Essential roles for imuA′- and imuB-encoded accessory
factors in DnaE2-dependent mutagenesis in
Digby F. Warnera,1, Duduzile E. Ndwandwea, Garth L. Abrahamsa, Bavesh D. Kanaa, Edith E. Machowskia,
?Ceslovas Venclovasb, and Valerie Mizrahia,1
aMedical Research Council/National Health Laboratory Service/University of the Witwatersrand Molecular Mycobacteriology Research Unit and Department of
Science and Technology/National Research Foundation Centre of Excellence for Biomedical Tuberculosis Research, Faculty of Health Sciences, University of the
Witwatersrand and the National Health Laboratory Service, Johannesburg 2000, South Africa; andbLaboratory of Bioinformatics, Institute of Biotechnology,
LT-02241 Vilnius, Lithuania
Edited by Stephen J. Benkovic, The Pennsylvania State University, University Park, PA, and approved June 14, 2010 (received for review March 2, 2010)
In Mycobacterium tuberculosis (Mtb), damage-induced mutagene-
sis is dependent on the C-family DNA polymerase, DnaE2. Included
with dnaE2 in the Mtb SOS regulon is a putative operon comprising
Rv3395c, which encodes a protein of unknown function restricted
primarily to actinomycetes, and Rv3394c, which is predicted to en-
code a Y-family DNA polymerase. These genes were previously
identified as components of an imuA-imuB-dnaE2–type mutagenic
cassette widespread among bacterial genomes. Here, we confirm
that Rv3395c (designated imuA′) and Rv3394c (imuB) are individu-
ally essential for induced mutagenesis and damage tolerance.
Yeast two-hybrid analyses indicate that ImuB interacts with both
ImuA′ and DnaE2, as well as with the β-clamp. Moreover, disrup-
tion of the ImuB-β clamp interaction significantly reduces induced
mutagenesis and damage tolerance, phenocopying imuA′, imuB,
and dnaE2 gene deletion mutants. Despite retaining structural fea-
tures characteristic of Y-family members, ImuB homologs lack con-
served active-site amino acids required for polymerase activity. In
contrast, replacement of DnaE2 catalytic residues reproduces the
dnaE2 gene deletion phenotype, strongly implying a direct role for
the α-subunit in mutagenic lesion bypass. These data implicate
differential protein interactions in specialist polymerase function
and identify the split imuA′-imuB/dnaE2 cassette as a compelling
target for compounds designed to limit mutagenesis in a pathogen
increasingly associated with drug resistance.
drug resistance|induced mutagenesis|Y-family polymerase|
family DNA polymerase implicated in error-prone bypass of
DNA lesions. Loss of DnaE2 activity renders Mtb hypersensitive
to DNA damage and eliminates induced mutagenesis. Moreover,
dnaE2 deletion attenuates virulence and reduces the frequency
of drug resistance in vivo. Mtb contains two DnaE-type poly-
merases; the other, DnaE1, provides essential, high-fidelity
replicative polymerase function (1). However, the basis for the
functional specialization of the DnaE subunits remains unclear
(2, 3). Although structural determinants such as active-site ar-
chitecture contribute significantly to inherent fidelity, it is pos-
sible that differential interactions with other DNA metabolic
proteins modulate polymerase function.
Bacterial genomes containing a DnaE2-type DNA polymerase
almost invariably encode a homolog of ImuB (4–6), a putative Y-
family polymerase that is usually present in a LexA-regulated
imuA-imuB-dnaE2 gene cassette (5). In Caulobacter crescentus,
both ImuB and ImuA are required for induced mutagenesis and
damage tolerance (6) whereas plasmid-encoded DnaE2 and
ImuB mediate UV-induced mutagenesis in Deinococcus deserti
(7). Although distributed widely across the bacterial domain, the
imuA-imuB-dnaE2 cassette is not found in organisms possessing
NA damage-induced base substitution mutagenesis in My-
cobacterium tuberculosis (Mtb) depends on DnaE2 (1), a C-
umuDC homologs (5). This suggests that the encoded proteins
perform an analogous function to DNA polymerase V (8), the
Y-family member required for damage-induced mutagenesis in
Escherichia coli (9).
Mtb contains a putative SOS-inducible operon, Rv3395c-
Rv3394c (1, 10), located ∼24.7 kb upstream of dnaE2. Rv3395c
homologs are found in a limited number of organisms and their
function is unknown, whereas Rv3394c exhibits significant ho-
mology to ImuB proteins (6), identifying the Rv3395c-Rv3394c
operon as part of a split imuA′-imuB/dnaE2 cassette (5). Rv3394c
(ImuB) contains a predicted β-clamp–binding motif, which des-
ignates the protein as a DinB3-type Y-family polymerase (11).
