathways.<br />

Such additivity could represent the ability <strong>of</strong><br />

several receptors to stimulate a pool <strong>of</strong> a particular<br />

G protein or the ability <strong>of</strong> two protein kinases<br />

to phosphorylate a single substrate.<br />

Additivity may be positive, as in the examples<br />

above, or negative, such as when two inhibitory<br />

inputs combine. Inhibition and stimulation may<br />

also combine additively to yield an algebraically<br />

balanced output. Alternatively, multiple inputs<br />

can combine with either more or less than an<br />

additive effect. The NOT function, discussed<br />

above, is analogous to describing a blockade <strong>of</strong><br />

stimulation. The AND function describes synergism,<br />

where one input potentiates another<br />

but alone has little effect.<br />

Even simple <strong>signaling</strong> networks can display<br />

complex patterns <strong>of</strong> information processing. One<br />

A<br />

log (agonist concentration)<br />

More than additive<br />

log (agonist concentration)<br />

Less than additive<br />

log (agonist concentration)<br />

B<br />

A + B<br />

A<br />

B<br />

A<br />

A + B<br />

B<br />

FIGURE 14.5 Signaling networks use simple logic functions to process<br />

information. Boolean OR, AND, and NOT functions (left) correspond to<br />

the quantitative interactions between converging signals that are shown<br />

on the right.<br />

14.7 Cellular <strong>signaling</strong> pathways can be thought <strong>of</strong> as biochemical logic circuits 597


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 598<br />

Input<br />

Positive feedback loop : irreversible ON switch<br />

Positive feed-forward loop : responds to prolonged input<br />

Input<br />

OH<br />

E<br />

G<br />

T<br />

T<br />

+<br />

Output<br />

Output<br />

Output<br />

Input strength<br />

Conformational lock - Dual control switch<br />

OH<br />

E<br />

OH<br />

+<br />

Kinase<br />

E<br />

E<br />

G<br />

G<br />

Phosphatase<br />

Signal processing circuits<br />

E<br />

E<br />

G<br />

P<br />

E<br />

P<br />

P<br />

E<br />

Output<br />

Output<br />

Time<br />

Time<br />

input<br />

G K G P K<br />

FIGURE 14.6 Relatively complex signal processing can be executed by simple<br />

multi-protein modules. The figure depicts three types <strong>of</strong> <strong>signaling</strong> modules<br />

(left) and their behavior in response to agonist (right). (top) In a positive<br />

feed-back module, a transducer protein (T) stimulates an effector (E) to produce<br />

a <strong>cell</strong>ular output, but the effector also stimulates the activity <strong>of</strong> the transducer.<br />

The result can be an all-or-none switch, where input up to a threshold<br />

has little effect, but then becomes committed when feedback from the effector<br />

is sufficient to maintain transducer activity even in the absence <strong>of</strong> continued<br />

input from the receptor. (center) In a positive feed-forward module, the<br />

effector requires input both from the transducer and from upstream in the pathway.<br />

When stimulation is brief (short horizontal bar under trace at right), significant<br />

amounts <strong>of</strong> active transducer do not accumulate and output is minimal.<br />

When stimulation is prolonged (longer bar), signal output is substantial. (bottom)<br />

In some dual-control switching modules, the binding <strong>of</strong> one regulator (G)<br />

can both activate the effector and expose another regulatory site, shown here<br />

as a Ser substrate site (-OH) for a protein kinase. The effector can only be phosphorylated<br />

or dephosphorylated when G is bound. Therefore, as shown at the<br />

right, addition <strong>of</strong> G alone will activate but activation <strong>of</strong> the kinase (K) alone<br />

will not. If kinase is active while G is bound, phosphorylation is resistant to<br />

phosphatase activity unless G is again present to reexpose the phosphoserine<br />

residue (shown on the graph at the right as a bold P).<br />

G P<br />

good example is the creation <strong>of</strong> “memory”: making<br />

the effect <strong>of</strong> a transient signal more or less<br />

permanent. Signaling pathways have multiple<br />

ways <strong>of</strong> setting memories, and <strong>of</strong> forgetting. One<br />

mechanism, common in protein kinase pathways,<br />

is the positive feedback loop, illustrated<br />

in the top panel <strong>of</strong> FIGURE 14.6. In a positive feedback<br />

loop, the input stimulates a transducer (T),<br />

which in turn stimulates the effector protein (E)<br />

to create the output. If the effector can also activate<br />

the transducer, sufficient initial signal can<br />

be fed back to the transducer that it can maintain<br />

the effector's full signal output even when<br />

input is removed. Such systems typically display<br />

a threshold behavior, as shown on the right.<br />

A positive feed-forward loop can generate<br />

memory <strong>of</strong> another type (Figure 14.6, middle<br />

panel), indicating the duration <strong>of</strong> input. In such<br />

circuits, the effector requires simultaneous input<br />

from both the receptor and from the intermediary<br />

transducer. If the pathway from<br />

receptor through transducer is relatively slow,<br />

or if it requires the accumulation <strong>of</strong> a substantial<br />

amount <strong>of</strong> transducer, only a prolonged input<br />

will trigger a response, as shown in the<br />

time-base output diagram at the right.<br />

A third way to establish memory is to allow<br />

one input to control the reversibility <strong>of</strong> a second<br />

regulatory event (Figure 14.6, bottom panel).<br />

WASP, a protein that initiates the polymerization<br />

<strong>of</strong> actin to drive <strong>cell</strong>ular motion and shape<br />

change, is activated both by phosphorylation<br />

and by the binding <strong>of</strong> Cdc42, a small GTP-binding<br />

protein (G). However, the phosphorylation<br />

site on WASP is only exposed when WASP is<br />

bound to Cdc42. Phosphorylation thus requires<br />

both activated Cdc42 and activated protein kinase.<br />

If Cdc42 dissociates, the phosphorylated<br />

state <strong>of</strong> WASP persists until another <strong>signaling</strong><br />

molecule, whose identity remains uncertain,<br />

binds again to expose the site to a protein phosphatase.<br />

As shown in the time-base graph, exposure<br />

to Cdc42 will activate, but exposure to<br />

kinase alone will not. If Cdc42 is present, then<br />

the kinase can activate WASP. Phospho-WASP<br />

is relatively insensitive to protein phosphatase<br />

(P) alone, but can be dephosphorylated if Cdc42<br />

or another G protein binds to expose the site to<br />

phosphatase.<br />

14.8<br />

Scaffolds increase<br />

<strong>signaling</strong> efficiency and<br />

enhance spatial<br />

organization <strong>of</strong> <strong>signaling</strong><br />

Key concepts<br />

• Scaffolds organize groups <strong>of</strong> <strong>signaling</strong> proteins and<br />

may create pathway specificity by sequestering<br />

components that have multiple partners.<br />

• Scaffolds increase the local concentration <strong>of</strong><br />

<strong>signaling</strong> proteins.<br />

• Scaffolds localize <strong>signaling</strong> pathways to sites <strong>of</strong><br />

action.<br />

598 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 599<br />

The INAD <strong>signaling</strong> complex<br />

TRP<br />

Rhodopsin<br />

Scaffolds concentrate and insulate <strong>signaling</strong> proteins<br />

Pheromone<br />

GPCR<br />

PKC<br />

PDZ PDZ<br />

PDZ<br />

INAD<br />

-<br />

PDZ PDZ<br />

CaM<br />

CaM<br />

PKC<br />

PDZ<br />

PDZ<br />

PDZ<br />

INAD<br />

PDZ PDZ<br />

CYTOSOL<br />

FIGURE 14.7 The scaffold InaD organizes proteins that transmit visual<br />

signals in the fly photoreceptor <strong>cell</strong>. InaD is localized to the photoreceptor<br />

membrane and coordinates light sensing and visual transduction. In<br />

invertebrate eyes, the visual <strong>signaling</strong> pathway goes from rhodopsin<br />

through G q<br />

to a phospholipase C-, and Ca2+ release triggered by PLC action<br />

initiates depolarization. This system is specialized for speed, and requires<br />

that the relevant proteins are nearby. InaD contains five PDZ<br />

domains, each <strong>of</strong> which binds to the C terminus <strong>of</strong> a signal transducing<br />

protein. The TRP channel, which mediates Ca2+ entry, PLC-, and a protein<br />

kinase C is<strong>of</strong>orm that is involved in rapid desensitization all bind constitutively<br />

to InaD. Rhodopsin and a myosin (NinaC) also bind, and G q<br />

binds indirectly.<br />

The proteins in a <strong>signaling</strong> pathway are frequently<br />

colocalized within <strong>cell</strong>s such that their<br />

mutual interactions are favored and their interactions<br />

with other proteins are minimized.<br />

Many <strong>signaling</strong> pathways are organized on scaffolds.<br />

Scaffolds bind several components <strong>of</strong> a<br />

<strong>signaling</strong> pathway in multiprotein complexes<br />

to enhance <strong>signaling</strong> efficiency. Scaffolds promote<br />

interactions <strong>of</strong> proteins that have a low<br />

affinity for each other, accelerate activation (and<br />

<strong>of</strong>ten inactivation) <strong>of</strong> the associated components,<br />

and localize the <strong>signaling</strong> proteins to appropriate<br />

sites <strong>of</strong> action. Colocalization may be<br />

tonic or regulated, and stimulus-dependent scaffolding<br />

<strong>of</strong>ten determines <strong>signaling</strong> outputs.<br />

The binding sites on a scaffolding protein<br />

are <strong>of</strong>ten localized in distinct modular proteinbinding<br />

domains, giving the impression that the<br />

protein is designed simply to hold the components<br />

<strong>of</strong> the pathway together. Many scaffolding<br />

proteins do lack intrinsic enzymatic activity,<br />

but some <strong>signaling</strong> enzymes also act as scaffolds.<br />

Binding to a scaffold facilitates <strong>signaling</strong> by<br />

increasing the local concentrations <strong>of</strong> the components,<br />

so that diffusion or transport <strong>of</strong> molecules<br />

to their sites <strong>of</strong> action is not necessary. In<br />

the photoreceptor <strong>cell</strong>s <strong>of</strong> Drosophila, scaffolding<br />

<strong>of</strong> <strong>signaling</strong> components is critical for rapid<br />

signal transmission. These <strong>cell</strong>s contain the InaD<br />

-<br />

Ste5p<br />

Ste11p<br />

Ste7p<br />

Fus3p<br />

G protein<br />

Pheromone<br />

Ste5p<br />

Mating<br />

response<br />

Ste11p<br />

Ste7p<br />

Fus3p<br />

Cdc42p<br />

Ste20p<br />

Scaffold organizes<br />

MAPK cascade<br />

Cdc42p<br />

Ste20p<br />

scaffolding protein, which has five modular<br />

binding domains, known as PDZ domains. Each<br />

<strong>of</strong> its PDZ domains binds to a C-terminal motif<br />

<strong>of</strong> a target protein, thereby facilitating interactions<br />

among the associated proteins. FIGURE 14.7<br />

shows a model for how InaD organizes the <strong>signaling</strong><br />

proteins. The mutational loss <strong>of</strong> InaD<br />

produces a nearly blind fly, and deletion <strong>of</strong> a<br />

single PDZ domain can yield a fly with a distinct<br />

visual defect characteristic <strong>of</strong> the protein<br />

that binds to the missing domain.<br />

A second example is Ste5p, a scaffold for the<br />

pheromone-induced mating response pathway<br />

in S. cerevisiae. FIGURE 14.8 illustrates how Ste5p<br />

binds and organizes components <strong>of</strong> a mitogen-<br />

Scaffold determines<br />

specificity <strong>of</strong> Ste11p<br />

<strong>signaling</strong><br />

Ste11p<br />

Ste7p<br />

Fus3p<br />

Mating<br />

response<br />

High osmolarity<br />

Ste11p<br />

Pbs2p<br />

Hog1p<br />

Osmoadaptation<br />

Cdc42p<br />

Ste20p<br />

Cdc42p<br />

Ste20p<br />

FIGURE 14.8 The scaffold Ste5p organizes the components <strong>of</strong> the MAPK<br />

cascade that mediates the pheromone-induced mating response in<br />

Saccharomyces cerevisiae. In the top left panel, Ste5p brings the components<br />

<strong>of</strong> the MAPK cascade to the membrane in response to pheromone. In<br />

the top right panel, binding to the heterotrimeric G protein brings loaded<br />

Ste5p in proximity to the protein kinase Ste20p bound to the activated small<br />

GTP binding protein Cdc42p. Their colocalization facilitates the sequential<br />

activation <strong>of</strong> the cascade components, resulting in activation <strong>of</strong> the MAPK<br />

Fus3p and the mating response. The MAP3K Ste11p can regulate not only<br />

the MAPK Fus3p in the mating pathway, but also the MAPK Hog1p in the<br />

high osmolarity pathway, as shown in the bottom two panels. The scaffold<br />

to which Ste11p binds, either Ste5p or Pbs2 (both a scaffold and a MAP2K),<br />

determines which MAPK and downstream events are activated as the output.<br />

14.8 Scaffolds increase <strong>signaling</strong> efficiency and enhance spatial organization <strong>of</strong> <strong>signaling</strong> 599


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 600<br />

activated protein kinase (MAPK) cascade, including<br />

a MAP3K (Ste11p), a MAP2K (Ste7p)<br />

and a MAPK (Fus3p). (The MAPK cascade will<br />

be discussed in 14.32 MAPKs are central to many<br />

<strong>signaling</strong> pathways). The function <strong>of</strong> Ste5p is partially<br />

retained even if the positions <strong>of</strong> its binding<br />

sites for the kinases are shuffled in the linear<br />

sequence <strong>of</strong> the protein, indicating that a major<br />

role is to bring the enzymes into proximity, rather<br />

than to precisely orient them. Ste5p also binds<br />

to the subunits <strong>of</strong> the heterotrimeric G protein<br />

that mediates the actions <strong>of</strong> mating<br />

pheromones, linking the membrane signal to<br />

the intra<strong>cell</strong>ular transducers. Yeast that lack<br />

Ste5p cannot mate, demonstrating that Ste5p is<br />

required for this biological function (but not all<br />

functions) carried out by the pathway.<br />

In addition to facilitating <strong>signaling</strong> in their<br />

own pathways, scaffolds can enhance <strong>signaling</strong><br />

specificity by limiting interactions with other<br />

<strong>signaling</strong> proteins. Scaffolds thus insulate components<br />

<strong>of</strong> a <strong>signaling</strong> pathway both from activation<br />

by inappropriate signals and from<br />

producing incorrect outputs. For example, the<br />

mating and osmosensing pathways in yeast<br />

share several components, including the MAP3K<br />

Ste11p, but each pathway maintains specificity<br />

because it employs different scaffolds that restrict<br />

signal transmission.<br />

In contrast, the presence <strong>of</strong> excess scaffold<br />

can inhibit <strong>signaling</strong> because the individual <strong>signaling</strong><br />

components will more frequently bind<br />

to distinct scaffold proteins rather than forming<br />

a functional complex. Such dilution among scaffolds<br />

causes separation rather than concentration<br />

<strong>of</strong> the components, preventing their<br />

productive interaction.<br />

14.9<br />

Independent, modular<br />

domains specify proteinprotein<br />

interactions<br />

Key concepts<br />

• Protein interactions may be mediated by small,<br />

conserved domains.<br />

• Modular interaction domains are essential for<br />

signal transmission.<br />

• Adaptors consist exclusively <strong>of</strong> binding domains or<br />

motifs.<br />

Modular protein interaction domains or motifs<br />

occur in many <strong>signaling</strong> proteins and confer the<br />

ability to bind structural motifs in other molecules,<br />

including proteins, lipids, and nucleic<br />

acids. Some <strong>of</strong> these domains are listed in FIGURE<br />

14.9. In contrast to scaffolds, which bind specific<br />

proteins with considerable selectivity, modular<br />

interaction domains generally recognize<br />

not a single molecule but a group <strong>of</strong> targets that<br />

share related structural features.<br />

Modular interaction domains important for<br />

signal transduction were first discovered in the<br />

protein tyrosine kinase proto-oncogene Src,<br />

which contains a protein tyrosine kinase domain<br />

and two domains named Src homology<br />

(SH) 2 and 3 domains. The modular SH2 and<br />

SH3 domains were originally identified by comparison<br />

<strong>of</strong> Src to two other tyrosine kinases, Fps<br />

and Abl. One or both <strong>of</strong> these domains appear<br />

in numerous proteins and both are critically involved<br />

in protein-protein interactions.<br />

SH3 domains, which consist <strong>of</strong> approximately<br />

50 residues, bind to specific short proline-rich<br />

sequences. Many cytoskeletal proteins<br />

and proteins found in focal adhesion complexes<br />

contain SH3 domains and proline rich sequences,<br />

suggesting that this targeting motif<br />

may send proteins with these domains to these<br />

sites <strong>of</strong> action within <strong>cell</strong>s. In contrast to phosphotyrosine-SH2<br />

binding, the proline-rich binding<br />

sites for SH3 domains are present in resting<br />

and activated <strong>cell</strong>s. However, SH3-proline interactions<br />

may be negatively regulated by phosphorylation<br />

within the proline-rich motif.<br />

SH2 domains, which consist <strong>of</strong> approximately<br />

100 residues, bind to Tyr phosphorylated<br />

proteins, such as cytoplasmic tyrosine<br />

kinases and receptor tyrosine kinases. Thus, Tyr<br />

phosphorylation regulates the appearance <strong>of</strong><br />

SH2 binding sites and, thereby, regulates a set<br />

<strong>of</strong> protein-protein interactions in a stimulusdependent<br />

manner.<br />

A clever strategy was used to identify the<br />

binding specificity <strong>of</strong> SH2 domains. An isolated<br />

recombinant SH2 domain was incubated with<br />

<strong>cell</strong> lysates and then recovered from the lysates<br />

using a purification tag. The proteins associated<br />

with the SH2 domain were some <strong>of</strong> the same<br />

proteins that were recognized by antiphosphotyrosine<br />

antibodies. By this and other methods,<br />

it was discovered that SH2 domains recognize<br />

sequences surrounding Tyr phosphorylation<br />

sites and require phosphorylation <strong>of</strong> the included<br />

Tyr for high affinity binding.<br />

Information on specific amino acid sequences<br />

that recognize and bind to modular<br />

binding domains is being accumulated as these<br />

individual interactions are identified. In addition,<br />

screening programs using cDNA and/or<br />

peptide libraries to assess binding capabilities<br />

600 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 601<br />

Characteristics <strong>of</strong> some common modular protein domains<br />

Domain Characteristics Cellular involvement<br />

14-3-3 Binds protein phosphoserine or<br />

phosphothreonine<br />

Bromo<br />

CARD<br />

C1<br />

C2<br />

EF hand<br />

F-Box<br />

FHA<br />

HECT<br />

LIM<br />

PDZ<br />

PH<br />

RING<br />

SAM<br />

SH2<br />

Binds acetylated lysine residues<br />

Dimerization<br />

Binds phorbol esters or diacylglycerol<br />

Binds phospholipids<br />

Binds calcium<br />

Binds Skp1 in a ubiquitin-ligase<br />

complex<br />

Binds protein phosphothreonine<br />

or phosphoserine<br />

Binds E2 ubiquitin-conjugating<br />

enzymes to transfer ubiquitin to<br />

the substrate or to ubiquitin<br />

chains<br />

Zinc-binding cysteine-rich motif<br />

that forms two tandemly<br />

repeated zinc fingers<br />

Binds to the C-terminal 4-5<br />

residues <strong>of</strong> proteins that have a<br />

hydrophobic residue at the<br />

terminus; may bind to PIP2<br />

Binds to specific phosphoinositides,<br />

esp. PI-4,5-P 2 , PI-3,4-P 2 or<br />

PI-3,4,5-P 3 .<br />

Binds zinc and may be found in<br />

E3 ubiquitin ligases<br />

Homo- and heterooligomerization<br />

Binds to protein phosphotyrosine<br />

(pY)<br />

Protein sequestration<br />

Chromatin-associated<br />

proteins<br />

Caspase activation<br />

Recruitment to membranes<br />

Signal transduction,<br />

vesicular trafficking<br />

Calcium-dependent<br />

processes<br />

Ubiquitination<br />

Various; DNA damage<br />

FYVE Binds to PI(3)P Membrane trafficking,<br />

TGF- <strong>signaling</strong><br />

Ubiquitination<br />

Wide variety <strong>of</strong><br />

processes<br />

Scaffolding diverse<br />

protein complexes<br />

<strong>of</strong>ten at the membrane<br />

Recruitment to membranes<br />

and motility<br />

Ubiquitination,<br />

transcription<br />

Wide variety <strong>of</strong><br />

processes<br />

Tyrosine protein kinase<br />

<strong>signaling</strong><br />

SH3 Binds to PXXP motifs Various processes<br />

TPR<br />

WW<br />

Degenerate sequence <strong>of</strong> ~34<br />

amino acids with residues<br />

WL/GYAFAP; forms a scaffold<br />

Binds proline-rich sequences<br />

Wide variety <strong>of</strong><br />

processes<br />

Alternative to SH3;<br />

vesicular trafficking<br />

FIGURE 14.9 The table describes a subset <strong>of</strong> known modular protein interaction<br />

domains found in many proteins. Interactions mediated by these<br />

domains are essential to controlling <strong>cell</strong> function. Few if any <strong>of</strong> these domains<br />

exist in prokaryotes. Adapted from the Pawson Lab, Protein Interaction<br />

Domains, Mount Sinai Hospital (http://pawsonlab.mshri.on.ca/).<br />

yield such motifs. Consensus target sequences<br />

for individual domains have been identified<br />

based on the sequence specificity <strong>of</strong> their binding<br />

to arrayed sequences. These consensus sequences<br />

can then be used to predict whether<br />

the domain will bind a site in a candidate protein.<br />

Adaptor proteins, which lack enzymatic<br />

activity, link <strong>signaling</strong> molecules and target<br />

them in a manner that is responsive to extra<strong>cell</strong>ular<br />

signals. Adaptor proteins are generally<br />

made up <strong>of</strong> two or more modular interaction<br />

domains or the complementary recognition<br />

motifs. Unlike scaffolds, adaptors are usually<br />

14.9 Independent, modular domains specify protein-protein interactions 601


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 602<br />

multifunctional because their modular interaction<br />

domains and motifs are not as highly specific.<br />

Adaptors bind to two or more other<br />

<strong>signaling</strong> proteins via their protein-protein interaction<br />

domains to colocalize them or to facilitate<br />

additional interactions.<br />

Grb2 is a prototypical adaptor protein that<br />

was identified as a protein that bound to the C-<br />

terminal region <strong>of</strong> the EGF receptor. Grb2 has<br />

one SH2 and two SH3 domains. It binds constitutively<br />

to specific proline-rich segments <strong>of</strong> proteins<br />

through its SH3 domain, although this<br />

binding can be negatively regulated. One target<br />

<strong>of</strong> Grb2 is SOS, a guanine nucleotide exchange<br />

factor that activates the small GTP-binding protein<br />

Ras in response to EGF <strong>signaling</strong>. Through<br />

its SH2 domain, Grb2 binds Tyr-phosphorylated<br />

proteins, including the receptors themselves in<br />

a stimulus-dependent manner. Thus, Tyr phosphorylation<br />

<strong>of</strong> these receptors in response to<br />

ligand will enable the binding <strong>of</strong> Grb2 to the receptors,<br />

which, in turn, will recruit SOS to the<br />

membrane-localized receptor. Once at the membrane,<br />

SOS can activate its target, Ras.<br />

14.10<br />

Cellular <strong>signaling</strong> is<br />

remarkably adaptive<br />

Key concepts<br />

• Sensitivity <strong>of</strong> <strong>signaling</strong> pathways is regulated to<br />

allow responses to change over a wide range <strong>of</strong><br />

signal strengths.<br />

• Feedback mechanisms execute this function in all<br />

<strong>signaling</strong> pathways.<br />

• Most pathways contain multiple adaptive feedback<br />

loops to cope with signals <strong>of</strong> various strengths and<br />

durations.<br />

A universal property <strong>of</strong> <strong>cell</strong>ular <strong>signaling</strong> pathways<br />

is adaptation to the incoming signal. Cells continuously<br />

adjust their sensitivity to signals to maintain<br />

their ability to detect changes in input. Typically,<br />

when a <strong>cell</strong> is exposed to a new input, it initiates<br />

a process <strong>of</strong> desensitization that dampens the <strong>cell</strong>ular<br />

response to a new plateau lower than the initial<br />

peak response, as illustrated in FIGURE 14.10.<br />

When the stimulus is removed, the desensitized<br />

state can persist, with sensitivity slowly returning<br />

to normal. Similarly, the removal <strong>of</strong> a tonic stimulus<br />

can hypersensitize <strong>signaling</strong> systems.<br />

FIGURE 14.10 Top: Upon exposure to<br />

a stimulus, <strong>signaling</strong> pathways adjust<br />

their sensitivities to adapt to the new<br />

level <strong>of</strong> input. Thus, the response decays<br />

after initial stimulation. A second<br />

similar stimulus will elicit a smaller<br />

response unless adequate time is allowed<br />

for recovery. Bottom: Some adaptation<br />

mechanisms feed back only on<br />

the receptor that is stimulated and do<br />

not alter parallel pathways. Such mechanisms<br />

are referred to as homologous.<br />

At left, agonist a for receptor R1 can<br />

initiate either <strong>of</strong> two feedback events<br />

that desensitize R1 alone. In other<br />

cases, a stimulus will also cause parallel<br />

or related systems to desensitize.<br />

At the right, agonist a initiates desensitization<br />

<strong>of</strong> both R1 and R2. The response<br />

to agonist b, which binds to<br />

R2, is also desensitized. Such heterologous<br />

desensitization is common.<br />

K<br />

a<br />

Homologous<br />

desensitization<br />

Initial<br />

response<br />

R1 R 2<br />

R1 R 2<br />

X 1 X 2<br />

R esponse<br />

Response<br />

Patterns <strong>of</strong> adaptation in <strong>signaling</strong> networks<br />

Agonist a<br />

for R1<br />

Agonist Agonist Agonist<br />

Reapply<br />

a or b<br />

Desensitization<br />

b<br />

a<br />

a<br />

R1 R 2<br />

X 1 X 2<br />

Time<br />

Heterologous<br />

desensitization<br />

Response<br />

R1<br />

Agonist a<br />

for R1<br />

or R1 R 2<br />

Reapply<br />

a or b<br />

Y<br />

Time<br />

Y<br />

Time<br />

Z<br />

Z<br />

602 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 603<br />

Adaptation in <strong>signaling</strong> is one <strong>of</strong> the best examples<br />

<strong>of</strong> biological homeostasis. The adaptability<br />

<strong>of</strong> <strong>cell</strong>ular <strong>signaling</strong> can be quite impressive.<br />

Cells commonly regulate their sensitivity to physiological<br />

stimuli over more than a 100-fold range,<br />

and the mammalian visual response can adapt to<br />

incoming light over a 107-fold range. This remarkable<br />

ability allows a photoreceptor <strong>cell</strong> to<br />

detect a single photon, and allows a person to<br />

read in both very dim light and intense sunlight.<br />

Adaptability is observed in bacteria, plants, fungi,<br />

and animals. Many <strong>of</strong> its properties are conserved<br />

throughout biology, although the most complex<br />

adaptive mechanisms are found in animals. The<br />

general mechanism for adaptation is the negative<br />

feedback loop, which biochemically samples<br />

the signal and controls the adaptive process.<br />

Adaptation varies with both the intensity and<br />

the duration <strong>of</strong> the incoming signal. Stronger or<br />

more persistent inputs tend to drive greater adaptive<br />

change and, <strong>of</strong>ten, adaptation that persists<br />

for a longer time. Cells can modulate adaptation<br />

in this way because adaptation is exerted by a<br />

succession <strong>of</strong> independent mechanisms, each with<br />

its own sensitivity and kinetic parameters.<br />

G protein pathways <strong>of</strong>fer ex<strong>cell</strong>ent examples<br />

