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

Подождите немного. Документ загружается.

for a short-term treatment. Recent experimental evi-

dence strongly suggests that such inhibitors may indeed

be beneﬁcial in certain pathologies, such as in multiple

myeloma, a malignancy that affects the bone marrow.

A completely different approach to drug development

can be the development of small molecules that interfere

with the activity of speciﬁc E3s. For example, peptides

that bind speciﬁcally to HPV-E6 can prevent its

association with p53 and interfere with its targeting by

E6-AP. Such peptides were able to induce p53 in HPV-

transformed cells with subsequent reversal of certain

malignant characteristics and induction of apoptosis.

FURTHER READING

Glickman, M. H., and Ciechanover, A. (2002). The ubiquitin –

proteasome proteolytic pathway: Destruction for the sake of

construction. Physiol. Rev. 82, 373–428.

Goldberg, A. L., Elledge, S. J., and Wade, J. (2001). The cellular

chamber of doom. Sci. Am. January, 68–73.

Hilt, W., and Wolf, D. H. (eds) (2000). Proteasomes: The World of

Regulatory Proteolysis. Eurekah.com/LANDES BIOSCIENCE

Publishing Company, Georgetown, Texas, USA.

Weissman, A. M. (2001). Themes and variations on ubiquitylation.

Nat. Rev. Cell Mol. Biol. 2, 169–179.

BIOGRAPHY

Aaron Ciechanover is a Distinguished Professor in the Cancer and

Vascular Research Center in the Faculty of Medicine at the

Technion, Haifa, Israel. His principal research interests are

regulation of transcriptional factors by the ubiquitin system.

He holds an M.D. from “Hadassah” Medical School of the

Hebrew University in Jerusalem, and a Ph.D. in biochemistry from

the Technion. While a graduate student of Dr. Avram Hershko, the

two discovered the principles and machinery of ubiquitination as a

degradation signal.

Michael Glickman is an Associate Professor in the Department of

Biology at the Technion. His principal research interest is mechanisms

of proteolysis by the proteasome, and proteasome composition.

He holds a Ph.D. in chemistry from the University of California at

Berkeley. As a research fellow with Dr. Dan Finley at Harvard Medical

School, he uncovered the basic structure of the regulatory complex of

the proteasome.

UBIQUITIN SYSTEM 303

Ubiquitin-Like Proteins

Edward T. H. Yeh

The University of Texas, Houston, Texas, USA

Ubiquitin is a small polypeptide that covalently attaches to

other proteins, which marks them for degradation by the

proteasome. Ubiquitin regulates many important cellular

processes, such as signal transduction, cell-cycle progression,

and the transformation of normal to malignant cells. Several

ubiquitin-like proteins, such as sentrin (also called SUMO)

and NEDD8/Rub1 have also been discovered. Similar to

ubiquitin, these ubiquitin-like proteins also modify other

proteins. However, they do not trigger proteasomal degra-

dation. Sentrin, in a process called sentrinization, usually

produces a change of the target protein’s cellular localization

and function. NEDD8, with cullin as its substrate, is a key

component of the ubiquitin ligase complex. Thus, NEDD8,

in a process called neddylation, indirectly regulates the

ubiquitin/proteasome system. Sentrinization and neddylation

require unique enzymes – E1, E2, and E3 – that are distinct

from those involved in the ubiquitin pathway. Furthermore,

both sentrinization and neddylation can be regulated by

proteases that speciﬁcally remove sentrin or NEDD8 from

their cellular targets.

Sentrin

There are three sentrins. Sentrin-1 is a 101-amino-acid

protein containing a ubiquitin homology domain

(residues 22–97) that is 48% homologous to human

ubiquitin (see Figure 1). Sentrin-2 is a 95-amino-acid

polypeptide that is 66% similar to sentrin-1 in its

ubiquitin homology domain. Sentrin-3 is a 103-amino-

acid polypeptide that is 97% identical to sentrin-2 in its

ubiquitin homology domain. All sentrins have distinct

N-terminal amino-acid sequences and C-terminal exten-

sions. In addition, the Gly–Gly residues required for

sentrinization are conserved in all sentrins (Figure 1).

ACTIVATION, CONJUGATION,

AND

LIGATION OF SENTRINS

The amino acids following the conserved Gly–Gly

residues in all sentrins are cleaved by C-terminal

hydrolases. The resultant sentrin monomer is then

transferred to a speciﬁc E1 complex (Figure 2). The

activating enzyme (E1) for sentrin is composed of two

subunits, AOS1 and UBA2. The human AOS1 protein is

56% similar to the N-terminal half of the ubiquitin E1.

