GDA201.QXD 03/22/2000 02:15 Page 162
`
`162
`
`Sloppier copier DNA polymerases involved in genome repair
`Myron F Goodman* and Brigette Tippin
`
`When chromosomal replication is impeded in the presence of
`DNA damage, members of a newly discovered
`UmuC/DinB/Rev1/Rad30 superfamily of procaryotic and
`eucaryotic DNA polymerases catalyze translesion synthesis at
`blocked replication forks. Although these polymerases share
`sequence elements essentially unrelated to the standard
`replication and repair enzymes, some of them (such as the
`SOS-induced Escherichia coli pol V) catalyze ‘error-prone’
`translesion synthesis leading to large increases in mutation,
`whereas others (an example being the Xeroderma
`pigmentosum variant gene product XPV pol h ) carry out
`aberrant, yet nonmutagenic translesion synthesis. Ongoing
`studies of these low fidelity polymerases could provide new
`insights into the mechanism of somatic hypermutation, a key
`element in the immune response.
`
`Addresses
`Department of Biological Sciences and Chemistry, University of
`Southern California, University Park, Los Angeles, California 90089-
`1340, USA
`*e-mail: mgoodman@mizar.usc.edu
`
`Current Opinion in Genetics & Development 2000, 10:162–168
`
`0959-437X/00/$ — see front matter © 2000 Elsevier Science Ltd.
`All rights reserved.
`
`Abbreviations
`HE
`holoenzyme complex
`pol
`DNA polymerase
`pol V Mut
`pol V mutasome
`SBB
`DNA single-stranded binding protein
`TLS
`translesion synthesis
`XPV
`Xeroderma pigmentosum variant gene product
`
`Introduction
`The first DNA polymerase (pol), Escherichia coli pol I, was
`discovered in 1957 by Arthur Kornberg (for review, see
`[1]). Twelve years later John Cairns [2] isolated a strain of
`E. coli containing a mutant pol I enzyme leading to the dis-
`coveries of pol II and pol III (for review, see [1]). Pol III is
`responsible primarily for replicating the bacterial genome,
`while pol I plays a major role in UV damage repair and in
`Okazaki fragment processing (for review, see [1]). The
`enigmatic pol II was recently shown to be involved in the
`reactivation of replication complexes stalled at DNA tem-
`plate lesions [3]. Thirty years have now passed since the
`discovery of the pol I mutant. Remarkably, the past
`18 months have witnessed the discovery of a variety of new
`procaryotic and eucaryotic DNA polymerases, including
`two more in E. coli.
`
`This review discusses these new DNA-damage tolerant
`polymerases with special emphasis placed on the role of
`the error-prone UmuD¢ 2C complex (E. coli pol V) in the
`well-documented SOS mutagenic response in E. coli. We
`provide an overview of the relationships between the novel
`
`UmuC/DinB/Rev1/Rad30 superfamily of DNA polymerases
`spanning procaryotic and eucaryotic organisms.
`
`A brief synopsis of the E. coli SOS-regulon
`E. coli’s SOS response involves the action of at least 25
`genes regulated at the transcriptional level by LexA
`repressor protein. Following damage to DNA, the LexA
`repressor undergoes proteolysis, mediated by RecA pro-
`tein acting as a coprotease, turning on SOS gene
`expression (Figure 1). Many of the proteins induced early
`in the SOS response are involved in nucleotide excision
`repair and recombination repair pathways [4]. These path-
`ways are ‘error-free’ meaning that they do not cause
`mutations above spontaneous background levels; however,
`it has been known since 1977 that a large, ~100-fold,
`increase in mutations accompanies SOS induction. This
`involves the action of RecA protein and UmuD¢ and
`UmuC proteins [5–7], where ‘Umu’ refers to UV mutage-
`nesis. RecA protein
`is
`required
`to process
`the
`mutagenically inactive UmuD to UmuD¢ — a shorter,
`mutagenically active form of the protein — in a reaction
`analogous to the cleavage of LexA protein [8–10]. Follow-
`ing cleavage of UmuD to UmuD¢ , UmuC and two
`molecules of UmuD¢ associate to form a UmuD¢ 2C com-
`plex, which, in the presence of activated RecA protein
`filament (RecA*), catalyzes ‘error-prone’ translesion syn-
`thesis (TLS) causing mutations at DNA damage sites [11].