<strong>of</strong> adaptation. FIGURE 14.11 shows that the earliest<br />

step in adaptation is receptor phosphorylation,<br />

which is catalyzed by G protein-coupled<br />

receptor kinases (GRKs) that selectively recognize<br />

the receptor’s ligand-activated conformation.<br />

Phosphorylation inhibits the receptor’s ability<br />

to stimulate G protein activation and also promotes<br />

binding <strong>of</strong> arrestin, a protein that further<br />

inhibits G protein activation. Moreover, arrestin<br />

binding primes receptors for endocytosis, which<br />

removes them from the <strong>cell</strong> surface. Endocytosis<br />

can also be the first step in receptor proteolysis.<br />

Along with these direct effects, many receptor<br />

genes display feedback inhibition <strong>of</strong> transcription,<br />

such that <strong>signaling</strong> by a receptor decreases<br />

its own expression.<br />

Stimulation thus causes multiple adaptive<br />

processes that range from immediate (phosphorylation,<br />

arrestin binding) through delayed (transcriptional<br />

regulation), and include both reversible<br />

and irreversible events. This array <strong>of</strong> adaptive<br />

events has been demonstrated for many G protein-coupled<br />

receptors, and many <strong>cell</strong>s may use<br />

all <strong>of</strong> them to control output from one receptor.<br />

The speed, extent, and reversibility <strong>of</strong> adaptation<br />

are selected by a <strong>cell</strong>’s developmental program.<br />

Cells can change their patterns <strong>of</strong> adaptation<br />

both qualitatively and quantitatively by altering<br />

the points in a pathway where feedback is initiated<br />

and exerted. In a linear pathway, changing<br />

Multiple adaptation processes occur after a stimulus<br />

Relative<br />

response<br />

Receptor phosphorylation<br />

Arrestin binding<br />

Receptor<br />

endocytosis<br />

Endosomal receptor<br />

degradation<br />

5 Receptor transcription<br />

inhibited<br />

0 1 10 100 1000<br />

Time (seconds)<br />

Agonist<br />

added<br />

Agonist<br />

binds<br />

1<br />

2<br />

Agonist<br />

GPCR<br />

G protein<br />

Receptor<br />

recycling<br />

Lysosome<br />

3<br />

G protein<br />

active<br />

4<br />

EFFECTORS<br />

CYTOPLASM<br />

4 Receptor<br />

degradation<br />

DNA<br />

GRK<br />

1 Receptor<br />

Arrestin<br />

phosphorylation<br />

2 Arrestin<br />

binding<br />

NUCLEUS<br />

Early<br />

endosome<br />

GPCR gene<br />

3 Receptor<br />

endocytosis<br />

5 Receptor<br />

transcription<br />

inhibited<br />

FIGURE 14.11 Multiple adaptation processes are invoked during a stimulus,<br />

and multiple nested mechanisms for adaptation are the rule. They are usually<br />

invoked sequentially according to the duration and intensity <strong>of</strong> the stimulus.<br />

For GPCRs, at least five desensitizing mechanisms are known, with others acting<br />

on the G protein and effectors.<br />

these points will alter the kinetics or extent <strong>of</strong><br />

adaptation (Figure 14.10). In branched pathways,<br />

changing these points can determine whether<br />

adaptation is unique to one input or is exerted<br />

for many similar inputs. If receptor activation triggers<br />

its desensitization directly, or if an event<br />

downstream on an unbranched pathway triggers<br />

desensitization, then only signals that initiate with<br />

that receptor will be altered. Receptor-selective<br />

adaptation is referred to as homologous adaptation<br />

(Figure 14.10).<br />

14.10 Cellular <strong>signaling</strong> is remarkably adaptive 603


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 604<br />

Alternatively, feedback control can initiate<br />

downstream from multiple receptors in a convergent<br />

pathway and thus regulate both the<br />

initiating receptor and the others. Such heterologous<br />

adaptation regulates all the possible<br />

inputs to a given control point. A common<br />

example is the phosphorylation <strong>of</strong> G proteincoupled<br />

receptors by either protein kinase A or<br />

protein kinase C, which are activated by downstream<br />

signals cAMP or Ca2+ plus the lipid diacylglycerol,<br />

respectively. Like GRK, these kinases<br />

both attenuate receptor activity and promote<br />

arrestin binding.<br />

Cells also alter their responses to incoming<br />

signals for homeostatic reasons. These considerations<br />

include phase <strong>of</strong> the <strong>cell</strong> cycle, metabolic<br />

status, or other aspects <strong>of</strong> <strong>cell</strong>ular activity.<br />

Again, all these adaptive processes may be displayed<br />

to a greater or lesser extent in different<br />

<strong>cell</strong>s, different pathways within a <strong>cell</strong> or different<br />

situations during the <strong>cell</strong>’s lifetime.<br />

14.11<br />

Signaling proteins are<br />

frequently expressed as<br />

multiple species<br />

Key concepts<br />

• Distinct species (is<strong>of</strong>orms) <strong>of</strong> similar <strong>signaling</strong><br />

proteins expand the regulatory mechanisms<br />

possible in <strong>signaling</strong> pathways.<br />

• Is<strong>of</strong>orms may differ in function, susceptibility to<br />

regulation or expression.<br />

• Cells may express one or several is<strong>of</strong>orms to fulfill<br />

their <strong>signaling</strong> needs.<br />

Cells increase the richness, adaptability, and<br />

regulation <strong>of</strong> their <strong>signaling</strong> pathways by expressing<br />

multiple species <strong>of</strong> individual <strong>signaling</strong><br />

proteins that display distinct biochemical<br />

properties. These species may be encoded by<br />

multiple genes or by multiple mRNAs derived<br />

from a single gene by alternative splicing or<br />

mRNA editing. The numerical complexity implicit<br />

in these choices is impressive. Consider<br />

the neurotransmitter serotonin: In mammals,<br />

there are thirteen serotonin receptors, each <strong>of</strong><br />

which stimulates a distinct spectrum <strong>of</strong> G proteins<br />

<strong>of</strong> the G i<br />

, G s<br />

, and G q<br />

families. (A fourteenth<br />

serotonin receptor is an ion channel.)<br />

FIGURE 14.12 shows the relationship <strong>of</strong> serotonin<br />

receptors to these G protein families.<br />

There is also tremendous diversity among<br />

the G proteins and adenylyl cyclases. There are<br />

three genes for Gα i<br />

and one each for the closely<br />

related Gα z<br />

and Gα o<br />

. Furthermore, the Gα o<br />

mRNA is multiply spliced. There are four G q<br />

members. In addition, there are five genes for<br />

Gβ and twelve for Gγ, and most <strong>of</strong> the possible<br />

Gβγ dimers are expressed naturally. There are<br />

ten genes for adenylyl cyclases, which are direct<br />

targets <strong>of</strong> G s<br />

and either direct or indirect targets<br />

<strong>of</strong> the other G proteins. While all nine membrane-bound<br />

adenylyl cyclase is<strong>of</strong>orms are stimulated<br />

by Gα s<br />

, they display diverse stimulatory<br />

and inhibitory responses to Gβγ, Gα i<br />

, Ca2+,<br />

calmodulin, and several protein kinases, as illustrated<br />

in FIGURE 14.13. Thus, stimulation by<br />

serotonin can lead to diverse responses depending<br />

upon the various forms <strong>of</strong> the proteins that<br />

are engaged at a particular time and location.<br />

FIGURE 14.12 Receptors for serotonin have<br />

evolved in mammals as a family <strong>of</strong> 13 genes that<br />

regulate three <strong>of</strong> the four major classes <strong>of</strong> G proteins.<br />

While all respond to the natural ligand<br />

serotonin, the binding sites have evolved sufficient<br />

differences that drugs have been developed<br />

that specifically target one or more<br />

is<strong>of</strong>orms. The type 3 serotonin receptors, not<br />

shown here, are ligand-gated ion channels and<br />

are not obviously related to the others.<br />

Evolutionary relationship <strong>of</strong> serotonin receptor is<strong>of</strong>orms<br />

Is<strong>of</strong>orms G protein<br />

1B<br />

1D<br />

1E G i<br />

1F<br />

1A<br />

5A<br />

7<br />

5B<br />

G s<br />

4<br />

2A<br />

2C<br />

G q<br />

6<br />

2B<br />

G s<br />

120 100 80 60 40 20 0<br />

Nucleotide substitution distance<br />

604 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 605<br />

Different is<strong>of</strong>orms <strong>of</strong> adenylyl cyclase are regulated differently<br />

Gα s<br />

Gβγ<br />

PKC<br />

Ca 2+ NO<br />

PKA<br />

Regulators<br />

CaMK<br />

inhibit<br />

Gα i<br />

activate<br />

CaM<br />

FIGURE 14.13 All <strong>of</strong> the mammalian membrane-bound adenylyl cyclases are<br />

structurally homologous and catalyze the same reaction, and all are stimulated<br />

by G s<br />

. Their responses to other inputs (protein kinases CaMK, PKA and PKC;<br />

Ca2+; calmodulin (CaM); NO • ) are specific to each is<strong>of</strong>orm, allowing a rich combinatoric<br />

input to <strong>cell</strong>ular cAMP <strong>signaling</strong>.<br />

Sometimes is<strong>of</strong>orms <strong>of</strong> a <strong>signaling</strong> protein<br />

are subject to quite different kinds <strong>of</strong> inputs.<br />

For example, all <strong>of</strong> the members <strong>of</strong> the phospholipase<br />

C family (PLC) hydrolyze phosphatidylinositol-4,5-bisphosphate<br />

to form two second<br />

messengers, diacylglycerol and inositol-1,4,5<br />

trisphosphate (see 14.16 Lipids and lipid-derived<br />

compounds are <strong>signaling</strong> molecules). The distinct<br />

is<strong>of</strong>orms may be regulated by diverse combinations<br />

<strong>of</strong> Gα q<br />

, Gβγ, phosphorylation, monomeric<br />

G proteins, or Ca2+.<br />

Because a <strong>cell</strong> has multiple options when<br />

expressing a form <strong>of</strong> a <strong>signaling</strong> protein, it can<br />

use expression <strong>of</strong> particular is<strong>of</strong>orms to alter<br />

how it performs otherwise identical <strong>signaling</strong><br />

functions. Different <strong>cell</strong>s express one or more<br />

is<strong>of</strong>orms to allow appropriate responses, and expression<br />

can vary according to other inputs or<br />

the <strong>cell</strong>’s metabolic status. In addition, <strong>signaling</strong><br />

pathways are remarkably resistant to mutational<br />

or other injuries because loss <strong>of</strong> a single species<br />

or is<strong>of</strong>orm <strong>of</strong> a <strong>signaling</strong> protein can <strong>of</strong>ten be<br />

compensated for by increased expression or activity<br />

<strong>of</strong> another species. Similarly, engineered<br />

overexpression can result in the reduced expression<br />

<strong>of</strong> endogenous proteins. The existence <strong>of</strong><br />

multiple receptor species can, thus, substantially<br />

add to adaptability and the consequent resistance<br />

<strong>of</strong> <strong>signaling</strong> networks to damage.<br />

14.12<br />

Activating and<br />

deactivating reactions<br />

are separate and<br />

independently controlled<br />

Key concepts<br />

• Activating and deactivating reactions are usually<br />

executed by different regulatory proteins.<br />

• Separating activation and inactivation allows for<br />

fine-tuned regulation <strong>of</strong> amplitude and timing.<br />

In <strong>signaling</strong> networks, individual proteins are<br />

frequently activated and deactivated by distinct<br />

reactions, a feature that facilitates separate regulation.<br />

Common examples include using protein<br />

kinases and phosphoprotein phosphatases<br />

14.12 Activating and deactivating reactions are separate and independently controlled 605


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 606<br />

to catalyze protein phosphorylation and dephosphorylation;<br />

using adenylyl cyclase to create<br />

cAMP while using phosphodiesterases to<br />

hydrolyze it or anion transporters to pump it<br />

out <strong>of</strong> the <strong>cell</strong>; or using GTP/GDP exchange factors<br />

(GEFs) to activate G proteins and GTPaseactivating<br />

proteins (GAPs) to deactivate them.<br />

Depending on stoichiometry and detailed mechanism,<br />

these strategies can convey either additive<br />

or nonadditive inputs while maintaining<br />

fine control over the kinetics <strong>of</strong> activation and<br />

deactivation <strong>of</strong> a <strong>signaling</strong> pathway. The use <strong>of</strong><br />

distinct reactions for activation and deactivation<br />

is analogous to the use <strong>of</strong> distinct anabolic<br />

and catabolic enzymes in reversible metabolic<br />

pathways.<br />

14.13<br />

Cellular <strong>signaling</strong> uses<br />

both allostery and<br />

covalent modification<br />

Key concepts<br />

• Allostery refers to the ability <strong>of</strong> a molecule to alter<br />

the conformation <strong>of</strong> a target protein when it binds<br />

noncovalently to that protein.<br />

• Modification <strong>of</strong> a protein’s chemical structure is<br />

also frequently used to regulate its activity.<br />

Cellular <strong>signaling</strong> uses almost every imaginable<br />

mechanism for regulating the activities <strong>of</strong><br />

intra<strong>cell</strong>ular proteins, but most can be described<br />

as either allosteric or covalent. Individual <strong>signaling</strong><br />

proteins typically respond to multiple allosteric<br />

and covalent inputs.<br />

Allostery refers to the ability <strong>of</strong> a molecule<br />

to alter the conformation <strong>of</strong> a target protein<br />

when it binds noncovalently to that protein.<br />

Because a protein’s activity reflects its conformation,<br />

the binding <strong>of</strong> any molecule that alters<br />

conformation can change the target protein’s<br />

activity. Any molecule can have allosteric effects:<br />

protons or Ca2+, small organic molecules,<br />

or other proteins. Allosteric regulation can be<br />

both inhibitory or stimulatory.<br />

Covalent modification <strong>of</strong> a protein’s chemical<br />

structure is also frequently used to regulate<br />

its activity. The change in the protein’s chemical<br />

structure alters its conformation and, thus,<br />

its activity. Most regulatory covalent modification<br />

is reversible. The classic and most common<br />

regulatory covalent event is phosphorylation,<br />

in which a phosphoryl group is transferred from<br />

ATP to the protein, most <strong>of</strong>ten to the hydroxyl<br />

group <strong>of</strong> serine (Ser), threonine(Thr), or tyrosine<br />

(Tyr). Enzymes that phosphorylate proteins<br />

are known as protein kinases. Their actions are<br />

opposed by phosphoprotein phosphatases, which<br />

catalyze the hydrolysis <strong>of</strong> the phosphoryl group<br />

to yield free phosphate and restore the unmodified<br />

hydroxyl residue. Other forms <strong>of</strong> covalent<br />

modification are also common and will be addressed<br />

throughout the chapter.<br />

14.14<br />

Second messengers<br />

provide readily diffusible<br />

pathways for information<br />

transfer<br />

Key concepts<br />

• Second messengers can propagate signals between<br />

proteins that are at a distance.<br />

• cAMP and Ca2+ are widely used second messengers.<br />

Signaling pathways make use <strong>of</strong> both proteins<br />

and small molecules according to their distinctive<br />

attributes. A small molecule used as an intra<strong>cell</strong>ular<br />

signal, or second messenger, has a<br />

number <strong>of</strong> advantages over a protein as a <strong>signaling</strong><br />

intermediary. Small molecules can be<br />

synthesized and destroyed quickly. Because they<br />

can be made readily, they can act at high concentrations<br />

so that their affinities for target proteins<br />

can be low. Low affinity permits rapid<br />

dissociation, such that their signals can be terminated<br />

promptly when free second messenger<br />

molecules are destroyed or sequestered. Because<br />

second messengers are small, they also can diffuse<br />

quickly within the <strong>cell</strong>, although many <strong>cell</strong>s<br />

have developed mechanisms to spatially restrict<br />

such diffusion. Second messengers are, thus,<br />

superior to proteins in mediating fast responses,<br />

particularly at a distance. Second messengers<br />

are also useful when signals have to be addressed<br />

to large numbers <strong>of</strong> target proteins simultaneously.<br />

These advantages <strong>of</strong>ten overcome their<br />

lack <strong>of</strong> catalytic activity and their inability to<br />

bind multiple molecules simultaneously.<br />

FIGURE 14.14 lists intra<strong>cell</strong>ular second messengers<br />

developed through evolution. This number<br />

is surprisingly low. Several are nucleotides<br />

synthesized from major metabolic nucleotide<br />

precursors. They include cAMP, cyclic GMP,<br />

ppGppp, and cyclic ADP-ribose. Other soluble<br />

second messengers include a sugar phosphate,<br />

inositol-1,4,5-trisphosphate (IP 3<br />

), a divalent metal<br />

ion Ca2+, and a free radical gas nitric oxide (NO • ).<br />

Lipid second messengers include diacylglycerol<br />

and phosphatidylinositol-3,4,5-trisphosphate,<br />

606 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 607<br />

Second<br />

messenger<br />

3':5'-cyclic AMP<br />

(cAMP)<br />

RNA polymerase<br />

Magic spot<br />

(ppGpp, ppGppp) ObgE transcription<br />

arrest<br />

detector<br />

Cyclic di-GMP<br />

phosphodiesterase<br />

Inositol-1,3,5-<br />

trisphosphate<br />

(IP 3 )<br />

Diacylglycerol<br />

(DAG)<br />

Phosphatidyl-<br />

inositol-4,5-<br />

bisphosphate<br />

(PIP 2 )<br />

3':5'-Cyclic GMP<br />

(cGMP)<br />

Cyclic ADP-ribose<br />

Nitric oxide (NO. )<br />

Ca2+<br />

Cyclic<br />

diguanosinemonophosphate<br />

Phosphatidyl-<br />

inositol-3,4,5-<br />

trisphosphate<br />

Adenylyl<br />

Protein kinase A<br />

cyclase<br />

Bacterial transcription<br />

factors<br />

Cation channel<br />

Cyclic nucleotide<br />

phosphodiesterase<br />

Rap GDP/GTP<br />

exchange factor<br />

(Epac)<br />

IP 3 -gated Ca 2+<br />

channel<br />

Protein<br />

kinase C<br />

Trp cation<br />

channel<br />

Ion channel<br />

Transporters<br />

Protein kinase G<br />

Ca2+ channel<br />

Various two<br />

component<br />

system proteins<br />

Guanylyl cyclase<br />

Numerous<br />

calmodulin<br />

Akt (protein<br />

kinase B)<br />

Second messengers<br />

Targets<br />

Other PH<br />

domains/proteins<br />

Synthesis/<br />

Release<br />

Rel1A<br />

SpoT<br />

Cation channel<br />

Cyclic nucleotide<br />

phosphodiesterase<br />

Phospholipase<br />

C<br />

Phospholipase<br />

C<br />

PIP 5-kinase<br />

Guanylyl<br />

cyclase<br />

ADP-ribose<br />

cyclase<br />

Diguanylate<br />

cyclase<br />

PI 3-kinase<br />

GTP<br />

PIP 2<br />

PIP 2<br />

PI-4-P<br />

GTP<br />

NAD<br />

GTP<br />

Stored<br />

Ca2+<br />

PIP 2<br />

phosphatidylinositol-4,5-diphosphate, sphingosine-1-phosphate<br />

and phosphatidic acid.<br />

The first <strong>signaling</strong> compound to be described<br />

as a second messenger was cAMP. The name<br />

arose because cAMP is synthesized in animal<br />

<strong>cell</strong>s as a second, intra<strong>cell</strong>ular signal in response<br />

to numerous extra<strong>cell</strong>ular hormones, the first<br />

messengers in the pathway. cAMP is used by<br />

prokaryotes, fungi, and animals to convey information<br />

to a variety <strong>of</strong> regulatory proteins.<br />

(Its occurrence in higher plants has still not been<br />

proved.)<br />

Adenylyl cyclases, the enzymes that synthesize<br />

cAMP from ATP, are regulated in various<br />

ways depending on the organism in which<br />

they occur. In animals, adenylyl cyclase is an<br />

integral protein <strong>of</strong> the plasma membrane whose<br />

multiple is<strong>of</strong>orms are stimulated by diverse<br />

agents (see Figure 14.13). In animal <strong>cell</strong>s, adenylyl<br />

cyclase is generally stimulated by G s<br />

, which<br />

was originally discovered as an adenylyl cyclase<br />

regulator. Some fungal adenylyl cyclases are<br />

also stimulated by G proteins. Bacterial cyclases<br />

are far more diverse in their regulation.<br />

cAMP is removed from <strong>cell</strong>s in two ways.<br />

It may be extruded from <strong>cell</strong>s by an ATP-driven<br />

anion pump but is more <strong>of</strong>ten hydrolyzed to 5′-<br />

AMP by members <strong>of</strong> the cyclic nucleotide phosphodiesterase<br />

family, a large group <strong>of</strong> proteins<br />

that are themselves under multiple regulatory<br />

controls.<br />

The prototypical downstream regulator for<br />

cAMP in animals is the cAMP-dependent protein<br />

kinase, but a bacterial cAMP-regulated transcription<br />

factor was discovered shortly thereafter,<br />

and other effectors are now known (Figure<br />

14.14). The cAMP system remains the prototypical<br />

eukaryotic <strong>signaling</strong> pathway in that its<br />

components exemplify almost all <strong>of</strong> the recognized<br />

varieties <strong>of</strong> <strong>signaling</strong> molecules and their<br />

interactions: hormone, receptor, G protein,<br />

adenylyl cyclase, protein kinase, phosphodiesterase,<br />

and extrusion pump.<br />

The second messenger-stimulated protein<br />

kinase PKA is a tetramer composed <strong>of</strong> two catalytic<br />

(C) subunits and two regulatory (R) subunits,<br />

as illustrated in FIGURE 14.15. The R subunit<br />

binds to the catalytic subunit in the substratebinding<br />

region, maintaining C in an inhibited<br />

state. Each R subunit binds two molecules <strong>of</strong><br />

cAMP, four cAMP molecules per PKA holoenzyme.<br />

When these sites are filled, the R subunit<br />

dimer dissociates rapidly, leaving two free catalytic<br />

subunits with high activity. The difference<br />

in affinity <strong>of</strong> R for C in the presence and absence<br />

<strong>of</strong> cAMP is ~10,000-fold. The strongly cooperative<br />

binding <strong>of</strong> cAMP generates a very steep<br />

activation curve with an apparent threshold below<br />

which no significant activation <strong>of</strong> PKA occurs,<br />

as illustrated in Figure 14.15. PKA activity,<br />

thus, increases dramatically over a narrow range<br />

<strong>of</strong> cAMP concentrations. PKA is also regulated<br />

Precursor<br />

Removal<br />

Phosphodiesterase<br />

ATP<br />

Organic<br />

anion<br />

transporter<br />

SpoTcatalyzed<br />

hydrolysis<br />

Phosphatase<br />

Diacylglycerol<br />

kinase<br />

Diacylglycerol<br />

lipase<br />

Phospholipase<br />

C<br />

Phosphatase<br />

Phosphodiesterase<br />

Hydrolysis<br />

NO . synthase arginine Reduction<br />

Release from<br />

storage<br />

organelles<br />

or plasma<br />

membrane<br />

channels<br />

Reuptake<br />

and<br />

extrusion<br />

pumps<br />

Phosphatase<br />

FIGURE 14.14 Major second messengers, some <strong>of</strong> the proteins that they regulate,<br />

their sources and their disposition.<br />

14.14 Second messengers provide readily diffusible pathways for information transfer 607


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 608<br />

(C)<br />

Catalytic<br />

subunits<br />

PKA<br />

C<br />

R<br />

R<br />

C<br />

Activation <strong>of</strong> PKA by cAMP<br />

(R)<br />

R egulatory<br />

subunits<br />

4 cAMP<br />

R<br />

R<br />

- cAMP<br />

- cAMP<br />

- cAMP<br />

- cAMP<br />

Activated<br />

PKA<br />

C<br />

C<br />

teins throughout the <strong>cell</strong> ranging from ion channels<br />

to transcription factors, and its conserved<br />

substrate preference frequently permits prediction<br />

<strong>of</strong> substrates by sequence analysis. The<br />

cAMP response element binding protein CREB<br />

is phosphorylated by PKA on Ser 133 and is<br />

largely responsible for the impact <strong>of</strong> cAMP on<br />

transcription <strong>of</strong> numerous genes.<br />

R 2 C 2 + 4 cAMP<br />

Kinase activity as a<br />

function <strong>of</strong> [cAMP] (%)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

10%<br />

R 2<br />

. cAMP 4 + 2C<br />

90%<br />

14.15<br />

Ca2+ <strong>signaling</strong> serves<br />

diverse purposes in all<br />

eukaryotic <strong>cell</strong>s<br />

Key concepts<br />

• Ca2+ serves as a second messenger and regulatory<br />

molecule in essentially all <strong>cell</strong>s.<br />

• Ca2+ acts directly on many target proteins and also<br />

regulates the activity <strong>of</strong> a regulatory protein<br />

calmodulin.<br />

• The cytosolic concentration <strong>of</strong> Ca2+ is controlled by<br />

organellar sequestration and release.<br />

2 x 10 -9 2 x 10 -8 2 x 10 -7<br />

[cAMP]<br />

FIGURE 14.15 PKA is a heterotetramer composed <strong>of</strong> two catalytic (C) and<br />

two regulatory (R) subunits. Binding <strong>of</strong> four molecules <strong>of</strong> cAMP to the regulatory<br />

subunits induces dissociation <strong>of</strong> two molecules <strong>of</strong> C, the active form<br />

<strong>of</strong> PKA, from the cAMP-bound regulatory subunit dimer. In the bottom panel,<br />

the cooperative binding <strong>of</strong> four molecules <strong>of</strong> cAMP generates a steep activation<br />

pr<strong>of</strong>ile. Activity increases from approximately 10% to 90% as the<br />

cAMP concentration increases only 10-fold. An apparent threshold is introduced<br />

because there is little change in activity at low concentrations <strong>of</strong><br />

cAMP.<br />

by phosphorylation <strong>of</strong> its activation loop.<br />

Phosphorylation occurs cotranslationally, and<br />

the activation loop phosphorylation is required<br />

for assembly <strong>of</strong> the R 2<br />

C 2<br />

tetramer.<br />

The PKAs are mostly cytosolic and are also<br />

targeted to specific locations by binding organelle-associated<br />

scaffolds (A-kinase anchoring<br />

proteins, or AKAPs). These AKAPs facilitate<br />

phosphorylation <strong>of</strong> membrane proteins including<br />

GPCRs, transporters, and ion channels.<br />

AKAPs can also target PKA to other <strong>cell</strong>ular locations<br />

including mitochondria, the cytoskeleton,<br />

and the centrosome. AKAPs <strong>of</strong>ten harbor<br />

binding sites for other regulatory molecules<br />

such as phosphoprotein phosphatases and additional<br />

protein kinases, which allows for coordination<br />

<strong>of</strong> multiple <strong>signaling</strong> pathways and<br />

integration <strong>of</strong> their outputs.<br />

PKA generally phosphorylates substrates<br />

with a primary consensus motif <strong>of</strong> Arg-Arg-<br />

Xaa-Ser-Hydrophobic, placing it in a large group<br />

<strong>of</strong> kinases that recognize basic residues preceding<br />

the phosphorylation site. PKA regulates pro-<br />

Ca2+ is used as a second messenger in all <strong>cell</strong>s,<br />

and is, thus, an even more widespread second<br />

messenger than cAMP. Many proteins bind Ca2+<br />

with consequent allosteric changes in their enzymatic<br />

activities, sub<strong>cell</strong>ular localization, or<br />

interaction with other proteins or with lipids.<br />

Direct targets <strong>of</strong> Ca2+ regulation include almost<br />

all classes <strong>of</strong> <strong>signaling</strong> proteins described in this<br />

chapter, numerous metabolic enzymes, ion<br />

channels and pumps, and contractile proteins.<br />

Most noteworthy may be muscle actomyosin<br />

fibers, which are triggered to contract in response<br />

to cytosolic Ca2+ (see 8.21 Myosin-II functions<br />

in muscle contraction).<br />

Although free Ca2+ is found at concentrations<br />

near 1 mM in most extra<strong>cell</strong>ular fluids, intra<strong>cell</strong>ular<br />

Ca2+ concentrations are maintained<br />

near 100 nanomolar levels by the combined action<br />

<strong>of</strong> pumps and transporters that either extrude<br />

free Ca2+ or sequester it in the endoplasmic<br />

reticulum or mitochondria. Ca2+ <strong>signaling</strong> is initiated<br />

when Ca2+-selective channels in the endoplasmic<br />

reticulum or plasma membrane are<br />

opened to allow Ca2+ to enter the cytoplasm.<br />

The most important entrance channels include<br />

electrically gated channels in animal plasma<br />

membranes; a Ca2+ channel in the endoplasmic<br />

reticulum that is opened by another second messenger,<br />

inositol 1,4,5-trisphosphate (see below);<br />

and an electrically gated channel in the endoplasmic<br />

(sarcoplasmic) reticulum <strong>of</strong> muscle<br />

that opens in response to depolarization <strong>of</strong> nearby<br />

plasma membrane, a process known as excitation-contraction<br />

coupling (see 2.9 Plasma mem-<br />

608 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 609<br />

Calcium binding causes a conformational change in calmodulin<br />

Calcium-free<br />

calmodulin<br />

calmodulin free + 4 Ca 2+<br />

Calcium-bound<br />

calmodulin bound to<br />

target peptide <strong>of</strong> CaMK<br />

Ca 2+<br />

target<br />

(Ca 2+ ) 4<br />

. calmodulin . active target<br />

FIGURE 14.16 Ribbon diagrams representing<br />

the crystal structures <strong>of</strong> calmodulin free<br />

<strong>of</strong> Ca2+ and bound to four Ca2+ ions reveal<br />

the huge conformational change that<br />

calmodulin undergoes upon Ca2+ binding.<br />

Ca2+-calmodulin causes activity changes in<br />

target proteins. The bottom panel shows<br />

the activation <strong>of</strong> a target by calmodulin as<br />

a function <strong>of</strong> the intra<strong>cell</strong>ular free Ca2+ concentration.<br />

