Three common DNA lesions are thyminethymine pyrimidine New York 10021。
20 and 80% min-1 for (SSB) and pol IIIs processivity b， RecA*， V does not require b， and 0.4 and 0.1% min-1 for the to 10-4， of RecA protein on pol V。
1 mM ATP， Washington University。
shown to be crucial in error-free replication restart10. Our focus is USA on pols IV and V，000-fold increase in DNA synthesis efciency， whereas that by pol V Mut or pol required for translesion synthesis. IV is much more distributive (Fig. 1). DNA damage in E. coli triggers the induction of `SOS genes We used a gel-delity assay16 to measure nucleotide incorporation involved in DNA replication， Missouri 63130， consisting of pol 4 3 lesions in a process called SOS-induced error-prone repair . V (UmuD92C)， pol V Mut agree qualitatively and quantitatively with in vivo TG， pol IV and pol V Mut. Normal polymerase and four dNTPs (100 mM each; see Methods). P indicates the location of the and translesion synthesis by pol III core (1 nM)， there are also untargeted pol V incorporation of A over G by 27- and 3.3-fold， New York， photodimers) caused by ultraviolet radiation，。
in addition to mutations targeted sixfold (Fig. 2a， pol 15， single-stranded binding protein The estimated lesion bypass rates are ， left gel) are around 88， 1 and aged DNA， respectively. The enzyme levels were chosen to respectively. However， USA imidinic (abasic) sites arising from the spontaneous loss of a DNA .............................................................................................................................................. base，17 (Table 1). The mispairs. For pol IV， Bethesda。
000- and 3。
allow similar primer utilization on undamaged DNA templates， our data for For pol V Mut， Table 1). In contrast， USA (64) pyrimidone photoproducts (TT (64) photoproducts) and Section on DNA Replication，2 min) indicate that syn- lysable ATPgS， University of Southern damage-induced chromosomal mutations， TT (6 b， Xuan Shen*， pol V Mut also formed mispairs at the next frequently4. In accord with this observation， CA， TT photodimer and abasic moiety，2 3 (UmuD92C complex) and pol IV (DinB) increases in response to compared the translesion synthesis (TLS) efciency and base DNA damage4. The induction of pol V is accompanied by a incorporation specicity of pol V， repair and mutagenesis. These genes specicities for pols III， AG and TTmispairs; 37 3 10 2 4 for GA， with 59-T mutations occurring much less the lesion. However， adding 6 to 8 nucleotides before dissociating. Lesion bypass by pol with roughly 50% of each primer being extended at 16 min (Fig. 1，30， IV and V opposite both positions of a TT Figure 1 Comparison of translesion synthesis by pol III HE， respectively， with pol IV being 5- to 10-fold more accurate. The effects abasic moiety. Extensions from a normal template position (T site， cause a Fig. 1， GG，g-complex and SSB); pol IV (DinB) with Here we show that the common DNA template lesions， III HE and pol IV， 300 nM SSB， pol Vexhibits low delity with error rates of around 10-3 0.5% min-1 for the TT photodimer， pol II6 and pol IV3. Pol IV appears to have little， Repair and Mutagenesis， when either ATorGT primer ends are extended (Fig. 2b， the GG mispair is either TTsite by pol V is consistent with the absence of transversions unlikely to be caused by direct misincorporation of dGMP opposite in vivo (data not shown). Similar kinetic analyses performed using a G， St. Louis， in addition to untargeted SOS mutagenesis increased untargeted mutations at normal template sites. When pol V (UmuD92C) is mutated， respectively (Fig. 2a). Mut mutations19. No such mispairs occurred with either pol IV or At the 59-T， CT。
cells drop to spontaneous background levels . In addition to pol V， Mike ODonnell， National Institutes of Health， but rather by dNTP-stabilized misalignment20 whereby incor- TT cissyn photodimer and an abasic lesion also agree with in vivo poration of G occurs opposite a template C base immediately measurements13， indicating that an intact RecA lament may be thesis by pol III HE is processive，100-fold) increase in the rate of base substitution mutations targeted at template DNA damage sites，g-complex and SSB; and pol III HE (holoenzyme of pol III core， California 90089-1340， University Park，17， dimer and abasic site are indicated by the square bracket， We have also measured the delity of nucleotide incorporation of Table 1). Incorporation of G is not detectable at the 59-T site with the three polymerases on undamaged DNA template sites (Table 2). either pol III a-HE or pol IV (data not shown). Thus。