The human UBA2 protein, which is homologous to the

C-terminal half of the ubiquitin E1, contains the active-

site cysteine residue required for the formation of a thiol

ester linkage with sentrins. The activated sentrin

molecule attached to E1 is then transferred to a carrier

protein, E2.

In contrast to the large number of E2s involved in

ubiquitination, sentrin utilizes only UBC9 as the carrier

protein. This exclusive usage of UBC9 is probably

explained by the existence of a 5-residue insertion that

forms an exposed tight

-hairpin and a 2-residue

insertion that forms a bulge in a loop close to the active

site of UBC9. The surface of UBC9 involved in sentrin-1

binding is positively charged, whereas the corresponding

surface in sentrin-1 is negatively charged. Moreover, the

face of UBC9 remote from the catalytic site possesses a

distinct electrostatic potential that may account for the

propensity of UBC9 to interact with other proteins.

There are at least three different E3 involved in

sentrinization. RanBP2, a protein that has been localized

in the nuclear pore complex, serves as the E3 for Sp100

and histone deacetylase 4. RanBP2 also binds to UBC9,

suggesting that sentrinization could take place at the

nuclear pore complex. Another E3, PIAS (protein

inhibitor of activated STAT), promotes sentrinization

of the androgen receptor, c-JUN, p53, and Sp3. Finally,

the human polycomb group protein, Pc2, functions as an

E3 for a generalized transcription repressor, CtBP.

PROTEINS MODIFIED BY SENTRIN

The list of sentrinized proteins has been expanding

rapidly. Three examples that have been characterized are

discussed here, viz., PML, RanGAP1, and I

PML, a RING ﬁnger protein with tumor suppressor

activity, has been implicated in the pathogenesis of acute

promyelocytic leukemia, which arises following a

reciprocal chromosomal translocation that fuses the

PML gene with the retinoic acid receptor

(RAR

)

gene. PML can be modiﬁed by all sentrins, but not by

ubiquitin or NEDD8. Two forms of PML-RAR

fusion

proteins have been found to be expressed in acute

promyelocytic leukemia. Remarkably, neither can be

sentrinized in vivo. Extensive mutational analysis of

PML has shown that Lys65 in the RING ﬁnger domain,

Lys160 in the B1 Box, and Lys490 in the nuclear

localization signal constitute three major sentrinization

sites. The PML mutant with Lys to Arg substitutions in

all three sites is expressed normally but cannot be

sentrinized. It also cannot recruit Daxx to the nuclear

body. The lack of sentrinization of PML eliminates

PML’s transcriptional coactivator activity.

RanGAP1, a major regulator of the Ras-like GTPase

Ran, which plays a critical role in the bidirectional

transport of proteins and ribonucleoproteins across the

nuclear pore complex was the ﬁrst protein shown to be

modiﬁed by sentrin-1. The 90 kDa sentrinized form of

RanGAP1 associates with RanBP2 of the nuclear pore

complex, whereas the 70 kDa unmodiﬁed form of

RanGAP1 is exclusively cytoplasmic. Thus, sentriniza-

tion is critical for RanGAP1 to translocate to the nuclear

pore. Interestingly, RanGAP1 cannot be modiﬁed by

sentrin-2/3.

Although most sentrinized proteins are localized in the

nucleus or associated with the nuclear envelope, a cyto-

solic protein, I

, can also be modiﬁed by sentrin-1.

FIGURE 1 Alignment of sentrin-1, sentrin-2, sentrin-3, NEDD8, and ubiquitin. The conserved C-terminal glycine–glycine residues were printed

in bold.

FIGURE 2 Schematic diagram of ubiquitination, sentrinization, and neddylation. E1: activating enzyme; E2: conjugating enzyme; E3: ligases.

The E1 for the sentrinization pathway consists of Aos1 and Uba2, whereas the E1 for the neddylation pathway consists of APP-BP1 and Uba3. The

E2 for the sentrinization pathway is Ubc9, whereas the E2 for the neddylation pathway is Ubc12. There are multiple E2s for the ubiquitination

pathway. Ub: ubiquitin; Sn: sentrin: Nd: NEDD8.

UBIQUITIN-LIKE PROTEINS 305

is a cytosolic inhibitor of NF

B, a transcription

factor involved in the induction of a large number of

proteins involved in inﬂammation. I

is phosphory-

lated on serine residues 32 and 36 following tumor

necrosis factor stimulation. The phosphorylated form of

is then polyubiquitinated on lysine residues 21 and

22 and degraded by the proteasome. Unlike ubiquitina-

tion, sentrinization of I

on lysine residue 21 is

inhibited by prior phosphorylation. Furthermore, sen-

trinized I

cannot be ubiquitinated and is resistant to

proteasomal degradation. Thus, sentrinization can com-

pete with ubiquitination for the same Lys residue on

target proteins, effectively constitutes an antiubiquitin

process.