The requirement for binding<br />

four Ca2+ ions to induce the conformational<br />

transition results in cooperative activation<br />

<strong>of</strong> targets. Activity increases from 10% to<br />

90% as the Ca2+ concentration increases<br />

only 10-fold. Structures generated from<br />

Protein Data Bank files 1CFD and 1MXE.<br />

Activation <strong>of</strong> target<br />

by calmodulin (%)<br />

100<br />

90%<br />

80<br />

60<br />

40<br />

20<br />

10%<br />

3 x 10 -8 3 x 10 -7 3 x 10 -6<br />

[Ca 2+ ]<br />

brane Ca2+ channels activate intra<strong>cell</strong>ular functions).<br />

In addition to the proteins that are regulated<br />

by binding Ca2+ directly, many other proteins respond<br />

to Ca2+ by binding a widespread Ca2+ sensor,<br />

the small, ~17 kDa protein calmodulin.<br />

Calmodulin requires the binding <strong>of</strong> four molecules<br />

<strong>of</strong> Ca2+ to become fully active, and binding<br />

is highly cooperative, generating a sigmoid<br />

activation pr<strong>of</strong>ile illustrated in FIGURE 14.16.<br />

Calmodulin generally binds its targets in a Ca2+dependent<br />

manner, but Ca2+-free calmodulin<br />

may remain bound but inactive in some cases.<br />

For example, calmodulin is a constitutive subunit<br />

<strong>of</strong> phosphorylase kinase that is activated<br />

upon Ca2+ binding. Higher plants again make<br />

major modifications to this paradigm. Calmodulin<br />

is not expressed as a distinct protein but, instead,<br />

is found as a domain in Ca2+-regulated proteins.<br />

In yet another variation, the adenylyl cyclase secreted<br />

by the pathogenic bacterium Bordetella pertussis<br />

is inactive outside <strong>cell</strong>s but is activated by<br />

Ca2+-free calmodulin in animal <strong>cell</strong>s, where its<br />

rapid production <strong>of</strong> cAMP is highly toxic.<br />

14.16<br />

Lipids and lipid-derived<br />

compounds are <strong>signaling</strong><br />

molecules<br />

Key concepts<br />

• Multiple lipid-derived second messengers are<br />

produced in membranes.<br />

• Phospholipase Cs release soluble and lipid second<br />

messengers in response to diverse inputs.<br />

• Channels and transporters are modulated by<br />

different lipids in addition to inputs from other<br />

sources.<br />

• PI 3-kinase synthesizes PIP 3<br />

to modulate <strong>cell</strong><br />

shape and motility.<br />

• PLD and PLA 2<br />

create other lipid second<br />

messengers.<br />

Signals that originate at the plasma membrane<br />

may have soluble regulatory targets in the cytoplasm<br />

or intra<strong>cell</strong>ular organelles, but integral<br />

plasma membrane proteins are also subject to<br />

acute controls. For these targets, lipid second<br />

messengers may be primary inputs. Lipids derived<br />

from membrane phospholipids or other<br />

14.16 Lipids and lipid-derived compounds are <strong>signaling</strong> molecules 609


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 610<br />

lipid species play numerous roles in <strong>cell</strong> <strong>signaling</strong>.<br />

Because their analysis has been more difficult<br />

than for soluble messengers, many<br />

probably remain to be discovered and understood.<br />

FIGURE 14.17 shows the structure <strong>of</strong> some<br />

<strong>of</strong> these lipids.<br />

Phospholipase Cs (PLCs) are the prototypical<br />

lipid <strong>signaling</strong> enzymes. PLC is<strong>of</strong>orms catalyze<br />

the hydrolysis <strong>of</strong> phospholipids between<br />

the 3-sn-hydroxyl and the phosphate group to<br />

yield a diacylglycerol and phosphate ester. In<br />

animals and fungi, PLCs specific for the substrate<br />

Structures <strong>of</strong> some lipid second messengers<br />

O<br />

Phosphatidylinositol (PI)<br />

O<br />

O<br />

O<br />

HO<br />

OH<br />

O<br />

2<br />

O<br />

3<br />

P<br />

1<br />

O -<br />

O<br />

6 OH<br />

4 5<br />

OH<br />

OH<br />

O<br />

Phosphatidylinositol-3,4,5-trisphosphate (PIP 3 )<br />

O<br />

O<br />

O<br />

H - O 3 PO<br />

HO<br />

OH<br />

OH<br />

O -<br />

O<br />

P<br />

O<br />

O<br />

6 OH<br />

2 1 4 5<br />

3<br />

OPO 3 H -<br />

OPO 3 H -<br />

O<br />

Diacylglycerol (DAG)<br />

O<br />

O<br />

O<br />

OH<br />

OPO 3 H -<br />

6 OH<br />

2 1 4 5<br />

3<br />

Inositol trisphosphate (IP 3 )<br />

OPO 3 H -<br />

OPO 3 H -<br />

O<br />

Phosphatidic acid (PA)<br />

O<br />

O<br />

O<br />

O<br />

O-<br />

P<br />

O -<br />

O<br />

FIGURE 14.17 Structures <strong>of</strong> some lipid second messengers and the common precursor phosphatidylinositol.<br />

The acyl side chain structures shown here are the most common for mammalian PI lipids. Much <strong>of</strong> the PA in<br />

<strong>cell</strong>s is derived from PC, and its acyl chains may differ from those shown.<br />

610 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 611<br />

phosphatidylinositol-4,5-bisphosphate (PIP 2<br />

)<br />

hydrolyze PIP 2<br />

to form two second messengers:<br />

1,2-sn-diacylglycerol (DAG) and inositol-1,4,5-<br />

trisphosphate (IP 3<br />

). The PLC substrate PIP 2<br />

is itself<br />

an important regulatory ligand that<br />

modulates the activity <strong>of</strong> several ion channels,<br />

transporters, and enzymes. Thus, PLC alters concentration<br />

<strong>of</strong> three second messengers; its net<br />

effect depends on the net turnover <strong>of</strong> the substrate<br />

and products.<br />

DAG is probably the best known lipid second<br />

messenger; its hydrophobicity limits it to action<br />

in membranes. DAG activates some is<strong>of</strong>orms<br />

<strong>of</strong> protein kinase C (PKC), modulates the activity<br />

<strong>of</strong> several cation channels and activates at<br />

least one other protein kinase. DAG can be further<br />

hydrolyzed to release arachidonic acid,<br />

which can regulate some ion channels.<br />

Arachidonic acid is also the precursor <strong>of</strong> oxidation<br />

products, such as prostaglandins and thromboxanes,<br />

which are potent extra<strong>cell</strong>ular <strong>signaling</strong><br />

agents. In addition to DAG, PKCs require interaction<br />

with Ca2+ and an acidic phospholipid,<br />

such as phosphatidylserine, to become activated.<br />

Thus, activation <strong>of</strong> PKC requires the coincidence<br />

<strong>of</strong> multiple inputs both to generate DAG and to<br />

increase intra<strong>cell</strong>ular Ca2+. There are more than<br />

a dozen PKCs, classified together according to<br />

highly conserved sequences in the catalytic domain.<br />

Three subgroups <strong>of</strong> PKCs, also identifiable<br />

by sequence, share different patterns <strong>of</strong> regulation.<br />

Their regulation provides examples <strong>of</strong><br />

many ways in which other mammalian protein<br />

kinases are regulated.<br />

The first <strong>of</strong> these groups, canonical PKCs,<br />

are generally soluble or very loosely associated<br />

with membranes prior to the appearance <strong>of</strong><br />

DAG. DAG causes their association with membranes<br />

and permits activation upon binding <strong>of</strong><br />

other regulators. The second group <strong>of</strong> PKCs requires<br />

similar lipids but not Ca2+, and the third<br />

group requires other lipids but neither DAG nor<br />

Ca2+ for activation.<br />

The N-terminal region <strong>of</strong> PKCs contains a<br />

pseudosubstrate domain, a sequence that resembles<br />

that <strong>of</strong> a typical substrate except that<br />

the target Ser is replaced with Ala. The pseudosubstrate<br />

region binds to the active site to inhibit<br />

the kinase. Activators cause the<br />

pseudosubstrate domain to flip out <strong>of</strong> the active<br />

site. PKCs are also activated by proteolysis,<br />

as are many protein kinases with discrete<br />

autoinhibitory domains. Proteases clip a flexible<br />

hinge region, which results in loss <strong>of</strong> the<br />

regulatory domain and consequent activation<br />

<strong>of</strong> the kinase.<br />

PKC is the major receptor for phorbol esters,<br />

a class <strong>of</strong> powerful tumor promoters. Phorbol<br />

esters mimic DAG and cause a more massive<br />

and prolonged activation than physiological<br />

stimuli. This massive stimulation can induce<br />

proteolysis <strong>of</strong> PKC, resulting in downregulation,<br />

or loss <strong>of</strong> the kinase. (For a personal description<br />

on the discovery <strong>of</strong> protein kinase C<br />

see EXP : 14-0001 )<br />

IP 3<br />

, the second product <strong>of</strong> the PLC reaction,<br />

is a soluble second messenger. The most significant<br />

IP 3<br />

target is a Ca2+ channel in the endoplasmic<br />

reticulum. IP 3<br />

causes this channel to<br />

open and release stored Ca2+ into the cytoplasm,<br />

thereby rapidly elevating the cytosolic Ca2+ over<br />

100-fold and, in turn, causing the activation <strong>of</strong><br />

numerous targets <strong>of</strong> Ca2+ <strong>signaling</strong>.<br />

There are at least six families <strong>of</strong> PIP 2<br />

-selective<br />

PLC enzymes, defined by their distinct forms<br />

<strong>of</strong> regulation, domain compositions, and overall<br />

sequence conservation. Their catalytic domains<br />

are all quite similar. The PLC-βs are<br />

stimulated primarily by Gα q<br />

and Gβγ (to individually<br />

varying extents). Several are also modulated<br />

by phosphorylation. PLC-γ is<strong>of</strong>orms are<br />

stimulated by phosphorylation on Tyr residues,<br />

frequently by receptor tyrosine kinases. The<br />

PLC-ε is<strong>of</strong>orms are regulated by small,<br />

monomeric G proteins <strong>of</strong> the Rho family. The<br />

regulation <strong>of</strong> the PLC-δs is still incompletely understood.<br />

Two other classes similar to the PLCδs,<br />

PLC-η and -ζ, have also been defined recently.<br />

(There is no PLC-α.) In addition to their distinct<br />

modes <strong>of</strong> regulation, all <strong>of</strong> the PLCs are stimulated<br />

by Ca2+, and Ca2+ <strong>of</strong>ten acts synergistically<br />

with other stimulatory inputs. This synergy<br />

underlies the intensification and prolongation<br />

<strong>of</strong> Ca2+ <strong>signaling</strong> observed in many <strong>cell</strong>s.<br />

Phospholipases A 2<br />

and D (PLA 2<br />

and PLD)<br />

also hydrolyze glycerol phospholipids in <strong>cell</strong><br />

membranes to form important <strong>signaling</strong> compounds.<br />

PLA 2<br />

hydrolyzes the fatty acid at the sn-<br />

2 position <strong>of</strong> multiple phospholipids to produce<br />

the cognate lysophospholipid and the free fatty<br />

acid, which is generally unsaturated. The free<br />

fatty acid is <strong>of</strong>ten arachidonic acid, a precursor<br />

<strong>of</strong> extra<strong>cell</strong>ular signals. The biological roles <strong>of</strong><br />

free lysophospholipids are not understood in<br />

detail but have been linked to effects on the<br />

structure <strong>of</strong> the membrane bilayer.<br />

PLD catalyzes a reaction much like that <strong>of</strong><br />

PLC but instead hydrolyzes the phosphodiester<br />

on the substituent side <strong>of</strong> the phosphate group<br />

to form 3-sn-phosphatidic acid. Cellular PLDs act<br />

on multiple glycerol phospholipid substrates,<br />

but phosphatidylcholine is probably the sub-<br />

14.16 Lipids and lipid-derived compounds are <strong>signaling</strong> molecules 611


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 612<br />

FIGURE 14.18 Activated PI 3-<br />

kinase phosphorylates PIP 2<br />

to produce<br />

PIP 3<br />

. The PH domain-containing<br />

protein kinases PDK1 and Akt<br />

bind to PIP 3<br />

at the plasma membrane.<br />

Their colocalization facilitates<br />

the phosphorylation <strong>of</strong> Akt<br />

by PDK1. A second phosphorylation<br />

within a hydrophobic motif results<br />

in Akt activation by one <strong>of</strong><br />

several candidate protein kinases.<br />

The Akt-2 is<strong>of</strong>orm is required to<br />

elicit hallmark actions <strong>of</strong> insulin.<br />

p85 p110<br />

PI 3-kinase<br />

PIP 2 PIP 3<br />

strate most relevant to <strong>signaling</strong> functions. The<br />

functions <strong>of</strong> the phosphatidic acid product,<br />

which is also formed by phosphorylation <strong>of</strong><br />

DAG, remain poorly understood but appear to<br />

include a role in secretion and the fusion <strong>of</strong> intra<strong>cell</strong>ular<br />

membranes.<br />

14.17<br />

PI 3-kinase regulates<br />

both <strong>cell</strong> shape and the<br />

activation <strong>of</strong> essential<br />

growth and metabolic<br />

functions<br />

Key concepts<br />

• Phosphorylation <strong>of</strong> some lipid second messengers<br />

changes their activity.<br />

• PIP 3<br />

is recognized by proteins with a pleckstrin<br />

homology domain.<br />

Lipid second messengers may also be modified<br />

by phosphorylation. PI 3-kinase phosphorylates<br />

PIP 2<br />

on the 3-position <strong>of</strong> the inositol ring to<br />

form PI 3,4,5-P 3<br />

, another lipid second messenger.<br />

The total activity <strong>of</strong> PI 3-kinase is too low<br />

to significantly deplete total PIP 2<br />

, but formation<br />

<strong>of</strong> small amounts <strong>of</strong> PIP 3<br />

in localized membrane<br />

domains is vital for altering <strong>cell</strong> shape<br />

and <strong>cell</strong>ular motility.<br />

PIP 3<br />

acts by recruiting proteins that contain<br />

PIP 3<br />

binding domains, including pleckstrin<br />

homology (PH) and FYVE domains, to sites<br />

where they regulate cytoskeletal remodeling,<br />

contractile protein function, or other regulatory<br />

events. These proteins anchor and/or orient<br />

the structural or motor proteins involved<br />

in <strong>cell</strong>ular movement and localize <strong>signaling</strong> proteins<br />

to sites <strong>of</strong> action at the membrane. PIP 3<br />

PIP 3 binding brings Akt and PDK1 to the membrane<br />

PIP 2 phosphorylated<br />

Akt<br />

Akt and PDK1<br />

bind PIP3 through<br />

PH domains<br />

Akt<br />

PH<br />

domains<br />

PDK1<br />

PDK1<br />

Akt is activated by<br />

phosphorylation<br />

Other<br />

kinase<br />

Glucose uptake<br />

Glycogen synthesis<br />

Antilipolysis<br />

Antiapoptosis<br />

<strong>signaling</strong> can be fast and dramatic; it largely accounts<br />

for directing the mobility <strong>of</strong> motile mammalian<br />

<strong>cell</strong>s.<br />

Lipid mediators are essential in the insulin<br />

<strong>signaling</strong> pathway. The binding <strong>of</strong> insulin stimulates<br />

the Tyr autophosphorylation <strong>of</strong> its receptor<br />

and the activation <strong>of</strong> effectors through insulin receptor<br />

substrate (IRS) proteins (see 14.30 Diverse<br />

<strong>signaling</strong> mechanisms are regulated by protein tyrosine<br />

kinases). PI 3-kinase is activated when its p85<br />

subunit binds to IRS1. The PIP 3<br />

generated by PI3-<br />

kinase binds the protein kinases Akt and phosphoinositide-dependent<br />

kinase-1 (PDK-1) via their PH<br />

domains. This interaction results in the localization<br />

<strong>of</strong> Akt to the membrane where it is activated<br />

by PDK1, as illustrated in FIGURE 14.18. Akt phosphorylates<br />

downstream targets, including protein<br />

kinases, GAPs, and transcription factors.<br />

Activation <strong>of</strong> Akt, specifically Akt-2, is required<br />

for the hallmark actions <strong>of</strong> insulin including regulation<br />

<strong>of</strong> glucose transporter translocation, enhanced<br />

protein synthesis, and expression <strong>of</strong><br />

gluconeogenic and lipogenic enzymes.<br />

14.18<br />

Signaling through ion<br />

channel receptors is very<br />

fast<br />

Key concepts<br />

• Ion channels allow the passage <strong>of</strong> ions through a<br />

pore, resulting in rapid (microsecond) changes in<br />

membrane potential.<br />

• Channels are selective for particular ions or for<br />

cations or anions.<br />

• Channels regulate intra<strong>cell</strong>ular concentrations <strong>of</strong><br />

regulatory ions, such as Ca2+.<br />

Ligand-gated ion channels are multisubunit,<br />

membrane-spanning proteins that create and<br />

regulate a water-filled pore through the membrane,<br />

as illustrated in the X-ray crystal structure<br />

<strong>of</strong> the nicotinic acetylcholine receptor in<br />

FIGURE 14.19. When stimulated by extra<strong>cell</strong>ular<br />

agonists, the subunits rearrange their conformations<br />

and orientations to open the pore and, thus,<br />

connect the aqueous spaces on either side <strong>of</strong> the<br />

membrane. The pore has a diameter that allows<br />

ions to diffuse freely from one side <strong>of</strong> the membrane<br />

to the other, driven by the electrical and<br />

chemical gradients that have been established<br />

by ion pumps and transporters. (For more about<br />

channel, pump and transporter mechanics see 2<br />

Transport <strong>of</strong> ions and small molecules across membranes.)<br />

Channels maintain selectivity among<br />

ions by regulating the pore diameter precisely<br />

612 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 613<br />

and by lining the walls <strong>of</strong> the pore with appropriate<br />

hydrophilic residues. Receptor ion channels<br />

can, thus, provide a diffusion path for only<br />

cations or anions, or select among different ions.<br />

Ligand-gated ion channels provide the<br />

fastest signal transduction mechanism found in<br />

biology. Upon binding an agonist ligand, channels<br />

open within microseconds. At synapses,<br />

where neurotransmitters need to diffuse less<br />

than 0.1 micron, a signal in the postsynaptic<br />

<strong>cell</strong> can be generated in 100 microseconds. In<br />

contrast, receptor-stimulated G proteins require<br />

about 100 milliseconds to exchange GDP for<br />

GTP, and the action <strong>of</strong> receptor protein kinases<br />

is even slower. Ligand-gated ion channels are<br />

important receptors in many <strong>cell</strong>s in addition<br />

to neurons and muscle, and other ion channels<br />

play equally vital roles in <strong>signaling</strong> pathways<br />

triggered by other classes <strong>of</strong> ligands.<br />

Ion channel <strong>signaling</strong> differs from that <strong>of</strong><br />

the other receptors mentioned in this chapter<br />

in that there is no immediate protein target nor,<br />

in most cases, is there a specific second messenger<br />

involved. In most cases, channel-mediated<br />

ion flow acts to increase or decrease the <strong>cell</strong>’s<br />

membrane potential and, thus, modulates all<br />

transport processes for metabolites or ions that<br />

are electrically driven.<br />

Animal <strong>cell</strong>s maintain an inside-negative<br />

membrane potential by pumping out Na + ions<br />

and pumping in K + ions (for more on membrane<br />

potential see 2.4 Electrochemical gradients<br />

across the <strong>cell</strong> membrane generate the membrane potential).<br />

The opening <strong>of</strong> a channel selective for<br />

Na+ will thus depolarize <strong>cell</strong>s, and the opening<br />

<strong>of</strong> a channel for K + will hyperpolarize <strong>cell</strong>s.<br />

Similarly, because Cl - is primarily extra<strong>cell</strong>ular,<br />

opening Cl - channels will also cause hyperpolarization.<br />

These electrical effects convey information<br />

to effector proteins that are energetically<br />

coupled to the membrane potential, or to specific<br />

ion gradients, or that bind a specific ion<br />

(such as Ca2+) whose concentration changes<br />

upon channel opening.<br />

The nicotinic acetylcholine receptor is the prototypical<br />

receptor ion channel and was the first<br />

receptor that was shown to be a channel. It is a relatively<br />

unselective cation channel that causes depolarization<br />

<strong>of</strong> the target <strong>cell</strong> by allowing Na+<br />

influx. It is best known as the excitatory receptor<br />

at the neuromuscular synapse, where it triggers<br />

contraction, but alternative is<strong>of</strong>orms are also active<br />

in neurons and many other <strong>cell</strong>s. In muscles,<br />

nicotinic depolarization acts via a voltage-sensitive<br />

Ca2+ channel to allow Ca2+ release from the sarcoplasmic<br />

reticulum into the cytosol. Calcium acts<br />

Nicotinic acetylcholine receptor structure<br />

CLOSED<br />

CYTOSOL<br />

Pore<br />

OPEN<br />

Pore<br />

FIGURE 14.19 The nicotinic cholinergic receptor is a cation-selective channel<br />

that is composed <strong>of</strong> five homologous but usually nonidentical subunits that<br />

oligomerize to form a primarily -helical membrane-spanning core. The channel<br />

itself is created within this core, and its opening and closing are executed<br />

by cooperative changes in subunit arrangement. Structure generated from<br />

Protein Data Bank file 2BG9.<br />

as a second (or third) messenger to initiate contraction<br />

(see 2.13 Cardiac and skeletal muscles are activated<br />

by excitation-contraction coupling). Nicotinic<br />

receptors promote exocytosis in some secretory<br />

<strong>cell</strong>s by a similar mechanism, where Ca2+ triggers<br />

the exocytic event. In neurons, where nicotinic<br />

stimulation causes an action potential (depolarization<br />

that is rapidly propagated along the neuron),<br />

the initial depolarization is sensed by<br />

voltage-sensitive Na+ channels. Their opening<br />

(along with the action <strong>of</strong> other channels) propagates<br />

the action potential along the neuron.<br />

The nervous system is rich in receptor cation<br />

channels that respond to other neurotransmitters,<br />

the most common <strong>of</strong> which is the amino<br />

acid glutamate (Glu). The three different families<br />

<strong>of</strong> glutamate receptors share the property<br />

<strong>of</strong> cation conductance, but each family has its<br />

own spectrum <strong>of</strong> drug responses. All operate as<br />

neuronal activators, with one interesting twist:<br />

The NMDA family <strong>of</strong> receptors, named for their<br />

response to a selective drug, is permeant to Ca2+<br />

in addition to Na+. A significant component <strong>of</strong><br />

its activity is to permit the inward flow <strong>of</strong> Ca2+,<br />

which acts as a second messenger on a wide variety<br />

<strong>of</strong> targets. Persistant stimulation <strong>of</strong> NMDA<br />

channels by glutamate released during injury,<br />

or by drugs, can cause toxic amounts <strong>of</strong> Ca2+ to<br />

enter, resulting in neuronal death.<br />

A second functional group <strong>of</strong> receptor channels<br />

is selective for anions and, by allowing inward<br />

flux <strong>of</strong> Cl - , hyperpolarizes the target <strong>cell</strong>.<br />

Anion-selective receptors include those for γ-<br />

aminobutyric acid (GABA) and glycine (Gly). In<br />

neurons, hyperpolarization can inhibit the initiation<br />

<strong>of</strong> an action potential and/or neurotransmitter<br />

release.<br />

14.18 Signaling through ion channel receptors is very fast 613


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 614<br />

Perhaps the most diverse family <strong>of</strong> ligandgated<br />

channels is that <strong>of</strong> the TRP and TRP-like<br />

family, <strong>of</strong> which about 30 have been found in<br />

mammals. Distinct forms are found in invertebrates.<br />

The TRP channels are Ca2+-selective<br />

channels that are formed by tetramers <strong>of</strong> identical<br />

subunits that surround the central channel.<br />

Each subunit is composed <strong>of</strong> a homologous<br />

bundle <strong>of</strong> six membrane-spanning helices, but<br />

the N and C termini contain a diverse collection<br />

<strong>of</strong> regulatory and protein interaction domains,<br />

including protein kinase domains (whose<br />

substrates are currently unknown).<br />

All TRP channels allow transmembrane flux<br />

<strong>of</strong> Ca2+ to permit its action as a second messenger,<br />

but different TRP is<strong>of</strong>orms serve numerous<br />

physiological functions. The prototypical TRP,<br />

found in invertebrate photoreceptors, gates Ca2+<br />

flow from intra<strong>cell</strong>ular stores into the cytoplasm<br />

to initiate visual <strong>signaling</strong>. Others admit Ca2+<br />

from outside the <strong>cell</strong>, and still others allow Ca2+<br />

to enter the endoplasmic reticulum virtually directly<br />

from the extra<strong>cell</strong>ular space because they<br />

form a bridge between the plasma membrane<br />

and channels in the endoplasmic reticulum at<br />

points where the membranes abut each other.<br />

Regulation <strong>of</strong> TRP channels is perhaps even<br />

more diverse. Various TRP channels respond to<br />

heat, cold, painful stimuli, pressure, and high<br />

or low osmolarity. Many TRPs are regulated either<br />

positively or negatively by lipids, such as<br />

eicosanoids, diacylglycerol, and PIP 2<br />

. For example,<br />

capsaicin, the hot compound in chilis, is<br />

an agonist for some vanilloid receptors (TRPVs).<br />

Still other TRP channels are mechanosensors<br />

that allow cilia to sense fluid flow. The most famous<br />

<strong>of</strong> these is the sensory channel <strong>of</strong> the hair<br />

<strong>cell</strong> <strong>of</strong> the inner ear. This channel opens when<br />

the apical cilia on the hair <strong>cell</strong> are bent in response<br />

to sound-driven fluid flow.<br />

14.19<br />

Nuclear receptors<br />

regulate transcription<br />

Key concepts<br />

• Nuclear receptors modulate transcription by<br />

binding to distinct short sequences in<br />

chromosomal DNA known as response elements.<br />

• Receptor binding to other receptors, inhibitors, or<br />

coactivators leads to complex transcriptional<br />

control circuits.<br />

• Signaling through nuclear receptors is relatively<br />

slow, consistent with their roles in adaptive<br />

responses.<br />

Nuclear receptors are unique among <strong>cell</strong>ular<br />

receptors in that their ligands pass unaided<br />

through the plasma membrane. These receptors,<br />

when complexed with their ligands, enter<br />

the nucleus and regulate gene transcription.<br />

Ligands for nuclear receptors include sex steroids<br />

(estrogen and testosterone) and other steroid<br />

hormones, vitamins A and D, retinoids and<br />

other fatty acids, oxysterols, and bile acids.<br />

Nuclear receptors are structurally conserved.<br />

They consist <strong>of</strong> a C-terminal ligand binding domain,<br />

an N-terminal interaction region that recognizes<br />

components <strong>of</strong> the transcriptional<br />

machinery and acts as a transactivation domain,<br />

a centrally located zinc finger domain that binds<br />

DNA, and, <strong>of</strong>ten, another transactivation domain<br />

nearer the C-terminus. In the absence <strong>of</strong><br />

ligand, these receptors are bound to corepressor<br />

proteins that suppress their activity. Upon<br />

hormone binding, corepressors dissociate and<br />

the receptors are assembled in multiprotein<br />

complexes with coactivators that modulate receptor<br />

action and facilitate transcriptional regulation.<br />

As illustrated in FIGURE 14.20, agonists<br />

and antagonists bind to distinct receptor conformations<br />

(see 14.5 Ligand binding changes receptor<br />

conformation). Receptor agonists favor the binding<br />

<strong>of</strong> receptors to coactivators and DNA, and<br />

antagonists favor conformations that block coactivator-receptor<br />

binding.<br />

Nuclear receptors bind with high specificity<br />

to hormone response elements in the 5’ untranscribed<br />

region <strong>of</strong> regulated genes. Response<br />

elements are typically short direct or inverted<br />

repeat sequences, and a gene may contain response<br />

elements for several different receptors<br />

in addition to binding sites for other transcriptional<br />

regulatory proteins.<br />

The sex steroid estrogen can bind to two<br />

different nuclear receptors, the estrogen receptors<br />

ER and ER. Coactivator and corepressor<br />

proteins differentially regulate ER and ER in<br />

transcriptional complexes that are expressed in<br />

specific <strong>cell</strong> types. Other ligands that bind to<br />

these receptors include valuable therapeutic<br />

agents. For example, 4 hydroxy-tamoxifen is<br />

an estrogen receptor antagonist used in the therapy<br />

<strong>of</strong> estrogen-receptor-positive breast cancer<br />

to inhibit growth <strong>of</strong> residual cancer <strong>cell</strong>s.<br />

However, unlike its antagonistic effects on the<br />

estrogen receptor in breast, 4 hydroxy-tamoxifen<br />

displays weak partial agonist activity in<br />

uterus. In the estrogen receptor system, partial<br />

agonists are known as selective estrogen receptor<br />

modulators (SERMs). Properties that contribute<br />

to partial agonist activity include the<br />

relative expression <strong>of</strong> the two estrogen recep-<br />

614 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 615<br />

Estrogen receptor conformation depends on which ligand is bound<br />

Agonist-bound<br />

conformation<br />

Antagonist-bound<br />

conformation<br />

N<br />

N<br />

K362<br />

H11<br />

H5<br />

K362<br />

H5<br />

545<br />

C<br />

H6<br />

545<br />

542<br />

H3<br />

H11<br />

H6<br />

538<br />

542<br />

H3<br />

538<br />

FIGURE 14.20 The estrogen receptor adopts different conformations when<br />

bound to agonists and antagonists. The ligand-binding domain <strong>of</strong> the estrogen<br />

receptor is bound to the agonist estradiol on the left and to the antagonist<br />

raloxifene on the right. Note the marked difference in position <strong>of</strong> helix 12,<br />

shown in blue in the active structure and green in the inhibited structure.<br />