`RecA plays a direct biochemical role during SOS mutage-
`nesis that is distinct from generalized recombination and
`coproteolysis [10,12,13] and that is apparently responsible
`for targeting UmuD¢ 2C to a template-lesion site proximal
`to the tip of the RecA* filament [14,15].
`
`Mutagenically inactive complexes formed with UmuD2C
`and UmuD¢ DC are thought to act as a regulatory switch to
`turn off mutagenesis once DNA damage sites have been
`either repaired or bypassed [16]. Early reviews on SOS
`were written by Witkin [17] and Walker [4], and Friedberg
`et al. [18] provide a recent comprehensive review of the
`SOS regulatory system, written prior to the discoveries of
`error-prone E. coli pol IV and pol V.
`
`SOS translesion synthesis reconstituted
`in vitro
`A replication complex confronting a damaged DNA tem-
`plate strand may be likened to a major train wreck
`resulting in ‘derailment’ of the core polymerase and its
`accessory subunits. When faced with excessive amounts of
`DNA damage, the cell sends out an SOS signal, perhaps in
`the form of a segment of single-stranded chromosomal
`DNA bound by RecA protein. A specialized group of pro-
`teins are induced that can copy damaged template sites,
`making errors along the way. The proteins required for
`SOS-induced mutation (also called SOS error-prone repair)
`
`Columbia Ex. 2085
`Illumina, Inc. v. The Trustees
`of Columbia University in the
`City of New York
`IPR2020-00988, -01065,
`-01177, -01125, -01323
`
`

`

`GDA201.QXD 03/22/2000 02:15 Page 162
`
`162
`
`Sloppier copier DNA polymerases involved in genome repair
`Myron F Goodman* and Brigette Tippin
`
`When chromosomal replication is impeded in the presence of
`DNA damage, members of a newly discovered
`UmuC/DinB/Rev1/Rad30 superfamily of procaryotic and
`eucaryotic DNA polymerases catalyze translesion synthesis at
`blocked replication forks. Although these polymerases share
`sequence elements essentially unrelated to the standard
`replication and repair enzymes, some of them (such as the
`SOS-induced Escherichia coli pol V) catalyze ‘error-prone’
`translesion synthesis leading to large increases in mutation,
`whereas others (an example being the Xeroderma
`pigmentosum variant gene product XPV pol h ) carry out
`aberrant, yet nonmutagenic translesion synthesis. Ongoing
`studies of these low fidelity polymerases could provide new
`insights into the mechanism of somatic hypermutation, a key
`element in the immune response.
`
`Addresses
`Department of Biological Sciences and Chemistry, University of
`Southern California, University Park, Los Angeles, California 90089-
`1340, USA
`*e-mail: mgoodman@mizar.usc.edu
`
`Current Opinion in Genetics & Development 2000, 10:162–168
`
`0959-437X/00/$ — see front matter © 2000 Elsevier Science Ltd.
`All rights reserved.
`
`Abbreviations
`HE
`holoenzyme complex
`pol
`DNA polymerase
`pol V Mut
`pol V mutasome
`SBB
`DNA single-stranded binding protein
`TLS
`translesion synthesis
`XPV
`Xeroderma pigmentosum variant gene product
`
`Introduction
`The first DNA polymerase (pol), Escherichia coli pol I, was
`discovered in 1957 by Arthur Kornberg (for review, see
`[1]). Twelve years later John Cairns [2] isolated a strain of
`E. coli containing a mutant pol I enzyme leading to the dis-
`coveries of pol II and pol III (for review, see [1]). Pol III is
`responsible primarily for replicating the bacterial genome,
`while pol I plays a major role in UV damage repair and in
`Okazaki fragment processing (for review, see [1]). The
`enigmatic pol II was recently shown to be involved in the
`reactivation of replication complexes stalled at DNA tem-
`plate lesions [3]. Thirty years have now passed since the
`discovery of the pol I mutant. Remarkably, the past
`18 months have witnessed the discovery of a variety of new
`procaryotic and eucaryotic DNA polymerases, including
`two more in E. coli.
`
`This review discusses these new DNA-damage tolerant
`polymerases with special emphasis placed on the role of
`the error-prone UmuD¢ 2C complex (E. coli pol V) in the
`well-documented SOS mutagenic response in E. coli. We
`provide an overview of the relationships between the novel
`
`UmuC/DinB/Rev1/Rad30 superfamily of DNA polymerases
`spanning procaryotic and eucaryotic organisms.