By comparing the amino-acid sequences surrounding

the acceptor lysine residues, a sentrinization consensus

sequence can be formulated. It appears that the acceptor

Lys residue is preceded by a hydrophobic amino acid (Ile

or Leu) and followed by a polar amino acid, such as Gln,

Thr, and a charged amino acid, Glu. The general

consensus formula is (I/L)K(Q/T)E.

DE-SENTRINIZATION

In yeast, there are two sentrin-speciﬁc proteases (SENPs)

that deconjugate sentrin from its substrate. Ulp1

possesses both isopeptidase and C-terminal hydrolase

activity and is essential for the transition from G2 to the

M phase of the cell cycle, the same phenotype observed

for the Ubc9 mutation in yeast. These observations

suggest that sentrinization and de-sentrinization are

required for cell-cycle progression in yeast. Ulp2, on

the other hand, is not essential for cell viability. The

Ulp2 mutant exhibits a pleiotropic phenotype that

includes temperature-sensitive growth, abnormal cell

morphology, and decreased plasmid and chromosome

stability. The mutant is also hypersensitive to DNA-

damaging agents, such as hydroxyurea. Ulp1 is localized

in the nuclear envelope, whereas Ulp2 is in the nucleus.

In humans, there are three well-characterized SENPs.

All human SENPs have conserved C-terminal region,

even though they vary in size. The similarity between

yeast Ulp1, Ulp2, and human SENPs is conﬁned

primarily to the C-terminal region of , 200 amino

acids, within which a ,90-residue segment forms a core

structure common to cysteine proteases. It appears that

the conserved C termini of the SENPs contain the

catalytic domain, whereas the N termini regulate their

cellular localization and substrate speciﬁcity. SENP1 is

localized in the nucleus, excluding the nucleolus; SENP2

is localized in the nuclear envelope and also forms

nuclear speckles; and SENP3 is localized predominantly

in the nucleolus. Both SENP1 and SENP2 can deconju-

gate sentrinized PML but not RanGAP1. In addition,

SENP1 and SENP2 can deconjugate many sentrinized

nuclear proteins. However, SENP3 appears to have more

restricted substrate speciﬁcity.

In summary, the sentrin system possesses its own E1,

E2, E3, and deconjugating enzymes, which serve to

regulate an increasingly recognized cellular processes.

NEDD8/Rub1

NEDD8 is a small protein consisting of 81 amino

acids that is 80% homologous to ubiquitin (see

Figure 1). The yeast and plant homologues of

NEDD8 have been termed Rub1 (related to ubiquitin

1) and are processed and conjugated to a small

number of cellular proteins. The overall structures of

NEDD8 and Rub1 are quite similar to those of

ubiquitin. The major difference occurs in two surface

regions with a different electrostatic surface potential.

Thus, unlike sentrin-1, NEDD8 is structurally more

closely related to ubiquitin.

ACTIVATION AND CONJUGATION

NEDD8 AND RUB1

UCH-L3, one of the ubiquitin C-terminal hydrolases,

also binds to NEDD8 and functions as a NEDD8

C-terminal hydrolase. Following processing of its C

terminus, NEDD8 is transferred to its E1.

The E1 for human NEDD8 is composed of two

subunits, APP-BP1 and UBA3. The human APP-BP1

protein is 56% similar to the N-terminal half of

ubiquitin’s E1. The human UBA3 protein, which is

homologous to the C-terminal half of ubiquitin’s E1,

contains the active-site cysteine residue required for the

formation of thiol ester linkage with NEDD8. A similar

E1 has also been identiﬁed in yeast and plants. The

counterpart of APP-BP1 has been named AXR1 in plant.

The importance of the Rub1 conjugation system is

highlighted by the requirement of the AXR1 protein for

the normal response to the plant hormone auxin in

Arabidopsis thaliana. The speciﬁcity of NEDD8’s E1 is

due, in part, to the preferential binding of UBA3 to

NEDD8, but not to ubiquitin or sentrin-1. The

conjugating enzyme for NEDD8 or Rub1 is UBC12. A

dominant-negative form of UBC12 could abolish

NEDD-8 conjugation in vivo and inhibit cell growth.

RUB1/NEDD8 MODIFICATION OF

CDC53/CULLINS

The major substrate for Rub1 in yeast is Cdc53 (also

known as yeast cullin), a 94 kDa protein required for the

to S phase of the cell-cycle progression. Cdc53 is a

common subunit of the SCF complex, a ubiquitin ligase

(E3) composed of Skp1, Cdc53/cullin, and an F box

306

UBIQUITIN-LIKE PROTEINS

protein. There are many F-box proteins, which play a

critical role in conferring substrate speciﬁcity to the

SCF complex. For example, the F-box protein, Cdc4,

forms a SCF

Cdc4

complex with Skip1 and Cdc53/cullin

that binds to phosphorylated Sic1 and catalyzes its

ubiquitination.

