Lennarz W.J., Lane M.D. (eds.) Encyclopedia of Biological Chemistry. Four-Volume Set . V. 4

Подождите немного. Документ загружается.

SUMO Activating Enzyme (E1 Enzyme)

SUMO E1 activating enzyme is composed of two

subunits, Aos1 and Uba2 (for alternative names see

Table I), both of which are essential for the function.

Aos1 and Uba2 share striking homology to the N- and

C- terminal halves of ubiquitin E1 enzyme, respectively.

SUMO E1 activating enzyme is present throughout the

cell, with a strong enrichment in the nucleoplasm. A

single SUMO E1 activating enzyme is expressed in most

organisms, even those containing several distinct SUMO

proteins. Current data suggest that mammalian SUMO

E1 works with equal efﬁciency on each of the three

mammalian SUMO proteins, although it has no afﬁnity

for other ubiquitin-related proteins.

SUMO Conjugating Enyzme (E2 Enzyme)

A single SUMO E2 conjugating enzyme, Ubc9, is

expressed in cells (see Table I for alternative names).

While it is strikingly similar to ubiquitin conjugating

enzymes, it works exclusively on SUMO proteins. As for

the E1 enzyme, Ubc9 does not appear to discriminate

between different SUMO proteins expressed within a

particular organism. Ubc9 localizes predominantly in

the nucleus, but is also present in the cytoplasm and at

nuclear pore complexes. A striking difference between

the ubiquitin-conjugating and the SUMO-conjugating

system is the ability of Ubc9 to directly bind to and

modify SUMO targets. However, with a few exceptions,

the efﬁciency of modiﬁcation in the absence of E3 ligases

is extremely poor.

SUMO Ligases (E3 Ligase)

Three types of SUMO E3 ligases are currently

known (Table I). The ﬁrst type is encoded by the

TABLE I

Nomenclature for SUMO and its Conjugating Enzymes

Human Baker’s yeast

SUMOs SUMO1/PIC 1/Ubl1/sentrin1/

GMP1/hSMT3

Smt3

SUMO2/sentrin3/Smt3a

SUMO3/sentrin2/Smt3b

E1 activating enzyme

Subunit 1 Aos1/SAE1/Sua Aos1

Subunit 2 Uba2/SAE2 Uba2

E2 conjugating enzyme Ubc9/UBE2I Ubc9

E3 ligating enzymes Pias1/GBP/DEAD/H box-

binding protein 1/ARIP

Siz1/UII 1

Pias3 Siz2/Nﬁ1

Piasx

/ARIP3

Piasx

/Miz1/Siz2

Piasy

RanBP2/Nup358

Pc2

Known proteins involved in the SUMO conjugation pathway. Since

most of them were independently identiﬁed by several groups, different

names exist in literature.

FIGURE 1 Ubiquitin and SUMO comparison. Although ubiquitin and SUMO1 share only 18% amino acid identity, the structural fold is very

similar. The indicated C-terminal double glycine motif is the attachment site to substrates. The images (generated using Moscript 2.1.2 and

Raster3D) were kindly provided by Dr. Peter Bayer.

SUMO MODIFICATION 131

“Protein inhibitor of activated STAT” (PIAS) family.

These proteins contain a predicted RING ﬁnger-like

structure that is essential for their function as E3 ligases.

They bind directly to Ubc9 and to selected SUMO

targets, and stimulate their modiﬁcation both in vivo

and in vitro. Baker’s yeast contains two PIAS type E3

ligases, Siz1 and Siz2. These proteins seem to be

responsible for most, but probably not for all, SUMO

targets, and have partially overlapping target speciﬁcity.

The second type is currently represented by a single

(vertebrate speciﬁc) 358 kDa protein, RanBP2/Nup358

(Ran binding protein/nucleoporin). RanBP2 is a com-

ponent of nuclear pore complexes, and thus may

couple sumoylation and nucleocytoplasmic transport.

Interestingly, the catalytic domain of RanBP2 does not

contain a RING ﬁnger motif, and also shows no other

homology to ubiquitin E3 ligases. The Polycomb group

protein Pc2 appears to represent a third type of an E3

ligase, as it bears no similarity to other E3 ligases.