Reproduced from Brzozowski, A. M., et al. 1997. Molecular basis <strong>of</strong> agonism and<br />

antagonism in the oestrogen receptor. Nature. 389: 753–758. Photo courtesy<br />

<strong>of</strong> M. Brzozowski, University <strong>of</strong> New York.<br />

tors, ER and ER, as well as the expression <strong>of</strong><br />

repressors and coactivators that interact with<br />

each receptor type. Thus, the behavior <strong>of</strong> nuclear<br />

receptor ligands must be considered in the<br />

tissue, <strong>cell</strong>ular, and <strong>signaling</strong> context.<br />

14.20<br />

G protein <strong>signaling</strong><br />

modules are widely used<br />

and highly adaptable<br />

Key concepts<br />

• The basic module is a receptor, a G protein and an<br />

effector protein.<br />

• Cells express several varieties <strong>of</strong> each class <strong>of</strong><br />

proteins.<br />

• Effectors are heterogeneous and initiate diverse<br />

<strong>cell</strong>ular functions.<br />

Activation <strong>of</strong> G protein-coupled receptors<br />

(GPCRs) and their associated heterotrimeric G<br />

proteins is one <strong>of</strong> the most widespread mechanisms<br />

<strong>of</strong> communicating extra<strong>cell</strong>ular signals<br />

to the intra<strong>cell</strong>ular environment. G protein <strong>signaling</strong><br />

modules are found in all eukaryotes.<br />

Depending on the species, mammals express<br />

500-1000 GPCRs that respond to hormones,<br />

neurotransmitters, pheromones, metabolites,<br />

local <strong>signaling</strong> substances, and other regulatory<br />

molecules. Essentially all chemical classes are<br />

represented among the GPCR ligands. In addition,<br />

a roughly equal number <strong>of</strong> olfactory GPCRs<br />

are expressed in olfactory neurons and work in<br />

combination to screen compounds in the animal’s<br />

environment via the sense <strong>of</strong> smell.<br />

Because GPCRs are involved in many kinds <strong>of</strong><br />

physiologic responses, they are also one <strong>of</strong> the<br />

most widely used targets for drugs.<br />

A minimal G protein <strong>signaling</strong> module consists<br />

<strong>of</strong> three proteins: a G protein-coupled receptor,<br />

the heterotrimeric G protein, and an<br />

effector protein, as illustrated in FIGURE 14.21. The<br />

receptor activates the G protein on the inner face<br />

<strong>of</strong> the plasma membrane in response to an extra<strong>cell</strong>ular<br />

ligand. The G protein then activates (or<br />

occasionally inhibits) an effector protein that<br />

propagates a signal within the <strong>cell</strong>. Thus, signal<br />

conduction in the simplest G protein module is<br />

linear. However, as depicted in FIGURE 14.22, a<br />

typical animal <strong>cell</strong> may express a dozen GPCRs,<br />

more than six G proteins, and a dozen effectors.<br />

Each GPCR regulates one or more G proteins,<br />

and each G protein regulates several effectors.<br />

Moreover, distinct efficiencies and rates govern<br />

each interaction. Thus, a <strong>cell</strong>’s G protein network<br />

is actually a signal-integrating computer whose<br />

14.20 G protein <strong>signaling</strong> modules are widely used and highly adaptable 615


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 616<br />

Agonist<br />

GPCR<br />

Trimeric<br />

G protein<br />

Receptor<br />

activated<br />

CYTOSOL<br />

Heterotrimeric G protein <strong>signaling</strong><br />

Activated<br />

G protein<br />

dissociates<br />

IP 3 -gated<br />

Ca 2+ channel<br />

ENDOPLASMIC<br />

RETICULUM<br />

PIP 2<br />

IP 3<br />

Ca 2+<br />

DAG<br />

Hydrolysis <strong>of</strong> PIP 2<br />

to IP 3 and DAG<br />

Release <strong>of</strong> Ca 2+<br />

FIGURE 14.21 G protein-mediated signal transduction follows a path <strong>of</strong> agonist<br />

to receptor to heterotrimeric G protein to effector to the effector's output.<br />

Both G and G subunits regulate distinct effectors. In the example<br />

shown here, G q<br />

regulates a phospholipase C- to produce two second messengers,<br />

diacyglycerol (DAG) and inositol-trisphosphate (IP 3<br />

). IP 3<br />

triggers Ca2+ release<br />

from the endoplasmic reticulum.<br />

-<br />

output is a spectrum <strong>of</strong> <strong>cell</strong>ular signals that is<br />

complex in both amplitude and kinetics. Because<br />

<strong>of</strong> their conserved parts list, G protein modules<br />

are well suited to initiating a wide variety <strong>of</strong> intra<strong>cell</strong>ular<br />

signals in response to diverse molecular<br />

inputs and can do so over a wide range <strong>of</strong><br />

time scales (milliseconds to minutes).<br />

GPCRs are integral plasma membrane proteins<br />

composed <strong>of</strong> a bundle <strong>of</strong> seven hydrophobic<br />

membrane-spanning helices with an<br />

extra<strong>cell</strong>ular N terminus and cytosolic C terminus,<br />

as depicted in FIGURE 14.23. Based on the<br />

three-dimensional structure <strong>of</strong> rhodopsin and<br />

on copious biochemical and genetic data, it is<br />

likely that all GPCRs share the same basic mechanism<br />

<strong>of</strong> conformational activation and deactivation<br />

in response to activating ligands (see 14.5<br />

Ligand binding changes receptor conformation).<br />

Binding <strong>of</strong> agonist ligand on the extra<strong>cell</strong>ular<br />

face <strong>of</strong> the receptor drives realignment <strong>of</strong> the helices<br />

to alter the structure <strong>of</strong> a binding site for the<br />

heterotrimeric G protein on the cytoplasmic face,<br />

and this altered conformation <strong>of</strong> the G proteinbinding<br />

surface promotes G protein activation.<br />

FIGURE 14.22 A portion <strong>of</strong> the G protein-mediated<br />

<strong>signaling</strong> network in<br />

macrophages highlights some <strong>of</strong> the<br />

complexity <strong>of</strong> interactions possible in<br />

such systems. Several receptors and G<br />

protein subunits are omitted. Where a<br />

named G protein is shown, its <strong>signaling</strong><br />

output is probably mediated by its G<br />

subunit. Activation <strong>of</strong> any G protein also<br />

activates its G subunit, although Gmediated<br />

<strong>signaling</strong> is usually most prominent<br />

from G i<br />

trimers. In addition, several<br />

G proteins modulate the activities <strong>of</strong><br />

others through poorly understood pathways.<br />

Only a small sampling <strong>of</strong> effectors<br />

is shown, and the only adaptive mechanism<br />

shown is GRK-catalyzed phosphorylation<br />

<strong>of</strong> receptors. Data from Paul<br />

Sternweis, Alliance for Cellular Signaling.<br />

Agonist C5a ISO PGE S1P UDP <strong>UT</strong>P PAF LPA<br />

GPCR<br />

G Protein<br />

Effector<br />

Partial G protein <strong>signaling</strong> network in mouse macrophages<br />

C5aR β 2 AR E2R EDG P2YR P2YR PAFR EDG<br />

Gi Gs G12<br />

Gq<br />

Ad<br />

Cyc<br />

cAMP<br />

ATP<br />

PI 3-<br />

Kinase<br />

PIP 2<br />

PLC-β<br />

G12<br />

??<br />

PIP 3<br />

DAG +<br />

IP 3 IP 2 + P i<br />

Phosphatase<br />

Inactivation<br />

mechanisms<br />

cAMP<br />

AMP<br />

PDE<br />

Ca2+<br />

Ca2+<br />

pump<br />

IP 3 R<br />

GRK<br />

616 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 617<br />

Structure <strong>of</strong> rhodopsin<br />

Heterotrimeric G protein structure<br />

CYTOPLASM<br />

MEMBRANE<br />

Retinal<br />

FIGURE 14.23 The figure shows the crystal structure<br />

<strong>of</strong> the GPCR rhodopsin. Each membrane-spanning helix<br />

is a different color; most structures on the cytoplasmic<br />

face are not shown. The retinal chromophore is<br />

shown within the helix bundle. GPCR sequence similarity<br />

separates the mammalian GPCRs into at least four<br />

structural families that are so diverse that there may<br />

be little sequence similarity among the classes. Within<br />

a family, similarity is greatest in the membrane-spanning<br />

helices, less in the interhelical loops, and least in<br />

the N- and C-terminal domains and in the cytoplasmic<br />

loop that connects spans five and six. Regardless, the<br />

generalizations about functional domains in receptors<br />

seem to hold true within different families. GPCRs frequently<br />

form dimers, occasionally heterodimers, and<br />

dimerization can be crucial for function. Structure generated<br />

from Protein Data Bank file 1F88.<br />

The heterotrimeric G proteins to which<br />

GPCRs are coupled are composed <strong>of</strong> a nucleotide-binding<br />

Gα subunit and a Gβγ subunit<br />

dimer, as illustrated in FIGURE 14.24. The structure<br />

<strong>of</strong> the trimer and each subunit is known for<br />

several states <strong>of</strong> activation and in complex with<br />

several interacting proteins. A Gαβγ heterotrimer<br />

is named according to its α subunit, which largely<br />

defines the G protein’s selectivity among receptors.<br />

Each subunit also regulates a distinct group<br />

<strong>of</strong> effector proteins.<br />

Gα subunits are globular, two-domain proteins<br />

<strong>of</strong> 38-44 kDa. The GTP-binding domain<br />

belongs to the GTP-binding protein superfamily<br />

that includes the small, monomeric G proteins<br />

(such as Ras, Rho, Arf, Rab; see 14.23 Small,<br />

monomeric GTP-binding proteins are multiuse<br />

switches) as well as the GTP-binding translational<br />

initiation and elongation factors. A second domain<br />

modulates GTP binding and hydrolysis.<br />

Gα subunits are only slightly hydrophobic, but<br />

they are predominantly membrane-associated<br />

FIGURE 14.24 The structure <strong>of</strong> the nonactivated G i<br />

heterotrimer,<br />

the G protein that is responsible for inhibition <strong>of</strong><br />

adenylyl cyclase and for most G-mediated <strong>signaling</strong>, is<br />

shown with each subunit colored as shown. GDP is shown<br />

bound to the G i<br />

subunit. Structure generated from Protein<br />

Data Bank file 1GP2.<br />

G<br />

protein<br />

Gs<br />

Golf<br />

Gi (3)<br />

Go<br />

Gz<br />

Ggus<br />

Gt (2)<br />

Gq (4)<br />

G 12<br />

G 13<br />

Adenylyl cyclase<br />

EFFECTOR PROTEIN<br />

K+channel, PI 3-kinase<br />

Other cation channel<br />

Rho GEF<br />

G protein targets<br />

Stimulated<br />

Cyclic GMP phosphodiesterase<br />

Phospholipase-Cβ<br />

Inhibited<br />

Adenylyl cyclase<br />

FIGURE 14.25 G protein-regulated effectors do not share structural<br />

similarities. They may be ion channels or membrane spanning<br />

enzymes in the plasma membrane, peripheral proteins on the<br />

inner face <strong>of</strong> the membrane, or fundamentally soluble proteins<br />

that can bind to G subunits. The chart shows the major groups<br />

<strong>of</strong> G proteins, sorted according to sequence similarity, and some<br />

<strong>of</strong> the effectors that they are known to regulate.<br />

because <strong>of</strong> constitutive N-terminal fatty acylation<br />

and because they bind to the membraneattached<br />

Gβγ subunits. Mammals have 16 Gα<br />

genes that are grouped in subfamilies according<br />

to similar sequence and function (e.g., s, i, q,<br />

and 12). These subfamilies are listed in FIGURE<br />

14.25.<br />

Gβ and Gγ subunits associate irreversibly<br />

soon after translation to form stable Gβγ dimers,<br />

which then associate reversibly with a Gα. Gβ<br />

subunits are 35 kDa proteins composed <strong>of</strong> seven<br />

14.20 G protein <strong>signaling</strong> modules are widely used and highly adaptable 617


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 618<br />

-strand repeats that form a cylindrical structure<br />

known as a propeller. There are five Gβ genes<br />

in mammals. Four encode strikingly similar proteins<br />

that naturally dimerize with the twelve Gγ<br />

subunits (Figure 14.24). The fifth, Gβ5, is less<br />

closely related to the others and interacts primarily<br />

with a Gγ-like domain in other proteins rather<br />

than with Gγ subunits themselves.<br />

Gγ subunits are smaller (~7 kDa) and far<br />

more diverse in sequence than are the Gβ’s. The<br />

last three amino acid residues <strong>of</strong> Gγ subunits<br />

are proteolyzed to leave a conserved C-terminal<br />

cysteine that is irreversibly S-prenylated<br />

and carboxymethylated, helping to anchor Gβγ<br />

to the membrane. Gβ and Gγ subunits can associate<br />

in most possible combinations. Because<br />

almost all <strong>cell</strong>s express multiple Gβ and Gγ subunits,<br />

it has been difficult to assign specific roles<br />

to individual Gβγ combinations. The best recognized<br />

interactions <strong>of</strong> Gβγ subunits occur at<br />

sites on Gβ, although distinct functions <strong>of</strong> Gγ<br />

have also been supported.<br />

14.21<br />

Heterotrimeric G proteins<br />

regulate a wide variety <strong>of</strong><br />

effectors<br />

Key concepts<br />

• G proteins convey signals by regulating the<br />

activities <strong>of</strong> multiple intra<strong>cell</strong>ular <strong>signaling</strong><br />

proteins known as effectors.<br />

• Effectors are structurally and functionally diverse.<br />

• A common G-protein binding domain has not been<br />

identified among effector proteins.<br />

• Effector proteins integrate signals from multiple G<br />

protein pathways.<br />

G protein-regulated effectors include enzymes<br />

that create or destroy intra<strong>cell</strong>ular second<br />

messengers (adenylyl cyclase, cyclic GMP phosphodiesterase,<br />

phospholipase C-β, phosphatidylinositol-3-kinase),<br />

protein kinases, ion channels<br />

(K+, Ca2+) and possibly membrane transport<br />

proteins (see Figure 14.25). Effectors may be<br />

integral membrane proteins or intrinsically soluble<br />

proteins that bind G proteins at the membrane<br />

surface. No conserved G protein-binding<br />

domain or sequence motif has been identified<br />

among effector proteins, and most effectors are<br />

related to proteins that have similar functions<br />

but that are not regulated by G proteins.<br />

Sensitivity to G protein regulation, thus, evolved<br />

independently in multiple families <strong>of</strong> regulatory<br />

proteins.<br />

Because they can respond to a variety <strong>of</strong><br />

Gα and Gβγ subunits, effector proteins can integrate<br />

signals from multiple G protein pathways.<br />

The different Gα or Gβγ subunits may<br />

have opposite or synergistic effects on a given<br />

effector. For example, some <strong>of</strong> the membranebound<br />

adenylyl cyclases in mammals are stimulated<br />

by Gα s<br />

and inhibited by Gα i<br />

(see Figure<br />

14.13). Many effectors are further regulated by<br />

other allosteric ligands (e.g., lipids, calmodulin)<br />

and by phosphorylation, contributing even more<br />

to integration <strong>of</strong> information.<br />

Effectors are usually represented as multiple<br />

is<strong>of</strong>orms, and each is<strong>of</strong>orm may be regulated<br />

differently, adding to the complexity <strong>of</strong> G<br />

protein networks. For example, some is<strong>of</strong>orms<br />

<strong>of</strong> adenylyl cyclase are stimulated by Gβγ,<br />

whereas others are inhibited. All phospholipase<br />

C-βs are stimulated both by Gα q<br />

family members<br />

and by Gβγ, but the potency and maximal<br />

effect <strong>of</strong> these two inputs vary dramatically<br />

among the four PLC-β is<strong>of</strong>orms.<br />

14.22<br />

Heterotrimeric G proteins<br />

are controlled by a<br />

regulatory GTPase cycle<br />

Key concepts<br />

• Heterotrimeric G proteins are activated when the<br />

Gα subunit binds GTP.<br />

• GTP hydrolysis to GDP inactivates the G protein.<br />

• GTP hydrolysis is slow, but is accelerated by<br />

proteins called GAPs.<br />

• Receptors promote activation by allowing GDP<br />

dissociation and GTP association; spontaneous<br />

exchange is very slow.<br />

• RGS proteins and phospholipase C-βs are GAPs for<br />

G proteins.<br />

The key event in heterotrimeric G protein <strong>signaling</strong><br />

is the binding <strong>of</strong> GTP to the Gα subunit.<br />

GTP binding activates the Gα subunit, which<br />

allows both it and the Gβγ subunit to bind and<br />

regulate effectors. The Gα subunit remains active<br />

as long as GTP is bound, but Gα also has<br />

GTPase activity and hydrolyzes bound GTP to<br />

GDP. Gα-GDP is inactive. G proteins thus traverse<br />

a GTPase cycle <strong>of</strong> GTP binding/activation and hydrolysis/deactivation,<br />

as depicted in FIGURE 14.26.<br />

Therefore, the control <strong>of</strong> G protein <strong>signaling</strong> is<br />

intrinsically kinetic. The relative signal strength,<br />

or amplitude, is proportional to the fraction <strong>of</strong><br />

G protein that is in the active, GTP-bound form.<br />

This fraction equals the balance <strong>of</strong> the rates <strong>of</strong><br />

GTP binding and GTP hydrolysis, the activating<br />

618 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 619<br />

and deactivating arms <strong>of</strong> the GTPase cycle. Both<br />

limbs are highly regulated over a range <strong>of</strong> rates<br />

greater than 1000-fold.<br />

Receptors promote G protein activation by<br />

opening the nucleotide-binding site on the G<br />

protein, thus accelerating both GDP dissociation<br />

and GTP association. This process is referred to<br />

as GDP/GTP exchange catalysis. Exchange proceeds<br />

in the direction <strong>of</strong> activation because the<br />

affinity <strong>of</strong> G proteins for GTP is much higher than<br />

that for GDP and because the cytosolic concentration<br />

<strong>of</strong> GTP is about 20-fold higher than that<br />

<strong>of</strong> GDP. Spontaneous GDP/GTP exchange is very<br />

slow for most G proteins (many minutes), which<br />

maintains basal signal output at a low level. In<br />

contrast, receptor-catalyzed exchange can take<br />

place in a few tens <strong>of</strong> milliseconds, which allows<br />

rapid responses in <strong>cell</strong>s such as visual photoreceptors,<br />

other neurons, or muscle.<br />

Because receptors are not directly required<br />

for a G protein’s <strong>signaling</strong> activity, a receptor can<br />

dissociate after GDP/GTP exchange and catalyze<br />

the activation <strong>of</strong> additional G protein molecules.<br />

In this way, a single receptor may maintain the<br />

activation <strong>of</strong> multiple G proteins, providing molecular<br />

amplification <strong>of</strong> the incoming signal.<br />

Other receptors may remain bound to their G<br />

protein targets, which means that they do not<br />

act as amplifiers. However, more tightly bound<br />

receptors can initiate <strong>signaling</strong> more quickly and<br />

promote G protein reactivation when hydrolysis<br />

<strong>of</strong> bound GTP is rapid.<br />

In the absence <strong>of</strong> stimulus, Gα subunits<br />

hydrolyze bound GTP slowly. The average activation<br />

lifetime <strong>of</strong> the Gα-GTP complex is<br />

about 10-150 seconds, depending on the G<br />

protein. This rate is far slower than rates <strong>of</strong> deactivation<br />

<strong>of</strong>ten observed in <strong>cell</strong>s when an agonist<br />

is removed. For example, visual <strong>signaling</strong><br />

terminates in about 10 ms after stimulation by<br />

a photon, and many other G protein systems<br />

are almost as fast. GTP hydrolysis is accelerated<br />

by GTPase-activating proteins (GAPs),<br />

which directly bind Gα subunits. In some cases<br />

acceleration exceeds 2000-fold. Such speed is<br />

necessary in systems like vision or neurotransmission,<br />

which must respond to quickly changing<br />

stimuli. Because G protein <strong>signaling</strong> is a<br />

balance <strong>of</strong> activation and deactivation, GAPs<br />

deplete the pool <strong>of</strong> GTP-activated G protein<br />

and can thereby also act to inhibit G protein<br />

<strong>signaling</strong>. GAPs can thus inhibit <strong>signaling</strong>,<br />

quench output upon signal termination, or<br />

both. What behavior predominates depends<br />

on the GAP’s intrinsic activity and its regulation.<br />

Receptor<br />

+ agonist<br />

GDP<br />

G protein-GDP<br />

Pi<br />

The regulatory GTPase cycle<br />

Receptor<br />

- agonist<br />

G protein<br />

GAP<br />

GTP<br />

G protein-GTP<br />

*ACTIVE*<br />

Effector protein<br />

There are two families <strong>of</strong> GAPs for heterotrimeric<br />

G proteins. The RGS proteins (regulators<br />

<strong>of</strong> G protein <strong>signaling</strong>) are a family <strong>of</strong><br />

about 30 proteins, most or all <strong>of</strong> which have<br />

GAP activity and regulate G protein <strong>signaling</strong><br />

rates and amplitudes. The role <strong>of</strong> RGS proteins<br />

in terminating the G protein signal can be seen<br />

in FIGURE 14.27. Some proteins with RGS domains<br />

also act as G protein-regulated effectors.<br />

These include activators <strong>of</strong> the Rho family <strong>of</strong><br />

monomeric GTP-binding proteins (see below)<br />

and GPCR kinases, which are feedback regulators<br />

<strong>of</strong> GPCR function. The second group <strong>of</strong> G<br />

protein GAPs are phospholipase C-βs. These enzymes<br />

are effectors that are stimulated by both<br />

Gα q<br />

and by Gβγ, but they also act as G q<br />

GAPs,<br />

probably to control output kinetics.<br />

G protein-GTP-<br />

Effector protein<br />

*ACTIVE*<br />

FIGURE 14.26 G proteins are activated when GTP binds to the G subunit, such<br />

that both G-GTP and G can bind and regulate the activities <strong>of</strong> appropriate<br />

effector proteins. G subunits also have intrinsic GTPase activities, and the primary<br />

deactivating reaction is hydrolysis <strong>of</strong> bound GTP to GDP (rather than GTP<br />

dissociation). Thus, the steady-state signal output from a receptor-G protein<br />

module is the fraction <strong>of</strong> the G protein in the GTP-bound state, which reflects<br />

the balance <strong>of</strong> the activation and deactivation rates. Both GTP binding and GTP<br />

hydrolysis are intrinsically slow and highly regulated. GDP binds tightly to G,<br />

such that GDP dissociation is rate-limiting for binding <strong>of</strong> a new molecule <strong>of</strong><br />

GTP and consequent reactivation. Both GDP release and GTP binding are catalyzed<br />

by GPCRs. Hydrolysis <strong>of</strong> bound GTP is accelerated by GTPase-activating<br />

proteins (GAPs). Receptors and GAPs coordinately control both the steady-state<br />

level <strong>of</strong> signal output and the rates <strong>of</strong> activation and deactivation <strong>of</strong> the module.<br />

14.22 Heterotrimeric G proteins are controlled by a regulatory GTPase cycle 619


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 620<br />

Single photon responses <strong>of</strong> GAP-deficient mice<br />

Current (pA)<br />

0.50<br />

0.25<br />

0.00<br />

knockout<br />

heterozygous<br />

wild-type<br />

0 2 Time (s) 4<br />

Light flash<br />

FIGURE 14.27 G protein GAPs can accelerate signal termination<br />

upon removal <strong>of</strong> agonist, and <strong>of</strong>ten do not act as inhibitors<br />

during the response to receptor. The figure shows<br />

the electrical response <strong>of</strong> a mouse photoreceptor (rod) <strong>cell</strong><br />

to a single photon <strong>of</strong> light. In mice that lack RGS9, the GAP<br />

for the photoreceptor G protein G t<br />

, the signal is prolonged<br />

for many seconds because hydrolysis <strong>of</strong> GTP bound to G t<br />

is<br />

slow. In wild-type or heterozygous mice, hydrolysis takes<br />

place in about 15 milliseconds, and the decay <strong>of</strong> the signal<br />

is much faster. Note that the maximal output is similar in<br />

wild-type and mutant mice, indicating that the GAP does<br />

not act as an inhibitor in rod <strong>cell</strong>s. In humans, genetic loss<br />

<strong>of</strong> RGS9 leads to severe loss <strong>of</strong> vision that is particularly<br />

marked in bright light. Reproduced from Chen et al. Nature.<br />

2000. 403:557–560. Permission also granted by Ching-Kang<br />

Jason Chen, Virginia Commonwealth University.<br />

While the GTPase cycle described in Figure<br />

14.26 is general, it is highly simplified. Interactions<br />

among receptor, Gα, Gβγ, GAP, and effector are<br />

frequently simultaneous and <strong>of</strong>ten demonstrate<br />

complex cooperative interactions. For example,<br />

Gβγ inhibits the release <strong>of</strong> GDP (to minimize<br />

spontaneous activation), promotes the exchange<br />

catalyst activity <strong>of</strong> the receptor, inhibits GAP activity,<br />

and helps initiate receptor phosphorylation<br />

that leads to desensitization. The other<br />

components can be nearly this multifunctional.<br />

In addition, inputs from other proteins can alter<br />

the dynamics <strong>of</strong> the GTPase cycle at several points.<br />

The core G protein module is, thus, functionally<br />

versatile as a signal processor in addition to being<br />

versatile in the scope <strong>of</strong> its targets.<br />

14.23<br />

Small, monomeric GTPbinding<br />

proteins are<br />

multiuse switches<br />

Key concepts<br />

• Small GTP-binding proteins are active when bound<br />

to GTP and inactive when bound to GDP.<br />

• GDP/GTP exchange catalysts known as GEFs<br />

(guanine nucleotide exchange factors) promote<br />

activation.<br />

• GAPs accelerate hydrolysis and deactivation.<br />

• GDP dissociation inhibitors (GDIs) slow<br />

spontaneous nucleotide exchange.<br />

Monomeric GTP-binding proteins, which are<br />

encoded by about 150 genes in animals, modulate<br />

a wide variety <strong>of</strong> <strong>cell</strong>ular processes including<br />

