Molecular Cell, Vol. 7, 571–579, March, 2001, Copyright 2001 by Cell Press
SOS Mutator DNA Polymerase IV
Functions in Adaptive Mutation
and Not Adaptive Amplification
ity is produced for growth on lactose medium without
acquisition of a Lac?point mutation. Adaptive point
mutation and amplification are separate adaptive re-
sponses and are both different from Lac?mutation in
The adaptive point mutation mechanism at lac can be
summarized as follows. The adaptive mutations occur
and require homologous recombination proteins of the
et al., 1994, 1996; Foster et al., 1996). DSBR is proposed
DNA polymerase errors occur (Harris et al., 1994).
Whereas growth-dependent Lac?mutations are hetero-
geneous, the adaptive mutations are nearly all ?1 dele-
tions in small mononucleotide repeats (Foster and Tri-
marchi, 1994; Rosenberg et al., 1994), resembling DNA
anism (Streisinger et al., 1966; reviewed in Ripley, 1990).
The adaptive mutations accumulate during a transient
vation (Longerich et al., 1995; Harris et al., 1997b, 1999).
The adaptive mutants, once formed, bear high frequen-
cies of unrelated mutations throughout their genomes,
indicating that some or all of the adaptive mutants arise
son et al., 1997; Rosche and Foster, 1999; Bull et al.,
2000a; Godoy et al., 2000; and see Bull et al., 2000b;
Cairns, 2000 for further discussion). Finally, efficient re-
combination-dependent adaptive mutation requires a
functional SOS response for upregulation of a protein(s)
other than or in addition to RecA (McKenzie et al., 2000).
One infers that both recombination and SOS are re-
quired because recombination genes are required that
are not also required for an SOS response (Foster et al.,
1996; Harris et al., 1996).
The enzymatic basis of the mutability underlying
adaptive mutation at lac has not been elucidated fully.
sible. On the one hand, the postreplicative mismatch
repair (MMR) system (reviewed by Modrich and Lahue,
1996) becomes limiting transiently during adaptive mu-
tation (Harris et al., 1997b, 1999), and genetic evidence
implicates the major replicative DNA polymerase, pol
III, in adaptive mutation (Foster et al., 1995; Harris et al.,
1997a). Therefore, a normal rate of DNA polymerase
error could lead to mutability because of failure to cor-
rect those errors. On the other hand, the involvement of
the SOS response suggests (among other possibilities)
that special mutator enzymes controlled by SOS could
be responsible (McKenzie et al., 2000). The umuDC-
(Cairns and Foster, 1991; McKenzie et al., 2000). This
study examines the other SOS mutator polymerase, pol
IV, encoded by dinB.
Pol IV is a poorly processive error-prone DNA poly-
merase (Wagner et al., 1999; but see Tang et al., 2000;
Wagner et al., 2000) and a member of the large, newly
elaborated DinB/UmuDC superfamily of DNA polymer-
ases in bacteria, archaea, and eukaryotes (reviewed by
Gregory J. McKenzie,* Peter L. Lee,*
Mary-Jane Lombardo,* P. J. Hastings,*
and Susan M. Rosenberg*†‡§
*Department of Molecular and Human Genetics
†Department of Biochemistry
‡Department of Molecular Virology and Microbiology
Baylor College of Medicine
Houston, Texas 77030-3411
Adaptive point mutation and amplification are induced
changes that can enhance survival. A specialized adap-
Escherichia coli assay, but its enzymatic basis re-
mained unclear. We report that the SOS-inducible, er-
ror-prone DNA polymerase (pol) IV, encoded by dinB,
is required for adaptive point mutation in the E. coli lac
operon. A nonpolar dinB mutation reduces adaptive
vival after oxidative or UV damage. We show that pol IV,
together with the major replicase, pol III, can account
tify a role for pol IV in inducible genetic change.
Radman (1975), Echols (1981), and others have sug-
gested that states of accelerated evolution might be
induced in response to stress and that enzymes might
amplification in bacteria (Hastings et al., 2000), support
the idea of differentiated states of hastened genetic
change (reviewed by Rosenberg, 2001). Adaptive muta-
tion is a process of increased mutability that occurs in
allowing survival. There are many assay systems for its
but in only one so far has adaptive mutation been dem-
onstrated to occur by a molecular mechanism different
from spontaneous mutation in growing cells (and so to
be a separate process). That assay measures reversion
of a lac ?1 frameshift allele carried on an F? episome
in Escherichia coli (Cairns and Foster, 1991). In the lac
system, one distinct mechanism produces adaptive
point mutations, conferring a Lac?phenotype via com-
pensatory frameshift mutations. Also in the lac system,
a separate adaptive response produces adaptive ampli-
amplification, the leaky lac mutant gene is amplified to
many copies such that sufficient ?-galactosidase activ-
Friedberg et al., 2000). The discoveries of multiple DNA
polymerases in all living organisms have raised the
question of why cells have so many (e.g., five are known
currently in E. coli). What are their functions? Some of
the DinB/UmuDC polymerases are translesion polymer-
ases known to promote DNA damage survival by allowing
replication to bypass otherwise replication-blocking le-
sions. The human XP-V (xeroderma pigmentosum vari-
ant) tumor suppressor protein (of the Rad30 subfamily)
However, the function(s) of pol IV (DinB subfamily) and
three ofits mammalian homologs(Friedberg etal., 2000)
have been elusive. Pol IV may participate in mutation of
undamaged phage ? DNA during infection of irradiated
E. coli (? untargeted mutagenesis; Brotcorne-Lannoye
and Maenhaut-Michel, 1986). Pol IV overproduction
causes hypermutation including –1 frameshifts and
2000). The purified pol IV enzyme makes similar errors
(Wagner et al., 2000).
point mutation at lac, but not for mutations in growing
cells, survival of UV or oxidative damage, or adaptive
amplification. Thus, one function of pol IV in E. coli in-
volves environmentally inducible genetic change.
cating that DNA pol IV function is required for most
be complemented with a single, ectopic chromosomal
copy of dinB?(Figure 1B), indicating that the decrease
in adaptive mutation is caused solely by the loss of pol
IV, and not other genes in the putative dinB operon.