DECONJUGATION

Overview

Isopeptide bonds between ubiquitin-related proteins and

their cellular targets are reversible in nature due to the

presence of speciﬁc enzymes that recognize and cleave

the bond. Currently, a single family of SUMO-speciﬁc

isopeptidases is known (Table II). All members contain a

200 amino acid catalytic core that is distantly related to

viral cysteine proteases but shows no similarity to

ubiquitin-speciﬁc enzymes.

Interestingly, SUMO isopeptidases serve two physio-

logical functions in vivo, cleavage of isopeptide bonds,

but also maturation of nascent SUMO (C-terminal

hydrolase activity). This distinguishes the SUMO system

from the ubiquitin system, where two distinct enzyme

families serve to fulﬁll these functions.

SUMO Isopeptidases

The founding member of this family is S. cerevisiae

Ulp1 (ubiquitin like protease). Bakers yeast contains

FIGURE 2 The SUMO modiﬁcation pathway. SUMO is synthesized as a precursor that requires proteolytic processing. A C-terminal

hydrolase/isopeptidase removes several C-terminal amino acids to expose the GlyGly motif essential for conjugation. SUMO conjugation involves

an enzymatic cascade: In an ATP-dependent step, mature SUMO ﬁrst forms a thioester with the E1 activating enzyme, the heterodimer Aos1/Uba2.

Subsequently, it is transferred to the E2 conjugating enzyme, Ubc9. SUMO is then either directly or in an E3 ligase-dependent step conjugated to the

substrate. SUMO de-modiﬁcation is performed by isopeptidases.

TABLE II

Nomenclature for SUMO Deconjugating Enzymes

Human Mouse Baker’s yeast

Senp1 SuPr-2 Ulp1

Senp2/KIAA1331 Senp2/Smt3IP2

/Axam2

and Supr-1

Ulp2/Smt4

Senp3/Susp3/SSP3 Senp3/Smt3IP1/SuPr-3

Senp5

Senp6/Susp1/SSP1/KIAA0797

Senp7/Susp2/SSP2/KIAA1707

SUMO isopeptidases share a 200 amino acid core domain. Known

proteins containing this core domain are listed in this table. The

proteins characterized as SUMO isopeptidases are indicated in bold.

Indicates different isoforms or alternative splice products.

132 SUMO MODIFICATION

two enzymes of this family, Ulp1 and Ulp2. Ulp1 is

required for viability in yeast, whereas Ulp2 is not

essential. They localize to different intracellular

compartments, with Ulp1 being enriched at nuclear

pore complexes, and Ulp2 being diffusely distributed

throughout the nucleus. Mammals on the other hand

have at least six distinct genes for SUMO isopeptidases

(see Table II). Due to alternative splicing, the number of

enzymes expressed is probably much larger. Only a few

of these proteins have been characterized in some detail.

They vary signiﬁcantly in size and intracellular localiz-

ation. It is plausible to assume that different isopepti-

dases exert different target speciﬁcity in large part

through their physical localization. Whether they also

differ in enzymatic properties, such as substrate speci-

ﬁcity or kinetics of cleavage, remains to be seen.

SUMO Target Proteins

Sumoylation as a means to regulate protein function

appears to be quite a common mechanism. Modiﬁcation

has been documented for more than 70 different target

proteins, and this number is expected to increase

signiﬁcantly. Based on current knowledge, some gener-

alizations can be made about the nature of the targets,

motifs required for modiﬁcation, consequences of

modiﬁcation, and regulation.

KNOWN TARGET PROTEINS

Based on immuno-ﬂuorescence analysis, most targets for

sumoylation are constitutively or transiently associated

with the nuclear compartment. Consistent with this,

most known targets can be associated with nuclear

processes such as chromatin remodeling, DNA repair,

transcription, or nucleocytoplasmic transport (amongst

them are histone deacetylases, topoisomerases, thymine-

DNA glycosylase, PCNA, p53, PML, heat shock

factors, steroid hormone receptors, I

, RanGAP1,

and many others). Other pathways to which SUMO has

been linked are, e.g., progression through mitosis

(mitotic arrest of yeast mutants defective in SUMO

conjugation or deconjugation; yeast septins, topoisome-

rase II and Pds5 are modiﬁed speciﬁcally in mitosis),

or infection with viruses (examples for viral SUMO

targets are: cytomegalovirus immediate early proteins

IE1 and IE2, adenovirus type 5 E1B-55 kDa, bovine

papillomavirus E1).