In mammalian cells, NEDD8 modiﬁes a limited

number of human cellular proteins that primarily

localized in the nucleus. Interestingly, all of the known

NEDD8 targets in mammalian cells are cullins. Human

Cul-1 is a major component of the SCF complex that is

involved in the degradation of I

-catenin, and p27.

The neddylation of Cul-1 dissociates CAND1, an

inhibitor of Cul-1–SKP1 complex formation. The

suppression of CAND1 therefore increases the level of

the CuL-1–SKP1 complex. Thus, the neddylation of

Cul-1 enhances the formation of the SCF complex,

which in turn stimulates protein polyubiquitination.

Human Cul-2 binds to the von Hippel-Lindau gene

product through elongin B and elongin C to form a

complex (Cul-2–VBC complex) that appears to have

ubiquitin ligase activity. The VBC complex itself

promotes, but is not essential for, NEDD8 conjugation

to Cullin-2. Human Cul-3 was initially identiﬁed as a

salicylate suppressible protein with unknown mechan-

ism of activation, and was recently shown to be involved

in the ubiquitination of cyclin E to control the S phase in

mammalian cells. Human Cul-4A associates with UV-

damaged DNA-binding protein and may play a role in

DNA repair. Thus, the cullins may constitute a family of

ubiquitin E3 ligases that have diverse biologic functions,

with the neddylation of cullins playing a crucial role in

ligase activity.

REGULATION OF NEDD8-CONJUGATES

There are several proteins that regulate NEDD8 con-

jugates. USP21 is a ubiquitin-deconjugating enzyme

that is also capable of removing NEDD8 from NEDD8

conjugates. The overexpression of USP21 profoundly

inhibits growth of U2OS cells. DEN1 is a speciﬁc protease

for NEDD8 conjugates, as it has no activity against

ubiquitin or sentrin conjugates. DEN1 can de-neddylate

NEDD8 conjugated Cul-1. Interestingly, DEN1 is highly

homologous to the catalytic domain of SENP. The

COP9 signalosome (CSN) is an evolutionarily conserved

multiprotein complex composed of eight subunits.

It was ﬁrst identiﬁed as an essential component in

the repression of light-regulated development in

Arabidopsis. CSN interacts with the SCF-type E3

ubiquitin ligases containing Cul1, which suggests that

CSN and SCF-type E3 ubiquitin ligases are closely

related. Indeed, CSN promotes the cleavage of NEDD8

from neddylated Cul1. The Jab1/MPN domain metal-

loenzyme motif in Csn5 subunit is responsible for the

de-neddylation activity of CSN. Thus, CSN binds to SCF-

type E3 ubiquitin ligases and probably regulates their

activity through the de-neddylation of conjugated Cul-1.

NUB1, an interferon-inducible protein, also regulates the

neddylation system. The overexpression of NUB1 leads

to the disappearance of most neddylated proteins. NUB1

appears to act by recruiting NEDD8 and its conjugates to

the proteasome for degradation, further highlighting the

close functional tie between neddylation and

ubiquitination.

FURTHER READING

Hay, R. T. (2001). Protein modiﬁcation by SUMO. Trends Biochem.

Sci. 26, 332 –333.

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

Dev. Biol. 16, 591–626.

Yeh, E. T. H., Gong, L., and Kamitani, T. (2000). Ubiquitin-like

proteins: New wines in new bottles. Gene 248, 1–14.

BIOGRAPHY

Edward T.H. Yeh is Chairman of the Department of Cardiology at the

University of Texas M.D. Anderson Cancer Center and Director of the

Research Center for Cardiovascular Diseases at the Institute of

Molecular Medicine at the University of Texas-Houston Health

Science Center. He holds an M.D. from the University of California,

Davis and received his postdoctoral training at Harvard University. His

laboratory was among the ﬁrst to show that sentrin and NEDD8 can be

covalently conjugated to other proteins.