We note that a single chromosomal copy of dinB?is
sufficient for adaptive mutation at lac (Figure 1B), con-
trary to the suggestion that expression of the extra copy
of dinB on the F? might be required (Godoy et al., 2000).
These results indicate a biological role for pol IV: it pro-
motes adaptive mutation.
The amount of adaptive point mutation requiring pol
in Figure 1. About 42% of the day 5 (i.e., adaptive) Lac?
colonies that remain in the pol IV-deficient strain carry
amplified arrays of the leaky lac?allele rather than a point
mutation, as compared with 9.5% for dinB?(Figure 1A).
These classes were distinguished by their colony color
after purification by streaking for single colonies onto
rich X-gal medium (Experimental Procedures). The fact
that amplified clones are about 40% of day 5 colonies
in pol IV-deficient cells indicates that the reduction in
adaptive point mutation in pol IV-deficient cells is actu-
ally about 85% (25% Lac?mutants seen, 60% of which
(Figures 1A and 2B). Thus, the vast majority of the adap-
tive point mutation is pol IV dependent.
In addition, the data show that pol IV is not required
for adaptive amplification. Amplified clones constitute
?10% of Lac?colonies in pol IV?cells (above) and
?40% of Lac?colonies in pol IV?, in which the total
number of Lac?colonies is reduced 4-fold (25% of that
seen in pol IV?). Thus, the number of amplified clones
in pol IV?cells is approximately the same as in pol IV?
(40% amplified of 25% total colonies equals 10%). Pol
IV is therefore required specifically for adaptive point
mutation and not for adaptive amplification.
To test whether pol IV is also required for growth-
dependent mutation, we measured the mutation rate in
dinB?and dinB?growing cells using fluctuation tests
(in which mutant frequencies determined in multiple in-
dependent cultures are used to calculate rates; Experi-
mental Procedures). To exclude adaptive mutants from
the counts of growth-dependent Lac?mutants, we ac-
quired ten independent Lac?mutant derivatives of the
dinB?and dinB10 strain. These were seeded at a known
experimental conditions, in parallel with the cultures in
which growth-dependent mutants were being enumer-
ated. These controls indicate the earliest possible time
which the seeded Lac?control colonies become visible)
(Harris et al., 1999). Failure to use these controls can give
uninterpretable results, because both growth-depen-
dent and adaptive mutants contribute to the colony
counts from which mutation rates are calculated (Harris
et al., 1994, 1996, 1997b, 1999). The results in Table 1
show that pol IV is not required for growth-dependent
mutation at lac.
We find that pol IV mutation does not affect the rate
of other growth-dependent mutations, including substi-
tutions, frameshifts, and other mutations in growing
a functional DNA pol IV, encoded by dinB, we con-
structed isogenic dinB?and mutant strains. dinB is the
first gene in an apparent operon of four damage-induc-
ible (Courcelle et al., 2001) SOS genes: dinB, yafN, yafO,
and yafP. The yaf genes have unknown functions,
though YafN has homology to the anti-toxin of the relBE
operon (Grønlund and Gerdes, 1999). All of these genes
are likely to be inactivated by previously published null
alleles of dinB: a deletion of dinB and part of yafN (Kim
To remove only pol IV function, we created a nonpolar
null allele of dinB identical to dinB10 (Wagner et al.,
1999), which replaces a highly conserved amino acid
(R49F), producing a mutant polymerase that is inactive
in vitro and does not enhance mutation when overpro-
duced in vivo. The lac frameshift–bearing strain carries
two copies of the dinB?gene, one on the F? and one in
the chromosome (Experimental Procedures). We con-
structed strains carrying dinB10 at both sites.
In adaptive mutation assays, Lac?cells are plated
onto lactose medium and incubated for several days
(Experimental Procedures). Lac?mutant colonies that
appear early (about day 2) represent growth-dependent
mutants formed before plating on lactose medium
(Cairns and Foster, 1991; see Harris et al., 1999). Colo-
nies that appear late (e.g., day 3–7) consist of a majority
of adaptive point mutants and a minority of adaptive
amplified clones, both formed after plating on lactose
medium (McKenzie et al., 1998; Hastings et al., 2000).
Pol IV Is Required Specifically for Adaptive Point
Mutation at lac
Replacement of both copies of dinB?with dinB10 re-
duces adaptive mutation about 4-fold (Figure 1A), indi-
Mutator Pol IV in Adaptive Mutation
Figure 1. DNA Polymerase IV Is Required for
Most Lac?Adaptive Mutation
(A) Total Lac?colonies are shown as open
symbols. Lac?point mutants (see text) are
plotted as filled symbols offset slightly from
colonies carrying amplification (Experimental
Procedures) was 9.5% (mean ? 2.6% SEM) in
the dinB?and 42% (? 5.6%) in the dinB10
(B) Decrease in mutation is complemented by
a single, ectopic, chromosomal copy of dinB?
controlled by its natural promoter. dinB?(open
squares), dinB10 (open diamonds), dinB?
?attB::dinB?(open circles), and (?) dinB10
?attB::dinB?(open triangles) strains SMR4562,
SMR5830, SMR5834, and SMR5851, respec-
tively. Means ? SEM (error bars) of ten inde-
pendent cultures tested are shown (except
for the filled symbols, mean ? SEM of four
cultures). Where not visible, error bars are
smaller than the plot symbol. Daily measurements of viable lac?cells on the plates (Relative cell number), shown normalized to the first day’s
count, show no net growth or death during the experiments (mean ? SEM, four cultures).
cells (Figure 3). We conclude that pol IV is required
specifically for adaptive mutation.
Our results disagree with a previous study, in which a
dinB mutation appeared to decrease the rate of growth-
dependent Lac?mutation slightly (Strauss et al., 2000).