CONSENSUS SITES FOR MODIFICATION

In contrast to ubiquitinylation, for which a unifying

motif has not been identiﬁed, sumoylation of most

targets seems to be speciﬁed by a short consensus

sequence in target proteins. This motif consists of

just four amino acids, CK £ E/D, and includes the

lysine residue that forms an isopeptide bond with

SUMO. C stands for a bulky aliphatic amino acid

residue. Additional structural features are probably

required to ensure accessibility to the conjugation

machinery. Some motifs are present at the very N- or

C-terminal end of a protein, others are ﬂanked by

proline residues that may induce a loop structure.

Consistent with this, RanGAP1, which is an extremely

efﬁcient SUMO target, presents its sumoylation motif

in an accessible loop.

FUNCTIONAL CONSEQUENCES

FOR

MODIFICATION

Similar to phosphorylation, sumoylation seems to have

many different functional consequences that depend on

the speciﬁc target protein. Considering SUMO’s size, it is

plausible to assume that conjugation can lead to

masking of binding sites, generation of novel binding

interphases, or conformational changes in the modiﬁed

protein. Examples have been reported for changes in

protein–protein or protein –DNA interactions, altera-

tion in subcellular localization, enhanced stability

through antagonizing ubiquitin/proteasome-mediated

degradation, and changes in enzymatic activity.

REGULATION

While some SUMO targets appear to be modiﬁed

constitutively, others are sumoylated only during a

speciﬁc period of the cell cycle, upon stress, or upon a

speciﬁc extracellular signal. Examples are sumoylation

of yeast septins during mitosis, sumoylation of topoi-

somerase upon treatment with DNA-damaging agents,

or Dictyostelium MEK1 sumoylation in response to

chemoattractant. While this suggests the existence of

elaborate regulatory mechanisms, current knowledge is

rather poor. Increased modiﬁcation of a speciﬁc target

may for example be due to changes in the target,

activation of a speciﬁc E3 ligase, or relocalization of a

speciﬁc isopeptidase. Evidence for cell-cycle-dependent

regulation of E3 ligases and isopeptidases is currently

only available in yeast: (1) ﬁssion yeast ulp1 resides at

the NPC during interphase, but is nuclear during

mitosis; (2) baker’s yeast E3 ligase Siz1 is intranuclear

during interphase, but partially relocalizes to the bud

neck in mitosis.

FURTHER READING

Bernier-Villamor, V., Sampson, D. A., Matunis, M. J., and Lima, C. D.

(2002). Structural basis for E2-mediated SUMO conjugation

revealed by a complex between ubiquitin-conjugating enzyme

Ubc9 and RanGAP1. Cell 108, 345–356.

Freiman, R. N., and Tjian, R. (2003). Regulating the regulators: Lysine

modiﬁcations make their mark. Cell 112, 1–17.

Goettsch, S., and Bayer, P. (2002). Structural attributes in the

conjugation of ubiquitin, SUMO and RUB to protein substrates.

Front Biosci. 7, a148–a162.

Johnson, E. S. (2002). Ubiquitin branches out. Nat. Cell Biol. 4,

E295–E298.

Kim, K. I., Baek, S. H., and Chung, C. H. (2002). Versatile protein tag,

SUMO: Its enzymology and biological function. J. Cell Physiol.

191, 257– 268.

Kurepa, J., Walker, J. M., Smalle, J., Gosink, M. M., Davis, S. J.,

Durham, T. L., Sung, D. Y., and Vierstra, R. D. (2003). The small

ubiquitin-like modiﬁer (SUMO) protein modiﬁcation system in

Arabidopsis. Accumulation of SUMO1 and -2 conjugates is

increased by stress. J. Biol. Chem. 278, 6862– 6872.

Melchior, F. (2000). SUMO – nonclassical ubiquitin. Annu. Rev. Cell

Develop. Biol. 16, 591– 626.

Melchior, F., Schergaut, M., and Pichler, A. (2003). SUMO: ligases,

isopeptidases and nuclear pores. Trends Biochem. Sci. 28,

612–618.