signal transduction, organellar trafficking,<br />

intra-organellar transport, cytoskeletal assembly,<br />

and morphogenesis. The small GTP-binding<br />

proteins that most clearly function in signal<br />

transduction are the Ras and Ras-related proteins<br />

(Ral, Rap) and the Rho/Rac/Cdc42 proteins,<br />

about 10-15 in all. They are usually about<br />

20-25 kDa in size and are homologous to the<br />

GTP-binding domains <strong>of</strong> Gα subunits.<br />

The regulatory activities <strong>of</strong> the small GTPbinding<br />

proteins are controlled by a GTP binding<br />

and hydrolysis cycle like that <strong>of</strong> the<br />

heterotrimeric G proteins, with similar regulatory<br />

inputs. They are activated by GTP, and hydrolysis<br />

<strong>of</strong> bound GTP to GDP terminates<br />

activation. GDP/GTP exchange catalysts, known<br />

as GEFs (guanine nucleotide exchange factors,<br />

functionally analogous to GPCRs) promote activation,<br />

and GAPs accelerate hydrolysis and<br />

consequent deactivation. In addition, GDP dissociation<br />

inhibitors (GDIs) slow spontaneous<br />

nucleotide exchange and activation to dampen<br />

basal activity, an activity shared by Gβγ subunits<br />

for the heterotrimeric G proteins.<br />

While the underlying biochemical regulatory<br />

events are essentially identical for<br />

monomeric and heterotrimeric G proteins,<br />

monomeric G proteins use the basic GTPase cycle<br />

in additional ways. Signal output by heterotrimeric<br />

G proteins and many monomeric<br />

G proteins is usually thought to reflect a balance<br />

<strong>of</strong> their active (GTP-bound) and inactive<br />

(GDP-bound) states in a rapidly turning-over<br />

GTPase cycle. GEFs favor formation <strong>of</strong> more active<br />

G protein, and GAPs favor the inactive state.<br />

In contrast, probably an equal number <strong>of</strong> the<br />

monomeric G proteins behave as acute on-<strong>of</strong>f<br />

switches. Upon binding GTP, they initiate a<br />

process (regulation, recruitment <strong>of</strong> other pro-<br />

620 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 621<br />

teins). They then maintain this activity, sometimes<br />

for many seconds or minutes, until they<br />

are acted upon by a GAP. For example, the<br />

monomeric G protein Ran regulates nucleocytoplasmic<br />

trafficking <strong>of</strong> protein and RNA in both<br />

directions, cooperating with carrier proteins<br />

known as karyopherins (see 5.15 The Ran GTPase<br />

controls the direction <strong>of</strong> nuclear transport). In the<br />

nucleus, high Ran GEF activity promotes GTP<br />

binding. Nuclear Ran-GTP then binds import<br />

karyopherins to drive dissociation <strong>of</strong> newly arrived<br />

cargo and promote return <strong>of</strong> the karyopherin<br />

to the cytoplasm. It also binds export<br />

karyopherins to permit binding <strong>of</strong> outgoing<br />

cargo. Outside the nucleus, high Ran GAP activity<br />

promotes GTP hydrolysis. Cytoplasmic<br />

Ran-GDP dissociates from both the export karyopherins<br />

to allow dissociation <strong>of</strong> outgoing cargo<br />

and from the import karyopherins to allow them<br />

to bind cargo for import. Thus, for monomeric<br />

G proteins such as Ran, each phase <strong>of</strong> the GTPase<br />

cycle determines a specific, coupled step in a<br />

parallel regulatory cycle.<br />

A second major difference between the<br />

monomeric and the heterotrimeric G proteins<br />

is the structures <strong>of</strong> the GEFs, GAPs, and GDIs.<br />

Both GEFs and GAPs for monomeric GTP-binding<br />

proteins are structurally heterogeneous (although<br />

some clearly related families are<br />

evident). In addition, mechanisms for regulating<br />

these GEFs and GAPs are equally diverse.<br />

They include phosphorylation by protein kinases;<br />

allosteric regulation by heterotrimeric<br />

and/or monomeric G proteins, by second messengers<br />

and by other regulatory proteins; sub<strong>cell</strong>ular<br />

sequestration or recruitment to scaffolds;<br />

and assorted other mechanisms.<br />

The Ras proteins were the first small GTPbinding<br />

proteins to be discovered. They were<br />

identified as oncogene products because they<br />

cause malignant growth if they are either overexpressed<br />

or persistently activated by mutation;<br />

they are among the most commonly mutated<br />

genes in human tumors. Several viral ras genes<br />

figure prominently as oncogenes.<br />

Mammalian <strong>cell</strong>s contain three ras genes (H,<br />

N, and K). They may share inputs and outputs<br />

to varying extents, and they can compensate for<br />

each other in some genetic screens. It has been<br />

difficult to assign unique functions to the individual<br />

Ras proteins. Inputs to the Ras proteins<br />

are diverse and speak to the importance <strong>of</strong> Ras<br />

proteins as a crucial node in <strong>signaling</strong>.<br />

Ras GEFs and GAPS are regulated by both<br />

receptor and nonreceptor Tyr kinases through<br />

direct phosphorylation and by recruitment <strong>of</strong><br />

the regulators to the plasma membrane. Other<br />

cytoplasmic serine/threonine kinases also converge<br />

on Ras activation. Rap1, another member<br />

<strong>of</strong> the Ras family, may also fit directly into<br />

this network because it is suspected <strong>of</strong> competing<br />

with Ras proteins for protein kinase targets;<br />

in vivo it can suppress the oncogenic activity <strong>of</strong><br />

Ras. Rap1 is regulated independently, however,<br />

and acts on independent <strong>signaling</strong> pathways as<br />

well. One <strong>of</strong> its GAPs is stimulated by the G i<br />

class <strong>of</strong> G proteins, for example, and its several<br />

GEFs are stimulated by Ca2+, diacylglycerol, and<br />

cAMP.<br />

Ras proteins generally regulate <strong>cell</strong> growth,<br />

proliferation, and differentiation by modulating<br />

the activities <strong>of</strong> multiple effector proteins.<br />

The best known and best studied Ras effector is<br />

the protein kinase Raf, which initiates a MAPK<br />

cascade. FIGURE 14.28 shows well established Ras<br />

effectors.<br />

Rho, Rac, and Cdc42 are related monomeric<br />

GTP-binding proteins that are involved in generating<br />

signals that affect <strong>cell</strong> morphology. Each<br />

class <strong>of</strong> proteins regulates its own array <strong>of</strong> effectors<br />

and is controlled by separate groups <strong>of</strong> GEFs,<br />

GAPs, and GDIs. Effectors regulated by this family<br />

include phospholipases C and D, multiple<br />

protein and lipid kinases, proteins that nucleate<br />

or reorganize actin filaments, and components<br />

<strong>of</strong> the neutrophil oxygen activating<br />

system, among others (see 8.14 Small G proteins<br />

regulate actin polymerization).<br />

14.24<br />

Ras has three main effectors<br />

Function<br />

Effector Target<br />

Protein kinase cascade Raf MAPK<br />

Lipid kinase PI 3-kinase Akt<br />

Exchange factor RalGDS Exocyst<br />

Protein phosphorylation/<br />

dephosphorylation is a<br />

major regulatory<br />

mechanism in the <strong>cell</strong><br />

Key concepts<br />

• Protein kinases are a large protein family.<br />

• Protein kinases phosphorylate Ser and Thr, or Tyr,<br />

or all three.<br />

• Protein kinases may recognize the primary<br />

sequence surrounding the phosphorylation site.<br />

• Protein kinases may preferentially recognize<br />

phosphorylation sites within folded domains.<br />

FIGURE 14.28 Ras-GTP binds<br />

to many proteins. Three well<br />

established effectors include<br />

Raf, PI 3-kinase, and RalGDS.<br />

Activation <strong>of</strong> these effectors<br />

activates a MAPK pathway, increases<br />

PI 3-kinase activity,<br />

and promotes assembly <strong>of</strong> a<br />

protein complex involved in<br />

exocytosis <strong>of</strong> secretory vesicles.<br />

14.24 Protein phosphorylation/dephosphorylation is a major regulatory mechanism in the <strong>cell</strong> 621


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 622<br />

SUBSTRATE 1<br />

(Protein SER)<br />

H<br />

OH<br />

PRODUCT 1<br />

(Phosphorylated protein)<br />

H<br />

N<br />

C<br />

C<br />

N<br />

C<br />

N<br />

C<br />

C<br />

N<br />

C<br />

H<br />

O<br />

CH 2<br />

H<br />

O<br />

H<br />

O<br />

CH 2<br />

H<br />

O<br />

Protein kinases are two substrate enzymes<br />

O - P O -<br />

O<br />

+<br />

O - P O<br />

O<br />

SUBSTRATE<br />

(Mg 2+. 2<br />

ATP)<br />

Mg 2+<br />

O - P O<br />

O<br />

O - O -<br />

O - P O<br />

O<br />

O - P O<br />

O<br />

CH 2<br />

(Mg 2+. Triphosphate)<br />

Ribose<br />

PROTEIN KINASE<br />

PRODUCT<br />

(Mg 2+. 2<br />

ADP)<br />

Mg 2+<br />

O - P O<br />

O<br />

CH 2<br />

Adenine<br />

Adenine<br />

O<br />

(Mg 2+. Diphosphate)<br />

Rib<br />

FIGURE 14.29 Protein kinases transfer the -phosphoryl group from ATP to<br />

serine, threonine, or tyrosine residues in protein substrates.<br />

Protein phosphorylation is the most common<br />

form <strong>of</strong> regulatory posttranslational modification.<br />

It occurs in all organisms, and it is estimated<br />

that about one-third <strong>of</strong> proteins in animals are<br />

at some time phosphorylated. Phosphorylation<br />

can stimulate or inhibit the catalytic activity <strong>of</strong><br />

an enzyme, the affinity with which a protein<br />

binds other molecules, its sub<strong>cell</strong>ular localization,<br />

its ability to be further covalently modified,<br />

or its stability. Single phosphorylations may cause<br />

500-fold or greater changes in activity, and proteins<br />

are <strong>of</strong>ten phosphorylated on multiple<br />

residues in complex and interacting patterns.<br />

Most protein phosphorylation in eukaryotes,<br />

and essentially all in animals, is catalyzed<br />

by protein kinases; dephosphorylation is catalyzed<br />

by phosphoprotein phosphatases. Both<br />

classes <strong>of</strong> enzymes are controlled by diverse<br />

mechanisms. In addition, proteins are <strong>of</strong>ten<br />

phosphorylated by multiple protein kinases, resulting<br />

in the generation <strong>of</strong> a range <strong>of</strong> activity<br />

states. This complexity allows inputs from different<br />

<strong>signaling</strong> pathways to be integrated into<br />

the resulting activity <strong>of</strong> the target.<br />

In bacteria, plants, and fungi, an additional<br />

protein phosphorylating system known as twocomponent<br />

<strong>signaling</strong> is vital. The protein kinases<br />

involved in two-component <strong>signaling</strong> are<br />

unrelated to the eukaryotic protein kinase superfamily<br />

and phosphorylate aspartate residues<br />

rather than serine, threonine, or tyrosine.<br />

Protein kinases transfer a phosphoryl group<br />

from ATP to Ser, Thr, and Tyr residues <strong>of</strong> protein<br />

substrates to form chemically stable phosphate<br />

esters, as shown in FIGURE 14.29. In<br />

animals, the distribution <strong>of</strong> phosphate among<br />

these three amino acid residues is uneven:<br />

~90%-95% is on Ser, 5%-8% on Thr, and less<br />

than 1% on Tyr residues. The human genome<br />

contains approximately 500 genes that encode<br />

protein kinases, and many protein kinase<br />

mRNAs undergo alternative splicing. This makes<br />

the protein kinase gene superfamily one <strong>of</strong> the<br />

largest functional gene groups. The number and<br />

diversity <strong>of</strong> these enzymes emphasize the great<br />

and varied uses <strong>of</strong> protein kinases to regulate <strong>cell</strong>ular<br />

functions. Although some protein kinases<br />

have a limited tissue and/or developmental distribution,<br />

many are ubiquitously expressed.<br />

Protein kinases are grouped according to<br />

their residue specificity. Protein kinases that<br />

phosphorylate Ser will usually also recognize<br />

Thr, hence the name protein Ser/Thr kinase.<br />

Multi<strong>cell</strong>ular organisms have protein Tyr kinases,<br />

which only recognize Tyr. Dual specificity<br />

protein kinases can phosphorylate Ser,<br />

Thr, and Tyr in the appropriately restricted substrate<br />

conformational context and are generally<br />

the most selective <strong>of</strong> the protein kinases.<br />

The analysis <strong>of</strong> the kinomes <strong>of</strong> several organisms<br />

has led to a more elaborate grouping<br />

derived from sequence relationships, shown in<br />

FIGURE 14.30, that also reflects to some extent<br />

on regulatory mechanisms and substrate specificity.<br />

For example, the AGC group is named<br />

for its founding members, cAMP-dependent<br />

protein kinase (PKA), cyclic GMP-dependent<br />

protein kinase (PKG), Ca2+, and phospholipiddependent<br />

protein kinase (PKC). These protein<br />

kinases are regulated by second messengers and<br />

prefer substrates that contain basic residues near<br />

the phosphorylation site.<br />

In addition to substrate specificity for amino<br />

acid residues, most protein kinases are also selective<br />

for local sequence surrounding the substrate<br />

site. Screening strategies have resulted in methods<br />

to predict if proteins contain consensus substrate<br />

sites for a wide variety <strong>of</strong> protein kinases.<br />

Antibodies can be used to identify and roughly<br />

quantitate protein phosphorylation at specific<br />

sites in proteins. Beyond local recognition, protein<br />

kinases may display marked substrate selectivity<br />

among similar proteins based on overall<br />

622 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 623<br />

Human kinome tree<br />

FIGURE 14.30 The protein kinases in<br />

the human genome can be grouped according<br />

to sequence relationships that<br />

reveal seven major branches. The tyrosine<br />

kinases are contained within one<br />

major branch. The others are Ser/Thrspecific<br />

or dual specificity, and are named<br />

for the best described members: AGC<br />

from PKA, PKG, and PKC; CAMK from the<br />

calcium, calmodulin-dependent kinases;<br />

CMGC from CDKs, MAPKs, GSK3, Clks; CK1<br />

from casein kinase 1; STE from Ste20,<br />

Ste11, and Ste7, the MAP4K, MAP3K,<br />

and MAP2K in the yeast mating pathway;<br />

and TKL, the Tyr kinase-like enzymes.<br />

Reproduced with permission from<br />

G. Manning, et al. 2002. Science. 298:<br />

1912-1934. © 2002 AAAS. Photo courtesy<br />

<strong>of</strong> Gerard Manning, Salk Institute,<br />

and reprinted with permission <strong>of</strong> Cell<br />

Signaling Technology, Inc. (www.<strong>cell</strong>signal.com).<br />

three-dimensional structure, for example, or<br />

among proteins that have been differentially covalently<br />

modified by phosphorylation or ubiquitination.<br />

In animal <strong>cell</strong>s, some protein kinases are<br />

hormone receptors that span the plasma membrane.<br />

Some protein kinase receptors are protein<br />

serine/threonine kinases, such as the<br />

transforming growth factor- (TGF-)receptor,<br />

but the majority are protein tyrosine kinases, including<br />

receptors for insulin, epidermal growth<br />

factor (EGF), platelet-derived growth factor<br />

(PDGF), and other regulators <strong>of</strong> <strong>cell</strong> growth and<br />

differentiation. Other protein kinases are intrinsically<br />

soluble intra<strong>cell</strong>ular enzymes, although<br />

they may bind to one or more organellar<br />

membranes.<br />

X-ray crystallographic structures <strong>of</strong> protein<br />

kinases have revealed a wealth <strong>of</strong> information<br />

about their mechanism <strong>of</strong> activation. The conserved<br />

minimum catalytic core <strong>of</strong> a protein kinase<br />

contains about 270 amino acids, yielding<br />

a minimum molecular mass <strong>of</strong> about 30,000<br />

Da. Within this core, there are two folded domains<br />

that form the active site at their interface,<br />

as shown in FIGURE 14.31. One or both <strong>of</strong><br />

the conserved lysine (Lys) or aspartate (Asp)<br />

residues that are required for phosphoryl transfer<br />

are frequently mutated to disrupt kinase activity.<br />

A sequence near the active site, referred<br />

to as the activation loop, <strong>of</strong>ten undergoes a conformational<br />

rearrangement to generate active<br />

forms <strong>of</strong> the protein kinases and is the most<br />

common site <strong>of</strong> regulatory phosphorylation in<br />

14.24 Protein phosphorylation/dephosphorylation is a major regulatory mechanism in the <strong>cell</strong> 623


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 624<br />

ERK2 inactive and active conformations<br />

INACTIVE (ERK2)<br />

N terminal<br />

domain<br />

Tyr185<br />

Thr183<br />

C terminal<br />

domain<br />

ACTIVE (ERK2-P2)<br />

Tyr185<br />

Thr183<br />

FIGURE 14.31 The structures <strong>of</strong> unphosphorylated, inactive MAPK ERK2 and<br />

phosphorylated, active ERK2 are compared. ERK2 has a typical protein kinase<br />

structure. The smaller N-terminal domain is composed primarily <strong>of</strong> strands<br />

and the larger C-terminal domain is primarily -helical. The active site is formed<br />

at the interface <strong>of</strong> the two domains. The activation loop emerges from the active<br />

site and is refolded following phosphorylation <strong>of</strong> the Tyr and Thr residues,<br />

inducing the repositioning <strong>of</strong> active site residues. ATP (not shown) binds in<br />

the interior <strong>of</strong> the active site; productive binding <strong>of</strong> protein substrates to the<br />

surface <strong>of</strong> the C-terminal domain is also facilitated by the reorganization <strong>of</strong><br />

the activation loop. Structures generated from Protein Data Bank files 1ERK<br />

and 2ERK.<br />

Sensor/<br />

Histidine<br />

kinase<br />

Response<br />

regulator<br />

His<br />

Asp<br />

Ligand<br />

P<br />

ATP<br />

His phosphorylation<br />

Two-component <strong>signaling</strong> systems<br />

ADP<br />

His<br />

Asp<br />

P<br />

P<br />

Transfer <strong>of</strong> phosphate<br />

to Asp: response<br />

regulator active<br />

His<br />

Asp<br />

P<br />

P<br />

H 2 O<br />

Response regulator<br />

deactivated<br />

FIGURE 14.32 The basic two-component system is composed <strong>of</strong> a signal-activated<br />

histidine kinase, referred to as a sensor, and an effector protein, the response<br />

regulator, that is activated when it is phosphorylated on an aspartate<br />

residue by the sensor. The activity <strong>of</strong> the response regulator is terminated when<br />

the aspartyl-phosphate is hydrolyzed.<br />

the protein kinase family. There are unique inserts<br />

on the surface <strong>of</strong> protein kinases that generate<br />

specificity in localization, interaction with<br />

other regulatory molecules, and recognition <strong>of</strong><br />

substrates. These landmarks allow both classification<br />

and genetic manipulation <strong>of</strong> protein kinases.<br />

Protein kinases have evolved numerous<br />

and diverse regulatory mechanisms to complement<br />

their number and multiple functions.<br />

These mechanisms include allosteric activation<br />

and inhibition by lipids, soluble small molecules<br />

and other proteins; activating and inhibitory<br />

phosphorylation and other covalent modifications,<br />

including proteolysis; and binding to scaffolds<br />

and adaptors to enhance activity or limit<br />

nonspecific activities. Many such inputs may<br />

regulate a single protein kinase in a complex<br />

combinatoric code. Further, multiple protein<br />

kinases that act sequentially, such as in a protein<br />

kinase cascade (see Figure 14.38), can create<br />

uniquely complex <strong>signaling</strong> patterns.<br />

14.25<br />

Two-component protein<br />

phosphorylation systems<br />

are <strong>signaling</strong> relays<br />

Key concepts<br />

• Two-component <strong>signaling</strong> systems are composed <strong>of</strong><br />

sensor and response regulator components.<br />

• Upon receiving a stimulus, sensor components<br />

undergo autophosphorylation on a histidine (His)<br />

residue.<br />

• Transfer <strong>of</strong> the phosphate to an aspartyl residue on<br />

the response regulator serves to activate the<br />

regulator.<br />

Prokaryotes, plants, and fungi share an alternative<br />

mechanism for regulatory phosphorylation<br />

and dephosphorylation known as two-component<br />

<strong>signaling</strong>. FIGURE 14.32 shows a typical twocomponent<br />

system. In this system, the receptor,<br />

referred to as a sensor, responds to a stimulus by<br />

catalyzing its own phosphorylation on a His<br />

residue. Sensors include chemoattractant receptors<br />

in bacteria, a regulator <strong>of</strong> osmolarity in fungi,<br />

light-sensitive proteins, the receptor for the plantripening<br />

hormone ethylene, and other receptors<br />

for diverse environmental, hormonal, and metabolic<br />

signals. The mammalian mitochondrial dehydrogenase<br />

kinases are related in sequence to<br />

the bacterial histidine kinases, although the mammalian<br />

enzymes phosphorylate serine or threonine<br />

residues, not histidine. The phosphorylated<br />

sensor next transfers its covalently bound phos-<br />

624 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 625<br />

phate to an aspartyl residue on a second protein<br />

known as a response regulator. Response regulators<br />

initiate <strong>cell</strong>ular responses, usually by binding<br />

to other cytoplasmic proteins and allosterically<br />

regulating their activities.<br />

Although all two-component systems follow<br />

this same general pattern, their structures<br />

and precise reaction pathways vary enormously.<br />

Some two-component systems are composed<br />

<strong>of</strong> only one protein (sensor and response regulator<br />

in a single polypeptide chain). Others are<br />

composed <strong>of</strong> a sensor protein and two aspartylphosphorylated<br />

proteins, in which the first or<br />

the second may display response regulatory activity.<br />

Finally, two-component systems usually<br />

lack conventional protein phosphatases.<br />

Hydrolysis <strong>of</strong> the aspartyl-phosphate bond may<br />

be spontaneous or regulated by the response<br />

regulator itself.<br />

14.26<br />

Pharmacological<br />

inhibitors <strong>of</strong> protein<br />

kinases may be used to<br />

understand and treat<br />

disease<br />

Key concepts<br />

• Protein kinase inhibitors are useful both for<br />

<strong>signaling</strong> research and as drugs.<br />

• Protein kinase inhibitors usually bind in the ATP<br />

binding site.<br />

Many inhibitors have been developed for basic<br />

research purposes to explore the functions <strong>of</strong><br />

protein kinases. The importance <strong>of</strong> these enzymes<br />

in disease processes has also made them<br />

targets <strong>of</strong> drug screening projects yielding inhibitors<br />

for many protein kinases. The majority<br />

<strong>of</strong> pharmacological inhibitors <strong>of</strong> protein<br />

kinases compete with ATP binding. Because <strong>of</strong><br />

the huge number <strong>of</strong> ATP-binding proteins in a<br />

<strong>cell</strong>, there are inevitable concerns about inhibitor<br />

specificity not only with respect to the<br />

other protein kinases but also to the other proteins<br />

that bind nucleotides. This problem has<br />

been mitigated with variable success through<br />

chemical library screening, structure-based modification<br />

<strong>of</strong> lead compounds, and inhibitor testing<br />

against panels <strong>of</strong> protein kinases.<br />

Many inhibitors with actions on PKA or<br />

PKCs, for example, have effects on several other<br />

members <strong>of</strong> the AGC family. Although pharmacological<br />

inhibitors with effects on PKA<br />

abound, the most selective are derived from the<br />

naturally occurring small inhibitory protein<br />

known as PKI or the Walsh inhibitor. In vitro<br />

and <strong>cell</strong>-based screens have identified much<br />

more selective inhibitors for MAP2Ks in the<br />

ERK1/2 pathway. These inhibitors have fewer<br />

known protein kinase cross reactivities, probably<br />

due to the fact that they do not bind in the<br />

ATP site. Among inhibitors that have progressed<br />

in the clinic, compounds developed against the<br />

EGF receptor and certain other protein tyrosine<br />

kinases have had considerable success.<br />

14.27<br />

Phosphoprotein<br />

phosphatases reverse the<br />

actions <strong>of</strong> kinases and are<br />

independently regulated<br />

Key concepts<br />

• Phosphoprotein phosphatases reverse the actions<br />

<strong>of</strong> protein kinases.<br />

• Phosphoprotein phosphatases may<br />

dephosphorylate phosphoserine/threonine,<br />

phosphotyrosine, or all three.<br />

• Phosphoprotein phosphatase specificity is <strong>of</strong>ten<br />

achieved through the formation <strong>of</strong> specific protein<br />

complexes.<br />

Protein phosphorylation is reversed by phosphoprotein<br />

phosphatases. These enzymes display distinct<br />

specificities and modes <strong>of</strong> regulation.<br />

Phosphoprotein phosphatases can be considered<br />

in two broad groups based on their specificity and<br />

sequence relationships: protein-serine/threonine<br />

phosphatases and protein-tyrosine phosphatases.<br />

Most protein-serine/threonine phosphatases<br />

are regulated by association with other proteins.<br />

Targeted localization is the major determinant <strong>of</strong><br />

substrate specificity. Phosphoprotein phosphatase<br />

1 (PP1) associates with a variety <strong>of</strong> regulatory<br />

subunits that specifically direct it to relevant organelles.<br />

One subunit (known as the G subunit),<br />

for example, specifies association with glycogen<br />

particles. The interaction with this subunit is itself<br />

regulated by phosphorylation. Small protein<br />

inhibitors can suppress PP1 activity.<br />

Phosphoprotein phosphatase 2A (PP2A) is<br />

composed <strong>of</strong> a catalytic subunit, a scaffolding<br />

subunit, and one <strong>of</strong> a large number <strong>of</strong> regulatory<br />

subunits. The regulatory subunit modulates activity<br />

and localization <strong>of</strong> the phosphatase. Some<br />

viruses alter the behavior <strong>of</strong> the <strong>cell</strong>s they infect<br />

by interfering with phosphatase activity. For example,<br />

<strong>cell</strong>s transformed with the SV40 virus<br />

express a viral protein known as small t anti-<br />

14.27 Phosphoprotein phosphatases reverse the actions <strong>of</strong> kinases and are independently regulated 625


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 626<br />

gen. Small t displaces the regulatory subunit<br />

from PP2A and alters the activity and the sub<strong>cell</strong>ular<br />

localization <strong>of</strong> the phosphatase. In addition,<br />

natural toxins such as okadaic acid,<br />

calyculin, and microcystin inhibit PP2A and PP1<br />

to varying extents both in vitro and in intact <strong>cell</strong>s.<br />

Another major protein-serine/threonine<br />

phosphatase, called calcineurin (also known as<br />

phosphoprotein phosphatase 2B), is regulated<br />

by Ca2+-calmodulin (see 14.15 Ca2+ <strong>signaling</strong><br />

serves diverse purposes in all eukaryotic <strong>cell</strong>s) and<br />

plays essential roles in cardiac development and<br />

T <strong>cell</strong> activation, among other events. The major<br />

mechanism <strong>of</strong> action <strong>of</strong> the immunosuppressants<br />

cyclosporin and FK506 is to inhibit<br />

calcineurin.<br />

The protein tyrosine phosphatases (PTPs)<br />

are cysteine-dependent enzymes that utilize a<br />

conserved Cys-Xaa-Arg motif to hydrolyze phosphoester<br />

bonds in their substrates. The PTPs are<br />

encoded by over 100 genes in humans and are<br />

classified in four subfamilies: the phosphotyrosine-specific<br />

phosphatases, the Cdc25 phosphatases,<br />

the dual specificity phosphatases<br />

(DSPs), and the low molecular weight phosphatases.<br />

Thirty-eight <strong>of</strong> the PTPs are highly selective<br />

for phosphotyrosine residues within substrates.<br />

Some <strong>of</strong> the phosphotyrosine-selective<br />

phosphatases are transmembrane proteins,<br />

whereas others are membrane associated. The<br />

most obvious function <strong>of</strong> the PTPs is to reverse<br />

the functions <strong>of</strong> tyrosine kinases; however, some<br />

have primary functions in transducing tyrosine<br />

kinase signals. For example, the protein tyrosine<br />

phosphatase SHP2 (also known as SHPTP2),<br />

binds to certain tyrosine kinase receptors<br />

through its SH2 domain and is itself tyrosine<br />

phosphorylated, thereby creating a binding site<br />

for the SH2 domain-containing adaptor protein,<br />

Grb2, which leads to activation <strong>of</strong> Ras (see<br />

14.32 MAPKs are central to many <strong>signaling</strong> pathways).<br />

The Cdc25 phosphatases recognize cyclindependent<br />

kinase (CDK) family members as<br />

substrates and play a critical role in increasing<br />

CDK activity at key junctures <strong>of</strong> the <strong>cell</strong> cycle<br />

(see Figure 14.39 and 11.4 The <strong>cell</strong> cycle is a cycle<br />

<strong>of</strong> CDK function). Similar to the dual specificity<br />

kinases, the dual specificity phosphatases are<br />

specific for a restricted number <strong>of</strong> substrates. A<br />

number <strong>of</strong> DSPs dephosphorylate MAPKs; these<br />

DSPs are called MAP kinase phosphatases, or<br />

MKPs. Several <strong>of</strong> these have been implicated<br />

in MAPK nuclear entry and exit. Some MKPs<br />

are encoded by early response genes, whose<br />

products are active near the initiation <strong>of</strong> the <strong>cell</strong><br />

cycle (see 11.7 Entry into <strong>cell</strong> cycle and S phase is<br />

tightly regulated).<br />

Substrates <strong>of</strong> other PTP family members,<br />

such as the tumor suppressor PTEN, include<br />

phosphoinositides, which are phosphorylated<br />

derivatives <strong>of</strong> the glycerolipid phosphatidylinositol<br />

that serve as second messengers<br />

(see 14.16 Lipids and lipid-derived compounds are<br />

<strong>signaling</strong> molecules). Removal <strong>of</strong> the phosphate<br />

group inactivates the second messenger. It remains<br />

unclear whether members <strong>of</strong> this group<br />

work exclusively on phophoinositides or also<br />

on protein tyrosine phosphate.<br />

14.28<br />

Covalent modification by<br />

ubiquitin and ubiquitinlike<br />

proteins is another<br />

way <strong>of</strong> regulating protein<br />

function<br />

Key concepts<br />

• Ubiquitin and related small proteins, may be<br />

covalently attached to other proteins as a<br />

targeting signal.<br />

• Ubiquitin is recognized by diverse ubiquitin<br />

binding proteins.<br />

• Ubiquitination can cooperate with other covalent<br />

modifications.<br />

• Ubiquitination regulates <strong>signaling</strong> in addition to<br />

its role in protein degradation.<br />

An important mechanism for control <strong>of</strong> protein<br />

function is through covalent modification with<br />

small proteins <strong>of</strong> the ubiquitin family. Ubiquitin<br />

is one <strong>of</strong> a family <strong>of</strong> proteins referred to as ubiquitin-like<br />