Notably, neither DnaE2 nor ImuA′ possesses an identifiable
clamp-binding motif (1, 12). The β-clamp modulates the re-
cruitment to the replication machinery of proteins involved in
specialist lesion bypass polymerases simultaneously to enable
rapid interchange (14, 15). The β-clamp–binding motif therefore
suggested that ImuB might be crucial for DnaE2 function.
dnaE2 cassette in mycobacteria. Our results establish essential roles
in ImuB and identify a role for the extended ImuB C-terminal do-
main in binding both DnaE2 and ImuA′. Although structurally
similar to Y-family polymerases, homology modeling indicates that
polymerase activity. Instead, disruption of the ImuB–β-clamp in-
teraction reproduces the imuB gene deletion phenotype, suggesting
an essential role for ImuB in mediating access of the other cassette
components to the replication fork. In contrast, through targeted
replacement of active-site residues, we elucidate a direct catalytic
role for DnaE2 in mutagenic lesion bypass. These observations
polymerase function in Mtb.
DNA Damage-Induced Expression of the Split imuA′-imuB/dnaE2
Cassette. All mycobacteria contain a split imuA′-imuB/dnaE2
Author contributions: D.F.W., D.E.N., G.L.A., B.D.K.,?C.V., and V.M. designed research;
D.F.W., D.E.N., G.L.A., B.D.K., and?C.V. performed research; D.F.W. and E.E.M. contributed
new reagents/analytic tools; D.F.W., D.E.N., G.L.A., B.D.K.,?C.V., and V.M. analyzed data;
and D.F.W.,?C.V., and V.M. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence may be addressed. E-mail: firstname.lastname@example.org or digby.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| July 20, 2010
| vol. 107
| no. 29
cassette (5), except Mycobacterium leprae, which has lost these
genes through genome decay (16). The genomic location varies
across mycobacterial species: in Mycobacterium smegmatis mc2155
(Msm), imuA′, and imuB are situated only 10.8 kb upstream of
dnaE2. However, the encoded proteins are highly conserved, and
a LexA-binding site is located in the promoter region upstream of
imuA′ in Mtb (17) and Msm (Fig. S1B). Previous studiesidentified
all three cassette components in the Mtb SOS regulon (1, 10), and
expression of dnaE2 was shown to parallel recA induction (1). We
analyzed transcript abundances of the cassette components rela-
tive to a sigA control (18) during mid log-phase growth of Mtb and
following UV exposure (Table S1) and compared with the non-
genes were significantly up-regulated within 6 h of UV treatment
and remained elevatedat 24h.dnaE1,whichencodestheessential
replicative α-subunit (1), was also up-regulated following UV ir-
radiation. Although there is no identifiable SOS box (17), a pro-
moter motif regulating the LexA-independent damage response
(10) is situated immediately upstream of dnaE1 (19), consistent
with damage induction.
ImuA′ and ImuB Are Individually Essential for DnaE2 Function in
Mycobacteria. To investigate the roles of imuA′ and imuB in
DNA damage-induced mutagenesis, allelic exchange mutants of
Mtb H37Rv were constructed in which imuB alone, or both imuA′
and imuB, were deleted and replaced by an antibiotic resistance
marker (Fig. S1A). Neither mutant exhibited a phenotype under
standard in vitro growth conditions. However, UV-induced mu-
tagenesis waseliminatedinbothstrains(Fig.1A) asdeterminedby
induced mutator phenotype was restored by complementation
with a fragment carrying both imuA′ and imuB genes integrated at
theattB site.Moreover,loss ofimuA′-imuB reproduced thednaE2
deletion phenotype (1), confirming the functionality of the split
imuA′-imuB/dnaE2 cassette. To determine the individual require-
ments for imuA′ and imuB, the similar genomic arrangement in
Msm was exploited (Fig. S1B). Deletion of imuB significantly re-
duced the UV-induced RifRfrequency in Msm (Fig. 1B), mim-
icking inactivation of dnaE2 (1). The same phenotype resulted
from deletion of imuA′ and was complemented by integration of
a wild-type copy of imuA′. These data established the individual
essentiality of ImuA′ and ImuB for induced mutagenesis and
confirmed the designation of the mycobacterial ImuA′ as a distant
homolog of ImuA (5, 6). Msm imuA′ and imuB deletion mutants
were hypersensitive to mitomycin C (MMC), a genotoxic agent
known to induce the SOS response in mycobacteria (1) (Fig. S1C).
increase damage sensitivity relative to individual deletion mutants
(ΔimuA′,ΔimuB,orΔdnaE2) (Fig.S1C). Moreover, integration of
wild-type copies of the deleted genes at the attB (imuA′-imuB) and
attL (dnaE2) sites reversed the MMC-hypersensitive phenotype of
genes operate in a single pathway.
ImuB Lacks Active Site Residues Required for DNA Polymerase Activity.
ImuB occupies a distinct branch of the UmuC subfamily of Y-
family polymerases (6) and possesses all structural domains typical
of Y-family members, including the defining little finger (20) (Fig.
2A). A putative β-clamp–binding motif (354QLPLWG359) that is
located between the Y polymerase-like N-terminal region and the
C-terminal extension characteristic of ImuB proteins identified
Mtb ImuB as founder of the DinB3-type Y-family polymerases
(11) (Fig. S2). Although classified as a Y-family member, com-
parative sequence analyses and homology modeling indicate that
the expected carboxylates (21) are replaced by other residues in the
ImuB active site (Fig. 2B). The absence of the complete triad of
active-site acidic residues is a feature of all ImuB homologs ana-
lyzed (Fig. S2) and strongly implies that these proteins cannot
function as DNA polymerases (22). In turn, this observation sug-
gests a model for translesion synthesis in which DnaE2 itself cat-
alyzes lesion bypass (6).