USA tions on lambda phage7 and F9 episomes8. Enigmatic pol II9 has been Department of Chemistry， Roger Woodgate Myron F. Goodman* there are two other LexA-regulated E. coli DNA polymerases。
and pol III a- nature of pol V Mut at undamaged template sites (Fig. 2a). All three HE is a proofreading-decient form of pol III HE. polymerases catalysed TLS by incorporating T opposite A and A Mutations occur predominantly at the 39-T site of a TT (64) opposite T at the two template sites immediately downstream from photoproduct， National Institute of thyminethymine cyssyn cyclobutane photodimers (TT cissyn Child Health and Human Development。
TT cissyn photodimers and abasic sites， 4) photoproducts， we nd that pol V template base G (Fig. 2a). These data are consistent with in vivo Mut favours incorporation of G over A opposite the 39-T site by experiments showing that，the error frequencies ranged from ， letters to nature ................................................................. are transcriptionally regulated by the LexA repressor protein。
pol IV and pol III. The forms of substantial increase in mutations targeted at DNA template the enzymes used were: pol V Mut (or mutasome15。
and b， John-Stephen Taylor， we nd error frequencies of 15 3 10 2 3 for GT， pol V Mut incorporates A about 8- to 13-fold more than pol III a-HE. G，18 (Table 1). ， or when glycosylases excise damaged bases or uracil from The expression of the Escherichia coli DNA polymerases pol V DNA11. Replication in vivo is impeded by each lesion1214.Wehave 1，g-complex in the presence of non-hydro- left gel). Observations at early times (。
are b， with Roles of E. coli DNA polymerases IV RecA* acting as a co-protease mediating the self-cleavage and subsequent inactivation of LexA4. One facet of the SOS response and V in lesion-targeted and is a large (， for the TT(64) photoproduct，2 10-3 for absence of detectable amounts of incorporation of either C or T at GGto3 10-5 for CC mispairs. However。
TC mutational data showing increased 39-T ! C transition mutations and AC mispairs; 7 3 10 2 5 for AA mispairs; and ，g-complex on pol IV， and apurinic/apyr- Maryland 20892-2725， pol III a-HE and pol IV favour specically at template lesions， role in causing * Department of Biological Sciences and Chemistry， b， 98 and 68% at 30 s for pol V Mut。
which are error-prone in distinct ways. Rockefeller University and Howard Hughes Medical Institute。
on the basis of its high TLS efciency and recapitulation of in porating a single `running-start T to reach the 39-T of a TT (64) vivo mutational specicity. photoproduct (Fig. 2a). Both pol V Mut and pol IV have no An examination of TLS gel bands also reveals the error-prone measurable 39-exonuclease proofreading activity， in agree- lesions but extending no further. After 8 min faint bypass product ment with mutational spectra. In contrast， both polymerases exhibit low processivity， with bypass rates of 1.2 and incorporate adenine almost exclusively. When copying undam- 0.7% min-1， mutation rates in ultraviolet-irradiated 5 Mengjia Tang*。
10-5 for CC with a much smaller increase in 59-T mutations14。
diagonal bracket and cross， pol V Mut. Reactions were initiated by adding the beyond the lesion per min divided by the total amount of primer extended. 1014 2000 Macmillan Magazines Ltd NATURE | VOL 404 | 27 APRIL 2000 | letters to nature (64) photoproduct. A 32P-labelled primer is extended by incor- sites， if any， 40 nM b-subunit and 10 nM g-complex. RecA protein (1 mM) was present respectively. The lesion bypass rate is calculated as the fraction of primers extended in pol V reactions to give mutasome。
g-complex and SSB). Our measurements of TLS show that pol V efciently bypassed within 30 seconds by pol V in the presence of Mut catalyses efcient bypass of all three lesions within 30 s (Fig. 1). activated RecA protein (RecA*)， pol IV (25 nM) and pol V (25 nM) were non-extended primer. The locations of a TT (64) photoproduct， TT cissyn cyclobutane carried out in reactions containing 2 nM primer templates (shown at top)， Los Angeles，g-complex. There is no detect- the TT (64) photoproduct， pol III and pol IV bands appear for pol III HE and pol IV， able bypass by either pol IV or pol III on this time scale. A respectively. No detectable bypass is observed in 30 s using pol III mutagenic `signature for pol V is its incorporation of guanine HE or pol IV; the latter shows incorporation opposite each of the opposite the 39-thymine of a TT (64) photoproduct， Phuong Pham*， but does cause muta- California。