`
`A brief synopsis of the E. coli SOS-regulon
`E. coli’s SOS response involves the action of at least 25
`genes regulated at the transcriptional level by LexA
`repressor protein. Following damage to DNA, the LexA
`repressor undergoes proteolysis, mediated by RecA pro-
`tein acting as a coprotease, turning on SOS gene
`expression (Figure 1). Many of the proteins induced early
`in the SOS response are involved in nucleotide excision
`repair and recombination repair pathways [4]. These path-
`ways are ‘error-free’ meaning that they do not cause
`mutations above spontaneous background levels; however,
`it has been known since 1977 that a large, ~100-fold,
`increase in mutations accompanies SOS induction. This
`involves the action of RecA protein and UmuD¢ and
`UmuC proteins [5–7], where ‘Umu’ refers to UV mutage-
`nesis. RecA protein
`is
`required
`to process
`the
`mutagenically inactive UmuD to UmuD¢ — a shorter,
`mutagenically active form of the protein — in a reaction
`analogous to the cleavage of LexA protein [8–10]. Follow-
`ing cleavage of UmuD to UmuD¢ , UmuC and two
`molecules of UmuD¢ associate to form a UmuD¢ 2C com-
`plex, which, in the presence of activated RecA protein
`filament (RecA*), catalyzes ‘error-prone’ translesion syn-
`thesis (TLS) causing mutations at DNA damage sites [11].
`RecA plays a direct biochemical role during SOS mutage-
`nesis that is distinct from generalized recombination and
`coproteolysis [10,12,13] and that is apparently responsible
`for targeting UmuD¢ 2C to a template-lesion site proximal
`to the tip of the RecA* filament [14,15].
`
`Mutagenically inactive complexes formed with UmuD2C
`and UmuD¢ DC are thought to act as a regulatory switch to
`turn off mutagenesis once DNA damage sites have been
`either repaired or bypassed [16]. Early reviews on SOS
`were written by Witkin [17] and Walker [4], and Friedberg
`et al. [18] provide a recent comprehensive review of the
`SOS regulatory system, written prior to the discoveries of
`error-prone E. coli pol IV and pol V.
`
`SOS translesion synthesis reconstituted
`in vitro
`A replication complex confronting a damaged DNA tem-
`plate strand may be likened to a major train wreck
`resulting in ‘derailment’ of the core polymerase and its
`accessory subunits. When faced with excessive amounts of
`DNA damage, the cell sends out an SOS signal, perhaps in
`the form of a segment of single-stranded chromosomal
`DNA bound by RecA protein. A specialized group of pro-
`teins are induced that can copy damaged template sites,
`making errors along the way. The proteins required for
`SOS-induced mutation (also called SOS error-prone repair)
`
`

`

`DNA polymerases involved in genome repair Goodman and Tippin 163
`
`GDA201.QXD 03/22/2000 02:15 Page 163
`
`Figure 1
`
`Model for UV induction of SOS genes in
`E. coli. (a) Binding of the LexA repressor (red
`squares) to regulatory operators upstream of
`SOS genes (black boxes) limits their
`expression under normal growth conditions.
`(b) Upon UV-induced DNA damage, RecA
`protein (green circles) becomes activated to
`RecA* by binding to regions of single-
`stranded DNA. RecA* can then act as a
`coprotease in the autocleavage of LexA
`allowing SOS genes to be turned on.
`
`have been known since the mid 1980s, thanks to extensive
`genetic data from many different laboratories [18]. These
`proteins include UmuC, UmuD¢ , RecA, and pol III
`holoenzyme complex (HE).
`
`In contrast to the extensive progress made in identifying
`the genetic elements required for SOS-induced mutation,
`attempts to identify biochemical roles for the SOS proteins
`were stymied by the insolubility of UmuC protein in
`aqueous solution. Nevertheless, Harrison Echols and co-
`workers [19] succeeded in purifying a denatured form of
`UmuC that, following renaturation, gave rise to low-level
`bypass of a site-directed abasic DNA-template lesion in
`vitro in the presence of UmuD¢ , RecA, and pol III HE.
`There remained considerable difficulties, however, obtain-
`ing reproducible yields and TLS activity using the
`denatured-renatured UmuC protein [20]. These difficulties
`were alleviated following purification of a soluble, native
`UmuD¢ 2C complex [21]. This complex actively catalyzed
`TLS [22••], as did a maltose-binding protein–UmuC
`(MBP-UmuC) fusion protein [23]. Both systems required
`RecA protein to catalyze TLS, with one surprising differ-
`ence: the native UmuD¢ 2C complex did not require the
`presence of pol III core to carry out TLS suggesting that it
`might contain an intrinsic DNA polymerase activity [22••].