Muller, S., Hoege, C., Pyrowolakis, G., and Jentsch, S. (2001). SUMO,

ubiquitin’s mysterious cousin. Nat. Rev. Mol. Cell Biol. 2, 202–210.

Schwartz, D. C., and Hochstrasser, M. (2003). A superfamily of

protein tags: ubiquitin, SUMO and related modiﬁers. Trends

Biochem. Sci. 28, 321– 328.

Seeler, J. S., and Dejean, A. (2003). Nuclear and unclear functions of

SUMO. Nat. Rev. Mol. Cell Biol. 4, 690–699.

Wilson, V. G., and Rangasamy, D. (2001). Viral interaction with the

host cell sumoylation system. Virus Res. 81, 17–27.

BIOGRAPHY

Frauke Melchior is a group leader at the Max-Planck Institute for

Biochemistry in Martinsried, Germany. Her principal research interests

are nucleocytoplasmic transport, and ubiquitin-related proteins of the

SUMO family. She holds a Ph.D. from the University of Marburg and

received her postdoctoral training at the Max-Planck Institute for

Biophysical Chemistry in Go

ttingen and the Scripps Research Institute

in La Jolla. She participated in the discovery of SUMO during her

postdoctoral time, and has since contributed to the ﬁeld with her own

group.

Andrea Pichler obtained her Ph.D. from the University of Vienna and

received a postdoctoral training at the Novartis Forschungsinstitut of

Vienna. Since 2000 she is a member of Dr. Frauke Melchior’s group at

the Max-Planck Institute of Biochemistry in Martinsried.

134 SUMO MODIFICATION

Superoxide Dismutase

Irwin Fridovich

Duke University Medical Center, Durham, North Carolina, USA

A small fraction of the O

consumed by living cells is

converted to superoxide O

. This free radical, and its

progeny, can damage a variety of biomolecules. This is the

cause of oxidation stress. Defenses are provided by the

superoxide dismutases that catalytically eliminate O

and by

the catalases and peroxidases that do the same for hydroper-

oxides. The varieties, distributions, and mode of action of the

superoxide dismutases are presented herein.

Introduction

Molecular oxygen (O

), while essential for the life of

aerobes, is potentially toxic. This toxicity is due to the

propensity of O

for reduction by a univalent pathway.

This facile univalent pathway of O

reduction generates

intermediates that lie between one O

and its four

electron reduction products – two molecules of water –

and it is the reactivity of these intermediates that is

responsible for the toxicity of O

. In order of their

production, these intermediates are the superoxide

anion radical (O

), hydrogen peroxide (H

), and

the hydroxyl radical (HO·). Superoxide dismutases

(SODs) catalyze the conversion of O

into H

plus

, i.e.,

þ O

þ 2H

SOD

þ O

and in so doing provide an important defense. H

,in

turn, is eliminated by the catalases and peroxidases. The

concerted action of the SODs with the catalases and

peroxidases prevents the formation of the very reactive

HO·. This is the case because HO· production in vivo

requires both O

and H

oxidizes the [4Fe–4S]

clusters of dehydratases, such as the aconitases,

causing release of Fe (II); the freed Fe (II) then reduces

to OH

plus HO· in a reaction known as the

Fenton reaction.

The SOD Family of Enzymes

The ﬁrst truly photosynthetic organisms were the

cyanobacteria and the oxygen they produced oxyge-

nated a previously anaerobic biosphere. This imposed a

common selection pressure on what must have been a

varied anaerobic biota. It is not surprising that several

different SODs were then evolved to deal with the

toxicity of the accumulating O

and that these persist to

this day. Thus, there are SODs based on Cu (II) plus

Zn (II), Mn (III), Fe (III), and Ni (II). They all catalyze

the dismutation of O

into H

þ O

All SODs work by a similar mechanism in which the

metal at the active site is reduced by one O

and then

reoxidized by the next O

. The active site metal thus

acts as a mediator passing an electron from one O

the next. In this way the electrostatic repulsion, which

would prevent close approach of one O

to another, is

bypassed by the SODs. All the SODs are very efﬁcient

catalysts and operate at close to the theoretical diffusion

limit. Their rate constants for interaction with O

are

, 3 £ 10

THE CU, ZN SODS

These SODs have been found in the cytosols of

eukaryotic cells, the intermembrane space of mitochon-

dria, the periplasm of gram-negative bacteria, and

in chloroplasts. The Cu (II) is linked to the Zn (II) at

the active site by the imidazolate moiety of a histidine

residue and this bridging imidazolate functions as a

proton supply during the catalytic cycle. Thus, when

the Cu (II)) is reduced by O

, the Cu (I) releases the

imidazolate that then protonates to imidazole. When

the Cu (I) is reoxidized to Cu (II), the imidazole

provides a proton to the nascent O

converting it

to HO

, as the bridging to the Zn (II) is reestablished.