The reason for the difference may be that the earlier
study did not account for adaptive mutations. Alterna-
tively, the small rate change may have been due to the
use of a polar dinB allele, which also disrupted genes
downstream of dinB.
Pol IV is also not required for survival of UV irradiation
and oxidative damage caused by hydrogen peroxide.
As seen in Figure 4, the dinB10 mutant is indistinguish-
able from an isogenic dinB?strain in UV survival and
hydrogen peroxide resistance. Control isogenic strains
carrying the lexA3(Ind?) mutation, blocking SOS gene
induction, or a mutation in xthA, encoding an exo-
nuclease required for repair of peroxide-induced dam-
age (Demple et al., 1983), show reduced resistance, as
SOS/LexA Induction Promotes Adaptive Point
Mutation Wholly via Pol IV
Because pol IV is one of the genes induced by the SOS
response (reviewed by Walker,1996), we asked whether
pol IV alone can account for the requirement for SOS
induction in adaptive point mutation (Cairns and Foster,
1991; McKenzie et al., 2000; Figure 2A). If induction
of additional SOS-induced genes were required, then
dinB10 lexA3(Ind?) cells (SOS noninducible due to an
uncleavable mutant LexA repressor) should produce
fewer adaptive mutations than dinB10 cells. However,
our experiments showed that the rate of adaptive muta-
tion in both genetic backgrounds is the same (Figure
2B), implying that induction of SOS genes that act inde-
pendently of pol IV is not required. Thus, genes such
Figure 2. Different Roles of SOS Induction in
Adaptive Amplification and Point Mutation
(A) Induction of the SOS/LexA regulon is not
required for adaptive amplification. Total
adaptive Lac?colonies (open symbols) are
decreased by the lexA3(Ind?) allele (open tri-
angles), whereas the fraction amplified (filled
symbols) is not. lexA?(squares) and lex-
A3(Ind?) (triangles) strains SMR583 and
(B) The contribution of SOS/LexA induction
Closed symbols display adaptive Lac?point
mutants for dinB?lexA?(squares), lex-
A3(Ind?) (triangles), dinB10 (diamonds), and
dinB10 lexA3(Ind?) (circles) strains SMR583,
SMR820, SMR5849, and SMR5850, respec-
tively. This is the same experiment shown in
(A) but with data from more of the strains
tested in parallel shown, and point mutation
displayed. Both sets of experiments were performed three times with similar results. In (B), the total adaptive Lac?colonies are also shown
for the dinB?lexA?control strain (open squares). Means ? SEM (error bars) of ten independent cultures tested are shown (except for the
filled symbols, mean ? SEM of four cultures). Where not visible, error bars are smaller than the plot symbol. Daily measurements of viable
lac?cells on the plates (Relative cell number), shown normalized to the first day’s count, show no net growth or death during the experiments
(mean ? SEM, four cultures).
Table 1. DNA Polymerase IV Does Not Affect lac Frameshift Reversion in Growing Cells
Mutation Rate to Lac?
of Mutants Experiment Mean (? SEM)
3.1 ? 10?9
1.9 ? 10?9
1.5 ? 10?9
1.8 ? 10?9
4.5 ? 10?9
1.2 ? 10?9
1.3 ? 10?9
1.1 ? 10?9
1.6 (? 0.3) ? 10?9
1.2 (? 0.3) ? 10?9
Strains are dinB?, SMR4562 and dinB10, SMR5830. See Experimental Procedures.
as the recA, ruvA, andruvB recombination genes, which
are required for adaptive mutation, appear to suffice
at their noninduced (constitutive) levels. These results
suggest that the requirement for SOS induction in adap-
tive point mutation (Cairns and Foster, 1991; McKenzie
et al., 2000; Figure 2A) may be accounted for solely by
Pol IV Contributes to –1 Deletions in a Variety of
Lac?adaptive mutations are nearly all –1 deletions in
small mononucleotide repeats (Foster and Trimarchi,
1994; Rosenberg et al., 1994). In the presence of wild-
type dinB?, most occur at a reversion hot spot (4Cs that
include the ?1 frameshift mutation inactivating lac), but
a significant portion (about one-third) occurs at other
mononucleotide repeats. We find that in the absence of
pol IV, –1 frameshifts occur mostly at the hot spot (24/
31 mutations sequenced, Figure 5), with other point mu-
tations being larger insertions and deletions or not at
mononucleotide repeats (Figure 5). The data imply that
pol IV facilitates –1 deletions at many different mono-
nucleotide repeats, mutations similar to the frameshift
component of the error spectrum of the purified poly-
merase (Wagner et al., 1999).
Induction of LexA/SOS Genes Is Not Required for
The SOS response was previously shown to be required
for adaptive point mutation. We tested whether SOS-
induced genes are also required for adaptive amplifica-
tion. We found that blocking induction of the SOS/LexA
able LexA repressor protein; Mount et al., 1972; Lin and
Little, 1989)decreases only point mutation,not adaptive
amplification (Figure 2A, filled symbols). Thus, only
adaptive point mutation, and not adaptive amplification,
requires induction of LexA controlled genes, supporting
the conclusion that these are separate pathways.
Overlapping Roles of Pol III and Pol IV
Previous data suggested that pol III may play a role in
adaptive point mutation. An antimutator pol III strain
decreased the total number of adaptive Lac?mutations
by about 4-fold (Foster et al., 1995; Harris et al., 1997a).
In agreement with these results, we find that the antimu-
tator pol III (encoded by dnaE915) reduces the number
of adaptive point mutations by about 80% (Figure 6).
Thus, neither pol IV mutation nor an antimutator pol III
inhibits all adaptive point mutation. However, in cells
carrying dnaE915 and a defective pol IV (circles), adap-
tive point mutation is essentially abolished (Figure 6).
both the pol IV-dependent and the pol IV-independent
adaptive point mutations, indicating overlapping roles
for pol III and pol IV in this process (discussed below).