UBIQUITIN-LIKE PROTEINS 307

UmuC, D Lesion Bypass

DNA Polymerase V

Zvi Livneh

Weizmann Institute of Science, Rehovot, Israel

DNA polymerase V from the bacterium Escherichia coli

(E. coli) is specialized to perform DNA synthesis across DNA

lesions. The latter are sites in DNA that were chemically

damaged by radiation or chemicals, and which usually block

DNA replication. The reaction catalyzed by pol V is called

translesion synthesis (TLS), translesion replication (TLR), or

lesion bypass. It is usually mutagenic, since the coding

properties of damaged nucleotides in DNA are often different

from those of the original undamaged nucleotides. Pol V is a

heterotrimer composed of one molecule of the UmuC protein

(48 kDa), and two molecules of the UmuD

protein (12 kDa

each). UmuD

is a proteolytic cleavage product of a larger

protein, termed UmuD (15 kDa). The UmuD and UmuC

proteins are encoded by an operon that is induced by DNA-

damaging agents, under regulation of the SOS response. Pol V

is a low-ﬁdelity DNA polymerase, which does not have an

exonuclease proofreading activity, and exhibits low processiv-

ity. Its lesion-bypass activity requires two additional proteins,

RecA, which is also the main recombinase in E. coli, and

single-strand DNA-binding protein (SSB), a protein essential

for replication. In addition, pol V is stimulated by the

processivity proteins of pol III, namely, the

-subunit DNA

sliding clamp, and the

-complex clamp loader. By virtue of its

mutagenic lesion-bypass activity, pol V is responsible for the

formation of mutations by a variety of DNA-damaging agents

that cause replication blocks, such as UV radiation, methy-

methanesulphonate (MMS), and more. Pol V homologues

were conserved in evolution from E. coli to humans, and

comprise the Y family of DNA polymerases.

Pol V and the SOS Response

The SOS response is a global stress response in E. coli,

induced by DNA damage that causes interruption of

DNA replication. It involves the induction of 30–40

genes, which collectively act to increase cell survival

under adverse environmental conditions. SOS genes

comprise a regulon, whose members are negatively

regulated at the transcriptional level by the LexA res-

pressor. Induction involves autocleavage of free LexA

repressor, promoted by an activated form of the RecA.

This cleavage shifts the binding equilibrium of LexA,

causing it to dissociate from its target genes, thereby

causing induction. RecA is usually activated by forming

a complex of multiple RecA molecules bound to a

stretch of single-stranded DNA (ssDNA), which is

formed as a result of exposure to damaging agents.

This RecA nucleoprotein ﬁlament has multiple roles in

cellular responses to DNA damage. In addition to the

induction of the SOS regulon, it promotes homologous

recombination repair, and it is essential for lesion bypass

by pol V. The umuD and umuC genes, encoding the two

subunits of pol V, are arranged in an operon that is part

of the SOS regulon (namely, its transcription is regulated

by LexA and RecA), and are therefore inducible by

DNA-damaging agents such as UV radiation or MMS.

Mutations that inactivate umuC or umuD cause a

drastic decrease in UV- or MMS-mutagenesis, with a

minor effect on cell survival. This demonstrated the

surprising phenomenon that mutagenesis caused by

DNA-damaging agents is an inducible process, which

requires speciﬁc gene products. In other words, muta-

genesis is not merely a by-product of replication, and the

presence of lesions in DNA does not cause mutations

without the activity of speciﬁc inducible proteins.

The Discovery of Pol V

It took nearly two decades from the time the umuC and

umuD genes were cloned, until the discovery that they

encode a specialized lesion-bypass DNA polymerase.

This long period of time is attributed to the great

difﬁculty in obtaining UmuC in soluble and active form,

the lack of any amino acid homology to classical DNA

polymerases, and the complexity of the in vitro lesion-

bypass assay system. In addition, genetic experiments

had indicated that the polC (dnaE), encoding the

subunit of pol III, was required for lesion bypass in vivo.

This led to the incorrect model that the UmuC and

UmuD

proteins were accessory proteins, which enabled

pol III to bypass DNA lesions. Finally, UmuCD

dependent lesion bypass systems were reconstituted

from puriﬁed components, and led to the ﬁnding that

UmuC has DNA polymerase activity, and together with

UmuD

it forms pol V. This discovery was made

simultaneously and independently in the laboratories

of Myron Goodman (University of Southern California,

Los Angeles) and Zvi Livneh (Weizmann Institute of

Science, Rehovot, Israel).

Model for the Mechanism of Lesion

Bypass by Pol V

Pol V is unable to bypass lesions on its own. It requires

two additional proteins, RecA and SSB, known to be

involved in other DNA transactions such as replication

and recombination. In addition, pol V is stimulated by

the processivity proteins, which are also required for

DNA replication. The RecA protein is required as a

nucleoprotein complex with ssDNA, in which multiple

RecA monomers form a protective helical sleeve around

the DNA (Figure 1, stage 1). The RecA ﬁlament prevents

nonproductive binding of pol V to ssDNA regions

remote to the primer terminus, and it targets pol V to

the primer-template region (Figure 1,stage2).