The HO

that leaves the active site protonates in the

water to H

. This mechanism can be depicted

as follows:

Both of these half-reactions of the catalytic cycle proceed

with a rate constant of , 3 £ 10

. The enzyme

surface contains charged groups that provide an

electrostatic funnel guiding the O

toward the active

site crevice. This electrostatic facilitation contributes to

the efﬁciency of these enzymes.

The Cu, Zn SODs found in eukaryotic cytosols are

homodimeric proteins with subunit weights , 16 kDa.

The corresponding enzyme from the periplasm of

Escherichia coli is monomeric. There is an extracellular

Cu, Zn SOD found in higher animals. It is usually

homotetrameric and glycosylated and has a subunit

weight of , 23 kDa.

THE MN SODS

These enzymes, whether from the matrix of mitochon-

dria or bacteria, exhibit marked sequence similarity,

reﬂecting a close evolutionary history and supporting

an endosymbiotic origin for mitochondria. The bac-

terial enzyme is usually a homodimer, whereas the

mitochondrial enzyme is a homotetramer. The subunit

weight is , 23 kDa. Some bacterial Mn SODs are

tetrameric. This is the case in Cryptococcus neofor-

mans. In E. coli the biosynthesis of Mn SOD is under

the control of the soxRS regulon, which coordinately

up-regulates the expression of a number of genes in

response to O

. The constitutively expressed SOX R

protein is transcriptionally inactive in its reduced form.

It can be oxidized by O

and then activates the

expression of the SOX S protein, which in turn

activates all the genes in the regulon. Thus, Mn SOD

is not measurable in extracts of anaerobically grown

E. coli, but exposure of cultures to aerobic conditions

elicits production of Mn SOD. Increasing production

of O

, by raising pO

, or by adding compounds such as

viologens, which can mediate enhanced production of

, increases the level of Mn SOD. It has been

possible to force E. coli to produce Mn SOD to 7%

of its soluble protein by aerobic exposure to the

viologen paraquat.

The nectar of tobacco ﬂowers has been found to

contain a stable Mn-protein named nectarin that

appears to be an Mn SOD.

THE FE SODS

These SODs, which are highly homologous to the

Mn SODs, are found in bacteria and in plants. Although

usually homodimeric, homotetrameric Fe SOD has been

found in Rhodococcus bronchialis and Mycobacterum

tuberculosis. The Fe SOD of E. coli is constitutive and

thus is found even in anaerobically grown cells. It can

thus be viewed as a standby defense against O

, which is

always maintained to protect in the event of a sudden

exposure to O

It is possible to reversibly remove the metals from

SODs. The apo enzymes so prepared are inactive but

can be reactivated by restoring the active site metal. It

is striking, that in spite of sequence and structural

homologies, the Fe SOD is inactive when made to

contain Mn in place of Fe and the Mn SOD is inactive

when similarly reconstituted with the non-native metal

Fe. There are, however, SODs that are active with

either Mn or Fe at their active sites. These SODs,

termed cambialistic SODs, are usually found in anae-

robes such as Propionibacterium shermanii or Strepto-

coccus mutans. These organisms produce the SOD

with iron when grown anaerobically but, when

exposed to low levels of O

, produce the same SOD

but with Mn. It is clear that even anaerobes have

need of a defense against the O

during transient

exposure to O

. Ferrous salts are soluble, and thus

available anaerobically, but become insoluble when

oxidized to the ferric state by O

. Manganous salts

remain soluble under both conditions – hence the

wisdom of a cambialistic SOD for an anaerobe. An

iron SOD has also been reported in the protozoan

Tetrahymena pyriformis.