Catalytic Activity of DnaE2 Is Required for Induced Mutagenesis in
Mtb. All three DNA polymerase IIIα active-site acidic residues
(23, 24) are present in Mtb DnaE2 (D439, D441, D579; Fig. S3). To
determine the requirement for DnaE2 polymerase activity in
damage tolerance and induced mutagenesis in mycobacteria,
a site-directed Msm mutant was generated in which catalytic
function was crippled by replacement of two corresponding ac-
tive-site aspartates with alanines (441DID443→441AIA443). The
chromosomal dnaE2AIAmutation eliminated UV-induced mu-
tagenesis (Fig. 3A) and rendered Msm hypersensitive to MMC
(Fig. 3B). The fact that the dnaE2AIAmutation phenocopied the
ΔdnaE2 knockout strongly implied a direct role for DnaE2 in
catalyzing translesion synthesis.
QLPLWG Clamp-Binding Motif Mediates the Interaction of ImuB with
β. A β-clamp–binding motif (946QFDLF950) (12) is situated in the
β-binding domain of Mtb DnaE1 (Fig. S3). The corresponding
region in Mtb DnaE2 (924RPDRLPGVG932) is invariant in my-
cobacterial genomes, but does not resemble known β-binding
motifs (11, 12). We showed recently (25) that the dnaN-encoded
β-subunit interacts with DnaE1 and with itself in yeast two-hy-
brid (Y2H) assays, but does not bind DnaE2 (Fig. 4). Similarly,
Mtb ImuA′ failed to bind β, whereas an interaction between
ImuB and the β-clamp was readily detected. In combination,
these data establish ImuB as the only cassette component with
β-binding capacity in Mtb. The observed interaction was con-
sistent with the β-binding motif (354QLPLWG359) in ImuB that
characterizes the DinB3 family (11) and is highly conserved
among mycobacterial ImuBs. Moreover, a Q354A mutation
(ImuBALPLWG) eliminated the ImuB–β interaction in the Y2H
system (Fig. 4), confirming a role for this motif in clamp binding.
(B) Msm at 4.5 h post UV treatment. Data represent single experiments performed in triplicate.
ImuA′ and ImuB are essential for induced mutagenesis. UV-induced mutation frequencies to rifampicin resistance (RifR) in (A) Mtb at 24 and 48 h and
| www.pnas.org/cgi/doi/10.1073/pnas.1002614107Warner et al.
Protein Interactions of the Split imuA′-imuB/dnaE2 Cassette Components
inMtb.Y2H analyses identified an ImuA′–ImuB interaction that was
retained in the ImuBALPLWGmutant (Fig. 4), confirming the speci-
ficity of the354QLPLWG359motif for binding β. In addition, both
ImuB and ImuBALPLWGinteracted with DnaE2, whereas DnaE2
and ImuA′ failed to interact with any protein analyzed other than
ImuB. The active-site mutant form of DnaE2, DnaE2AIA, retained
the ability to bind ImuB (Fig. 4), confirming that loss of DnaE2
catalytic activity—and not disruption of the inferred ImuB–DnaE2
interaction—was responsible for the phenotypes of the Msm
dnaE2AIAmutant strain (Fig. 3). Interactions of ImuB with DnaE1
and with itself were also elucidated (Fig. 4), identifying multiple
partners for ImuB (Fig. S4A).
C-Terminal Region of ImuB Is Required for Interaction with Other
Cassette Components.Wecouldnot detectarelationshipbetween
the C-terminal region of Mtb ImuB and any known protein
structure. However, this region contains stretches of predicted
structural disorder (Fig. S4B) that are a hallmark of protein–
protein interaction sites (26). We generated Y2H constructs in
which nonsense codons were introduced into ImuB to preserve
N-terminal Y-family structural domains while progressively
eliminating C-terminal segments (Fig. 4). In Mtb ImuBC168, the
entire 168-amino-acid C-terminal region is eliminated down-
stream of the intact354QLPLWG359motif. The truncated ImuB
retained its interactionwith β, confirmingtheidentity oftheclamp-
binding motif. In contrast, the ability to bind β was eliminated in
motif were deleted (Fig. 4). Both truncation mutants, ImuBC168and
ImuBCB, lost the ability to bind DnaE1, DnaE2, and ImuA′,
confirming the involvement of the C-terminal region of ImuB in
multiple protein interactions. To evaluate the requirement for
we constructed integrating vectors carrying wild-type Msm imuA′
and mutant imuB alleles with equivalent C-terminal truncations
(imuA′-imuBC168and imuA′-imuBCB). Both constructs restored
the induced mutagenesis (Fig. S5A) and damage tolerance (Fig.
S5B) phenotypes of the single ΔimuB and double ΔimuA′ ΔimuB
knockout strains, thereby supporting a critical role for ImuB–
DnaE2 and/or ImuB–ImuA′ interaction(s) in pathway function.