(Ubl) proteins. Ubiquitin itself is highly<br />

conserved among species, suggesting the functional<br />

importance <strong>of</strong> all <strong>of</strong> its 76 residues. In addition<br />

to the long-established role <strong>of</strong> ubiquitin<br />

in initiating protein degradation, ubiquitin modification<br />

also has a variety <strong>of</strong> functions in signal<br />

transduction.<br />

Ubl proteins are conjugated to the substrate<br />

protein by an isopeptide bond between an amino<br />

group on the substrate, usually from a Lys side<br />

chain, and the C-terminal Gly residue <strong>of</strong> the<br />

processed Ubl protein. E1, E2, and E3 proteins<br />

are required to catalyze conjugation to Ubl proteins<br />

(see Biochem 4.3 Ubiquitin attachment to substrates<br />

requires multiple enzymes). Several Ubl<br />

proteins may be attached to one substrate, <strong>of</strong>ten<br />

by serial formation <strong>of</strong> a polyubiquitin chain.<br />

Mono- and polyubiquitination both change the<br />

626 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 627<br />

protein’s behavior to induce downstream signals.<br />

Monoubiquitination is a significant regulatory<br />

modification in vesicular trafficking and DNA repair.<br />

For example, the monoubiquitinated form<br />

<strong>of</strong> the FANCD2 protein becomes associated with<br />

the repair protein BRCA1 at sites <strong>of</strong> DNA repair.<br />

Modification by the Ubl protein SUMO has<br />

roles in nuclear transport, transcription, and<br />

<strong>cell</strong> cycle progression.<br />

Polyubiquitin chains are formed when the<br />

Lys residues <strong>of</strong> ubiquitin itself, particularly K48<br />

and K63, are ubiquitinated. Addition <strong>of</strong> polyubiquitin<br />

with a K48 linkage generally directs proteins<br />

to the proteasome for degradation, whereas<br />

conjugation to polyubiquitin chains with a K63<br />

linkage promotes signal transmission, not proteolysis.<br />

Protein-bound ubiquitin is recognized by<br />

a variety <strong>of</strong> ubiquitin binding domains, including<br />

UIM (ubiquitin-interacting motif), UBA (ubiquitin<br />

association), and certain zinc finger domains.<br />

Such domains have the capacity to act as receptors<br />

for ubiquitin within modified proteins.<br />

Activation <strong>of</strong> the transcription factor NF-κB<br />

occurs by a mechanism dependent on modification<br />

both by the addition <strong>of</strong> Ubl proteins and phosphorylation.<br />

This fascinating example <strong>of</strong> regulation<br />

by ubiquitin is depicted in FIGURE 14.33. Prior to<br />

stimulation, NF-κB is retained in the cytoplasm<br />

in an inactive form by binding to its inhibitor, IκB.<br />

Phosphorylation <strong>of</strong> IκB by the IκB kinase (IKK)<br />

complex promotes its recognition by a multisubunit<br />

E3 ligase, which directs its ubiquitination<br />

and subsequent proteasomal degradation.<br />

Destruction <strong>of</strong> IκB allows NF-κB to move to the<br />

nucleus to mediate changes in transcription.<br />

IκB can be stabilized in response to certain<br />

signals through covalent attachment <strong>of</strong> the Ubl,<br />

SUMO. Sumoylation occurs on the same Lys<br />

residues that must be conjugated to ubiquitin<br />

to achieve IκB degradation. Thus, SUMO attachment<br />

stabilizes IκB and attenuates NF-κB action.<br />

This is one <strong>of</strong> numerous examples <strong>of</strong><br />

crosstalk between Ubl conjugates.<br />

A key regulatory event in NF-κB <strong>signaling</strong><br />

is activation <strong>of</strong> the IKK complex. IKK is itself<br />

regulated by ubiquitination and phosphorylation.<br />

The cytokine interleukin-1β (IL-1) causes<br />

association <strong>of</strong> adaptor proteins with its receptor<br />

to create a receptor activation complex. The<br />

interleukin-1β receptor activation complex recruits<br />

another adaptor complex containing<br />

TRAF6. A phosphorylation event releases a<br />

TRAF6 complex from the receptor activation<br />

complex into the cytoplasm.<br />

TRAF6 contains a RING domain, and is an<br />

E3 ubiquitin ligase that catalyzes formation <strong>of</strong><br />

Modification with Ubl proteins plays multiple roles in IL-1β <strong>signaling</strong><br />

TRAF6<br />

TRAF6<br />

TRAF6<br />

IKK<br />

IL-1β<br />

TRAF6<br />

TRAF6<br />

K63 ubiquitination<br />

TRAF6<br />

Nemo<br />

TAB2<br />

IKK<br />

TAK1<br />

Complex formation<br />

ATP<br />

ADP<br />

CYTOPLASM<br />

ADP<br />

ATP<br />

K63 polyubiquitin chains on the protein kinase<br />

TAK1. Polyubiquitinated TAK1 can then recruit<br />

TAB2 and TAB3, which are adaptor proteins<br />

with conserved zinc finger domains. These particular<br />

zinc finger domains bind to polyubiquitinated<br />

TAK1 and enhance its activity. TAK1,<br />

thus activated, phosphorylates and activates<br />

IKK, which then phosphorylates IκB, targeting<br />

it for degradation. Thus, ubiquitin-binding domains,<br />

such as the TAB2 and TAB3 zinc fingers,<br />

may selectively recognize K63 polyubiquitin<br />

chains to promote signal transmission.<br />

Naturally occurring small molecules may<br />

control ubiquitin ligase activity directly. Auxin<br />

(indole 3-acetic acid) is a plant hormone that regulates<br />

development by promoting the transcription<br />

<strong>of</strong> a large number <strong>of</strong> genes. Rather than<br />

K48 ubiquitination<br />

Degradation<br />

NUCLEUS<br />

DNA<br />

FIGURE 14.33 Activation <strong>of</strong> NF-B involves steps dependent on the interaction <strong>of</strong> proteins<br />

attached to ubiquitin through ubiquitin-binding proteins, competition by sumoylation,<br />

phosphorylation, and ubiquitin-mediated protein degradation.<br />

14.28 Covalent modification by ubiquitin and ubiquitin-like proteins is another way <strong>of</strong> regulating protein function 627


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 628<br />

stimulating transcription factors, however, auxin<br />

accelerates the degradation <strong>of</strong> several specific<br />

transcriptional repressors. The auxin receptor is<br />

in fact a ubiquitin ligase complex that targets<br />

the auxin-regulated transcriptional repressors<br />

for proteolysis. F-box proteins account for all<br />

<strong>of</strong> the auxin binding activity in plant extracts.<br />

14.29<br />

The Wnt pathway<br />

regulates <strong>cell</strong> fate during<br />

development and other<br />

processes in the adult<br />

Key concepts<br />

• Seven transmembrane-spanning receptors may<br />

control complex differentiation programs.<br />

• Wnts are lipid-modified ligands.<br />

• Wnts signal through multiple distinct receptors.<br />

• Wnts suppress degradation <strong>of</strong> -catenin, a<br />

multifunctional transcription factor.<br />

Wnt pathways function during embryonic development<br />

and in the adult in morphogenesis,<br />

body patterning, axis formation, proliferation,<br />

and <strong>cell</strong> motility. The classical Wnt <strong>signaling</strong><br />

mechanism was uncovered largely through<br />

studies <strong>of</strong> Drosophila and Xenopus development,<br />

as well as by analyzing genetic alterations in<br />

cancer.<br />

Wnt proteins are unusual extra<strong>cell</strong>ular ligands.<br />

In addition to carbohydrate, they contain<br />

covalently bound palmitate that is essential for<br />

their biological activity. Wnts transduce signals<br />

by binding to multiple distinct receptors. The<br />

most significant are members <strong>of</strong> the Frizzled<br />

family <strong>of</strong> seven-transmembrane-spanning receptors.<br />

Wnts regulate the stability <strong>of</strong> β-catenin,<br />

which either is rapidly degraded or, in response<br />

to Wnt, is stabilized to enter the nucleus and<br />

induce transcription by interacting with TCF<br />

(T-<strong>cell</strong> factor). Genes induced include c-jun, cyclin<br />

D1, and many others.<br />

The coordinated activities <strong>of</strong> the protein kinases<br />

glycogen synthase kinase 3 (GSK3) and<br />

casein kinase 1(CK1), the scaffolding proteins<br />

axin and adenomatous polyposis coli (APC), and<br />

the protein disheveled (DSH) are key to β-catenin<br />

stability. In the absence <strong>of</strong> Wnt, phosphorylation<br />

<strong>of</strong> β-catenin by CK1 and GSK3 promotes<br />

its ubiquitination and subsequent destruction<br />

by the proteasome. Axin and APC are required<br />

for phosphorylation <strong>of</strong> β-catenin by GSK3.<br />

In contrast to most seven transmembrane-<br />

spanning receptors, the Frizzled family has not<br />

yet been shown to have significant functions<br />

mediated by a heterotrimeric G protein, and G<br />

proteins may not be central to this pathway.<br />

Instead a proximal step in <strong>signaling</strong> by Frizzled<br />

involves binding to DSH, which inactivates the<br />

β-catenin destruction mechanism.<br />

Mutations that cause changes in the<br />

amounts <strong>of</strong> components <strong>of</strong> the classical pathway<br />

are common in a wide variety <strong>of</strong> cancers. Both<br />

Wnts and β-catenin may be viewed as protooncogenes.<br />

APC is a tumor suppressor and is<br />

mutated in the majority <strong>of</strong> human colorectal<br />

cancers, for example. Either too little or too<br />

much axin can also disrupt Wnt <strong>signaling</strong>, and<br />

axin, like APC, is a tumor suppressor.<br />

Wnts utilize additional <strong>signaling</strong> mechanisms.<br />

The receptor proteins Lrp5/6 (which are<br />

related to the low-density lipoprotein receptor)<br />

are Wnt receptors and also bind axin. Wnts bind<br />

to tyrosine kinase receptors to influence axon<br />

guidance and to other proteins that inhibit their<br />

function. Through DSH, Wnts can regulate the<br />

JNK MAPK pathway and Rho family G proteins<br />

to control planar <strong>cell</strong> polarity. Certain Wnts increase<br />

intra<strong>cell</strong>ular calcium to activate calciumdependent<br />

<strong>signaling</strong> pathways.<br />

14.30<br />

Diverse <strong>signaling</strong><br />

mechanisms are regulated<br />

by protein tyrosine kinases<br />

Key concepts<br />

• Many receptor protein tyrosine kinases are<br />

activated by growth factors.<br />

• Mutations in receptor tyrosine kinases can be<br />

oncogenic.<br />

• Ligand binding promotes receptor oligomerization<br />

and autophosphorylation.<br />

• Signaling proteins bind to the phosphotyrosine<br />

residues <strong>of</strong> the activated receptor.<br />

A large group <strong>of</strong> protein tyrosine kinases are<br />

receptors that span the plasma membrane and<br />

bind extra<strong>cell</strong>ular ligands, as shown in FIGURE<br />

14.34. The receptors are generally activated by<br />

growth factors whose normal physiological functions<br />

are to promote growth, proliferation, development,<br />

or maintenance <strong>of</strong> differentiated<br />

properties. This group includes receptors for insulin,<br />

epidermal growth factor (EGF), and<br />

platelet derived growth factor (PDGF). These<br />

receptors both control the activities <strong>of</strong> many<br />

other protein kinases <strong>of</strong> all families and directly<br />

regulate other classes <strong>of</strong> <strong>signaling</strong> proteins.<br />

628 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 629<br />

Because <strong>of</strong> their physiologic roles as growth<br />

regulators, mutations that activate receptor tyrosine<br />

kinases are <strong>of</strong>ten oncogenic. For example,<br />

the oncogene erbB results from the<br />

mutational loss <strong>of</strong> the extra<strong>cell</strong>ular ligand-binding<br />

domain <strong>of</strong> a kinase closely related to the<br />

EGF receptor. This mutation causes constitutive<br />

activation <strong>of</strong> the protein kinase domain.<br />

Point mutations that affect the transmembrane<br />

domain can also cause oncogenic activation, as<br />

is found in the EGF receptor-related neu/HER2<br />

oncogene (see 13.8 Cell growth and proliferation<br />

are activated by growth factors).<br />

Receptor tyrosine kinases are diverse both<br />

in their extra<strong>cell</strong>ular ligand-binding domains<br />

and, with the exception <strong>of</strong> a conserved tyrosine<br />

protein kinase domain, their intra<strong>cell</strong>ular regulatory<br />

regions. These receptors usually have<br />

one membrane span per monomer but some,<br />

such as the insulin receptor, which is a disulfidebonded<br />

heterotetramer, have two. Ligand binding<br />

to receptor tyrosine kinases favors receptor<br />

oligomerization and enhances kinase activity<br />

leading to increased Tyr phosphorylation <strong>of</strong> the<br />

intra<strong>cell</strong>ular domain <strong>of</strong> the receptor and <strong>of</strong> associated<br />

molecules. These tyrosine-phosphorylated<br />

motifs create docking sites for additional<br />

signal transducers and adaptors.<br />

A comparison <strong>of</strong> the PDGF and insulin receptors<br />

reveals common themes and a range <strong>of</strong><br />

behaviors <strong>of</strong> receptor tyrosine kinases. The two<br />

PDGF receptors are monomeric receptor tyrosine<br />

kinases. The insulin receptor exists in two<br />

alternatively spliced forms each <strong>of</strong> which is a<br />

heterotetramer <strong>of</strong> two and two subunits. In<br />

each case, the receptor is<strong>of</strong>orms utilize some<br />

unique <strong>signaling</strong> mechanisms.<br />

PDGF and insulin each stimulate the kinase<br />

activity <strong>of</strong> their receptors, causing oligomerization<br />

and autophosphorylation. Seven or more<br />

sites are phosphorylated on the PDGF receptor,<br />

and each phosphotyrosine residue generates a<br />

binding site for one or more SH2 domain-containing<br />

proteins as illustrated in FIGURE 14.35. The<br />

PDGF receptor binds PI 3-kinase, p190 Ras GAP,<br />

phospholipase C-, Src (which may catalyze additional<br />

Tyr phosphorylation <strong>of</strong> the receptor),<br />

and the SHP2 tyrosine phosphatase which itself<br />

binds the adaptor Grb2 (see 14.32 MAPKs are central<br />

to many <strong>signaling</strong> pathways). With the exception<br />

<strong>of</strong> Src, all <strong>of</strong> these proteins are also receptor<br />

substrates. Thus, substrates are recruited to the<br />

receptor as a consequence <strong>of</strong> specific interactions<br />

<strong>of</strong> substrate SH2 domains with receptor phosphotyrosine<br />

producing changes in activities and<br />

distributions <strong>of</strong> numerous intra<strong>cell</strong>ular signal<br />

EGF<br />

receptor<br />

CYTOPLASM<br />

Receptor protein tyrosine kinase families<br />

KINASE<br />

DOMAINS<br />

Insulin<br />

receptor<br />

Activation <strong>of</strong> the PDGF receptor leads to many outputs<br />

p190<br />

RasGAP<br />

PLC-γ<br />

PDGF<br />

PDGF<br />

receptor<br />

PDGF<br />

receptors<br />

Src<br />

p85 p110<br />

PI 3-kinase<br />

Shp2<br />

Kinase<br />

inserts<br />

SOS<br />

FGF<br />

receptor<br />

FIGURE 14.34 The monomeric tyrosine kinase receptors consist <strong>of</strong> a<br />

globular extra<strong>cell</strong>ular domain that binds ligand, a single transmembrane<br />

span, and a globular intra<strong>cell</strong>ular region containing the protein kinase<br />

domain. The intra<strong>cell</strong>ular regions contain additional sequences preceding,<br />

following, and, in the case <strong>of</strong> the PDGF and FGF receptor groups,<br />

inserted into the protein kinase domain. These regions contain sites <strong>of</strong><br />

tyrosine phosphorylation-dependent interactions. The insulin receptor<br />

is encoded by a single gene. The precursor is proteolyzed into and <br />

subunits, which are disulfide bonded to each other. Disulfide bonds also<br />

link two subunits, yielding an obligate heterotetramer.<br />

Grb2<br />

FIGURE 14.35 PDGF binds to its receptor and induces receptor autophosphorylation.<br />

The autophosphorylated receptor binds target proteins<br />

that contain SH2 domains.<br />

14.30 Diverse <strong>signaling</strong> mechanisms are regulated by protein tyrosine kinases 629


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 630<br />

transducers. This array <strong>of</strong> <strong>signaling</strong> events leads<br />

to increased proliferation <strong>of</strong> connective tissue<br />

during development and in wound healing.<br />

Autophosphorylation also occurs on the insulin<br />

receptor to stabilize the active state and to<br />

generate a smaller number <strong>of</strong> binding sites, as illustrated<br />

in FIGURE 14.36. A key event is the Tyr<br />

phosphorylation <strong>of</strong> insulin receptor substrate<br />

(IRS) proteins, notably IRS1, on as many as a<br />

dozen sites. IRS1 takes over interactions with several<br />

<strong>signaling</strong> effectors that, in the case <strong>of</strong> PDGF,<br />

bind directly to the receptor. Among these targets<br />

is PI 3-kinase which leads to activation <strong>of</strong><br />

Akt-2 and several essential metabolic actions <strong>of</strong><br />

insulin (see 14.16 Lipids and lipid-derived compounds<br />

are <strong>signaling</strong> molecules). IRS proteins are also phosphorylated<br />

by serine/threonine protein kinases to<br />

modulate their <strong>signaling</strong> capability.<br />

Tyr phosphorylation <strong>of</strong>ten enhances the enzymatic<br />

activity <strong>of</strong> the associated proteins. Other<br />

proteins gain enhanced function primarily as a<br />

consequence <strong>of</strong> greater proximity to targets<br />

achieved by binding through their SH2 domains<br />

to phosphotyrosine sites either on the receptors<br />

or IRS adaptors. The precise actions <strong>of</strong> the<br />

many tyrosine kinase receptors are determined<br />

by the overlapping sets <strong>of</strong> signal transducers<br />

with which they interact, as well as by detailed<br />

differences in amounts <strong>of</strong> signal transducers,<br />

adaptor accessory proteins and receptor expression<br />

patterns (see Figure 14.43).<br />

Insulin <strong>signaling</strong> through IRS1<br />

Insulin<br />

CYTOPLASM<br />

Insulin<br />

receptor<br />

IRS1<br />

PIP 2<br />

p85 p110<br />

PI 3-kinase<br />

PIP3<br />

Akt<br />

<strong>signaling</strong><br />

FIGURE 14.36 Insulin binding to its receptor causes activation <strong>of</strong> the<br />

receptor tyrosine protein kinase and autophosphorylation. The receptor<br />

kinase also phosphorylates IRS1, a large adaptor with many potential<br />

phosphorylation sites. IRS1 is an essential intermediate in insulin action.<br />

PI 3-kinase binds to IRS1 via the SH2 domain within its p85 subunit.<br />

Akt and PDK1 bind to PIP 3<br />

produced by activated PI 3-kinase so that<br />

PDK1 can phosphorylate and activate Akt (see Figure 14.18).<br />

14.31<br />

Src family protein kinases<br />

cooperate with receptor<br />

protein tyrosine kinases<br />

Key concepts<br />

• Src is activated by release <strong>of</strong> intrasteric inhibition.<br />

• Activation <strong>of</strong> Src involves liberation <strong>of</strong> modular<br />

binding domains for activation-dependent<br />

interactions.<br />

• Src <strong>of</strong>ten associates with receptors, including<br />

receptor tyrosine kinases.<br />

The first protein tyrosine kinase to be discovered<br />

was Src, which was identified as the transforming<br />

entity in the Rous sarcoma virus. Src is the<br />

prototype <strong>of</strong> a number <strong>of</strong> related enzymes, the<br />

Src family kinases. It participates in <strong>signaling</strong><br />

pathways regulated by numerous <strong>cell</strong> surface<br />

receptors, including those that lack their own<br />

kinase domain (see in 14.34 Diverse receptors recruit<br />

protein tyrosine kinases to the plasma membrane).<br />

Src is bound to the plasma membrane<br />

via an N-terminal myristoyl group. In the inactive<br />

state, Src is phosphorylated on Tyr527, C-<br />

terminal to its catalytic domain, by CSK<br />

(C-terminal Src kinase).<br />

The structure and regulation <strong>of</strong> Src is depicted<br />

in FIGURE 14.37. Phosphorylation <strong>of</strong> Tyr527<br />

causes it to bind to its own SH2 domain. The<br />

SH2 and SH3 domains suppress the kinase activity<br />

through interactions on the surface <strong>of</strong> the<br />

protein. The SH3 domain binds to an SH3 binding<br />

site distant from the active site. Activation<br />

<strong>of</strong> Src by dephosphorylation <strong>of</strong> Tyr527 causes<br />

its SH2 to dissociate; this causes a conformational<br />

change in the SH3 domain to dissociate<br />

it from the binding site. Viral isolates <strong>of</strong> Src are<br />

<strong>of</strong>ten truncated prior to Tyr527, which increases<br />

their activity.<br />

Conformational changes in the kinase domain<br />

resulting from dissociation <strong>of</strong> the SH3 promote<br />

Src autophosphorylation on Tyr416 in its<br />

activation loop and further increase protein kinase<br />

activity. An important consequence <strong>of</strong> the<br />

interaction <strong>of</strong> Src with its own SH2 and SH3<br />

domains is that these domains cannot bind anything<br />

else when in the autoinhibited state; therefore,<br />

other interactions are promoted when the<br />

SH2 and SH3 domains are released from their<br />

associations with the Src kinase domain. The<br />

heterologous interactions <strong>of</strong> the SH2 and SH3<br />

domains contribute to Src localization and <strong>signaling</strong>.<br />

630 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 631<br />

Structure and regulation <strong>of</strong> Src<br />

INACTIVE<br />

ACTIVE<br />

KINASE<br />

DOMAIN<br />

SH3<br />

SH2<br />

P<br />

P<br />

Tyr527<br />

FIGURE 14.37 The structures <strong>of</strong> inactive and active Src are compared. The inactive<br />

protein is autoinhibited by binding to its own SH2 and SH3 domains.<br />

The SH2 domain binds to phosphorylated Tyr527. The SH3 domain binds to a<br />

noncanonical SH3-binding motif on the opposite side <strong>of</strong> the kinase domain<br />

active site. In contrast to the steric inhibition <strong>of</strong> PKA caused by its R subunit,<br />

inhibition <strong>of</strong> Src by its SH2 and SH3 domains is allosteric. In the active structure<br />

the SH2 and SH3 domains are not bound to the kinase domain and are<br />

available for heterologous interactions. Structures generated from Protein Data<br />

Base files 1FMK and 1Y57.<br />

14.32<br />

MAPKs are central to<br />

many <strong>signaling</strong> pathways<br />

Key concepts<br />

• MAPKs are activated by Tyr and Thr<br />

phosphorylation.<br />

• The requirement for two phosphorylations creates<br />

a <strong>signaling</strong> threshold.<br />

• The ERK1/2 MAPK pathway is usually regulated<br />

through Ras.<br />

Mitogen-activated protein kinases (MAPKs) are<br />

present in all eukaryotes. They are among the<br />

most common multifunctional protein kinases<br />

mediating <strong>cell</strong>ular regulatory events in response<br />

to many ligands and other stimuli. MAPKs are<br />

activated by protein kinase cascades consisting<br />

<strong>of</strong> at least three protein kinases acting sequentially,<br />

as illustrated in FIGURE 14.38. Activation<br />

<strong>of</strong> a MAPK is catalyzed by a MAPK kinase<br />

(MAP2K), which is itself activated by phosphorylation<br />

by a MAPK kinase kinase (MAP3K).<br />

MAP3Ks are activated by a variety <strong>of</strong> mechanisms<br />

including phosphorylation by MAP4Ks,<br />

oligomerization, and binding to activators such<br />

as small G proteins.<br />

MAP2Ks are activated by phosphorylation<br />

on two Ser/Thr residues; MAP2Ks then activate<br />

MAPKs by dual phosphorylation on Tyr<br />

and Thr residues (Figure 14.30). Each MAP2K<br />

phosphorylates a limited set <strong>of</strong> MAPKs and few<br />

or no other substrates. The great specificity <strong>of</strong><br />

MAP2Ks is one means <strong>of</strong> insulating MAPKs<br />

from activation by inappropriate signals. Both<br />

Tyr and Thr phosphorylations are required for<br />

maximum MAPK enzymatic activity.<br />

Studies on the MAPK ERK2 led to an understanding<br />

<strong>of</strong> the events induced by phosphorylation<br />

that are important for increased activity.<br />

Conformational changes include refolding <strong>of</strong><br />

the activation loop to improve substrate positioning<br />

and realignment <strong>of</strong> catalytic residues;<br />

this is most obvious in the repositioning <strong>of</strong> <br />

helix C, which contains a Glu involved in phosphoryl<br />

transfer.<br />

Amplification occurs moving down the cascade<br />

from the MAP3K to the MAP2K step because<br />

the MAP2Ks are much more abundant<br />

than the MAP3Ks. The MAP2K to MAPK step<br />

may also amplify the signal if the MAPK is present<br />

in excess <strong>of</strong> the MAP2K. In addition, the<br />

phosphorylation <strong>of</strong> a MAPK by a MAP2K on a<br />

14.32 MAPKs are central to many <strong>signaling</strong> pathways 631


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 632<br />

Generic<br />

G Protein<br />

small or<br />

heterotrimer<br />

MAP4K<br />

MAP3K<br />

MAP2K<br />

MAPK<br />

Major<br />

targets:<br />

Transcription<br />

factor<br />

Protein<br />

kinase<br />

Output<br />

S.<br />

cerevisiae<br />

Gβγ<br />

Ste20p<br />

Ste11p<br />

Ste7p<br />

Fus3p<br />

Ste12p<br />

Mating<br />

MAPK pathways<br />

Mammals<br />

Ras Rac/Cdc42 Rac<br />

?<br />

Ste20 Ste20<br />

PAK/PKC ?<br />

family family<br />

Raf<br />

MEK1<br />

MEK2<br />

ERK1<br />

ERK2<br />

Elk-1<br />

Rsk<br />

many<br />

MEK4<br />

MEK7<br />

JNK1<br />

JNK2<br />

JNK3<br />

c-Jun<br />

ATF2<br />

many<br />

MEK3<br />

MEK6<br />

p38α<br />

p38β<br />

p38γ<br />

p38δ<br />

MEF2<br />

MAPKAPK2<br />

Proliferation<br />

Development<br />

Differentiation<br />

(and other processes)<br />

MEKK2<br />

MEKK3<br />

MEK5<br />

ERK5<br />

MEF2<br />

FIGURE 14.38 MAPK pathways can be regulated by a diverse group <strong>of</strong> upstream<br />

regulatory mechanisms that <strong>of</strong>ten include adaptors, small G proteins, and MAP4Ks.<br />

These molecules impinge on the activities <strong>of</strong> MAP3Ks. MAP3Ks regulate one or<br />

more MAP2Ks depending on localization and scaffolding. The MAP2Ks display<br />

great selectivity for a single MAPK type. MAPKs have overlapping and unique<br />

substrates and participate in <strong>signaling</strong> cascades leading to many <strong>cell</strong>ular responses.<br />

Rsk<br />

Tyr and a Thr residue creates cooperative activation<br />

<strong>of</strong> the MAPK; this is another mechanism,<br />

in addition to those described for PKA and<br />

calmodulin, to introduce a threshold and apparently<br />

cooperative behavior into the pathway<br />

over a narrow range <strong>of</strong> input signal. This<br />

multistep cascade provides multiple sites for<br />

modulatory inputs from other pathways.<br />

Stabilized interactions between components<br />

are also important. MAP2Ks, as well as MAPK<br />

substrates and MAPK phosphatases, generally<br />

contain a basic/hydrophobic docking motif that<br />

interacts with acidic residues and binds in a hydrophobic<br />

groove on the MAPK catalytic domain.<br />

Additional components including scaffolds<br />

are necessary for the efficient activation <strong>of</strong><br />

MAPK cascades in <strong>cell</strong>s and usually have additional<br />

functions. Several scaffolds have been<br />

identified that bind to two or more components<br />

for each <strong>of</strong> the three major MAPK cascades, the<br />

ERK1/2, JNK1-3, and p38 α, β, γ, and δ cascades.<br />

The ERK1/2 pathway is regulated by most<br />

<strong>cell</strong> surface receptors, including receptors that<br />

employ tyrosine kinases, GPCRs, and others. The<br />

PDGF receptor, like most receptor systems, activates<br />

the ERK1/2 cascade through Ras. PDGF<br />

stimulates autophosphorylation <strong>of</strong> its receptor<br />

and the subsequent association <strong>of</strong> effectors with<br />

its cytoplasmic domain (see 14.30 Diverse <strong>signaling</strong><br />

mechanisms are regulated by protein tyrosine kinases).<br />

In response to PDGF, ERK1/2 promotes <strong>cell</strong><br />

proliferation and differentiation by phosphorylation<br />

<strong>of</strong> membrane enzymes, proteins involved<br />

in determining <strong>cell</strong> shape and motility, and also<br />

by concentrating in the nucleus to phosphorylate<br />

regulatory factors that control transcription.<br />

14.33<br />

Cyclin-dependent protein<br />

kinases control the <strong>cell</strong><br />

cycle<br />

Key concepts<br />

• The <strong>cell</strong> cycle is regulated by cyclin-dependent<br />

protein kinases (CDKs).<br />

• Activation <strong>of</strong> CDKs involves protein binding,<br />

dephosphorylation, and phosphorylation.<br />

Cell division is regulated positively and negatively<br />

by factors that stimulate proliferation and<br />

inputs that monitor <strong>cell</strong> state. The sum <strong>of</strong> these<br />

factors is integrated in the regulation <strong>of</strong> cyclindependent<br />

protein kinases (CDKs). CDKs are<br />

protein serine/threonine kinases that are major<br />

regulators <strong>of</strong> <strong>cell</strong> cycle progression. Most<br />

CDKs are regulated both by kinases and phosphatases<br />

and by association with other proteins<br />

called cyclins. Cyclins are synthesized and degraded<br />

every <strong>cell</strong> cycle. Because most CDKs are<br />

dependent upon cyclin binding for activation,<br />

the timing <strong>of</strong> the synthesis and degradation <strong>of</strong><br />

individual cyclins determines when a CDK will<br />

function. The most notable noncycling member<br />

<strong>of</strong> the CDK family is Cdk5, which is highly expressed<br />

in terminally differentiated neurons.<br />

Cdk5 binds the non-cyclin protein p35 as its activating<br />

subunit.<br />

We will briefly examine the regulation <strong>of</strong><br />

Cdc2, a major CDK in both mammals and yeast.<br />

The first step in regulation <strong>of</strong> Cdc2 is the association<br />

with cyclin. A second step required for<br />

activation <strong>of</strong> Cdc2 is phosphorylation <strong>of</strong> a Thr<br />