Deletions and Consequences

E.COLI

Mutants of E. coli lacking both the Mn SOD and

the Fe SOD (SodA SodB) have been prepared.

Anaerobically they grow as well as the parental strain.

However, aerobically they grow slowly, exhibit a

high rate of mutagenesis, and require branched-chain,

N Zn(II)

–

Cu(I)

N Zn(II)

Cu(I)

N Zn(II)

Cu(II) N

N Zn(II)

Cu(II)

–

136 SUPEROXIDE DISMUTASE

sulfur-containing, and aromatic amino acids. The null

mutants are also hypersensitive towards paraquat. All

of these phenotypic deﬁcits disappear when a functional

gene encoding any active SOD is inserted. The slow

growth of the null mutant can be understood in terms

of the energy devoted to repairing or replacing

oxidatively damaged molecules and to the inactivation

of such O

sensitive enzymes as aconitase. The high rate

of O

-dependent mutagenesis reﬂects oxidative DNA

damage, and repair, which is error-prone. The nutri-

tional auxotrophies exhibited by the null mutant have

several explanations. The need for branched amino

acids arises because the penultimate step in the

biosynthesis of these amino acids is catalyzed by a

dihydroxy acid dehydratase that contains a [4Fe –4S]

cluster that is rapidly oxidized, and inactivated, by O

The need for sulfur-containing amino acids evidently

reﬂects leakiness of the cell envelope towards sulﬁte,

which is an intermediate on the pathway from sulfate

to cysteine. Finally, the aromatic amino acid auxotro-

phy reﬂects the inability to make erythrose-4-phos-

phate, which is one of the starting materials on the

pathway leading to these amino acids. Erythrose-4-

phosphate, in turn, depends on the sequential actions of

transketolase and transaldolase, and the transketolase

intermediate 1, 2-dihydroxyethylthiamine pyropho-

sphate is rapidly oxidized by O

. The varied expla-

nations for these phenotypic deﬁcits reﬂect the variety

of damage that can be caused by O

and the multiple

targets that are protected by the SODs.

YEAST

Mutants of the yeast Saccharomyces cerevisia lacking

either the cytosolic Cu, Zn SOD or the mitochondrial

Mn SOD have been prepared. Their phenotypic deﬁcits,

which are all oxygen-dependent, include slow growth,

hypersensitivity towards paraquat or quinones, vacuolar

fragmentation, sensitivity towards oxygen, and, in the

case of the Cu, Zn SOD-null, lysine auxotrophy. All of

these problems can be explained on the basis of O

damage, as was the case for the E. coli SOD-null mutants.

MICE

Murine mutants lacking Mn SOD are severely affected

and those that survive to term are low-birth-weight and

only survive for 4–14 days. Their problems can be seen

as a deﬁcit of mitochondrial functions and that in turn

can be explained on the basis of damage to mitochon-

drial components by O

. The heterozygotes that have

half the normal level of Mn SOD are viable but have

been shown to be hypersensitive towards oxidative

inactivation of aconitase and towards release of

cytochrome c from mitochondria when challenged

with tertiary butyl hydroperoxide. The deﬁcit imposed

by halving the Mn SOD was also seen as a greater

susceptibility to reperfusion injury.

In contrast to the embryonic and neonatal fatality

imposed by Mn SOD deletion, the homozygous Cu, Zn

SOD-null mice were apparently normal under labora-

tory conditions. They are reported to be relatively

intolerant of neuronal injury and prone to develop

hepatoma at . 9 months of age. It seems likely that

there is some redundancy in scavenging O

in the

cytosol, so that the lack of Cu, Zn SOD may be

compensated by up-regulation of another O

scaven-

ger, perhaps a superoxide reductase, such as has been

found in anaerobes.

SOD and Amyotrophic

Lateral Sclerosis

ALS or Lou Gherig’s disease is a late onset, progressive,

and fatal paralytic disease due to death of motor

neurons. There are sporadic and familial types of ALS

and these are not distinguishable on the basis of clinical

symptoms. Approximately 20% of the familial cases of

ALS have been associated with point mutations causing

single amino acid replacements in Cu, Zn SOD.