`UmuD¢ 2C is a novel error-prone DNA
`polymerase, E. coli pol V
`To resolve the discrepancy between the two studies,
`UmuD¢ 2C was purified from a pol III temperature-sensi-
`tive strain containing a pol II deletion [24••] and was found
`
`to copy undamaged DNA at nonpermissive temperatures
`but required RecA to carry out TLS. A purified mutant
`complex, UmuD¢ 2C104 (Asp101fi Asn), failed to catalyze
`TLS. These data demonstrated that UmuD¢ 2C contained
`an intrinsic error-prone DNA polymerase activity, E. coli
`pol V. It was subsequently confirmed
`that
`the
`MBP–UmuC fusion protein also contained polymerase
`activity in the absence of the pol III core [25].
`
`Biochemical basis of SOS mutagenesis
`Having an in vitro assay available enables the following
`four basic questions to be addressed. What are the roles of
`each of the proteins required to catalyze TLS and most
`importantly what is the biochemical mechanism of pol V in
`relation to RecA protein? What are the efficiencies for
`bypassing diverse types of template DNA damage? How
`does the specificity of nucleotide incorporation measured
`in vitro compare with in vivo mutation spectra for different
`DNA lesions? What can be said about UmuD¢ 2C-catalyzed
`mutations at undamaged template sites?
`
`The proteins involved in lesion bypass are pol V
`(UmuD¢ 2C), RecA, b processivity clamp, g clamp-loading
`complex, and DNA single-stranded binding protein (SSB)
`[22••,24••]. Although pol V alone can form W–C base pairs
`with relatively low efficiency opposite undamaged tem-
`plate sites, it cannot catalyze incorporation opposite the
`commonly occurring abasic, cis-syn T–T dimer, or 6-4 T–T
`photoproduct lesions (M Tang, MF Goodman, unpub-
`lished data). While addition of RecA, b processivity clamp,
`g clamp-loading complex, or SSB stimulates pol V activity,
`
`

`

`GDA201.QXD 03/22/2000 02:15 Page 164
`
`164 Chromosomes and expression mechanisms
`
`Figure 2
`
`Pol V (UmuD¢ 2C) error-prone lesion bypass. Immediately following
`DNA damage and induction of the SOS response, E. coli attempt to
`repair their genome by various error-free mechanisms. (a) If any
`damage escapes these pathways and the replicative pol III HE
`complex encounters a DNA lesion, the pol III core is effectively blocked
`from further DNA synthesis and (b) dissociates from the DNA leading
`to uncoupling of the replication fork. Activated RecA* forms a filament
`on the damaged template and (c) ~40 minutes post-induction of SOS,
`the mutagenically active (UmuD¢ 2C) pol V is formed. The assembly of
`(UmuD¢ 2C) pol V on the 3¢ -OH vacated by pol III core at the site of the
`lesion is believed to be targeted by RecA*. (d) Pol V Mut, consisting of
`UmuD¢ 2C, RecA, b sliding clamp, g clamp loading complex, and SSB
`(not shown), subsequently catalyzes error-prone TLS past a 6–4 T–T
`photoproduct incorporating G preferentially at the 3¢ T leading to Tfi C
`transitions, consistent with genetic data. (e) Synthesis by pol V is
`distributive in the presence of RecA leading to its dissociation
`following the incorporation of only a few nucleotides beyond the lesion.
`Pol III core can then re-assemble on the primer terminus and resume
`replication of the remaining chromosome.
`
`TLS requires the presence of all of the above proteins
`[22••,24••]. We will refer to the UmuD¢ 2C, RecA, b pro-
`cessivity clamp, g
`clamp-loading complex, SSB protein
`
`combination by the term pol V Mut (mutasome), as origi-
`nally suggested by Harrison Echols [11]. Remarkably, RecA
`stimulates pol V activity by 14,000 fold, reflecting its essen-
`tial contribution to a functional pol V Mut [26••]. TLS by
`pol V occurs in the absence of b processivity clamp and g
`clamp-loading complex when non-hydrolyzable ATPg S
`replaces ATP in the reaction, suggesting that RecA filament
`disassembly eradicates pol V’s ability to copy past template
`damage sites. As pol V Mut and the pol III core compete for
`the same 3¢ -primer termini, pol-V-catalyzed TLS is actually
`inhibited in the presence of pol III core [24••]. Neverthe-
`less, pol III HE has a vital role to play in taking over from
`the distributive (i.e. rapidly dissociating) pol V, once the
`lesion is bypassed, to carry out processive replication on the
`next undamaged stretch of template. A model depicting
`pol V Mut catalyzed TLS is shown in Figure 2.