By mapping the ImuA′ sequence onto the RecA structure (Fig.
assessed their effects on the interaction with ImuB (Fig. 4). Re-
moval of 31 (ImuA′N31) or 48 (ImuA′N48) amino acids from the N
terminus (eliminating the predicted N-terminal subdomain; Fig.
in Y2H assays (Fig. 4). In contrast, a 44-amino-acid C-terminal
truncation (ImuA′C) eliminated ImuB binding. The correspond-
ing truncation in Msm imuA′ abrogated the ability of a com-
plementing vector to restore damage tolerance to the ΔimuA′
deletion mutant (Fig. 5B). However, as the C-terminal truncation
might have affected the structural integrity of the protein, it is
uncertain whether this region alone or the entire structural do-
main is required for the ImuA′–ImuB interaction.
ImuB–β-Clamp Interaction Is Required for Induced Mutagenesis. A
survey of representative imuB/dnaE2-containing bacteria revealed
clamp-binding motifs in ImuB and DnaE2 proteins (Figs. S2 and
cassette function. To test this hypothesis, we constructed an in-
tegration vector carrying wild-type imuA and imuBALPLWG, which
encodes ImuB with a mutated β-binding motif (352ALPLWG357),
and evaluated its ability to complement the Msm ΔimuA′ ΔimuB
interaction in Y2H assays (Fig. 4). However, a single Q352A sub-
restore damage tolerance and induced mutagenesis to wild-type
system to reproduce all elements contributing to protein–protein
interactions in the natural mycobacterial host. Therefore, we con-
structed additional alleles containing multiple mutations that tar-
geted hydrophobic and/or aromatic residues in the putative ImuB
β-binding motif. Two alleles containing double mutations, imu-
BAAPLWGand imuBALPLGG, retained ImuB function (Fig. 6).
However, mutation of the first five β-binding motif residues in
imuBAAAAGGeliminated the ability of the resulting vector to com-
plement the induced mutagenesis (Fig. 6A) and damage tolerance
ImuB–β-clamp interaction is required for cassette operation.
mutagenesis and damage tolerance. (A) UV-induced
mutation frequencies to rifampicin resistance (RifR). (B)
Sensitivity to MMC treatment. Log-fold dilutions were
plated on antibiotic-free medium or on medium con-
taining MMC at 0.02 and 0.04 μg·ml−1. Data are from
a representative experiment performed in triplicate.
DnaE2 catalytic activity is required for induced
residues. (A) Comparison of the Mtb ImuB homology model with the X-ray
structure of Sulfolobus solfataricus Dpo4 complexed with DNA and incoming
nucleotide (PDB id: 1jx4) (21). ImuB is colored from N terminus (blue) to C ter-
of Dpo4 and correspondingresiduesinImuB.Only thepalmdomain isshownin
both structures. The incoming nucleotide in the Dpo4–DNA complex is colored
pink, and the magnesium ion is shown as a cyan sphere.
ImuB is a homolog of Y-family polymerases but lacks active-site acidic
Warner et al. PNAS
| July 20, 2010
| vol. 107
| no. 29
The phenotypes associated with knockout of Mtb dnaE2 sug-
gested a functional analogy with E. coli DNA polymerase V (1).
Here, we extended those observations, demonstrating the essen-
tiality of ImuA′ and ImuB for damage tolerance and induced
mutagenesis and establishing that these proteins act in the same
pathway as DnaE2. We also confirmed the inducible expression
of dnaE2, imuA′, and imuB and identified dnaE1 as a component
of the Mtb damage response. Previous studies included DnaE2
and ImuB among three damage-responsive DNA polymerases
in Mtb (1, 10), the other being a putative DNA polymerase X
(Rv3856c). However, the absence of active-site carboxylates
appears incompatible with catalytic function in ImuB, whereas
a truncated polymerase domain probably precludes polymerase
function in Rv3856c (Fig. S8). Therefore, in contrast to E. coli
DNA polymerase V, which is part of an SOS regulon that
and IV (Y family) (reviewed in ref. 27), the polymerases induced
as part of the Mtb damage response seem to be limited to two
α-subunits, DnaE1 and DnaE2.
Point mutations in the DnaE2 active site eliminated induced
mutagenesis and damage tolerance, reproducing the dnaE2 de-
letion phenotype. In combination with the prediction that ImuB
lacks polymerase activity, this observation implies that DnaE2
structures are adapted to specialist lesion bypass (20), sequence
analysis reveals few clues to DnaE2 function. All major DNA
polymerase IIIα structural/functional domains (23, 24) are readily
E. coli has been implicated in the interaction of α with the clamp-
loader subunit, τ (28, 29). A strong α–τ interaction enables simul-
taneous leading and lagging-strand synthesis by the DNA poly-
merase III holoenzyme (13), and the absence of this region in
DnaE2 and all other nonessential dnaE-type α-subunits (Fig. S3)
might correlate with an inability to substitute essential replicative
function. Evidence implicating a defective α–τ interaction in an
E. coli mutator phenotype (30) further reinforces the potential
contribution of τ-binding to the functional specialization of DnaE
Mtb DnaE2doesnotbind β.However,a highlyconserved motif
is located immediately downstream of the predicted OB fold do-
main in DnaE2proteins (Fig. S3), which suggests the potential for
other intermolecular interactions. In E. coli, replicative poly-
the dnaQ-encoded proofreading exonuclease, and disruptions to
bypass in the absence of DNA polymerases IV and V (31, 32). In
combination, these observations suggest that differential inter-
actions with DnaQ-like proteins might determine DnaE subunit
functioninMtb.Additional 3′-5′exonuclease activitywasrecently
located to the PHP domain of DnaE-type polymerases (33), al-
though the contribution of this function to polymerase fidelity
remains to be determined. The corresponding domain of Mtb
(34–36), whereas key residues are absent in DnaE2. Notably, the
residues present in Mtb DnaE1 are conserved in the sole DnaE
polymerase in M. leprae (Fig. S3), an organism marked by a com-
plete lack of DnaQ homologs (37). Therefore, although the
functional consequence of these differences remains to be eluci-
dated, it is tempting to speculate that loss of conserved residues in
the DnaE2 PHP domain impacts relative fidelity.