Figure 3. Rates of Frameshift and Substitution Mutations in dinB?
and dinB10 Cells during Growth
The various frameshift and substitution mutation assays (see La-
cine/valine biosynthesis genes, conferring valine resistance; Strep
and Spec, substitution mutations in two ribosomal protein genes
conferring streptomycin and spectinomycin resistance, respec-
tively; Nal, substitution mutations in the gyr genes conferring nali-
dixic acid resistance; and Tet, reversion of a ?1 frameshift mutation
(4G to 5G, Experimental Procedures) in a chromosomal tetA gene
conferring tetracycline resistance. This is similar to the 3G to 4G
lacI33 frameshift allele used in these adaptive mutation studies.
dinB?(filled bars) and dinB10 (hatched bars) strains are SMR4596
andSMR6049, respectively. Error bars, one SEM of three indepen-
The data presented in this paper imply that the SOS
mutator DNApolymerase polIV isa mutation-promoting
enzyme required specifically for most (about 85% of)
adaptive point mutation (Figure 1), but not for growth-
dependent Lac?(Table 1) or other (Figure 3) mutation.
at a variety of mononucleotide repeats (Figure 5), similar
to the frameshift component of the error spectrum of
the purified enzyme (Wagner et al., 1999). Further, pol
IV can account for the requirement for SOS induction
in the lac system (Figure 2B, Cairns and Foster, 1991;
McKenzie et al., 2000). Finally, pol IV is not required for
Mutator Pol IV in Adaptive Mutation
Figure 4. Loss of Pol IV Confers No Detect-
able Change in Survival of UV or Oxidative
(A) UV sensitivity. Four cultures per strain
were tested, and the means ? SEM (error
bars) are shown. DinB?(open squares),
dinB10 (open diamonds), and lexA3(Ind?)
(open circles), strains SMR4562, SMR5830,
and FC231, respectively.
(B) Sensitivity to hydrogen peroxide. Four cul-
tures of each strain were tested in parallel,
and the mean ? SEM are shown. Strains are
asin (A)withthe additionof SMR5287lacking
exonuclease III (encoded by xthA), used in
base excision repair of oxidatively damage
resistance to UV light (Kenyon and Walker, 1980; Figure
4) or hydrogen peroxide (Figure 4).
to be required for both amplification and point mutation
The results also reveal that neither pol IV nor induction
of SOS/LexA-controlled genes is required for adaptive
amplification of lac (Figures 1 and 2A). These data add
distinct by showing that they require different proteins.
These data also suggest that the role of pol IV (and SOS
induction) is in error-prone DNA synthesis that gener-
ates adaptive point mutations, but not generally in DNA
synthesis in stationary phase, which would be expected
Contributions of Pol IV and MMR Limitation to
Mutability and the Characteristic Sequences of lac
Adaptive Point Mutations
The requirement for an error-prone polymerase, pol IV,
in adaptive point mutation supports models in which
special error-prone synthesis leads to mutation, making
previous models invoking depressed mismatch repair
(MMR) as the sole basis of mutability implausible. How-
ever, limiting MMR also appears to contribute. First,
apart from resembling the frameshift errors made by pol
Figure 5. DNA Pol IV Promotes –1 Deletions
at a Variety of Mononucleotide Repeat Sites
in Lac?Adaptive Mutation
A roughly300 nucleotide (nt) segmentof DNA
spanning the lac frameshift allele was se-
quenced from PCR-amplified DNA from day
5 dinB10 Lac?point mutants (primers lacIL2
5??AGGCTATTCTGGTGGCCGGA, and lacD2-
ing was performed by Lone Star Labs, Inc.
(Houston, TX). Compensatory frameshift mu-
tations in a possible 130 nt region between
the two out-of-frame stop codons (boxed)
can restore gene function. In dinB?cells,
adaptive reversions are –1 deletions at a hot
spot (nt 1039) and at many different mono-
1059, 1064, 1071, 1075, and 1109, data from
Rosenberg et al. 1994). In dinB10 cells, only
the hot spot repeat is appreciably active for
–1 repeat deletions, and other insertions and
deletions are also prevalent. The other muta-
tions include a –1 frameshift with an adjacent
substitution (at nt 1094–5); a ?2 insertion (nt
1092); an insertion of ?40 bp (from 3? of the
sequenced area to nt 1120); and three large
deletions of 103 bp (nt 1017–1119), 103 bp
(979–1081), and 211 bp (nt 878–1088). Nt re-
peat positions are indicated above the left-
most base covered by the number, and the
additional base of the original ?1 frameshift
mutation in the repeat at nt 1039 is not num-
mutation as follows. First, the error-free SOS DNA pol
II appears to compete with the polymerase(s) making
adaptive mutations, in that pol II-deficiency increases
adaptivemutation (Fosteretal.,1995; Harris,1997).Per-
haps pol II competes with pol IV at the replisome. Sec-
ond, an anti-mutator pol III allele decreases Lac?adap-
tive mutation ?3-fold (Foster et al., 1995; Harris et al.,
1997a), decreasing both pol IV-dependent and pol IV-
independent point mutation (Figure 6). The apparent
overlap between pol III and pol IV (Figure 6) can be
understood by hypotheses in which pol III and pol IV
compete with and/or substitute for each other on DNA
(e.g., Friedberg et al., 2000; Tang et al., 2000). In one
general model, pol IV makes the errors that become
mutations. This is supported by the similarity of the
sequence spectrum of adaptive mutations attributable
to pol IV (Figure 5) with the frameshift error spectrum
of the purified polymerase (Wagner et al. 1999). The pol
III antimutator protein might exclude pol IV from DNA
(and might then lower pol IV-independent point muta-
tions by excluding some other polymerase). Alterna-
tively, pol IIImight correct errors made by polIV. It could
also do both. In another general model, pol III could
make errors that are fixed as mutations by pol IV (see
Tang, et al. 2000). Other hypotheses are also possible.
Whichever may be the case, the data indicate involve-
ment of both polymerases and suggest that replisomes
may exchange pols II (see above), III, and IV.