`
`Several important results have begun to emerge from in
`vitro assays using pol V Mut. A comparison of TLS using
`pols III, IV, and V revealed that only pol V Mut was able to
`catalyze efficient bypass of abasic, cis-syn T–T and 6-4
`T–T lesions [26••]. Another key issue concerns the speci-
`ficity of incorporation opposite each lesion. For example,
`in accordance with the in vivo mutation spectra, pol V
`favors incorporation of G opposite the 3¢ -T of the 6-4 pho-
`toproduct resulting in Tfi C transition mutations, whereas
`pols III and IV favor incorporation of A almost exclusively
`[26••]. The TLS and incorporation specificity data taken
`together for these three lesions suggest that pol V Mut is
`responsible for generating most, if not all, SOS mutations
`targeted at DNA damage sites.
`
`Pol V Mut also exhibits remarkably low fidelity when
`copying undamaged DNA, with error rates of about 10–3
`for most transition and transversion base mispairs [26••].
`This observation is consistent with the requirement for
`UmuD¢ and C in order to observe mutations in the absence
`of DNA damage in RecA730 cells with constitutive induc-
`tion of SOS [27]. The recently discovered pol IV (encoded
`by dinB) is also induced as part of the SOS regulon but its
`only known phenotype is in causing an increase in simple
`frameshift mutations on undamaged lambda phage DNA
`[28]. A deletion of the gene encoding pol IV (D dinB) has no
`measurable effect on either targeted on untargeted chro-
`mosomal mutations; however, an increase in F¢ episomal
`frameshift mutations accompanying the overproduction of
`pol IV [29] suggests that pol IV might also act on chromo-
`somal DNA. A recent in vitro study shows that pol IV is
`able to extend mismatched primer 3¢ -ends with unusually
`high efficiency [30••], a property also exhibited by pol V
`Mut [22••].
`
`Like pol V, pol IV can utilize the b processivity clamp and
`g clamp-loading complex resulting in a 3000 fold increase
`in pol IV activity [26••]. E. coli pol IV (DinB) and pol V
`(UmuD¢ 2C) share common sequence elements with two
`yeast polymerases, Rev1 and Rad30, and with their animal
`cell counterparts. These ‘parent’ enzymes make up a
`
`

`

`GDA201.QXD 03/22/2000 02:15 Page 165
`
`DNA polymerases involved in genome repair Goodman and Tippin 165
`
`superfamily of aberrant DNA polymerases that has little in
`common with the well-known polymerase families A
`(e.g. E. coli pol I and Bacteriophage T7 pol), B (e.g. E. coli
`pol II and eucaryotic pols a
`, d , e ), C (e.g. E. coli pol III
`subunit) and X (e.g. eucaryotic pol b ) involved in DNA
`replication and repair [31].
`
`Sequence domains of the UmuC/DinB/Rev1/
`Rad30 superfamily of DNA polymerases
`The UmuC/DinB/Rev1/Rad30 superfamily members
`(Figure 3) possess five highly conserved regions, domains
`I–V, presumed to be involved in binding and catalysis for
`template directed nucleotide incorporation. Site-directed
`mutations eliminating polymerase activity have been
`found in the most conserved residues in domain I
`(Asp8fi His) and domain II (Arg49fi Phe) in E. coli DinB,
`as well as mutations in domain III, including the double
`mutant Asp155fi Ala Glu156fi Ala of Rad30 from Saccha-
`romyces cerevisiae, and Asp103fi Asn in DinB and UmuC
`Asp101fi Asn from E. coli [24••,30••,32••]. It remains to be
`determined whether the substrate-binding or catalysis
`
`steps are affected. Two helix-hairpin-helix DNA-binding
`motifs are found in domains IV and V. A carboxy-termi-
`nal deletion of S. cerevisiae Rad30 resulting in truncation
`of domain V is also devoid of detectable polymerase
`activity [32••].