Once targeted to the primer-terminus site, the RecA

ﬁlament needs to locally dissociate, to allow productive

binding of pol V. This requires, in addition to pol V, also

the SSB protein. The pol V initiation complex is

stabilized by RecA–UmuD

interactions, UmuC– DNA

interactions, and perhaps interactions with SSB. Once

this pol V complex has formed an initiation complex,

DNA synthesis can occur. However, to get full pol V

holoenzyme activity, the processivity proteins are

needed, including the

-subunit DNA sliding clamp,

and the

-complex clamp loader (Figure 1, stages 3, 4).

These proteins are essential for highly processive DNA

replication by DNA polymerase III; however, they also

increase the processivity of DNA polymerases II, IV, and

V, although to a much lesser extent than pol III. DNA

synthesis by pol V holoenzyme causes dissociation of

RecA monomers, as the polymerase progresses. Pol V is

a major lesion-bypass machine, which can bypass a wide

variety of DNA lesions. This includes abasic sites, and

the two major lesions caused by UV light, a cyclobutyl

thymine– thymine dimer, and a thymine–thymine 6–4

adduct. Most remarkably, pol V is able to bypass a

stretch of 12 methylene residues inserted into the DNA.

This artiﬁcial lesion is an extreme form of DNA damage,

since it bears no resemblance to any of the regular DNA

constitutents. Despite this remarkable bypass ability, pol

V is poor in bypassing some DNA lesions, most notably

benzo[a]pyrene–guanine adducts. The ability of pol V

to bypass DNA lesions correlates with its low ﬁdelity.

However, low ﬁdelity by itself is not sufﬁcient to ensure

effective lesion bypass, since there are enzymes with

ﬁdelity similar to pol V, which nevertheless are not

lesion-bypass enzymes (e.g., mammalian DNA poly-

merase

). The crystal structure of pol V was not

determined yet. However, based on several crystal

structures of other homologous lesion-bypass DNA

polymerases, a spacious and ﬂexible active site is likely

to be responsible, at least in part, for the ability to bypass

DNA lesions.

The Fidelity of Pol V

The ﬁdelity of DNA synthesis by pol V undamaged DNA

templates is 10

–10

errors/nucleotide replicated.

This is 100–1000-fold lower than the ﬁdelity of the

replicative pol III holoenzyme. DNA sequence analysis

of pol V errors showed that pol V produces all types of

errors, including frameshifts and base substitutions.

However, it has a propensity to produce transversion

mutations, namely, changing a pyrimidine into a purine,

or vice versa. Pol V forms transversions at a frequency

up to 300-fold higher than pol III holoenzyme (which

tends to form transition mutations; Table I). It should be

4. Lesion bypass

UmuC

UmuD'

1. Formation of a RecA-ssDNA nucleoprotein filament

Pol V

RecA

2. Targeting of pol V to the primer terminus

SSB

3. Loading of the -subunit DNA sliding clamp

-subunit

FIGURE 1 Outline of the mechanism of lesion bypass by pol V. The

red square represents a blocking lesion. The “M” represents an incorrect

insertion opposite the lesion. Reproduced from Livneh, Z. (2001). DNA

damage control by novel DNA polymerases: Translesion replication

and mutagenesis. J. Biol. Chem. 276, 25639–25642, with permission

of the American Society for Biochemistry and Molecular Biology.

UmuC, D LESION BYPASS DNA POLYMERASE V 309

noted that base –base mismatches produced by DNA

polymerases during synthesis on undamaged DNA are

potential substrates for the mismatch repair system. This

system primarily corrects mismatches that are precur-

sors for transition mutations, namely, the types of

mismatches that the replicative pol III holoenzyme

tends to make. This signiﬁcantly reduces mutations

that can occur during DNA replication, and contributes

to the high replication ﬁdelity. The repair by the

mismatch repair system of mismatches that are pre-

cursors of transversions, such as pol V tends to make, is

20-fold less effective than transition mismatches. This

implies that pol V generates mismatches that are largely

immune to mismatch repair and yield mutations. This

provides the mechanistic basis for the phenomenon of

untargeted SOS mutagenesis, also termed SOS mutator

activity. In essence, this phenomenon involves the

formation of mutations in undamaged regions of DNA

in cells in which the SOS response was induced. This

process is umuDC- and recA- dependent, and primarily

yields transversion mutations. Based on the ﬁdelity of

pol V, untargeted mutagenesis can be explained by the

activity of pol V under SOS conditions in undamaged

regions of the E. coli chromosome. It should be noted

that there is a second branch of untargeted mutagenesis

that depends on pol IV rather than pol V.