Our evidence indicates a central role for ImuB in cassette
function that is independent of polymerase activity. Structurally,
ImuB resembles Y-family polymerases; however, the active-site
architecture is inconsistent with catalysis and instead reinforces
the likely importance of protein interactions for ImuB function.
Consistent with this idea, we identified multiple partners for
ImuB, including DnaE1 and DnaE2, as well as ImuA′ and the
β-clamp. We also established the capacity of ImuB to interact
with itself and implicated the C-terminal extension in this self-
association and in the interaction of ImuB with DnaE1, DnaE2,
and ImuA′. The fact that the C terminus binds multiple partners
suggests that this region mediates protein interactions that distin-
guish ImuB from other Y-family members. Our Y2H experiments
also confirmed the role of the DinB3-type QLPLWG motif in the
interaction of ImuB with β. Notably, disruption of this motif in the
ImuA′ is colored according to truncation variants: blue and cyan indicate N-terminal truncations, and red is the C-terminal truncation. (B) Deletion of 44 C-
terminal amino acids (imuA′C) renders Msm sensitive to MMC treatment. Log-fold dilutions were plated on antibiotic-free medium (Left and Right, Rows 1
and 4), 0.02 μg·ml−1(Left and Right, Rows 2 and 5), and 0.04 μg·mL−1(Left and Right, Rows 3 and 6) MMC. Data are from a representative experiment
performed in duplicate.
ImuA′ is required for damage tolerance. (A) Comparison of the Mtb ImuA′ homology model and the X-ray structure of E. coli RecA (PDB id: 1u94) (57).
identified in this study and previously (25). ND, not determined. An open circle (o) indicates interactions maintained where bait or prey vector contains full-
length ImuB; ImuB self-association is eliminated when both vectors carry the ImuBC168allele. A blue bar indicates the position of the β-clamp–binding motif
(354QLPLWG359) deleted in ImuBCBand mutated (XXX) in ImuBALPLWG. The439AIA441mutation in Mtb DnaE2 is indicated (XXX). Lines are to scale.
Interactions of cassette components. Summary of Mtb protein interactions detected (●) or absent (×) on highest stringency growth medium as
| www.pnas.org/cgi/doi/10.1073/pnas.1002614107 Warner et al.
that ImuB—through β—is required to mediate access of DnaE2
to the replication fork. It is difficult, therefore, to reconcile the
conflicting reports of biological function: specifically, the observa-
putida (22), whereas the homologous Pseudomonas aeruginosa
proteins are dispensable for damage tolerance (38) but not for in-
consistent with the data presented here in that they indicate es-
damage tolerance and induced mutagenesis.
In E. coli, RecA constitutes a key component of the DNA poly-
merase V “mutasomal complex” comprising UmuD′2C-RecA-ATP
(8). Mtb ImuA′ resembles RecA at a structural level, but lacks the
characteristic C-terminal domain. We observed the direct in-
teraction of ImuA′ with ImuB involving the C-terminal regions of
both proteins; however, no evidence of ImuA′ self-association was
detected that might indicate RecA-like filamentation (40). Instead,
structural peculiarities such as the absence of a RecA-like nucleo-
tide-binding motif, as well as extensive differences in the regions
corresponding to RecA DNA-binding loops, suggest the possible
functional specialization of ImuA′. Therefore, although broadly
consistent with the identification of the imuA′-imuB-dnaE2 cassette
as the nonorthologous replacement of the umuDC system (5), our
data reveal a “mutasome” comprising distinct components: a dam-
synthesis, a putative polymerase-inactive Y-family homolog that
binds β as well as the other cassette components, and a predicted
DNA-binding protein. Fundamental questions therefore remain
regarding the relevance of the inferred protein–protein interactions
to cassette function, the temporal order of those interactions, the
potential for complex formation, and the possibility that the com-
ponents are subject to additional regulation.
Mtb engages multiple strategies to subvert tuberculosis (TB)
chemotherapy (41), which favors the emergence of antibiotic re-
sistance (42). One approach to the development of antimicrobials
is to disarm mechanisms of induced mutagenesis (43). The
encoded components of the split imuA′-imuB/dnaE2 cassette are
essential for induced mutagenesis and so present compelling
targets for the discovery of anti-TB drugs.