`
`The four major subgroups of this family can be
`distinguished by their unique domains. The Rev1 subfam-
`ily — so far found only in eucaryotes — has three distinct
`domains. The most amino-terminal is the BRCT (BRca1
`C-terminal) domain, found in many other eucaryotic
`enzymes such as XRCC1 (X-ray cross complementing 1)
`and the breast cancer gene BRCA-1, and is believed to
`mediate protein–protein interactions necessary to form
`coordinated complexes involved in both cell cycle check-
`points and DNA repair [33]. Rev1 proteins require
`association with Rev3–Rev7 (pol z ) in order to perform
`error-prone TLS by incorporating C opposite abasic sites
`[34]. Perhaps the unique carboxy-terminal end of the Rev1
`proteins has been acquired and conserved to mediate this
`specific interaction.
`
`Figure 3
`
`Alignment of some members of the UmuC/DinB/Rev1/Rad30
`superfamily. A schematic representation of the conserved and unique
`domains present in the UmuC/DinB/Rev1/Rad30 superfamily is
`shown. The highly conserved domains I–V containing probable
`catalytic residues that have been mutated in several studies and helix-
`hairpin-helix DNA-binding motifs are denoted above by Roman
`numerals. E. coli UmuC is the least conserved family member followed
`by the newly discovered human Rad30B, which shares the small extra
`region of homology (light blue) found in both the DinB and Rad30
`subgroups. UmuC and human Rad30B both have unique carboxy-
`terminal ends (thin black lines). The DinB subgroup shows remarkable
`
`conservation of three short motifs (shown in purple), which are
`present from E. coli to humans. The C2H2 and C2HC zinc binding
`motifs (shown as green and yellow diamonds respectively) are
`presumed to be involved in DNA binding and perhaps in selective
`targeting. The BRCT domain is shown (pink oval) at the amino-
`terminal end of the Rev1 subgroup. Conserved regions of unknown
`function are found in the amino (pink ovals) and carboxyl termini
`(peach squares) of human and C. elegans DinB. Additional motifs
`conserved within subgroups are indicated by arrows. Amino acid
`lengths are indicated in parenthesis. Ce (C. elegans), Ec (E. coli),
`h (human), Sc (S. cerevisiae).
`
`a
`

`

`GDA201.QXD 03/22/2000 02:15 Page 166
`
`166 Chromosomes and expression mechanisms
`
`Both the DinB and Rad30 subfamilies have acquired extra
`DNA-binding domains represented by zinc fingers or clus-
`ters, although each has done so in its own way. The Rad30
`members in yeast and humans have added the C2H2 zinc
`finger DNA binding motif while the DinB group in the
`higher eucaryotes (Caenorhabditis elegans and humans) have
`selected the C2HC type of zinc cluster ([35•]; B Tippin,
`MF Goodman, unpublished data). These acquisitions
`might increase DinB’s stability on DNA or to lend speci-
`ficity to the type of template and lesion to which it can
`bind. DinB in C. elegans and humans has also acquired two
`other unique amino- and carboxy-terminal motifs, the lat-
`ter containing putative nuclear localization sequences.
`Human Rad30B is an interesting case in that it shares a
`small extra region of homology between domains II and III
`with the DinB proteins but shows slightly more identity at
`the amino-acid level to yeast Rad30, and thus cannot be
`classified definitively into either subfamily [36••]. Like
`UmuC, human Rad30B has an extensive carboxy-terminal
`region that cannot be characterized by any known motifs,
`while maintaining the minimal domain conservation seen
`across the entire family.
`
`Similar folks, different strokes: a functional
`family of aberrant polymerases
`UmuD¢ 2C (pol V) and Rad30 (pol h ) homologs appear to
`have a similar biological role in replicating past template-
`damage sites, but with profoundly different biological
`consequences. E. coli strains with deletions of pol V are
`nonmutable by UV radiation and are incapable of perform-
`ing error-prone TLS, whereas yeast with deletion of the
`gene encoding pol h exhibit a slight increase in UV muta-
`bility [37] and are no longer able to catalyze error-free TLS
`[38]. Therefore, diametrically opposite TLS phenotypes
`are generated: error-prone with E. coli pol V Mut and error-
`free with yeast pol h
`. The human Rad30 homolog has been
`identified as the Xeroderma pigmentosum variant gene
`product, XPV pol h
`[39••], which by analogy with yeast
`pol h
`[32••] bypasses T–T cyclobutane dimers by correct-
`ly incorporating two A’s. Skin cancers occurring in the
`absence of pol h are caused, presumably, by errant copying
`of T–T dimers by the human homolog of yeast pol z
`(Rev3–Rev7) or perhaps using Rad30B.