Regulation of Pol V

The activity of pol V is tightly regulated at several levels,

to ensure that this mutator polymerase acts only under

situations, and at sites, when it is really needed. As

described above, the umuDC operon is regulated at the

transcriptional level by the LexA repressor and the RecA

activator like other SOS genes. Within the timed

expression of SOS genes, the umuDC operon is induced

late, nearly an hour after the onset of induction. When

SOS genes become repressed again, as the SOS response

is shut down, the umuDC operon is repressed early. This

late induction, early shutdown expression pattern

provides an extra layer of regulation on pol V. It limits

the activity of pol V to a stage after error-free repair

mechanisms had a chance to repair DNA, and even at

that late stage in SOS, the activity of pol V is limited in

time, as it is shut off early. Additional regulation is at the

posttranslational level. The UmuD

protein is cleaved to

yield a shorter protein termed UmuD

, which is the

active protein in pol V. The proteolytic processing occurs

via the self-cleavage of the UmuD

protein, promoted by

activated RecA, in a mechanism that is similar to the

cleavage of the LexA repressor. Finally, there is

regulation by turnover, as UmuD is degraded by the

Lon and ClpXP proteases. The steady-state level of

UmuC in noninduced cells is low and undetected by

currently available antibodies. Upon SOS induction,

UmuC accumulates to a level of , 200 molecules/cell.

UmuD is present in higher amounts, starting with 200

molecules in the noninduced cell, and going up to 2400

molecules/cell after induction. Finally, pol V is regulated

at the activity level, in its requirement for a RecA

nucleoprotein ﬁlament. Such ﬁlaments are formed in

DNA only when replication is arrested (e.g., at a lesion)

and a persistent single-stranded region is generated.

In vivo Function of Pol V

The major phenotype of E. coli cells lacking pol V is a

strong decrease in mutations caused by DNA-damaging

agents such as UV radiation or MMS. Survival is

decreased too, but to a lower degree compared to

defects in other DNA repair pathways, such as excision

repair. Based on these phenotypes, there are two views

on the in vivo role of pol V. One is that pol V repairs

replication gaps and this leads to increased survival.

According to this model, the mutations formed by pol V

are a by-product of the gap-ﬁlling activity, and can be

viewed as the price paid for this gap repair (repair at the

expense of mutations). The other view portrays the

generation of mutations as the main role of pol V.

According to this view, when a population of E. coli

cells is under hostile environmental conditions, e.g., lack

of nutrients, or presence of DNA-damaging agents such

as UV light, it increases its mutations rate via pol V, to

facilitate adaptation and increase ﬁtness. Even one

successful mutation in a cell (e.g., conferring higher

resistance to sunlight) may have a great impact on the

TABLE I

Mutations Produced during In Vitro DNA Synthesis by Pol V and

Pol III Holoenzyme

Mutation

frequency 3 10

/gene

Mutation Mismatch

Pol III Pol V Pol V/pol III

Transition

T ! CT:G, 1.4 368.3 . 263

Transversion

A ! C A:G 2.8 161.1 58

A ! T A:A , 1.4 414.4 . 296

C ! A C:T , 1.4 92.1 . 66

T ! A T:T 2.8 230.2 82

T ! GT:C, 1.4 69.1 . 49

The mutations were formed in the cro repressor gene of phage

during gap-ﬁlling DNA synthesis.

Reproduced from Maor-Shoshani, A., Reuven, N. B., Tomer, G.,

and Liuneb, Z. (2000). Highly mutagenic replication by DNA

polymerase V (UmuC) provides a mechanistic basis for SOS untargeted

mutagenesis. Proc. Natl Acad. Sci. USA 97, 565–570, with permission

of the National Academy of Sciences, U.S.A.

310 UmuC, D LESION BYPASS DNA POLYMERASE V

entire population, since in time these cells, which have a

selective advantage, may proliferate to a state where

they take over the population. Consistent with such an

idea is the ﬁnding that when mutator strains are grown

for many generations along with nonmutator strains, the

mutator strain takes over. Similarly, umuDC-deﬁcient

mutants suffer a severe reduction in ﬁtness when

competing against cells with normal pol V. Since

mutations form randomly in relation to their functional

outcome, a mutator acting long enough will start

accumulating deleterious mutations, and will lose its

adaptive advantage. The fact that the activity of pol V is

transient (namely, as long as the SOS response is on)

helps to preserve successful mutations from being

neutralized by the formation of other mutations. It

was also suggested that pol V initially evolved as a

generic bypass DNA polymerse, which functioned to

increase survival by overcoming replication blocks, but

as more sophisticated DNA repair mechanisms evolved,

the importance of the bypass function decreased, and

what kept pol V during evolution was its mutator effect.

Of course, the various hypotheses are not mutually

exclusive, and the in vivo role of pol V may be both in

survival and mutagenesis.