Materials and Methods
Bacterial Strains and Growth Conditions. Strains, plasmids, and oligonucleo-
tides are described in Tables S2 and S3. Mtb was grown on Middlebrook 7H10
(Merck) supplemented with 0.5% glycerol and Middlebrook Oleic Acid Dex-
trose Catalase (OADC) enrichment (Merck) or in Middlebrook 7H9 supple-
mented with 0.2% glycerol, Middlebrook OADC, and 0.05% Tween 80. For
Msm, 0.1% Tween 80 was used.
Yeast Two-Hybrid Analyses. Protein–protein interactions were assessed using
the Clontech Matchmaker Y2H system. Fusion of Mtb ImuB to the binding
domain resulted in autoactivation on low-stringency growth medium;
therefore, all interactions involving Mtb ImuB were inferred from experi-
ments using an ImuB–activation domain fusion construct only. All other
proteins were cloned as activation domain and binding domain fusions.
Construction of Mutant Strains of Mtb and Msm. Mtb and Msm mutants were
constructed by allelic exchange (44) using suicide plasmids described in Table
S2. Alleles for site-directed mutagenesis were generated by the megaprimer
method (45). Genetically complemented derivatives were generated by in-
tegration of L5 or Tweety-based vectors at attB (46) or attL (47).
DNA-Damaging Treatments and Determination of Mutation Frequencies. UV-
induced mutation frequencies were determined as described (1). Survival
assays were performed by plating 10-fold serial dilutions of log-phase Msm
cultures on solid media containing MMC.
Gene Expression Analysis. Bacteria were grown in 7H9 medium and aliquots
sampled during different phases of growth. Real-time quantitative reverse
transcription–PCR was carried out with the primer pairs detailed in Table S3,
as described (25).
Homology Modeling. Models were generated using previously described
methods (48). Briefly, structural templates were selected by PSI-BLAST (49),
and the position-dependent reliability of alignments was estimated by PSI-
BLAST-ISS (50). In addition, HHsearch (51), COMPASS (52), and COMA (53)
were used to reduce uncertain regions in alignments with structural tem-
plates. The agreement of all three algorithms was required before regions
were considered reliably aligned. Three-dimensional models were refined by
iterative steps of construction and assessment of different alignment var-
iants for unreliable regions and/or combinations of structural templates.
Models corresponding to different sequence-structure alignment variants
were built by Modeler (54), and residue side chains were positioned with
SCWRL3 (55). Modeled structures were assessed iteratively with Prosa2003
(56) until no further improvement in energy scores was possible and the
visual assessment revealed no significant flaws.
ACKNOWLEDGMENTS. This work was funded by grants from the Medical
Research Council of South Africa (to V.M.), the National Research Founda-
tion (V.M.), the Howard Hughes Medical Institute (International Research
Scholar’s grants to V.M. and?C.V.), the University of the Witwatersrand
(V.M., D.F.W. and D.E.N), and the South African Tuberculosis and AIDS
Training (SATBAT) program (National Institutes of Health/Fogarty Interna-
tional Center 1U2RTW007370/3) (D.E.N.). We thank Sabine Ehrt, Dirk Schap-
pinger, Graham Hatfull, and Diane Kuhnert for providing vectors and
Helena Boshoff and members of the Mizrahi laboratory for valuable advice
1. Boshoff HI, Reed MB, Barry CE, III, Mizrahi V (2003) DnaE2 polymerase contributes to
in vivo survival and the emergence of drug resistance in Mycobacterium tuberculosis.
2. Bruck I, Goodman MF, O’Donnell M (2003) The essential C family DnaE polymerase is
error-prone and efficient at lesion bypass. J Biol Chem 278:44361–44368.
3. Le Chatelier E, et al. (2004) Involvement of DnaE, the second replicative DNA
polymerase from Bacillus subtilis, in DNA mutagenesis. J Biol Chem 279:1757–1767.
4. Abella M, et al. (2004) Widespread distribution of a lexA-regulated DNA damage-
inducible multiple gene cassette in the Proteobacteria phylum. Mol Microbiol 54:
5. Erill I, Campoy S, Mazon G, Barbé J (2006) Dispersal and regulation of an adaptive
mutagenesis cassette in the bacteria domain. Nucleic Acids Res 34:66–77.
6. Galhardo RS, Rocha RP, Marques MV, Menck CF (2005) An SOS-regulated operon
involved in damage-inducible mutagenesis in Caulobacter crescentus. Nucleic Acids
duced mutagenesis and damage tolerance. (A) UV-induced
mutation frequencies to rifampicin resistance (RifR). (B)
Sensitivity to MMC treatment. Log-fold dilutions of Msm
ΔimuA′ ΔimuB complemented with full-length imuA′-imuB
containing wild-type (WT) and mutant imuB alleles were
plated on 0.02 μg·mL−1(Rows 1, 3, 5, 7, and 9) and 0.04
μg·ml−1(Rows 2, 4, 6, 8, and 10) MMC; at right is the un-
treated control. Data are from a representative experiment
performed in duplicate.