Figure 6. Overlapping Roles of Pol III and Pol IV in Adaptive Point
Open symbols are total Lac?colonies, and filled symbols point
mutants only for strains carrying dinB?dnaE?(squares), dinB10
(diamonds), dnaE915 (triangles), and dinB10 dnaE915 (circles):
SMR6113, SMR5945, SMR6114, and SMR5944, respectively. The
experiment was performed three times with similar results. Means ?
SEM (error bars) of ten independent cultures tested are shown.
Where not visible, error bars are smaller than the plot symbol. Daily
measurements of viable lac?cells on the plates (Relative cell num-
ber), shown normalized to the first day’s count, show no net growth
or death during the experiments (mean ? SEM, four cultures).
Function of Pol IV for E. coli
A biological function can now be assigned to pol IV, a
polymerases, in adaptive mutation. Is this its only func-
tion? Other polymerases in the UmuDC, Rad30, and
Rev1 branches of this superfamily are translesion poly-
merases (Friedberg et al., 2000), but the evidence for
pol IV is ambiguous. Purified pol IV deals poorly with
tive cells are not sensitive to UV (Kenyon and Walker,
1980, and Figure 4A) or hydrogen peroxide (Figure 4B).
Although, together with pol V, pol IV was implicated in
synthesis across benzo(a)pyrene adducts (Napolitano
et al., 2000), that study used a deletion of dinB and
part of yafN (probably also polar on yafO and yafP, see
Experimental Strategy), making the conclusion uncer-
of pol IV, it is a different role than the one pol IV plays
in adaptive mutation because the former requires pol
V (Napolitano et al., 2000), whereas the latter is pol V
independent (Cairns and Foster, 1991; McKenzie et al.,
2000). Pol IV might facilitate DNA replication promoted
by DSBR recombination, the proposed source of repli-
cation in adaptive mutation (Harris et al., 1994). Yeast
Rev3, or pol zeta (Rev1 subfamily), promotes substitu-
tion mutations associated with yeast DSBR (Holbeck
and Strathern, 1997). Regardless of other possible func-
tions of pol IV, its central role in adaptive mutability
recalls suggestions of enzymes specialized for mutabil-
ity (Radman, 1975; Echols, 1981, and others subse-
quently), accelerating evolution when needed.
IV enzyme, the Lac?adaptive mutation sequences are
postsynthesis MMR (Longerich et al., 1995). Second,
MMR limitation has been demonstrated to occur, and
to be required for, efficient adaptive mutation in this
and limiting MMR are therefore both implicated and
might possibly be related. For example, Wagner and
Nohmi (2000) report that pol IV overproduction causes
an insufficiency of MMR activity that can be alleviated
by overproducing MutL. MutL also becomes limiting for
MMR during adaptive mutation (Harris et al., 1997b,
1999) and in mutants with an error-prone DNA polymer-
ase III (Schaaper and Radman, 1989). In all these cases,
it could be that excess polymerase errors titrate MMR,
for Kim et al. (1997), pol IV overproduction did not pro-
cells. This implies that MMR was not limiting in their
overproduction experiments. Whether the demonstrated
MMR limitation during adaptive mutation (Harris et al.,
1997b, 1999) is caused by, or independently of, pol IV-
produced errors, the combination is likely to interact
synergistically to produce a condition of hypermutation.
Role of This Adaptive Mutation Mechanism
in Bacterial Evolution
ing genes. Is recombination-dependent adaptive muta-
Roles of Other DNA Polymerases
E. coli has five DNA polymerases. Pols II, III, and IV
have been implicated in the synthesis during adaptive
Mutator Pol IV in Adaptive Mutation
NR9918 (Fijalkowska et al., 1993). SMR4576 and SMR6049 carrying
upp::Tn10dtet?1 (with a 4G to 5G frameshift at bp 331 of tetA;
Foster, 1997) are described by H. J. Bull, M.-J. Lombardo, and S. M.
Rosenberg (unpublished data).
tion generally relevant to bacterial evolution? First, in
adaptive mutation at lac, substitutions probably also
tions as well as frameshifts (Kim et al., 1997; Wagner
and Nohmi, 2000). Second, many pathogenic bacteria
stress) by frequent frameshift mutation events that turn
gene functions off and on (e.g., Deitsch et al., 1997;
Saunders et al., 2000). These bacteria might employ
adaptive mutation strategies similar to those discussed
here. In fact, the pathogens Neisseria meningitidis and
N. gonorrhoeae have one or more genes homologous
to dinB (open reading frame NMB1448 in strain MC58,
Tettelin et al., 2000; and NMA1661 in strain Z2491, Park-
hill et al., 2000). Third, regarding the relative importance
of inducible mutation mechanisms, versus selection of
preexisting mutator strains, we note that the mutator
(LeClerc et al., 1996; Matic et al., 1997; Denamur et al.,
2000; Oliver et al., 2000). The majority of wild bacteria
(80%–99%) are not mutators, such that adaptive muta-
tion strategies may contribute appreciably (Rosenberg
et al., 1998; Hastings et al., 2000).
Mutation and Amplification Assays
et al., 1996). Daily measurements of viable lac?cells on the plates
ments. Growth-dependent Lac?mutation measurements used 40
tube fluctuation tests, as described (Harris et al., 1999). Mutation
rates were calculated by the method of the median (Lea and Coul-
son, 1949; as modified by von Borstel, 1978). Other mutations rate
assays used 30 tube fluctuation tests with TetR, ValR, and NalRcalcu-
lated by the method of the median and StrepRand SpecRby the P0
method (Lea and Coulson, 1949; von Borstel, 1978; correction for
P0as per Rosche and Foster, 2000). Because TetRcolonies continue
to appear over time, TetRassays were done with TetRcontrols as
described for Lac (Harris et al., 1999, Results), to exclude mutants
formed on the Tet plates and were scored at 12 hr (90%–100% of
the control colonies visible). Selection agents were tetracycline, 10
?g/ml;valine, 5?g/ml;streptomycin, 100?g/ml; spectinomycin,100
?g/ml; and nalidixic acid, 10 ?g/ml.