Bacterial Homologues of Pol V

E. coli contains a homologue of pol V, termed pol IV,

which is the product of the SOS gene dinB. Like pol V,

pol IV is an SOS-inducible, low-ﬁdelity, and low-

processivity DNA polymerase, which is able to bypass

some DNA lesions. As in pol V, the processivity and

lesion-bypass ability of pol IV are increased by the

-subunit clamp, and the

-subunit clamp loader.

However, unlike pol V, pol IV is a single subunit

enzyme, and its bypass activity requires neither RecA

nor SSB. On undamaged DNA, pol IV tends to form

frameshifts, primarily minus-one deletions. The in vivo

role of pol IV is not as clear as that of pol V. Pol IV is

required for untargeted mutagenesis of phage

. This

mutagenic pathway is observed in nonirradiated phage

when it infects a UV-irradiated E. coli host. In addition,

pol IV is required for the bypass of a benzo[a]pyrene–G

adducts. It is also required for stationary phase

mutations.

Homologues of pol V exist in many, but not all

bacteria. Pol V homologues were also found on native

bacterial conjugative plasmids, such as R46. These are

large plasmids, in the range of 100 kbp, with broad host-

range speciﬁcity, which often carry multiple antibiotics-

resistance markers. In the case of the R46 plasmid, the

UmuD, UmuD

, and UmuC homologues are MucA,

MucA

and MucB, respectively. MucA

and MucB were

shown to form a lesion-bypass DNA polymerase, pol RI,

with properties similar to those of pol V. Pol RI,

like pol V, required both RecA and SSB for bypassing

an abasic site. MucA and MucB are present on plasmid

pKM101, a natural deletion derivative of plasmid R46.

Plasmid pKM101 was introduced into the Salmonella

strains that are used in the Ames test for mutagens, in

order to improve its mutagenic sensitivity.

The emergence of pathogenic bacteria that are

resistant to multiple types of antibiotics is associated in

part of the cases with the presence of native conjugative

plasmids. The fact that native plasmids, which have a

limited genome size, and rely heavily on host factors,

carry mutator lesion-bypass DNA polymerase genes

indicates that the latter have a special role in the life

cycle of such plasmid. One possibility is that the mutator

activity of the polymerases is required when the

plasmids enter a new foreign bacterial host, and need

rapid adaptation to the new intracellular environment.

Eukaryotic Homologues of Pol V

Homologues of pol V were found to be conserved in

evolution, and comprise the Y family of DNA poly-

merases. The yeast S. cerevisiae contains two homol-

ogues of UmuC, pol

(product of the RAD30 gene), and

REV1, a G-template speciﬁc DNA polymerase. Human

cells contain no less than four homologues of pol V: pol

, pol

, and REV1. The DNA polymerases of

the Y family were discovered in 1999 by several

investigators, simultaneously and independently.

The S. cerevisiae pol

was reported ﬁrst, discovered

by Satya Prakash and Louis Prakash (University of Texas

Medical Branch, Galveston). Both the S. cerevisiae and

the human pol

have the remarkable property that they

replicate across an unmodiﬁed DNA TT sequence, with

similar efﬁciency and speciﬁcity to that of a TT

cyclobutyl dimer. In the absence of pol

in the cell,

another polymerase takes over, but performs the

reaction with a lower ﬁdelity, leading to increased UV

mutagenesis. This is the situation in the human

hereditary disease Xeroderma Pigmentosum Variant

(XP-V). Unlike other forms of XP, which are defective

in components of the error-free nucleotide excision

repair, XP-V patients are deﬁcient in DNA pol

(the

XP-V gene product). XP-V patients lacking this enzyme

show extreme sensitivity to sunlight and a high

susceptibility to skin cancer. This led to the unexpected

conclusion that lesion bypass may be functionally

nonmutagenic, and that lesion-bypass DNA poly-

merases may function to protect mammals from at

least certain types of cancer.

S. cervisiae has an additional lesion-bypass polymer-

ase, pol

, encoded by the REV3 and REV7 genes. It is

homologous to the replicative pol

, and not to the

Y-family DNA polymerases. Historically, pol

was the

ﬁrst lesion-bypass DNA polymerase, discovered in 1996

UmuC, D LESION BYPASS DNA POLYMERASE V 311

by David Hinkle and Christopher Lawrence (Rochester

University), followed soon thereafter by the discovery, in

the same year, and by the same investigators, that REV1

has nucleotidyl transferase activity. Human cells also

contain additional polymerases, which can bypass

lesions in vitro. This includes pol

and pol

,and

homologues of the yeast REV3 and REV7 genes,

encoding a putative pol

. Overall this suggests that

lesion bypass has an important role in the response to

DNA damage in mammalian cells.