The ImuB–β-clamp interaction is required for in-
Warner et al.PNAS
| July 20, 2010
| vol. 107
| no. 29
7. Dulermo R, Fochesato S, Blanchard L, de Groot A (2009) Mutagenic lesion bypass and Download full-text
two functionally different RecA proteins in Deinococcus deserti. Mol Microbiol 74:
8. Jiang Q, Karata K, Woodgate R, Cox MM, Goodman MF (2009) The active form of
DNA polymerase V is UmuD′(2)C-RecA-ATP. Nature 460:359–363.
9. Tang M, et al. (2000) Roles of E. coli DNA polymerases IV and V in lesion-targeted and
untargeted SOS mutagenesis. Nature 404:1014–1018.
10. Rand L, et al. (2003) The majority of inducible DNA repair genes in Mycobacterium
tuberculosis are induced independently of RecA. Mol Microbiol 50:1031–1042.
11. Dalrymple BP, Wijffels G, Kongsuwan K, Jennings P (2003) Towards an understanding
of protein-protein interaction network heirarchies. Analysis of DnaN-binding peptide
motifs in members of protein families interacting with the eubacterial processivity
clamp, the subunit of DNA Polymerase III. First Asia-Pacific Bioinformatics Conference
(APBC2003), ed Chen Y-PP (Australian Computer Society, Sydney, Australia), pp
12. Dalrymple BP, Kongsuwan K, Wijffels G, Dixon NE, Jennings PA (2001) A universal
protein-protein interaction motif in the eubacterial DNA replication and repair
systems. Proc Natl Acad Sci USA 98:11627–11632.
13. Johnson A, O’Donnell M (2005) Cellular DNA replicases: Components and dynamics at
the replication fork. Annu Rev Biochem 74:283–315.
14. Heltzel JM, Maul RW, Scouten Ponticelli SK, Sutton MD (2009) A model for DNA
polymerase switching involving a single cleft and the rim of the sliding clamp. Proc
Natl Acad Sci USA 106:12664–12669.
15. Pagès V, Fuchs RP (2002) How DNA lesions are turned into mutations within cells?
16. Cole ST, et al. (2001) Massive gene decay in the leprosy bacillus. Nature 409:
17. Davis EO, Dullaghan EM, Rand L (2002) Definition of the mycobacterial SOS box and
use to identify LexA-regulated genes in Mycobacterium tuberculosis. J Bacteriol 184:
18. Manganelli R, Dubnau E, Tyagi S, Kramer FR, Smith I (1999) Differential expression of
10 σ factor genes in Mycobacterium tuberculosis. Mol Microbiol 31:715–724.
19. Gamulin V, Cetkovic H, Ahel I (2004) Identification of a promoter motif regulating the
major DNA damage response mechanism of Mycobacterium tuberculosis. FEMS
Microbiol Lett 238:57–63.
20. Yang W, Woodgate R (2007) What a difference a decade makes: Insights into
translesion DNA synthesis. Proc Natl Acad Sci USA 104:15591–15598.
21. Ling H, Boudsocq F, Woodgate R, Yang W (2001) Crystal structure of a Y-family DNA
polymerase in action: A mechanism for error-prone and lesion-bypass replication. Cell
22. Koorits L, et al. (2007) Study of involvement of ImuB and DnaE2 in stationary-phase
mutagenesis in Pseudomonas putida. DNA Repair (Amst) 6:863–868.
23. Bailey S, Wing RA, Steitz TA (2006) The structure of T. aquaticus DNA polymerase III is
distinct from eukaryotic replicative DNA polymerases. Cell 126:893–904.
24. Lamers MH, Georgescu RE, Lee SG, O’Donnell M, Kuriyan J (2006) Crystal structure of
the catalytic α subunit of E. coli replicative DNA polymerase III. Cell 126:881–892.
25. Kana BD, et al. (2010) Role of the DinB homologs Rv1537 and Rv3056 in Mycobacterium
tuberculosis. J Bacteriol 192:2220–2227.
26. Dyson HJ, Wright PE (2005) Intrinsically unstructured proteins and their functions. Nat
Rev Mol Cell Biol 6:197–208.
27. Goodman MF (2002) Error-prone repair DNA polymerases in prokaryotes and
eukaryotes. Annu Rev Biochem 71:17–50.
28. Kim DR, McHenry CS (1996) Biotin tagging deletion analysis of domain limits involved
in protein-macromolecular interactions. Mapping the τ binding domain of the DNA
polymerase III α subunit. J Biol Chem 271:20690–20698.
29. López de Saro FJ, Georgescu RE, O’Donnell M (2003) A peptide switch regulates DNA
polymerase processivity. Proc Natl Acad Sci USA 100:14689–14694.
30. Gawel D, Pham PT, Fijalkowska IJ, Jonczyk P, Schaaper RM (2008) Role of accessory
DNA polymerases in DNA replication in Escherichia coli: Analysis of the dnaX36
mutator mutant. J Bacteriol 190:1730–1742.
31. Borden A, et al. (2002) Escherichia coli DNA polymerase III can replicate efficiently
past a T-T cis-syn cyclobutane dimer if DNA polymerase V and the 3′ to 5′ exonuclease
proofreading function encoded by dnaQ are inactivated. J Bacteriol 184:2674–2681.