The fraction of Lac?colonies carrying amplification rather than
point mutation was determined in dinB?and dinB10 day 5 Lac?
colonies (40 colonies/culture, four independent cultures) of each
strain as previously described (Hastings et al., 2000) by picking and
restreaking Lac?colonies to LBH X-gal rifampicin medium to test
instability of the Lac?phenotype. Unstable Lac?carry roughly 30
copies of lac?amplified DNA in direct repeats of 7–40 kb (Hastings
et al., 2000). This method was also used for Figures 2 and 6.
its mammalian homologs function similarly? The mouse
pol IV homologs, pol ? and pol ?, and true ortholog,
DinB1 or pol ? (each also in humans), are abundant in
lymphoid (?) and germline cells (? and ?), respectively
(Friedberg et al., 2000). Their functions are unknown,
although roles in somatic hypermutation (Friedberg et
ulin and/or T cell receptor genes seem possible. Could
there be programmed mutation, driving evolution, in
germ cells of mammals? As with the immune system,
selections against deleterious mutations are stringent
in germ cells (successful completion of development)
such that programmed germline mutation/evolution
might not be impossible.
UV and Oxidative Damage Survival Assays
Diluted saturated cultures (four/strain) in LBH medium (e.g., Torkel-
son et al., 1997) were plated on LBH plates and irradiated in a
ide (H202) was measured as described (Demple et al., 1983), splitting
log phase LBH cultures, exposing half to 5.6 mM H2O2(and half to
H202-free control medium) for 15 min, and plating for viable cells.
We thank R.S. Harris for identifying the merodiploidy of dinB, G.
Church for plasmid pKO3, H. Ohmori for discussion, our friend M.
Radman for discussion of our results on this study over the past 4
years, A. Beaudet, H.J. Bull, E. Friedberg, P. Hanawalt, J. Lupski,
and D. Nelson for comments on the manuscript, and M. Price, I.
Aponyi, and J. Ross for technical assistance. We acknowledge B.A.
Roe, L. Song, S. P. Lin, X. Yuan, S. Clifton, T. Ducey, L. Lewis, D.W.
Dyer, and the Gonococcal Genome Sequencing Project supported
by National Institutes of Health (NIH) AI38399. This work was sup-
ported by a Department of Defense Breast Cancer Research Gradu-
ate Fellowship (G. J. M.) and NIH grants F32-GM19909 (to M.-J. L.),
R01-GM53158, and R01-AI43917.
Bacterial Strains and Mutant Alleles
see also for FC231) and were constructed using standard P1 trans-
duction methods (Miller, 1992). dinB10 (Wagner et al., 1999) was
constructed by PCR site-directed mutagenesis, replaced in the
chromosome (Link et al., 1997) and transduced into a proAB?strain
to link it with proAB?. proAB?dinB10 was transduced into the F?
1991) was also transduced to carry dinB10, then mated with the F?
lac carrying dinB10 to make the dinB10 homozygous strain,
SMR5830. dinB10 was identified by (positive) DraI digestion of PCR
products. Ectopic expression of dinB?in SMR5834 and SMR5851
was accomplished by replacement of the bacterial attB site with
dinB?including its natural promoter (basepairs 249,092–255,436 of
the E. coli genome sequence, as described; L. Gumbiner-Russo,
M.-J. Lombardo, and S. M. Rosenberg, unpublished data). SMR583
(FC40 malB::Tn9), SMR820 (FC40 malB::Tn9 lexA3(Ind?)), SMR5849
carry malB::Tn9 from D. Ennis (Lafayette, LA) and lexA3(Ind?) from
FC231 (Cairns and Foster, 1991). SMR5287 carries ?(xthA-pncA)90
zdi-201::Tn10 from BW9116 (E. coli Genetic Stock Center, Yale Uni-
versity). SMR6113 (FC40 zae::Tn10dcam zae-502::Tn10), SMR6114
(FC40 zae::Tn10dcam dnaE915 zae-502::Tn10), SMR5944 (SMR5830
zae::Tn10dcam dnaE915 zae-502::Tn10), and SMR5945 (SMR5830
zae::Tn10dcam zae-502::Tn10) carry alleles from NR9915 and
Received December 6, 2000; revised February 26, 2001.
Brotcorne-Lannoye, A., and Maenhaut-Michel, G. (1986). Role of
RecA protein in untargeted UV mutagenesis of bacteriophage
lambda: evidence for the requirement for the dinB gene. Proc. Natl.
Acad. Sci. USA 83, 3904–3908.
Bull, H.J., McKenzie, G.J., Hastings, P.J., and Rosenberg, S.M.
richia coli chromosome is promoted by recombination. Genetics
Bull, H.J., McKenzie, G.J., Hastings, P.J., and Rosenberg, S.M.
(2000b). The contribution of transiently hypermutable cells to muta-
tion in stationary phase. Genetics 156, 925–926.
Cairns, J. (2000). The contribution of bacterial hypermutators to
mutation in stationary phase. Genetics 156, 923.
Cairns, J., and Foster, P.L. (1991). Adaptive reversion of a frameshift
mutation in Escherichia coli. Genetics 128, 695–701.
Courcelle, J., Khodursky, A., Peter, B., Brown, P.O., and Hanawalt,
P.C. (2001). Comparative gene expression profiles following UV ex-
posure in wild-type and SOS-deficient Escherichia coli. Genetics,
Deitsch, K.W., Moxon, E.R., and Wellems, T.E. (1997). Shared
themes of antigenic variation and virulence in bacterial, protozoal,
and fungal infections. Microbiol. Mol. Biol. Rev. 61, 281–293.