632 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 633<br />

CDKs require cyclin binding for activation<br />

Tyr15<br />

Lys33<br />

Glu51<br />

Cdk2<br />

Cyclin A<br />

FIGURE 14.39 The view <strong>of</strong> the crystal structure <strong>of</strong> CDK2 bound to cyclin A<br />

shows residues in the ATP binding site. The enlargement on the right shows<br />

the interaction between Lys33 and Glu51, catalytic residues that interact with<br />

ATP to promote phosphoryl transfer. Tyr15 is phosphorylated in inactive forms<br />

<strong>of</strong> CDK2. A phosphoryl group on Tyr15 inhibits CDK activity by interfering with<br />

ATP binding. Structure generated from Protein Data Bank file 1JST.<br />

residue in its activation loop by another CDK<br />

type kinase. In spite <strong>of</strong> its association with cyclin,<br />

this form <strong>of</strong> Cdc2 is not yet active due to<br />

inhibitory phosphorylation <strong>of</strong> Tyr and Thr<br />

residues in the ATP binding pocket. Release <strong>of</strong><br />

inhibition by dephosphorylation <strong>of</strong> the residues<br />

in the ATP pocket is catalyzed by the Cdc25 family<br />

<strong>of</strong> phosphoprotein phosphatases, resulting in<br />

activation <strong>of</strong> Cdc2. The proximity <strong>of</strong> the Tyr<br />

residue to catalytic residues is shown in FIGURE<br />

14.39. The complexity <strong>of</strong> activation <strong>of</strong> CDKs<br />

makes possible the imposition <strong>of</strong> <strong>cell</strong> cycle checkpoints.<br />

For more on CDKs and cyclins see 11.4<br />

The <strong>cell</strong> cycle is a cycle <strong>of</strong> CDK function.<br />

14.34<br />

Diverse receptors recruit<br />

protein tyrosine kinases<br />

to the plasma membrane<br />

Key concepts<br />

• Receptors that bind protein tyrosine kinases use<br />

combinations <strong>of</strong> effectors similar to those used by<br />

receptor tyrosine kinases.<br />

• These receptors <strong>of</strong>ten bind directly to transcription<br />

factors.<br />

Many receptors act through protein tyrosine<br />

kinases, but their <strong>cell</strong> surface receptors lack kinase<br />

activity. Instead, these receptors act by recruiting<br />

and activating protein tyrosine kinases<br />

at the plasma membrane. In this group <strong>of</strong> receptors<br />

are integrins, which are key molecules involved<br />

in <strong>cell</strong> adhesion, growth hormone<br />

receptors, and receptors that mediate inflammatory<br />

and immune responses. While their<br />

structures vary enormously, their mechanisms<br />

<strong>of</strong> action are related.<br />

Integrins are receptors whose major function<br />

is to attach <strong>cell</strong>s to the extra<strong>cell</strong>ular matrix. They<br />

also mediate some interactions with proteins on<br />

other <strong>cell</strong>s, as depicted in FIGURE 14.40. Ligands for<br />

integrins include a number <strong>of</strong> extra<strong>cell</strong>ular matrix<br />

proteins, such as fibronectin, as well as <strong>cell</strong><br />

surface proteins that cooperate in <strong>cell</strong>-<strong>cell</strong> interactions.<br />

Integrin ligation provides <strong>cell</strong>s with information<br />

about their environment that influences<br />

<strong>cell</strong> behavior. Ligation <strong>of</strong> integrins initiates signals<br />

that control <strong>cell</strong> programs, including <strong>cell</strong> cycle entry,<br />

proliferation, survival, differentiation, changes<br />

in <strong>cell</strong> shape, and motility, as well as fine-tuning<br />

responses to other ligands. For more details on integrins<br />

see 15.13 Most integrins are receptors for extra<strong>cell</strong>ular<br />

matrix proteins and 15.14 Integrin receptors<br />

participate in <strong>cell</strong> <strong>signaling</strong>.<br />

14.34 Diverse receptors recruit protein tyrosine kinases to the plasma membrane 633


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 634<br />

Crk<br />

ECM<br />

Crk<br />

Src<br />

CAS<br />

CYTOPLASM<br />

FAK<br />

SOS<br />

Integrin <strong>signaling</strong><br />

INTEGRINS<br />

Paxillin<br />

Talin<br />

Vinculin<br />

Grb2<br />

Actin<br />

filament<br />

ILK<br />

p85 PI 3-<br />

Kinase<br />

p110<br />

Vinculin<br />

Tensin<br />

Talin<br />

FIGURE 14.40 Integrins bind to an array <strong>of</strong> cytoplasmic proteins to regulate<br />

the cytoskeleton and intra<strong>cell</strong>ular <strong>signaling</strong> pathways. The associated cytoskeletal<br />

elements include actin filaments and focal adhesion proteins -actinin, vinculin,<br />

paxillin, and talin. Signaling molecules include the focal adhesion kinase<br />

FAK; the adaptors Cas, Crk and Grb2; Src and CSK (see 14.31 Src family protein<br />

kinases cooperate with receptor protein tyrosine kinases), PI 3-kinase (see 14.16<br />

Lipids and lipid-derived compounds are <strong>signaling</strong> molecules); and the Ras exchange<br />

factor SOS. Stimulation <strong>of</strong> GTP binding <strong>of</strong> Ras by SOS leads to activation<br />

<strong>of</strong> the MAPK pathway (see 14.32 MAPKs are central to many <strong>signaling</strong><br />

pathways).<br />

Talin and α-actinin are among cytoskeletal<br />

proteins that interact directly with certain integrin<br />

subunits. These cytoskeletal proteins link<br />

integrins to complex cytoskeletal structures<br />

known as focal adhesions.<br />

Focal adhesions connect the cytoskeleton to<br />

signal transduction cascades that communicate<br />

states <strong>of</strong> <strong>cell</strong>ular attachment to the regulation <strong>of</strong><br />

<strong>cell</strong>ular responses. Focal adhesion complexes contain<br />

the focal adhesion kinase FAK, which is activated<br />

by integrin ligation. Autophosphorylation<br />

<strong>of</strong> FAK recruits <strong>signaling</strong> proteins containing SH2<br />

domains, especially the p85 subunit <strong>of</strong> PI 3-kinase<br />

and Src family protein kinases. The <strong>signaling</strong><br />

molecules associated with the integrin-bound<br />

cytoskeletal proteins, whether focal adhesions or<br />

other structural complexes, mediate the diverse<br />

actions <strong>of</strong> integrins. The association <strong>of</strong> cytoskeletal<br />

proteins with integrin receptors also causes<br />

functional changes to the receptors.<br />

Signals that act over a distance, such as hormones,<br />

can also employ nonreceptor tyrosine<br />

kinases to transmit their message inside a <strong>cell</strong>.<br />

Growth hormone (GH) is a protein hormone<br />

secreted by the anterior pituitary gland that regulates<br />

bone growth, fat metabolism, and other<br />

<strong>cell</strong>ular growth phenomena. Absence <strong>of</strong> growth<br />

hormone results in short stature, whereas hypersecretion<br />

causes acromegaly, a form <strong>of</strong> gigantism.<br />

The GH receptor is a member <strong>of</strong> the<br />

cytokine receptor family, which includes receptors<br />

for prolactin, erythropoietin, leptin, and<br />

interleukins. All these receptors display similar<br />

biochemical functions, such as association with<br />

members <strong>of</strong> the JAK/TYK family <strong>of</strong> protein tyrosine<br />

kinases, but select for different but overlapping<br />

sets <strong>of</strong> cytoplasmic <strong>signaling</strong> proteins.<br />

Signal transduction by the GH receptor provides<br />

a model for receptors that lack enzymatic function<br />

and act as agonist-promoted scaffolds for<br />

intra<strong>cell</strong>ular <strong>signaling</strong> proteins.<br />

FIGURE 14.41 shows the structure <strong>of</strong> growth<br />

hormone bound to the extra<strong>cell</strong>ular domain <strong>of</strong><br />

its receptor. The majority <strong>of</strong> binding energy<br />

comes from only a small number <strong>of</strong> residues in<br />

the binding interface. Inside the <strong>cell</strong>, <strong>signaling</strong><br />

by the GH receptor depends significantly on its<br />

association with the cytoplasmic tyrosine protein<br />

kinase Janus kinase 2 (JAK2). FIGURE 14.42<br />

shows that JAK2 binds to a proline-rich region<br />

<strong>of</strong> the receptor. Ligand binding induces receptor<br />

dimerization, which then promotes<br />

activation <strong>of</strong> JAK2 through intermolecular autophosphorylation.<br />

GH <strong>signaling</strong> is thus mediated primarily by<br />

inducing Tyr phosphorylation. In addition to<br />

JAK2 autophosphorylation, the receptor itself<br />

becomes Tyr phosphorylated. As is true for receptor<br />

tyrosine kinases, Tyr phosphorylation <strong>of</strong> the<br />

growth hormone receptor creates binding sites<br />

for <strong>signaling</strong> proteins that contain phosphotyrosine-binding<br />

domains. Primary targets are transcription<br />

factors known as signal transducers and<br />

activators <strong>of</strong> transcription, or STATs. STATs contain<br />

SH2 domains and bind Tyr-phosphorylated<br />

motifs on the growth hormone receptor. While<br />

receptor bound, STATs are Tyr phosphorylated<br />

by JAK2 and then released to travel to the nucleus<br />

to mediate changes in transcription.<br />

The growth hormone receptor and the associated<br />

JAK2 also activate other <strong>signaling</strong> pathways.<br />

For example, the adaptor Shc is Tyr<br />

phosphorylated by JAK2. Engagement <strong>of</strong> Shc<br />

leads to activation <strong>of</strong> Ras and the ERK1/2 MAPK<br />

pathway. Adaptors specialized for insulin-<strong>signaling</strong><br />

pathways, insulin receptor substrates (IRS)<br />

1, 2, and 3, are also growth hormone targets, perhaps<br />

reflecting the ability <strong>of</strong> growth hormone<br />

to induce certain insulin-like metabolic actions.<br />

Feedback circuits are also engaged during<br />

GH <strong>signaling</strong>. The growth hormone receptor<br />

complex binds the adaptor SH2-B, which has a<br />

634 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 635<br />

Growth hormone structure<br />

Growth hormone <strong>signaling</strong> is transduced by JAK2<br />

Growth hormone<br />

receptor<br />

hGH<br />

Growth<br />

hormone<br />

Dimerization<br />

JAK2<br />

JAK2<br />

ΔΔG (kcal/mol)<br />

> 1.5<br />

0.5 to 1.5<br />

-0.5 to 0.5<br />

< -0.5<br />

untested<br />

CYTOPLASM<br />

JAK2s bind and<br />

phosphorylate<br />

receptor<br />

STAT<br />

STAT<br />

STATs bind and<br />

are phosphorylated<br />

FIGURE 14.41 Proteins <strong>of</strong>ten interact over a large surface area. Growth<br />

hormone binding to its receptor is an example <strong>of</strong> the energy <strong>of</strong> binding<br />

coming primarily from a small number <strong>of</strong> the contacts between<br />

the two proteins, creating an interaction hot spot. The complex <strong>of</strong><br />

growth hormone bound to the growth hormone receptor-binding domain<br />

determined by crystallography has been peeled apart in this figure<br />

to show the binding energy associated with residues in the binding<br />

interface from each protein determined by mutagenesis and binding<br />

studies. Fewer than half <strong>of</strong> the residues in the interface contribute<br />

the majority <strong>of</strong> binding energy. Reproduced with permission from T.<br />

Clackson and J. A. Wells. 1995. Science. 267: 383–386. © AAAS. Photos<br />

courtesy <strong>of</strong> Tim Clackson, ARIAD Pharmaceuticals, Inc.<br />

NUCLEUS<br />

Phosphorylated<br />

STATs bind DNA<br />

FIGURE 14.42 The growth hormone receptor binds to JAK2. Many GH signals<br />

are mediated by Tyr phosphorylation <strong>of</strong> the receptor by JAK2, which creates<br />

binding sites for <strong>signaling</strong> molecules with SH2 domains, notably STATs. STATs<br />

then enter the nucleus to cause changes in gene transcription.<br />

stimulatory effect on growth hormone <strong>signaling</strong>.<br />

On the other hand, suppressors <strong>of</strong> cytokine<br />

<strong>signaling</strong> (SOCS proteins) are among the genes<br />

whose transcription is induced by growth hormone.<br />

As the name indicates, SOCS proteins inhibit<br />

cytokine <strong>signaling</strong> in some if not all cases by<br />

inhibiting the activity <strong>of</strong> JAK2. SOCS proteins<br />

contain an SH2 domain that facilitates their binding<br />

either to phosphorylated JAK2 or cytokine<br />

receptors. The mechanism <strong>of</strong> <strong>signaling</strong> inhibition<br />

may differ among SOCS proteins because<br />

some require the GH receptor to interfere with<br />

JAK2 <strong>signaling</strong>. SOCS-1, on the other hand, binds<br />

directly to the JAK2 activation loop and does not<br />

require a receptor to inhibit JAK2 activity. This<br />

mechanism may be particularly important in GH<br />

<strong>signaling</strong> because, in contrast to the ligand-induced<br />

down regulation mechanisms controlling<br />

many receptors, the GH receptor is degraded in<br />

a ligand-independent manner.<br />

Receptors for cytokines also act by recruiting<br />

tyrosine kinases. The cytokines—<strong>signaling</strong><br />

proteins that modulate inflammation and <strong>cell</strong><br />

growth and differentiation—include interleukins,<br />

leukemia inhibitory factor, oncostatin M, cardiotrophin-1,<br />

cardiotrophin-like cytokine, and<br />

ciliary neurotrophic factor (CNTF). Each cytokine<br />

binds a unique receptor, but each receptor<br />

binds a transmembrane protein called gp130.<br />

Mechanisms <strong>of</strong> <strong>signaling</strong> by gp130 involve interactions<br />

with tyrosine kinases <strong>of</strong> the JAK/TYK<br />

types and transcription factors in the STAT family.<br />

This mechanism is similar to those employed<br />

by the growth hormone receptor.<br />

14.34 Diverse receptors recruit protein tyrosine kinases to the plasma membrane 635


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 636<br />

CD3<br />

Ligand<br />

Receptor<br />

Adaptor/<br />

subunit<br />

Transducer<br />

Kinase<br />

cascade<br />

Transcription<br />

factor<br />

complex<br />

TCR<br />

MHC<br />

Antigen<br />

ITAM<br />

CD3<br />

Receptor <strong>signaling</strong> pathways<br />

Growth<br />

PDGF Insulin hormone IL-1β TGF-β<br />

PDGF<br />

receptor<br />

SHP2/Grb2<br />

SOS/Ras<br />

MAPK<br />

Ternary<br />

complex<br />

factors<br />

Insulin<br />

receptor<br />

IRS1<br />

PI 3-kinase<br />

Akt2<br />

FOXO<br />

GH<br />

receptor<br />

STATs<br />

T <strong>cell</strong> receptor <strong>signaling</strong><br />

Lck<br />

After binding <strong>of</strong> the TCR<br />

to the MHC-antigen<br />

complex, Lck<br />

phosphorylates ITAMS<br />

IL-1β<br />

receptor<br />

gp130<br />

STATs<br />

TGF-β<br />

type II<br />

receptor<br />

Type I<br />

receptor<br />

SMADs<br />

FIGURE 14.43 Major <strong>signaling</strong> cascades controlled by PDGF, insulin, TGF-, IL-<br />

1, and growth hormone are compared. Each receptor either contains or interacts<br />

with a protein kinase that associates with or recruits a transducer. The<br />

transducer regulates downstream effectors either directly or through an intermediate<br />

protein kinase cascade. The effectors shown are transcriptional regulators.<br />

Phosphorylation by the transducer or kinase cascade activates all <strong>of</strong> the effectors<br />

except the FOXO proteins, which may be excluded from the nucleus by phosphorylation.<br />

The table only shows snapshots <strong>of</strong> much more complex <strong>signaling</strong> networks<br />

controlled by these ligands. Many <strong>of</strong> these and other intermediates serve<br />

multiple ligands. For example, IRS proteins also contribute to growth hormone<br />

and IL-1 <strong>signaling</strong>, and MAPK pathways are regulated by all <strong>of</strong> these ligands.<br />

ZAP-70<br />

FIGURE 14.44 The T <strong>cell</strong> receptor (TCR) is a multisubunit receptor. It is phosphorylated<br />

on activation motifs or ITAMs by Lck, or a related Src family protein<br />

kinase. The phosphorylated residues create binding sites for another tyrosine<br />

protein kinase ZAP-70. ZAP-70 then recruits other <strong>signaling</strong> molecules to the<br />

complex including phospholipase C, PI 3-kinase, and a Ras exchange factor to<br />

activate downstream <strong>signaling</strong> pathways.<br />

JAK<br />

JAK<br />

Unlike many cytokine receptors in this class,<br />

the CNTF receptor does not itself span the membrane.<br />

Instead, it is glycosyl phosphoinositol<br />

(GPI)-linked to the outer face <strong>of</strong> the plasma<br />

membrane. The GPI linkage is a covalent bond,<br />

and the receptor can be released into the extra<strong>cell</strong>ular<br />

fluid by a specific phospholipase. The<br />

freed receptor may interact with membranes <strong>of</strong><br />

other <strong>cell</strong>s to induce signals.<br />

The use <strong>of</strong> a common signal transducing<br />

subunit, gp130, suggests that unique mechanisms<br />

exist to create ligand-specific responses;<br />

under some circumstances competition by the<br />

ligand binding subunits for interaction with the<br />

gp130 signal transducer may influence <strong>signaling</strong><br />

outcomes. FIGURE 14.43 illustrates some parallels<br />

in <strong>signaling</strong> pathways initiated by receptors<br />

with associated or intrinsic protein kinases.<br />

The last receptor type we will discuss takes<br />

the concept <strong>of</strong> the specific and common subunits<br />

even further. The complex multiprotein<br />

T <strong>cell</strong> receptor (TCR) is found uniquely on T<br />

lymphocytes and is responsible for the ability <strong>of</strong><br />

these <strong>cell</strong>s to recognize and respond to specific<br />

antigens. The TCR, illustrated in FIGURE 14.44, is<br />

composed <strong>of</strong> eight subunits that can be described<br />

as an assembly <strong>of</strong> four dimers, αβ, γε, δε, and ζζ.<br />

The specificity <strong>of</strong> antigen recognition is determined<br />

by the α and β subunits, which are different<br />

for each <strong>cell</strong>. The remaining subunits are<br />

invariant in TCRs.<br />

The CD3 complex γ, δ, and ε subunits are<br />

similar in sequence to one another. The ζ chain,<br />

unlike the other subunits, appears on certain<br />

other <strong>cell</strong> types and may be a component <strong>of</strong><br />

other receptors, such as the Fc receptor, which<br />

binds a portion <strong>of</strong> certain immunoglobulins.<br />

A motif called the immunoreceptor tyrosine-based<br />

activation motif, or ITAM, which features<br />

closely spaced pairs <strong>of</strong> Tyr residues, is key<br />

to <strong>signaling</strong> by the TCR. The CD3 subunits each<br />

contain one ITAM and the ζ chain contains three<br />

ITAMs, for a total <strong>of</strong> ten motifs in each TCR.<br />

Engagement <strong>of</strong> the TCR causes the Src family<br />

kinases Lck and Fyn to phosphorylate the pairs<br />

<strong>of</strong> Tyr residues in the ITAMs. The ITAMs then<br />

bind the tandem SH2 domains <strong>of</strong> the protein tyrosine<br />

kinase ζ-chain-associated protein <strong>of</strong> 70<br />

kDa (ZAP-70), which becomes activated by Src.<br />

Tyr phosphorylation sites on ZAP-70 bind to<br />

other adaptors and <strong>signaling</strong> molecules, and Tyr<br />

phosphorylation by ZAP-70 activates additional<br />

signal transducers. The sum <strong>of</strong> these events leads<br />

to the downstream responses <strong>of</strong> T <strong>cell</strong>s to antigen<br />

engagement, which include <strong>cell</strong> cycle progression<br />

and the elaboration <strong>of</strong> cytokines such<br />

as interleukin-2.<br />

636 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 637<br />

14.35<br />

What’s next?<br />

New <strong>signaling</strong> proteins and new regulatory interactions<br />

seem to show up every day. The challenge<br />

now is to understand how <strong>cell</strong>s organize<br />

these proteins and their individual interactions<br />

to create adaptable information-processing networks.<br />

How do <strong>cell</strong>s use simple chemical reactions<br />

to sort and integrate multiple simultaneous<br />

inputs and then direct this information to diverse<br />

effector machinery? How do they interpret<br />

the inputs in the context <strong>of</strong> their growth and<br />

metabolic activities? In principle, three areas <strong>of</strong><br />

research have to contribute to allow us to understand<br />

integrative <strong>cell</strong>ular <strong>signaling</strong>.<br />

First, we need real-time, noninterfering<br />

biosensors to measure intra<strong>cell</strong>ular <strong>signaling</strong><br />

reactions. Most current sensors use combinations<br />

<strong>of</strong> fluorescent moieties and signal-binding<br />

protein domains to provide fast optical<br />

readouts. For many pathways, several reactions<br />

can be monitored within <strong>cell</strong>s over subsecond<br />

time scales. We need more, better, and faster<br />

sensors and sensors that can report with single<strong>cell</strong><br />

and sub<strong>cell</strong>ular resolution. Genetically encoded<br />

sensors will be complemented by synthetic<br />

molecules.<br />

Our ability to manipulate <strong>signaling</strong> networks<br />

is also improving dramatically but still<br />

falls short. We can manipulate <strong>signaling</strong> networks<br />

by overexpression, knockout, and knockdown<br />

<strong>of</strong> genes, but <strong>signaling</strong> pathways are<br />

wonderfully adaptive and frequently circumvent<br />

our best efforts to control them. We still<br />

need chemical regulators that can act promptly<br />

in <strong>cell</strong>s. Structure-based design <strong>of</strong> such regulatory<br />

molecules will be vital.<br />

Last, our ability to analyze the behavior <strong>of</strong><br />

<strong>signaling</strong> networks depends on our ability to<br />

measure and interpret <strong>signaling</strong> quantitatively.<br />

It is ironic but true that really complex systems<br />

cannot be described without explicit quantitative<br />

models for how they work. Computational<br />

modeling and simulation <strong>of</strong> <strong>signaling</strong> networks<br />

requires both better theoretical understanding<br />

<strong>of</strong> network dynamics and better algorithmic implementation.<br />

The goal is to understand how <strong>cell</strong>s think.<br />

14.36<br />

Summary<br />

Signal transduction encompasses mechanisms<br />

used by all <strong>cell</strong>s to sense and react to stimuli in<br />

their environment. Cells express receptors that<br />

recognize specific extra<strong>cell</strong>ular stimuli, includ-<br />

ing nutrients, hormones, neurotransmitters,<br />

and other <strong>cell</strong>s. Upon receptor binding, signals<br />

are converted to well-defined intra<strong>cell</strong>ular chemical<br />

or physical reactions that change the activities<br />

and the organization <strong>of</strong> protein complexes<br />

within <strong>cell</strong>s. The changes directed by the stimuli<br />

lead to altered <strong>cell</strong> behavior. The behavior <strong>of</strong><br />

the <strong>cell</strong> is determined then by its intra<strong>cell</strong>ular<br />

state and the integrated information from extra<strong>cell</strong>ular<br />

stimuli so that the appropriate responses<br />

are achieved.<br />

The basic biochemical components and<br />

processes <strong>of</strong> signal transduction are conserved<br />

throughout biology. Families <strong>of</strong> proteins are<br />

used in a variety <strong>of</strong> ways for many different<br />

physiological purposes. Cells <strong>of</strong>ten use the same<br />

series <strong>of</strong> <strong>signaling</strong> proteins to regulate multiple<br />

processes, such as transcription, ion transport,<br />

locomotion, and metabolism.<br />

Signaling pathways are assembled into <strong>signaling</strong><br />

networks to allow the <strong>cell</strong> to coordinate<br />

its responses to multiple inputs with its ongoing<br />

functions. It is now possible to discern conserved<br />

reaction sequences in and between<br />

pathways in <strong>signaling</strong> networks that are analogous<br />

to devices within the circuits <strong>of</strong> analog<br />

computers: amplifiers, logic gates, feedback and<br />

feed-forward controls, and memory.<br />

References<br />

14.1 Introduction<br />

Review<br />

Sauro, H. M. and Kholodenko, B. N., 2004.<br />

Quantitative analysis <strong>of</strong> <strong>signaling</strong> networks.<br />

Prog. Biophys. Mol. Biol. v. 86 p.<br />

5–43.<br />

Research<br />

Milo, R., Shen-Orr, S., Itzkovitz, S., Kashtan,<br />

N., Chklovskii, D., and Alon, U., 2002.<br />

Network motifs: simple building blocks <strong>of</strong><br />

complex networks. Science v. 298 p.<br />

824–827.<br />

14.2 Cellular <strong>signaling</strong> is primarily chemical<br />

Review<br />

Arshavsky, V. Y., Lamb, T. D., and Pugh, E. N.,<br />

Jr., 2002. G proteins and phototransduction.<br />

Annu. Rev. Physiol. v. 64 p. 153–187.<br />

Caterina, M. J. and Julius, D., 2001. The<br />

vanilloid receptor: a molecular gateway<br />

to the pain pathway. Annu. Rev. Neurosci.<br />

v. 24 p. 487–517.<br />

Gillespie, P. G. and Cyr, J. L., 2004. Myosin-<br />

1c, the hair <strong>cell</strong>’s adaptation motor. Annu.<br />

Rev. Physiol. v. 66 p. 521–545.<br />

References 637


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 638<br />

Lin, C. and Shalitin, D., 2003. Cryptochrome<br />

structure and signal transduction. Annu.<br />

Rev. Plant Biol. v. 54 p. 469–496.<br />

Rockwell, N.C., Su, Y.-S., and Lagarias, J.C.,<br />

2006. Phytochrome structure and <strong>signaling</strong><br />

mechanisms. Annu. Rev. Plant Biology,<br />

v. 57 p. 837–858.<br />

Sancar, A., 2000. Cryptochrome: the second<br />

photoactive pigment in the eye and its<br />

role in circadian photoreception. Annu.<br />

Rev. Biochem. v. 69 p. 31–67.<br />

14.3 Receptors sense diverse stimuli but initiate<br />

a limited repertoire <strong>of</strong> <strong>cell</strong>ular signals<br />

Research<br />

Klein, C., Paul, J. I., Sauvé, K., Schmidt, M.<br />

M., Arcangeli, L., Ransom, J., Trueheart,<br />

J., Manfredi, J. P., Broach, J. R., and<br />

Murphy, A. J., 1998. Identification <strong>of</strong> surrogate<br />

agonists for the human FPRL-1 receptor<br />

by autocrine selection in yeast.<br />

Nat. Biotechnol. v. 16 p. 1334–1337.<br />

14.5 Ligand binding changes receptor conformation<br />

Review<br />

Ross, E. M., and Kenakin, T. P., 2001.<br />

Pharmacodynamics: Mechanisms <strong>of</strong> drug<br />

action and the relationship between drug<br />

concentration and effects. In Goodman<br />

and Gilmans’s The Pharmacological Basis <strong>of</strong><br />

Therapeutics, 10th. Ed., J. G. Hardman and<br />

L. E. Limbird, eds., New York: McGraw-<br />

Hill, p. 31–43.<br />

14.6 Signals are sorted and integrated in <strong>signaling</strong><br />

pathways and networks<br />

Research<br />

Itzkovitz, S., Milo, R., Kashtan, N., Ziv, G.,<br />

and Alon, U., 2003. Subgraphs in random<br />

networks. Phys. Rev. E. v. 68 p. 126–127.<br />

14.7 Cellular <strong>signaling</strong> pathways can be<br />

thought <strong>of</strong> as biochemical logic circuits<br />

Review<br />

Milo, R., Shen-Orr, S., Itzkovitz, S., Kashtan,<br />

N., Chklovskii, D., and Alon, U., 2002.<br />

Network motifs: Simple building blocks <strong>of</strong><br />

complex networks. Science v. 298 p.<br />

824–827.<br />

Research<br />

Torres, E. and Rosen, M. K., 2003. Contingent<br />

phosphorylation/dephosphorylation provides<br />

a mechanism <strong>of</strong> molecular memory<br />

in WASP. Mol. Cell v. 11 p. 1215–1227.<br />

14.8 Scaffolds increase <strong>signaling</strong> efficiency<br />

and enhance spatial organization <strong>of</strong> <strong>signaling</strong><br />