32. Vandewiele D, Borden A, O’Grady PI, Woodgate R, Lawrence CW (1998) Efficient
translesion replication in the absence of Escherichia coli Umu proteins and 3′-5′
exonuclease proofreading function. Proc Natl Acad Sci USA 95:15519–15524.
33. Stano NM, Chen J, McHenry CS (2006) A coproofreading Zn(2+)-dependent exonuclease
within a bacterial replicase. Nat Struct Mol Biol 13:458–459.
34. Baños B, Lázaro JM, Villar L, Salas M, de Vega M (2008) Editing of misaligned 3′-
termini by an intrinsic 3′-5′ exonuclease activity residing in the PHP domain of
a family X DNA polymerase. Nucleic Acids Res 36:5736–5749.
35. Nakane S, Nakagawa N, Kuramitsu S, Masui R (2009) Characterization of DNA
polymerase X from Thermus thermophilus HB8 reveals the POLXc and PHP domains
are both required for 3′-5′ exonuclease activity. Nucleic Acids Res 37:2037–2052.
36. Wieczorek A, McHenry CS (2006) The NH2-terminal php domain of the α subunit of
the Escherichia coli replicase binds the ε proofreading subunit. J Biol Chem 281:
37. Dawes SS, Mizrahi V (2001) DNA metabolism in Mycobacterium leprae. Lepr Rev 72:
38. Cirz RT, O’Neill BM, Hammond JA, Head SR, Romesberg FE (2006) Defining the
Pseudomonas aeruginosa SOS response and its role in the global response to the
antibiotic ciprofloxacin. J Bacteriol 188:7101–7110.
39. Sanders LH, Rockel A, Lu H, Wozniak DJ, Sutton MD (2006) Role of Pseudomonas
aeruginosa dinB-encoded DNA polymerase IV in mutagenesis. J Bacteriol 188:
40. Datta S, et al. (2003) Crystal structures of Mycobacterium smegmatis RecA and its
nucleotide complexes. J Bacteriol 185:4280–4284.
41. Warner DF, Mizrahi V (2006) Tuberculosis chemotherapy: The influence of bacillary
stress and damage response pathways on drug efficacy. Clin Microbiol Rev 19:
42. Kohanski MA, DePristo MA, Collins JJ (2010) Sublethal antibiotic treatment leads to
multidrug resistance via radical-induced mutagenesis. Mol Cell 37:311–320.
43. Smith PA, Romesberg FE (2007) Combating bacteria and drug resistance by inhibiting
mechanisms of persistence and adaptation. Nat Chem Biol 3:549–556.
44. Parish T, Stoker NG (2000) Use of a flexible cassette method to generate a double
unmarked Mycobacterium tuberculosis tlyA plcABC mutant by gene replacement.
45. Smith AM and Klugman KP (1997) “Megaprimer” method of PCR-based mutagenesis:
The concentration of megaprimer is a critical factor. Biotechniques 22:438, 442.
46. Lee MH, Pascopella L, Jacobs WR, Jr, Hatfull GF (1991) Site-specific integration of
mycobacteriophage L5: Integration-proficient vectors for Mycobacterium smegmatis,
Mycobacterium tuberculosis, and bacille Calmette-Guérin. Proc Natl Acad Sci USA 88:
47. Pham TT, Jacobs-Sera D, Pedulla ML, Hendrix RW, Hatfull GF (2007) Comparative
construction of compatible site-specific integration vectors for mycobacteria.
48. Venclovas?C, Margelevi? cius M (2009) The use of automatic tools and human expertise
in template-based modeling of CASP8 target proteins. Proteins 77 (Suppl 9):81–88.
49. Altschul SF, et al. (1997) Gapped BLAST and PSI-BLAST: A new generation of protein
database search programs. Nucleic Acids Res 25:3389–3402.
50. Margelevi? cius M, Venclovas?C (2005) PSI-BLAST-ISS: An intermediate sequence search
tool for estimation of the position-specific alignment reliability. BMC Bioinformatics
51. Söding J (2005) Protein homology detection by HMM-HMM comparison. Bioinformatics
52. Sadreyev R, Grishin N (2003) COMPASS: A tool for comparison of multiple protein
alignments with assessment of statistical significance. J Mol Biol 326:317–336.
53. Margelevi? cius M, Venclovas?C (2010) Detection of distant evolutionary relationships
between protein families using theory of sequence profile-profile comparison. BMC
54. Šali A, Blundell TL (1993) Comparative protein modelling by satisfaction of spatial
restraints. J Mol Biol 234:779–815.
55. Canutescu AA, Shelenkov AA, Dunbrack RL, Jr (2003) A graph-theory algorithm for
rapid protein side-chain prediction. Protein Sci 12:2001–2014.
56. Sippl MJ (1993) Recognition of errors in three-dimensional structures of proteins.
57. Xing X, Bell CE (2004) Crystal structures of Escherichia coli RecA in a compressed
helical filament. J Mol Biol 342:1471–1485.
Tweety: Evolutionaryinsights and
| www.pnas.org/cgi/doi/10.1073/pnas.1002614107Warner et al.