`
`Nucleotide insertion fidelity appears to be governed by a
`geometrical selection principle in which polymerase active
`clefts are specifically designed to accommodate W–C base
`pairs [40,41]. The low fidelity of the Umu-like enzymes
`might result from a strategic trade-off in which active site
`geometrical constraints have been relaxed allowing TLS to
`occur at a cost of perhaps a 100-fold reduction in
`nucleotide insertion fidelity. Indeed, E.
`coli pol V
`[22••,24••] and yeast [42] and human pol h
`(L Prakash,
`personal communication) make excessive numbers of
`errors when copying undamaged DNA templates in vitro,
`with base substitution error rates of about 10–2 to 10–3. Sev-
`eral possible explanations may account for how the low
`fidelity pol h
`can catalyze error-free TLS. Assuming that
`
`most UV-induced T–T dimers undergo repair prior to
`replication, then the remaining few can still be copied
`accurately 95 to perhaps 99.5% of the time by errant pol h
`.
`It’s also possible that pol h
`is designed specifically to copy
`T–T cyclobutane dimers accurately. Compared with pol h
`,
`which appears targeted to a specific type of DNA damage
`site, E. coli pol V Mut can copy a variety of lesions [26••],
`and although pol V Mut is clearly not meant to replicate
`undamaged DNA, it nonetheless does so occasionally
`causing a marked increase in untargeted mutagenesis [27].
`
`In contrast to pol V Mut, the absence of an effect of dinB
`(pol IV) deletions on SOS lesion-targeted mutations
`implies that E. coli pol IV is not involved in TLS. Instead,
`pol IV might play a role in relieving replication forks
`stalled at misaligned primer–template structures in
`homopolymer runs [43], or at forks impeded perhaps by
`transiently slipped mispairs that are difficult either to
`extend or to proofread [27,44]. Such a role is consistent
`with the observation that frameshift mutation rates in
`phage lambda and on F¢ episomes are dependent on pol IV
`levels [28,29].
`
`Speculation on the biochemical basis of
`somatic hypermutation
`The discovery of the UmuC/DinB/Rev1/Rad30 superfam-
`ily of aberrant DNA polymerases may provide new
`impetus in addressing a vexing question in the field of
`immunology: what is the molecular basis of B cell somatic
`hypermutation? Hypermutation in the variable region of
`immunoglobulin genes is responsible, in part, for generat-
`ing an astounding diversity of antibodies. Yet, the
`molecular mechanisms governing this process remain a
`completely open question [45]. On the basis of studies
`using mouse B cells, at least two components are required
`to generate somatic hypermutation. An arbitrary promoter
`must be located immediately upstream from any arbitrary
`target gene, and at least one downstream V-gene specific
`enhancer element is also required, although two enhancers
`are somewhat more effective [46]. Mutations in the DNA
`target sequence are mostly transitions that are located
`proximal to the end of the promoter and diminish further
`downstream. The hypermutation rate is estimated to be
`about 10–3 per base pair [47], reminiscent of pol V Mut,
`pol IV [26••] and pol h
`[42]. Somatic hypermutation might
`therefore involve the action of an errant DNA polymerase
`targeted to the V-gene by a combination of enhancer and
`promoter binding proteins in a manner loosely analogous
`to RecA targeting of pol V to the site of a lesion [22••,24••].
`
`Conclusions and future directions
`The next stage in the study of the UmuC/DinB/Rev1/
`Rad30 superfamily of polymerases will be to elucidate with
`what proteins the polymerases interact and the mecha-
`nisms of selective targeting. Lesion bypass by pol V has
`been shown to require interaction with the b processivity
`factor of the replicative E. coli pol III HE [22••,24••], while
`RecA is thought to target pol V to lesion sites [13–15]. It is
`
`

`

`GDA201.QXD 03/22/2000 02:15 Page 167
`
`DNA polymerases involved in genome repair Goodman and Tippin 167
`
`not known, however, how the other homologs are targeted
`exclusively to sites of DNA damage nor whether or not
`they interact with replication forks or DNA repair com-
`plexes. Further biochemical studies using mutants of this
`novel family of polymerases should help answer these
`questions. We also anticipate that the recent discovery of
`this superfamily of ‘sloppier copier’ polymerases will
`reveal new insights into the mechanism of somatic hyper-
`mutation.