Transgenic mice expressing the wild-type human Cu,

Zn SOD are normal, but those expressing one of the

ALS-associated mutant Cu, Zn SODs develop progress-

ive paralysis at , 3 months of age. Thus, the mutations

in the SOD cause the death of motor neurons because

of some as yet unknown gain of function, rather than

because of a loss of SOD activity. This is certainly the

case because: most of the mutant SODs retain full SOD

activity; the transgenic mice retain the active murine

Cu, Zn SOD; the disease is genetically dominant, i.e.,

heterozygotes develop the disease; and ﬁnally, Cu, Zn

SOD knockout mice do not become paralyzed. The

transgenic mice provide a useful model of the human

disease and are being used to explore both the toxic

gain of function of the mutant Cu, Zn SOD and

treatment modalities.

Mimics

has been found to play causative roles in numerous

inﬂammatory diseases and reperfusion injuries. For

this reason non-enzymic catalysts of the dismutation

of O

are being explored as pharmaceutical agents and

appear to have promise.

FURTHER READING

Hassan, H. M. (1997). Cytotoxicity by oxyRadicals and the evolution

of superoxide dismutases. In Lung Biology in Health and Disease

(L. B. Clerch and D. J. Massaro, eds.) Vol 105, pp. 27–47. Marcel

Dekker, New York.

Maier, C. M., and Chan, R. H. (2002). Role of superoxide dismutases

in oxidative damage and neurodegenerative disorders. Neuroscien-

tist 8, 323–334.

McCord, J. M. (2002). Superoxide dismutases in aging and disease.

Methods Enzymol. 349, 331 –341.

Melov, S. (2002). Therapeutics against mitochondrial oxidative

stress in animal models of aging. Ann. N.Y. Acad. Sci. 959,

330–340.

Storz, G., and Imlay, J. A. (1999). Oxidative stress. Curr. Opin.

Microbiol. 2, 188–194.

BIOGRAPHY

Irwin Fridovich is a James B. Duke Professor of Biochemistry,

Emeritus. He is a member of the National Academy of Science and

of the National Academy of Arts and Sciences. He and his students

discovered and characterized the superoxide dismutase.

138 SUPEROXIDE DISMUTASE

Syk Family of Protein

Tyrosine Kinases

Andrew C. Chan

Genentech, Inc., San Francisco, California, USA

The Syk (spleen tyrosine kinase) family of protein tyrosine

kinases (PTKs) encode essential signaling components required

for normal immunity. Their functions have been most intensely

studied within mammalian immune cells. While not found

within the Caenorhabditis elegans genome, this family of

PTKs is represented earlier within the phylogenetic tree, in the

hydra vulgaris, as a single gene product expressed in epithelial

cells and plays an important function in the recognition of

foreign cells. In mammalian systems, this family of PTKs

appears to have evolved from a gene duplication event to give

rise to its two family members – ZAP-70 and Syk in

mammalian cells. Genetic studies in humans and mice have

demonstrated their essential roles in the function and devel-

opment of T cells, B cells, mast cells, monocytes/macrophages,

and the lymphatic system. Studies further underscore their

importance in T cell antigen receptor (TCR), B cell antigen

receptor (BCR), IgG and IgE receptors (FcRs) and integrin

receptor signaling. This review will discuss our current

understanding of this family of PTKs in mammalian immune

cell function.

Zeta-Chain-Associated Protein

of 70K Mr (ZAP-70)

ZAP-70 was initially identiﬁed as a 70K Mr tyrosine

phosphorylated protein that associates with the TCR

following receptor activation. Biochemical and molecu-

lar characterization revealed that ZAP-70 is a PTK with

the characteristic Syk-family signature of tandem Src-

homology (tSH2) domains at its amino (N) terminus and

a carboxy (C)-terminal catalyticdomain. The domainbet-

ween the two SH2 domains has been termed Interdomain

A, while Interdomain B spans between the C-terminal

SH2 and the C-terminal kinase domain (Figure 1).