Demple, B., Halbrook, J., and Linn, S. (1983). Escherichia coli xth
mutants are hypersensitive to hydrogen peroxide. J. Bacteriol. 153,
Denamur, E., Lecointre, G., Darlu, P., Tenaillon, O., Acquaviva, C.,
Sayada, C., Sunjevaric, I., Rothstein, R., Elion, J., Taddei, F., et al.
(2000). Evolutionary implications of the frequent horizontal transfer
of mismatch repair genes. Cell 103, 711–721.
Echols, H. (1981). SOS functions, cancer and inducible evolution.
Cell 25, 1–2.
Fijalkowska, I.J., Dunn, R.L., and Schaaper, R.M. (1993). Mutants of
Escherichia coli with increased fidelity of DNA replication. Genetics
Foster, P.L. (1997). Nonadaptive mutations occur in the F? episome
during adaptive mutation conditions in Escherichia coli. J. Bacteriol.
Foster, P.L. (1999). Mechanisms of stationary phase mutation: a
decade of adaptive mutation. Annu. Rev. Genet. 33, 57–88.
Foster, P.L., and Trimarchi, J.M. (1994). Adaptive reversion of a
frameshift mutation in Escherichia coli by simple base deletions in
homopolymeric runs. Science 265, 407–409.
Foster, P.L., Gudmundsson, G., Trimarchi, J.M., Cai, H., and Good-
man, M.F. (1995). Proofreading-defective DNA polymerase II in-
creases adaptive mutation in Escherichia coli. Proc. Natl. Acad. Sci.
USA 92, 7951–7955.
Foster, P.L., Trimarchi, J.M., and Maurer, R.A. (1996). Two enzymes,
both of which process recombination intermediates, have opposite
effects on adaptive mutation in Escherichia coli. Genetics 142,
Friedberg, E.C., Walker, G.C., and Siede, W. (1995). DNA Repair and
Mutagenesis (Washington, D.C.: ASM Press).
Friedberg, E.C., Feaver, W.J., and Gerlach, V.L. (2000). The many
tional avoidance. Proc. Natl. Acad. Sci. USA 97, 5681–5683.
Godoy, V.G., Gizatullin, F.S., and Fox, M.S. (2000). Some features
of the mutability of bacteria during nonlethal selection. Genetics
ogous with relBE of Escherichia coli plasmid P307 are ubiquitous
in prokaryotes. J. Mol. Biol. 285, 1401–1415.
Harris, R.S.(1997). On amolecular mechanism ofadaptive mutation.
Ph.D. thesis (Edmonton: University of Alberta).
Adaptive amplification: an inducible chromosomal instability mech-
anism. Cell 103, 723–731.
esis during double-strand break repair in Saccharomyces cerevis-
iae. Genetics 147, 1017–1024.
late gene expression at specific loci in Escherichia coli. Proc. Natl.
Acad. Sci. USA 77, 2819–2823.
K., Sofuni, T., Nohmi, T., and Ohmori, H. (1997). Multiple pathways
of dinB/dinP results in strongly enhancing mutagenesis in the ab-
sence of any exogenous treatment to damage DNA. Proc. Natl.
Acad. Sci. USA 94, 13792–13797.
LaRossa, R.A. (1996). Mutant selections linking physiology, inhibi-
tors, and genotypes. In Escherichia coli and Salmonella: Cellular
and Molecular Biology, F.C. Neidhardt, R. Curtiss III, J.L. Ingraham,
E.C.C. Lin, K.B. Low, B. Magasanik, W.S. Reznikoff, M. Riley, M.
Schaechter, and H. E. Umbarger, eds. (Washington, D.C.: ASM
Press), pp. 2527–2587.
Lea, D.E., and Coulson, C.A. (1949). The distribution of the numbers
of mutants in bacterial populations. J. Genet. 49, 264–285.
LeClerc, J.E., Li, B., Payne, W.L., and Cebula, T.A. (1996). High
gens. Science 274, 1208–1211.
Lin, L.L., and Little, J.W. (1989). Autodigestion and RecA-dependent
cleavage of Ind- mutant LexA proteins. J. Mol. Biol. 210, 439–452.
ing precise deletions and insertions in the genome of wild-type
J. Bacteriol. 179, 6228–6237.
Longerich, S., Galloway, A.M., Harris, R.S., Wong, C., and Rosen-
berg,S.M.(1995). Adaptivemutationsequencesreproduced bymis-
Matic, I., Radman, M., Taddei, F., Picard, B., Doit, C., Bingen, E.,
Denamur, E., and Elion, J. (1997). Highly variable mutation rates in
commensal and pathogenic Escherichia coli. Science 277, 1833–
McKenzie, G.J., Lombardo, M.-J., and Rosenberg, S.M. (1998). Re-
combination-dependent mutation in Escherichia coli occurs in sta-
tionary phase. Genetics 149, 1163–1165.
McKenzie, G.J., Harris, R.S., Lee, P.L., and Rosenberg, S.M. (2000).
The SOS response regulates adaptive mutation. Proc. Natl. Acad.
Sci. USA 97, 6646–6651.
Miller, J.H. (1992). A Short Course in Bacterial Genetics (Cold Spring
Harbor, NY: Cold Spring Harbor Laboratory Press).
Modrich, P., and Lahue, R. (1996). Mismatch repair in replication
fidelity, genetic recombination, and cancer biology. Annu. Rev. Bio-
chem. 65, 101–133.
Mount, D.W., Low, K.B., and Edmiston, S.J. (1972). Dominant muta-
tions (lex) in Escherichia coli K-12 which affect radiation sensitivity
and frequency of ultraviolet light-induced mutations. J. Bacteriol.
Napolitano,R., Janel-Bintz,R.,Wagner, J.,andFuchs, R.P.P.(2000).
All three SOS-inducible DNA polymerases (pol II, pol IV, and pol V)
are involved in induced mutagenesis. EMBO J. 19, 6259–6265.
Oliver, A., Canto ´n, R., Campo, P., Baquero, F., and Bla ´zquez, J.
(2000). High frequency of hypermutable Pseudomonas aeruginosa
in cystic fibrosis lung infection. Science 288, 1251–1254.