`
`Acknowledgements
`The comprehensive, standard-setting studies of Evelyn Witkin and
`Harrison Echols in SOS regulation and mutagenesis require special
`mention. We want to express our sincere gratitude to past and present
`collaborators Hatch Echols, Roger Woodgate, Mike O’Donnell, Ramon
`Eritja, John Petruska, Kevin McEntee and to the many graduate and
`postdoctoral students for their important contributions in elucidating
`mechanisms of DNA repair and mutagenesis. This work was supported by
`National Institutes of Health Grants GM42554, GM21422 and AG00093.
`
`References and recommended reading
`Papers of particular interest, published within the annual period of review,
`have been highlighted as:
`• of special interest
`•• of outstanding interest
`Kornberg A, Baker TA: DNA Replication, edn 2. New York:
`WH Freeman and Company; 1992:113-196.
`
`1.
`
`2.
`
`3.
`
`DeLucia P, Cairns J: Isolation of sn E. coli strain with a mutation
`affecting DNA polymerase. Nature 1969, 224:1164-1166.
`
`Rangarajan S, Woodgate R, Goodman MF: A phenotype for
`enigmatic DNA polymerase II: a pivotal role for pol II in replication
`restart in UV-irradiated Escherichia coli. Proc Natl Acad Sci USA
`1999, 96:9224-9229.
`
`4. Walker GC: Inducible DNA repair systems. Annu Rev Biochem
`1985, 54:425-457.
`
`5.
`
`6.
`
`7.
`
`8.
`
`9.
`
`Kato T, Shinoura Y: Isolation and characterization of mutants of
`Escherichia coli deficient in induction of mutagenesis by
`ultraviolet light. Mol Gen Genet 1977, 156:121-131.
`
`Steinborn G: Uvm mutants of Escherichia coli K12 deficient in UV
`mutagenesis. I. Isolation of uvm mutants and their phenotypical
`characterization in DNA repair and mutagenesis. Mol Gen Genet
`1978, 165:87-93.
`
`Sommer S, Knezevic J, Bailone A, Devoret R: Induction of only one
`SOS operon, umuDC, is required for SOS mutagenesis in E. coli.
`Mol Gen Genet 1993, 239:137-144.
`
`Shinagawa H, Iwasaki T, Kato T, Nakata A: RecA protein-dependent
`cleavage of UmuD protein and SOS mutagenesis. Proc Natl Acad
`Sci USA 1988, 85:1806-1810.
`
`Burckhardt SE, Woodgate R, Scheuremann RH, Echols H: UmuD
`mutagenesis protein of Escherichia coli: overproduction,
`purification, and cleavage by RecA. Proc Natl Acad Sci USA 1988,
`85:1811-1815.
`
`10. Nohmi T, Battista JR, Dodson LA, Walker GC: RecA-mediated
`cleavage activates UmuD for mutagenesis: mechanistic
`relationship between transcriptional derepression and
`posttranslational activation. Proc Natl Acad Sci USA 1988,
`85:1816-1820.
`
`11. Echols H, Goodman MF: Mutation induced by DNA damage: a
`many protein affair. Mutation Res 1990, 236:301-311.
`
`12. Dutreix M, Moreau PL, Bailone A, Galibert F, Battista JR, Walker GC,
`Devoret R: New recA mutations that dissociate the various RecA
`protein activities in Escherichia coli provide evidence for an
`additional role for RecA protein in UV mutagenesis. J Bacteriol
`1989, 171:2413-2415.
`
`13. Sweasy JB, Witkin EM, Sinha N, Roegner-Maniscalco V: RecA
`protein of Escherichia coli has a third essential role in SOS
`mutator activity. J Bacteriol 1990, 172:3030-3036.
`
`14. Bailone A, Sommer S, Knezevic J, Dutreix M, Devoret R: A RecA
`protein mutant deficient in its interaction with the UmuDC
`complex. Biochemie 1991, 73:479-484.
`
`15. Frank EG, Hauser J, Levine AS, Woodgate R: Targeting of the
`UmuD, UmuD¢¢ , and MucA¢¢ mutagenesis proteins to DNA by RecA
`protein. Proc Natl Acad Sci USA 199