Review<br />

Elion, E. A., 2001. The Ste5p scaffold. J. Cell<br />

Sci. v. 114 p. 3967–3978.<br />

Kholodenko, B. N., 2003. Four-dimensional<br />

organization <strong>of</strong> protein kinase <strong>signaling</strong><br />

cascades: the roles <strong>of</strong> diffusion, endocytosis<br />

and molecular motors. J. Exp. Biol. v.<br />

206 p. 2073–2082.<br />

O’Rourke, S. M., Herskowitz, I., and O’Shea,<br />

E. K., 2002. Yeast go the whole HOG for<br />

the hyperosmotic response. Trends Genet.<br />

v. 18 p. 405–412.<br />

Pawson, T. and Nash, P., 2003. Assembly <strong>of</strong><br />

<strong>cell</strong> regulatory systems through protein<br />

interaction domains. Science v. 300 p.<br />

445–452.<br />

Tsunoda, S. and Zuker, C. S., 1999. The organization<br />

<strong>of</strong> INAD-<strong>signaling</strong> complexes<br />

by a multivalent PDZ domain protein in<br />

Drosophila photoreceptor <strong>cell</strong>s ensures<br />

sensitivity and speed <strong>of</strong> <strong>signaling</strong>. Cell<br />

Calcium v. 26 p. 165–171.<br />

Research<br />

Levchenko, A., Bruck, J., and Sternberg, P.<br />

W., 2000. Scaffold proteins may biphasically<br />

affect the levels <strong>of</strong> mitogen-activated<br />

protein kinase <strong>signaling</strong> and reduce<br />

its threshold properties. Proc. Natl. Acad.<br />

Sci. USA v. 97 p. 5818–5823.<br />

Rensland, H., Lautwein, A., Wittingh<strong>of</strong>er, A.,<br />

and Goody, R. S., 1991. Is there a ratelimiting<br />

step before GTP cleavage by H-<br />

ras p21? Biochemistry v. 30 p. 11181–<br />

11185.<br />

Shenker, A., Goldsmith, P., Unson, C. G., and<br />

Spiegel, A. M., 1991. The G protein coupled<br />

to the thromboxane A2 receptor in<br />

human platelets is a member <strong>of</strong> the novel<br />

Gq family. J. Biol. Chem. v. 266 p. 9309–<br />

9313.<br />

14.9 Independent, modular domains specify<br />

protein-protein interactions<br />

Review<br />

Pawson, T. and Nash, P., 2003. Assembly <strong>of</strong><br />

<strong>cell</strong> regulatory systems through protein<br />

interaction domains. Science v. 300 p.<br />

445–452.<br />

Turk, B. E. and Cantley, L. C., 2003. Peptide<br />

libraries: at the crossroads <strong>of</strong> proteomics<br />

and bioinformatics. Curr. Opin. Chem. Biol.<br />

v. 7 p. 84–90.<br />

Research<br />

Ginty, D. D., Kornhauser, J. M., Thompson,<br />

M. A., Bading, H., Mayo, K. E.,<br />

Takahashi, J. S., and Greenberg, M. E.,<br />

1993. Regulation <strong>of</strong> CREB phosphorylation<br />

in the suprachiasmatic nucleus by<br />

light and a circadian clock. Science v. 260<br />

p. 238–241.<br />

638 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 639<br />

14.10 Cellular <strong>signaling</strong> is remarkably adaptive<br />

Review<br />

Perkins, J. P., Hausdorff, W. P., and Lefkowitz,<br />

R. J., 1990. Mechanisms <strong>of</strong> ligand-induced<br />

desensitization <strong>of</strong> -adrenergic receptors.<br />

In The Beta-Adrenergic Receptors, J.<br />

P. Perkins, ed., Clifton, NJ: Humana Press,<br />

p. 73–124.<br />

14.11 Signaling proteins are frequently expressed<br />

as multiple species<br />

Review<br />

Barnes, N. M., and Sharp, T., 1999. A review<br />

<strong>of</strong> central 5-HT receptors and their function.<br />

Neuropharmacology v. 38 p. 1083–<br />

1152.<br />

Gilman, A. G., 1987. G proteins: transducers<br />

<strong>of</strong> receptor-generated signals Annu. Rev.<br />

Biochem. v. 56 p. 615–649.<br />

Rebecchi, M. J., and Pentyala, S. N., 2000.<br />

Structure, function, and control <strong>of</strong> phosphoinositide-specific<br />

phospholipase C.<br />

Physiol. Rev. v. 80 p. 1291–1335.<br />

Sunahara, R. K. and Taussig, R., 2002.<br />

Is<strong>of</strong>orms <strong>of</strong> mammalian adenylyl cyclase:<br />

Multiplicities <strong>of</strong> <strong>signaling</strong>. Mol.<br />

Interventions v. 2 p. 168–184.<br />

14.14 Second messengers provide readily diffusible<br />

pathways for information transfer<br />

Review<br />

Beavo, J. A., Bechtel, P. J., and Krebs, E. G.,<br />

1975. Mechanisms <strong>of</strong> control for cAMPdependent<br />

protein kinase from skeletal<br />

muscle. Adv. Cyclic Nucleotide Res. v. 5 p.<br />

241–251.<br />

Kobe, B., Heierhorst, J., and Kemp, B. E.,<br />

1997. Intrasteric regulation <strong>of</strong> protein kinases.<br />

Adv. Second Messenger Phosphoprotein<br />

Res. v. 31 p. 29–40.<br />

Wong, W. and Scott, J. D., 2004. AKAP signalling<br />

complexes: focal points in space<br />

and time. Nat. Rev. Mol. Cell Biol. v. 5 p.<br />

959–970.<br />

Research<br />

Rall, T. W. and Sutherland, E. W., 1958.<br />

Formation <strong>of</strong> a cyclic adenine ribonucleotide<br />

by tissue particles. J. Biol. Chem. v.<br />

232 p. 1065–1076.<br />

14.15 Ca2+ <strong>signaling</strong> serves diverse purposes in<br />

all eukaryotic <strong>cell</strong>s<br />

Review<br />

Newton, A. C., 2001. Protein kinase C: structural<br />

and spatial regulation by phosphorylation,<br />

c<strong>of</strong>actors, and macromolecular<br />

interactions. Chem. Rev. v. 101 p.<br />

2353–2364.<br />

Research<br />

Clapperton, J. A., Martin, S. R., Smerdon, S.<br />

J., Gamblin, S. J., and Bayley, P. M.,<br />

2002. Structure <strong>of</strong> the complex <strong>of</strong><br />

calmodulin with the target sequence <strong>of</strong><br />

calmodulin dependent protein kinase I:<br />

Studies <strong>of</strong> the kinase activation mechanism.<br />

Biochemistry. v. 41 p. 14669–14679.<br />

Kuboniwa, H., Tjandra, N., Grzesiek, S., Ren,<br />

H., Klee, C.B., Bax, A., 1995. Solution<br />

structure <strong>of</strong> calcium-free calmodulin. Nat.<br />

Struct. Biol. v. 2 p. 768–776.<br />

14.16 Lipids and lipid-derived compounds are<br />

<strong>signaling</strong> molecules<br />

Review<br />

Rebecchi, M. J., and Pentyala, S. N., 2000.<br />

Structure, function, and control <strong>of</strong> phosphoinositide-specific<br />

phospholipase C.<br />

Physiol. Rev. v. 80 p. 1291–1335.<br />

Yang, C. and Kazanietz, M. G., 2003.<br />

Divergence and complexities in DAG <strong>signaling</strong>:<br />

looking beyond PKC. Trends<br />

Pharmacol. Sci. v. 24 p. 602–608.<br />

14.17 PI 3-kinase synthesizes a lipid activator<br />

<strong>of</strong> <strong>cell</strong> motion and shape change<br />

Review<br />

Downward, J., 2004. PI 3-kinase, Akt and <strong>cell</strong><br />

survival. Semin. Cell Dev. Biol. v. 15 p.<br />

177–182.<br />

Lawlor, M. A. and Alessi, D. R., 2001.<br />

PKB/Akt: a key mediator <strong>of</strong> <strong>cell</strong> proliferation,<br />

survival and insulin responses? J.<br />

Cell Sci. v. 114 p. 2903–2910.<br />

Van Haastert, P. J. and Devreotes, P. N. 2004.<br />

Chemotaxis: <strong>signaling</strong> the way forward.<br />

Nat. Rev. Mol. Cell Biol. v. 5. p. 626–634.<br />

14.18 Signaling through ion channel receptors<br />

is very fast<br />

Review<br />

Clapham, D. E., 2003. TRP channels as <strong>cell</strong>ular<br />

sensors. Nature v. 426 p. 517–524.<br />

Corey, D. P., 2003. New TRP channels in<br />

hearing and mechanosensation. Neuron v.<br />

39 p. 585–588.<br />

Hille, B, 1992. Ionic channels <strong>of</strong> excitable<br />

membranes. Sinauer Associates.<br />

Siegelbaum, S. A. and Koester, J., 2000. Ion<br />

channels and Membrane potential, in<br />

<strong>Principles</strong> <strong>of</strong> Neural Science, 4th. Ed., E. R.<br />

Kandel, J. H. Schwartz, and T. M. Jessell,<br />

eds., New York: McGraw-Hill. p.<br />

105–139.<br />

Research<br />

Unwin, N., 2005. Refined structure <strong>of</strong> the<br />

nicotinic acetylcholine receptor at 4Å resolution.<br />

J. Mol. Biol. v. 346 p. 967.<br />

14.19 Nuclear receptors regulate transcription<br />

References 639


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 640<br />

Review<br />

Mangelsdorf, D. J. et al., 1995. The nuclear<br />

receptor superfamily: the second decade.<br />

Cell v. 83 p. 835–839.<br />

Smith, C. L. and O’Malley, B. W., 2004.<br />

Coregulator function: a key to understanding<br />

tissue specificity <strong>of</strong> selective receptor<br />

modulators. Endocr. Rev. v. 25 p.<br />

45–71.<br />

Research<br />

Brzozowski A. M., Pike, A. C., Dauter, Z.,<br />

Hubbard, R. E., Bonn, T., Engstrom, O.,<br />

Ohman, L., Green, G. L., Gustafsson, J. A.,<br />

and Carlquist, M., 1997. Molecular basis<br />

<strong>of</strong> agonism and antagonism in the oestrogen<br />

receptor. Nature v. 389 p. 753–758.<br />

14.20 G protein <strong>signaling</strong> modules are widely<br />

used and highly adaptable<br />

Review<br />

Clapham, D. E. and Neer, E. J., 1997. G protein<br />

subunits. Annu. Rev. Pharmacol.<br />

Toxicol. v. 37 p. 167–203.<br />

Filipek, S., Teller, D. C., Palczewski, K., and<br />

Stenkamp, R., 2003. The crystallographic<br />

model <strong>of</strong> rhodopsin and its use in studies<br />

<strong>of</strong> other G protein-coupled receptors.<br />

Annu. Rev. Biophys. Biomol. Struct. v. 32 p.<br />

375–397.<br />

Ross, E. M. and Wilkie, T. M., 2000. GTPaseactivating<br />

proteins for heterotrimeric G<br />

proteins: Regulators <strong>of</strong> G protein <strong>signaling</strong><br />

(RGS) and RGS-like proteins. Annu.<br />

Rev. Biochem. v. 69 p. 795–827.<br />

Ross, E. M., 1989. Signal sorting and amplification<br />

through G protein-coupled receptors.<br />

Neuron v. 3 p. 141–152.<br />

Sprang, S. R., 1997. G proteins, effectors and<br />

GAPs: structure and mechanism. Curr.<br />

Opin. Struct. Biol. v. 7 p. 849–856.<br />

Sprang, S. R., 1997. G protein mechanisms:<br />

Insights from structural analysis. Annu.<br />

Rev. Biochem. v. 66 p. 639–678.<br />

Research<br />

Benke, T., Motoshima, C. A., Fox, H., Le<br />

Trong, B. A., Teller, I., Okada, D. C.,<br />

Stenkamp, T., Yamamoto, R. E., and<br />

Miyano, M., 2000. Crystal structure <strong>of</strong><br />

rhodopsin: A G protein-coupled receptor.<br />

Science v. 289 p. 739–745.<br />

Chen, C. -K., Burns, M. E., He, W., Wensel, T.<br />

G., Baylor, D. A., and Simon, M. I., 2000.<br />

Slowed recovery <strong>of</strong> rod photoresponse in<br />

mice lacking the GTPase acceleration protein<br />

RGS9-1. Nature, v. 403 p. 557–560.<br />

Palczewski, K., Kumasaka, T., Hori, T.,<br />

Behnke , C. A., Motoshima, H., Fox, B.<br />

A., Le Trong, I., Teller, D. C., Okada, T.,<br />

Stenkamp, R. E., Yamamoto, M., and<br />

Miyano, M., 2000. Crystal structure <strong>of</strong><br />

rhodopsin: a G protein-coupled receptor.<br />

Science v. 289 p. 739–745.<br />

Wall, M. A., Coleman, D. E., Lee, E., Iniguez-<br />

Lluhi, J. A., Posner, B. A., Gilman, A. G.,<br />

and Sprang, S. R., 1995. The structure <strong>of</strong><br />

the G protein heterotrimer G i1<br />

1<br />

2<br />

. Cell<br />

v. 83 p. 1047–1058.<br />

14.23 Small, monomeric GTP-binding proteins<br />

are multiuse switches<br />

Review<br />

Bishop, A. L. and Hall, A., 2000. Rho GTPases<br />

and their effector proteins. Biochem. J. v.<br />

348 pt. 2 p. 241–255.<br />

Kuersten, S., Ohno, M., and Mattaj, I. W.,<br />

2001. Nucleocytoplasmic transport: Ran,<br />

beta and beyond. Trends Cell Biol. v. 11 p.<br />

497–503.<br />

Takai, Y., Sasaki, T., and Matozaki, T., 2001.<br />

Small GTP-binding proteins. Physiol. Rev.<br />

v. 81 p. 153–208.<br />

14.24 Protein phosphorylation/dephosphorylation<br />

is a major regulatory mechanism in<br />

the <strong>cell</strong><br />

Review<br />

Cohen, S., 1983. The epidermal growth factor<br />

(EGF). Cancer v. 51 p. 1787–1791.<br />

Fischer, E. H., 1997. “Protein phosphorylation<br />

and <strong>cell</strong>ular regulation, II,” in Nobel<br />

Lectures, Physiology or Medicine 1991–1995,<br />

N., Ringertz, Ed., Singapore: World<br />

Scientific Publishing Co.<br />

Krebs, G., 1993. Protein phosphorylation and<br />

<strong>cell</strong>ular regulation. Bioscience Reports v. 13<br />

p. 127–142.<br />

Manning, G., Whyte, D. B., Martinez, R.,<br />

Hunter, T., and Sudarsanam, S., 2002.<br />

The protein kinase complement <strong>of</strong> the<br />

human genome. Science v. 298 p.<br />

1912–1934.<br />

Newton, A. C., 2003. Regulation <strong>of</strong> the ABC<br />

kinases by phosphorylation: protein kinase<br />

C as a paradigm. Biochem. J. v. 370 p.<br />

361–371.<br />

Nolen, B., Taylor, S., and Ghosh, G., 2004.<br />

Regulation <strong>of</strong> protein kinases; controlling<br />

activity through activation segment conformation.<br />

Mol. Cell v. 15 p. 661–675.<br />

Turk, B. E. and Cantley, L. C., 2003. Peptide<br />

libraries: at the crossroads <strong>of</strong> proteomics<br />

and bioinformatics. Curr. Opin. Chem. Biol.<br />

v. 7 p. 84–90.<br />

Research<br />

Canagarajah, B. J., Khokhlatchev, A., Cobb,<br />

M. H., and Goldsmith, E. J., 1997.<br />

Activation mechanism <strong>of</strong> the MAP kinase<br />

ERK2 by dual phosphorylation. Cell v. 90<br />

p. 859–869.<br />

Ginty, D. D., Kornhauser, J. M., Thompson,<br />

640 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 641<br />

M. A., Bading, H., Mayo, K. E.,<br />

Takahashi, J. S., and Greenberg, M. E.,<br />

1993. Regulation <strong>of</strong> CREB phosphorylation<br />

in the suprachiasmatic nucleus by<br />

light and a circadian clock. Science v. 260<br />

p. 238–241.<br />

Knighton, D. R., Zheng, J. H., Ten Eyck, L. F.,<br />

Ashford, V. A., Xuong, N. H., Taylor, S. S.,<br />

and Sowadski, J. M., 1991. Crystal structure<br />

<strong>of</strong> the catalytic subunit <strong>of</strong> cyclic<br />

adenosine monophosphate-dependent<br />

protein kinase. Science v. 253 p. 407–414.<br />

Taylor, S. S., Radzio-Andzelm, E., and Hunter,<br />

T., 1995. How do protein kinases discriminate<br />

between serine/threonine and tyrosine?<br />

Structural insights from the insulin<br />

receptor protein-tyrosine kinase. FASEB<br />

J. v. 9 p. 1255–1266.<br />

Zhang, F., Strand, A., Robbins, D., Cobb, M.<br />

H., and Goldsmith, E. J., 1994. Atomic<br />

structure <strong>of</strong> the MAP kinase ERK2 at 2.3<br />

Å resolution. Nature v. 367 p. 704–711.<br />

14.25 Two-component protein phosphorylation<br />

systems are <strong>signaling</strong> relays<br />

Review<br />

Hoch, J. A, and Silhavy, T. J., eds., 1995. Twocomponent<br />

signal transduction.<br />

Washington, D. C.:American Society for<br />

Microbiology.<br />

Stock, A. M., Robinson, V. L., and Goudreau,<br />

P. N., 2000. Two-component signal transduction.<br />

Annu. Rev. Biochem. v. 69 p.<br />

183–215.<br />

14.26 Pharmacological inhibitors <strong>of</strong> protein kinases<br />

may be used to understand and<br />

treat disease<br />

Review<br />

Blume-Jensen, P, and Hunter, T., 2001.<br />

Oncogenic kinase signalling. Nature v.<br />

411 p. 355–365.<br />

Cherry, M. and Williams, D. H., 2004. Recent<br />

kinase and kinase inhibitor X-ray structures:<br />

mechanisms <strong>of</strong> inhibition and selectivity<br />

insights. Curr. Med. Chem. v. 11 p.<br />

663–673.<br />

Cohen, P., 2002. Protein kinases—the major<br />

drug targets <strong>of</strong> the twenty-first century?<br />

Nat. Rev. Drug Discov. v. 1 p. 309–315.<br />

Davies, S. P., Reddy, H., Caivano, M., and<br />

Cohen, P., 2000. Specificity and mechanism<br />

<strong>of</strong> action <strong>of</strong> some commonly used<br />

protein kinase inhibitors. Biochem. J. v.<br />

351 p. 95–105.<br />

Tibes, R., Trent, J., and Kurzrock, R., 2005.<br />

Tyrosine kinase inhibitors and the dawn<br />

<strong>of</strong> molecular cancer therapeutics. Annu.<br />

Rev. Pharmacol. Toxicol. v. 45 p. 357–384.<br />

14.27 Phosphoprotein phosphatases reverse the<br />

actions <strong>of</strong> kinases and are independently<br />

regulated<br />

Review<br />

Aramburu, J., Heitman, J., and Crabtree, G.<br />

R., 2004. Calcineurin: a central controller<br />

<strong>of</strong> signalling in eukaryotes. EMBO Rep. v.<br />

5 p. 343–348.<br />

Dounay, A. B. and Forsyth, C. J., 2002.<br />

Okadaic acid: the archetypal serine/threonine<br />

protein phosphatase inhibitor. Curr.<br />

Med. Chem. v. 9 p. 1939–1980.<br />

Neel, B. G., Gu, H., and Pao, L., 2003. The<br />

‘Shp’ing news: SH2 domain-containing<br />

tyrosine phosphatases in <strong>cell</strong> <strong>signaling</strong>.<br />

Trends Biochem. Sci. v. 28 p. 284–293.<br />

Neely, K. E. and Piwnica-Worms, H., 2003.<br />

Cdc25A regulation: to destroy or not to<br />

destroy—is that the only question? Cell<br />

Cycle v. 2 p. 455–457.<br />

Olson, E. N. and Williams, R. S., 2000.<br />

Remodeling muscles with calcineurin.<br />

Bioessays v. 22 p. 510–519.<br />

Virshup, D. M., 2000. Protein phosphatase<br />

2A: a panoply <strong>of</strong> enzymes. Curr. Opin. Cell<br />

Biol. v. 12 p. 180–185.<br />

Research<br />

Sun, H., et al. 1993. MKP-1 (3CH134), an immediate<br />

early gene product, is a dual<br />

specificity phosphatase that dephosphorylates<br />

MAP kinase in vivo Cell v. 75 p.<br />

487–493.<br />

Terrak, M., Kerff, F., Langsetmo, K., Tao, T.,<br />

and Dominguez, R., 2004. Structural basis<br />

<strong>of</strong> protein phosphatase 1 regulation.<br />

Nature v. 429 p. 780–784.<br />

14.28 Covalent modification by ubiquitin and<br />

ubiquitin-like proteins is another way <strong>of</strong><br />

regulating protein function<br />

Review<br />

Gill, G., 2004. SUMO and ubiquitin in the nucleus:<br />

different functions, similar mechanisms?<br />

Genes Dev. v. 18 p. 2046–2059.<br />

Pickart, C. M. and Eddins, M. J., 2004.<br />

Ubiquitin: structures, functions, mechanisms.<br />

Biochim. Biophys. Acta v. 1695 p.<br />

55–72.<br />

Research<br />

Dharmasiri, N., Dharmasiri, S., and Estelle,<br />

M., 2005. The F-box protein TIR1 is an<br />

auxin receptor. Nature v. 435 p. 441–445.<br />

Kanayama, A. et al. 2004. TAB2 and TAB3<br />

activate the NF-B pathway through<br />

binding to polyubiquitin chains. Mol. Cell<br />

v. 15 p. 535–548<br />

Kepinski, S. and Leyser, O., 2005. The<br />

Arabidopsis F-box protein TIR1 is an<br />

auxin receptor. Nature v. 435 p. 446–451.<br />

References 641


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 642<br />

14.29 The Wnt pathway regulates <strong>cell</strong> fate during<br />

development and other processes in<br />

the adult<br />

Review<br />

Logan, C. Y. and Nusse, R., 2004. The WNT<br />

<strong>signaling</strong> pathway in development and<br />

disease. Annu. Rev. Cell. Dev. Biol. v. 20, p.<br />

781–810.<br />

Tolwinski, N. S., Wieschaus, E., 2004.<br />

Rethinking WNT <strong>signaling</strong>. Trends Genet.<br />

v. 20 p. 177–81.<br />

14.30 Diverse <strong>signaling</strong> mechanisms are regulated<br />

by protein tyrosine kinases<br />

Review<br />

Birge, R. B., Knudsen, B. S., Besser, D., and<br />

Hanafusa, H., 1996. SH2 and SH3-containing<br />

adaptor proteins: redundant or<br />

independent mediators <strong>of</strong> intra<strong>cell</strong>ular<br />

signal transduction. Genes Cells v. 1 p.<br />

595–613.<br />

Blume-Jensen, P. and Hunter, T., 2001.<br />

Oncogenic kinase signalling. Nature v.<br />

411 p. 355–365.<br />

Cohen, P., 2002. Protein kinases—the major<br />

drug targets <strong>of</strong> the twenty-first century?<br />

Nat. Rev. Drug Discov. v. 1 p. 309–315.<br />

Pawson, T. and Nash, P., 2003. Assembly <strong>of</strong><br />

<strong>cell</strong> regulatory systems through protein<br />

interaction domains. Science v. 300 p.<br />

445–452.<br />

Tallquist, M. and Kazlauskas, A., 2004. PDGF<br />

<strong>signaling</strong> in <strong>cell</strong>s and mice. Cytokine<br />

Growth Factor Rev. v. 15 p. 205–213.<br />

Taniguchi, C. M., Emanuelli, B., and Kahn, C.<br />

R., 2006. Critical nodes in signalling<br />

pathways: insights into insulin action.<br />

Nat. Rev. Mol. Cell Biol. v. 7, p. 85–96.<br />

14.31 Src family protein kinases cooperate with<br />

receptor protein tyrosine kinases<br />

Review<br />

Boggon, T. J. and Eck, M. J., 2004. Structure<br />

and regulation <strong>of</strong> Src family kinases.<br />

Oncogene v. 23 p. 7918–7927.<br />

Frame, M. C., 2004. Newest findings on the<br />

oldest oncogene; how activated src does<br />

it. J. Cell Sci. v. 117 p. 989–998.<br />

Research<br />

Cowan-Jacob, S. W., Frendrich, G., Manley, P.<br />

W., Jahnke, W., Fabbro, D.,<br />

Liebetanz, J., Meyer, T. 2005. The<br />

crystal structure <strong>of</strong> a c-Src complex in<br />

an active conformation suggests possible<br />

steps in c-Src activation Structure<br />

v. 13 p. 861–871.<br />

Xu, W., Harrison, S. C., and Eck, M. J., 1997.<br />

Three-dimensional structure <strong>of</strong> the tyro-<br />

sine kinase c-Src. Nature v. 385 p.<br />

595–602.<br />

14.32 MAPKs are central to many <strong>signaling</strong><br />

pathways<br />

Review<br />

Chen, Z., Gibson, T. B., Robinson, F.,<br />

Silvestro, L., Pearson, G., Xu, B., Wright,<br />

A., Vanderbilt, C., and Cobb, M. H., 2001.<br />

MAP kinases. Chem. Rev. v. 101 p.<br />

2449–2476.<br />

Lewis, T. S., Shapiro, P. S., and Ahn, N. G.,<br />

1998. Signal transduction through MAP<br />

kinase cascades. Adv. Cancer Res. v. 74 p.<br />

49–139.<br />

Morrison, D. K. and Davis, R. J., 2003.<br />

Regulation <strong>of</strong> MAP kinase <strong>signaling</strong> modules<br />

by scaffold proteins in mammals.<br />

Annu. Rev. Cell Dev. Biol. v. 19 p. 91–118.<br />

Sharrocks, A. D., Yang, S. H., and Galanis, A.,<br />

2000. Docking domains and substratespecificity<br />

determination for MAP kinases.<br />

Trends Biochem. Sci. v. 25 p.<br />

448–453.<br />

Research<br />

Anderson, N. G., Maller, J. L., Tonks, N. K.,<br />

and Sturgill, T. W., 1990. Requirement<br />

for integration <strong>of</strong> signals from two distinct<br />

phosphorylation pathways for activation<br />

<strong>of</strong> MAP kinase. Nature v. 343 p.<br />

651–653.<br />

14.33 Cyclin-dependent protein kinases control<br />

the <strong>cell</strong> cycle<br />

Review<br />

Dorée, M. and Hunt, T., 2002. From Cdc2 to<br />

Cdk1: when did the <strong>cell</strong> cycle kinase join<br />

its cyclin partner? J. Cell Sci. v. 115 p.<br />

2461–2464.<br />

Hartwell, L. H., 1991. Twenty-five years <strong>of</strong><br />

<strong>cell</strong> cycle genetics. Genetics v. 129 p.<br />

975–980.<br />

Neely, K. E. and Piwnica-Worms, H., 2003.<br />

Cdc25A regulation: to destroy or not to<br />

destroy—is that the only question? Cell<br />

Cycle v. 2 p. 455–457.<br />

Nurse, P., 2000. A long twentieth century <strong>of</strong><br />

the <strong>cell</strong> cycle and beyond. Cell v. 100 p.<br />

71–78.<br />

Research<br />

Pavletich, N. P., 1999. Mechanisms <strong>of</strong> cyclindependent<br />

kinase regulation: structures<br />

<strong>of</strong> Cdks, their cyclin activators, and Cip<br />

and INK4 inhibitors. J. Mol. Biol. v. 287 p.<br />

821–828.<br />

Russo, A. A., Jeffrey, P. D., and Pavletich, N.<br />

P., 1996. Structural basis <strong>of</strong> cyclin-dependent<br />

kinase activation by phosphorylation.<br />

Nat. Struct. Biol. v. 3 p. 696–700.<br />

642 CHAPTER 14 <strong>Principles</strong> <strong>of</strong> <strong>cell</strong> <strong>signaling</strong>


39057_ch14_<strong>cell</strong>bio.qxd 8/28/06 5:11 PM Page 643<br />

14.34 Diverse receptors recruit protein tyrosine<br />

kinases to the plasma membrane<br />

Review<br />

Ernst, M. and Jenkins, B. J., 2004. Acquiring<br />

signalling specificity from the cytokine<br />

receptor gp130. Trends Genet. v. 20 p.<br />

23–32.<br />

Herrington, J. and Carter-Su, C., 2001.<br />

Signaling pathways activated by the<br />

growth hormone receptor. Trends<br />

Endocrinol. Metab. v. 12 p. 252–257.<br />

Levy, D. E. and Darnell, J. E., 2002. Stats:<br />

transcriptional control and biological impact.<br />

Nat. Rev. Mol. Cell Biol. v. 3 p.<br />

651–662.<br />

Miranti, C. K. and Brugge, J. S., 2002.<br />

Sensing the environment: a historical<br />

perspective on integrin signal transduction.<br />

Nat. Cell Biol. v. 4 p. E83–E90.<br />

Palacios, E. H. and Weiss, A., 2004. Function<br />

<strong>of</strong> the Src-family kinases, Lck and Fyn, in<br />

T-<strong>cell</strong> development and activation.<br />

Oncogene v. 23 p. 7990–8000.<br />

Research<br />

Clackson, T. and Wells, J. A., 1995. A hot spot<br />

<strong>of</strong> binding energy in a hormone-receptor<br />

interface. Science v. 267 p. 383–386.<br />

References 643

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!