Parkhill, J., Achtman, M., James, K.D., Bentley, S.D., Churcher, C.,
Klee, S.R., Morelli,G., Basham, D., Brown, D., Chillingworth,T., et al.
(2000). CompleteDNA sequence ofa serogroup A strainof Neisseria
meningitidis Z2491. Nature 404, 502–506.
Radman, M. (1975). SOS repair hypothesis: phenomenology of an
inducible DNA repair which is accompanied by mutagenesis. Basic
Life Sci. 5A, 355–367.
Ripley, L.S. (1990). Frameshift mutation: determinants of specificity.
Annu. Rev. Genet. 24, 189–213.
Harris, R.S.,Longerich, S.,and Rosenberg, S.M.(1994). Recombina-
tion in adaptive mutation. Science 264, 258–260.
Harris, R.S., Ross, K.J., and Rosenberg, S.M. (1996). Opposing roles
of the Holliday junction processing systems of Escherichia coli in
recombination-dependent adaptive mutation. Genetics 142, 681–691.
Harris, R.S., Bull, H.J., and Rosenberg, S.M. (1997a). A direct role
forDNA polymeraseIII inadaptivereversion ofa frameshiftmutation
in Escherichia coli. Mutat. Res. 375, 19–24.
Harris, R.S., Feng, G., Ross, K.J., Sidhu, R., Thulin, C., Longerich, S.,
repair protein MutL becomes limiting during stationary-phase muta-
tion. Genes Dev. 11, 2426–2437.
Harris, R.S., Feng, G., Ross, K.J., Sidhu, R., Thulin, C., Longerich,
S., Szigety, S.K., Hastings, P.J., Winkler, M.E., and Rosenberg, S.M.
tion. Mutat. Res. 437, 51–60.
Hastings, P.J., Bull, H.J., Klump, J.R., and Rosenberg, S.M. (2000).
Mutator Pol IV in Adaptive Mutation
tators in adaptive mutation in Escherichia coli. Proc. Natl. Acad. Sci.
USA 96, 6862–6867.
Rosche, W.A., and Foster, P.L. (2000). Determining mutation rates
in bacterial populations. Methods 20, 4–17.
Rosenberg, S.M. (1997). Mutation for survival. Curr. Opin. Genet.
Dev. 7, 829–834.
Rosenberg, S.M. (2001). Evolution at will: adaptive mutation. Nat.
Rev. Genet., in press.
Rosenberg, S.M., Longerich, S., Gee, P., and Harris, R.S. (1994).
Adaptive mutation by deletions in small mononucleotide repeats.
Science 265, 405–407.
Rosenberg, S.M., Thulin, C., and Harris, R.S. (1998). Transient and
heritable mutators in adaptive evolution in the lab and in nature.
Genetics 148, 1559–1566.
Saunders, N.J., Jeffries, A.C., Peden, J.F., Hood, D.W., Tettelin, H.,
Rappuoli, R., and Moxon, E.R. (2000). Repeat-associated phase
variable genes in the complete genome sequence of Neisseria men-
ingitidis strain MC58. Mol. Microbiol. 37, 207–215.
Schaaper, R.M., andRadman, M. (1989). Theextreme mutator effect
of Escherichia coli mutD5 results from saturation of mismatch repair
by excessive DNA replication errors. EMBO J. 8, 3511–3516.
Strauss, B.S., Roberts, R., Francis, L., and Pouryazdanparast, P.
(2000). Role of the dinB gene product in spontaneous mutation in
Escherichia coli with an impaired replicative polymerase. J. Bacte-
riol. 182, 6742–6750.
Streisinger, G., Okada, Y., Emrich, J., Newton, J., Tsugita, A., Ter-
zaghi, E., and Inouye, M. (1966). Frameshift mutations and the ge-
netic code. Cold Spring Harbor Symp. Quant. Biol. 31, 77–84.
Tang, M., Pham, P., Shen, X., Taylor, J.S., O’Donnell, M., Woodgate,
R., and Goodman, M.F. (2000). Roles of E. coli DNA polymerases
Tettelin, H., Saunders, N.J., Heidelberg, J., Jeffries, A.C., Nelson,
K.E., Eisen, J.A., Ketchum, K.A., Hood, D.W., Peden, J.F., Dodson,
R.J., et al. (2000). Complete genome sequence of Neisseria menin-
gitidis serogroup B strain MC58. Science 287, 1809–1815.
Torkelson, J., Harris, R.S., Lombardo, M.-J., Nagendran, J., Thulin,
C., and Rosenberg, S.M. (1997). Genome-wide hypermutation in a
subpopulation of stationary-phase cells underlies recombination-
dependent adaptive mutation. EMBO J. 16, 3303–3311.
von Borstel, R.C. (1978). Measuring spontaneous mutation rates in
yeast. Methods Cell Biol. 20, 1–24.
Wagner, J., and Nohmi, T. (2000). Escherichia coli DNA polymerase
IV mutator activity: genetic requirements and mutational specificity.
J. Bacteriol. 182, 4587–4595.
Wagner, J., Gruz, P., Kim, S.R., Yamada, M., Matsui, K., Fuchs, R.P.,
and Nohmi, T. (1999). The dinB gene encodes a novel E. coli DNA
polymerase, DNA pol IV, involved in mutagenesis. Mol. Cell 4,
Wagner, J., Fujii, S., Gruz, P., Nohmi, T., and Fuchs, R.P.P. (2000).
The ? clamp targets DNA polymerase IV to DNA and strongly in-
creases its processivity. EMBO Reports 1, 484–488.
richia coli and Salmonella: Cellular and Molecular Biology, F.C.
Neidhardt, R. Curtiss III, J.L. Ingraham, E.C.C. Lin, K.B. Low, B.
Magasanik, W.S. Reznikoff, M. Riley, M. Schaechter, and H. E. Um-
barger, eds. (Washington, D.C.: ASM Press), pp. 1400–1416.