INTERACTION OF ZAP-70

WITH THE TCR ITAM

Binding studies and, ultimately, the solution of the crystal

structure encoding the tandem SH2 domains revealed

that both SH2 domains of ZAP-70 cooperate to bind the

doubly phosphorylated immunoreceptor tyrosine based

activation (dP-ITAM) motifs (Figure 2). The ITAMs

consist of the consensus sequence D/E x x Y x x L/I X

6–8

Y x x L/I that is represented in the cytoplasmic domains

of integrins as well as each of the signaling subunits

of the TCR, BCR, and FcRs. Both tyrosine residues

within the ITAM are phosphorylated by Src-family PTKs

that, in turn, provide the high-avidity binding sites for the

tSH2 domains of ZAP-70. While the C-terminal SH2

domain binds the N-terminal ITAM tyrosine, the

C-terminal SH2 domain forms a portion of the pocket

for the N-terminal SH2 domain that correspondingly

binds the C-terminal ITAM tyrosine residue. This inter-

dependence of the two SH2 domains explains the rigid

requirements for the tandem SH2 domains of ZAP-70 as

well as the little ﬂexibility of spacing between the two

tyrosine residues within the ITAM sequence in binding

ZAP-70. Mutation of either SH2 domains or mutation of

either ITAM tyrosine results in . 100-fold decrease in

binding avidities and, in turn, a non-functional TCR.

While ZAP-70 was initially found to be inducibly

associated with the TCR in Jurkat T cells, subsequent

studies in human and murine thymocytes and peripheral

T cells revealed that the TCR

-subunit is already

phosphorylated in its basal unactivated state, though the

degree of phosphorylation is further augmented follow-

ing TCR activation. In turn, in resting thymocytes and

peripheral T cells, ZAP-70 is constitutively associated

with the TCR, though the degree of association appears

to be augmented concomitant with the degree of

-phosphorylation following receptor activation.

A model of sequential phosphorylation of the

-chain

was proposed based on observations in the 3.L2 T cell

hybridoma, but this sequence does not appear to apply

to thymocytes nor normal peripheral T cells. In addition

to localizing ZAP-70 to the TCR complex, the tSH2:I-

TAM interaction has also been proposed to regulate

ZAP-70 enzymatic activity. In vitro binding of ZAP-70

to a dpITAM peptide results in enhanced ZAP-70

activity. This model of ZAP-70 activation, however, is

at odds with the pre-existing ZAP-70:ITAM interaction

described in normal human thymocytes, murine thymo-

cytes, and peripheral T cells.

In contrast to the classical tSH2:dPITAM interaction,

the crystal structure and NMR studies of the tSH2

domains in the absence of an ITAM peptide reveals two

highly ﬂexible independent SH2 domains. These

additional forms of interaction may provide the struc-

tural basis for more non-classical interactions. For

example, the N-terminal SH2 domain alone in conjunc-

tion with Interdomain A of both ZAP-70 and Syk bind

the NXXY motifs within the cytoplasmic domain of the

3 integrin. Of note, tyrosine phosphorylation of the

NXXY motif inhibits binding to the SH2 domain. These

more non-classical interactions may, in part, explain

immunoﬂuorescence studies that demonstrate targeting

of ZAP-70 to the T cell cortex independent of the tSH2

domains. In contrast, studies utilizing ﬂuorescence

imaging and immunoﬂuorescence of green ﬂuorescent

protein-tagged ZAP-70 is more consistent with a

cytosolic distribution of ZAP-70 in resting cells with

redistribution to the plasma membrane following

TCR engagement.

SH2

Kinase

SH2

Kinase

SH2 SH2

Kinase

ZAP-70

Syk

SykB

492

493

292

315

319

525

526

Vav

PLCg

Lck

Cbl

323

348

352

Vav

PLCg

Cbl

323

348

352

Vav

PLCg

Cbl

Interdomain A Interdomain B

FIGURE 1 Schematic representation of ZAP-70 and Syk PTKs. Schematic structural representation of the structural domains within the

Syk PTKs. T-loop tyrosine residues are represented in red; positive regulatory Interdomain B tyrosine residues are represented in blue;

negative regulatory Interdomain B tyrosine residues are represented in black. Numbering utilizes the human protein sequences for both ZAP-70

and Syk.

a /b

mIg

Syk

YPO4

Src-PTKs

ZAP-70

YPO4

Syk

YPO4

Lc k

CD4

BCR

TCR

FIGURE 2 Schematic representation of T and B cell antigen receptors with Src- and Syk-PTKs.

140 Syk FAMILY OF PROTEIN TYROSINE KINASES