JOURNAL OF BACTERIOLOGY, Sept. 2007, p. 6655–6664
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Vol. 189, No. 18
Investigation of Transcription Repression and Small-Molecule
Responsiveness by TetR-Like Transcription Factors Using a
Heterologous Escherichia coli-Based Assay?
Sang Kyun Ahn,1Kapil Tahlan,1Zhou Yu,2and Justin Nodwell1*
Department of Biochemistry and Biomedical Sciences, McMaster University, 1200 Main St. W., Hamilton, Ontario,
Canada L8N 3Z5,1and Department of Medical Genetics and Microbiology, University of Toronto,
1 King’s College Circle, Toronto, Ontario, Canada M5S 1A82
Received 7 May 2007/Accepted 10 July 2007
The SCO7222 protein and ActR are two of ?150 TetR-like transcription factors encoded in the Streptomyces
coelicolor genome. Using bioluminescence as a readout, we have developed Escherichia coli-based biosensors
that accurately report the regulatory activity of these proteins and used it to investigate their interactions with
DNA and small-molecule ligands. We found that the SCO7222 protein and ActR repress the expression of their
putative target genes, SCO7223 and actII-ORF2 (actA), respectively, by interacting with operator sequence in
the promoters. The operators recognized by the two proteins are related such that O7223(an operator for
SCO7223) could be bound by both the SCO7222 protein and ActR with similar affinities. In contrast, Oact(an
operator for actII-ORF2) was bound tightly by ActR and more weakly by the SCO7222 protein. We demon-
strated ligand specificity of these proteins by showing that while TetR (but not ActR or the SCO7222 protein)
interacts with tetracyclines, ActR (but not TetR or the SCO7222 protein) interacts with actinorhodin and
related molecules. Through operator-targeted mutagenesis, we found that at least two nucleotide changes in
O7223were required to disrupt its interaction with SCO7222 protein, while ActR was more sensitive to changes
on Oact. Most importantly, we found that the interaction of each protein with wild-type and mutant operator
sequences in vivo and in vitro correlated perfectly. Our data suggest that E. coli-based biosensors of this type
should be broadly applicable to TetR-like transcription factors.
Three mechanisms commonly confer resistance to the anti-
biotic tetracycline: antibiotic degradation by TetX-like en-
zymes (31), ribosome protection by TetO-like proteins (7, 23),
and, most frequently, antibiotic export by TetA-like efflux
pumps (30). Many of the genes encoding these resistance de-
terminants are under the direct control of repressor proteins
referred to as the TetR-like transcription factors. Like most
repressors, these proteins bind operator sequences in their
target promoters preventing transcription initiation by RNA
polymerase holoenzyme. A smaller number of TetR-like pro-
teins have been implicated in transcriptional activation (6, 11).
TetR, the best-characterized member of this family, controls
the expression of the efflux pump-encoding gene tetA in the
transposon Tn10 (reviewed in reference 13). The tetR and tetA
genes are divergently transcribed and separated by ?80 bp of
DNA that includes their promoters. The TetR protein binds
tightly to two nearly identical 15-bp palindromic operator se-
quences (Otet) to interfere with transcription of both genes.
When tetracycline enters the cell, it binds the C-terminal li-
gand-binding domain in TetR as a complex with Mg2?, causing
its release from Otet(19, 25, 26). This relieves repression,
permitting the expression of tetA and the export of the antibi-
otic out of the cell (16).
Genes encoding TetR-like transcription factors are common
in bacterial genomes (reviewed in reference 20). For example,
the filamentous antibiotic-producing bacterium Streptomyces
coelicolor encodes at least 150 TetR-like transcription factors.
It is noteworthy that many of these are closely linked (either as
divergently transcribed genes or in operons) to genes that
encode proteins similar to TetA-like efflux pumps and TetX-
like monooxygenases. It is striking therefore, that relatively low
concentrations of tetracycline (1 ?g/ml) can impair this organ-
ism’s growth (21). This constitutes considerable sensitivity
given that 30 ?g/ml is required to inhibit the growth of resistant
Escherichia coli strains (1), and it suggests that the TetR-like
transcription factors control many physiological processes un-
related to tetracycline. Only a few of the S. coelicolor TetR-like
proteins have been linked to specific processes: ActR regulates
export of the polyketide actinorhodin (3, 8, 24), Pip controls a
putative multidrug resistance gene (10), PqrA controls a para-
quat efflux pump (5), CprA and CprB may interact with ?-bu-
tyrolactones (17, 18), and ScbR controls production of the
?-butyrolactone SCB1 (27).
All of the TetR-like transcription factors that have been
studied in molecular detail interact with small-molecule li-
gands that are chemically related or identical to the sub-
strates of the proteins encoded by their target genes. TetR
binds tetracycline (15), ActR interacts with actinorhodin
and actinorhodin biosynthetic intermediates (24), and QacR
binds various cationic lipophilic drugs (12). It is our view,
therefore, that identification of the small-molecule ligands
that interact with the C-terminal regulatory domains of
TetR-like proteins of unknown function is a powerful means
* Corresponding author. Mailing address: Department of Bio-
chemistry and Biomedical Sciences, HSC 4H21, McMaster Univer-
sity, 1200 Main Street West, Hamilton, Ontario, Canada L8N 3Z5.
Phone: (905) 525-9140, ext. 27335. Fax: (905) 522-9033. E-mail:
?Published ahead of print on 20 July 2007.
of deciphering their roles as well as those of the genes they
To facilitate this endeavor, we have developed a biosensor
mechanism that we believe can be applied to many members of
this family (24). The biosensors are based on synthetic pro-
moters consisting of ?10 and ?35 promoter elements, sepa-
rated by a putative binding site for a TetR-like transcription
factor. These promoters are cloned upstream of the luxCDABE
operon of Photorhabdus luminescens. LuxA and -B encode a
luciferase enzyme, while LuxC, -D, and -E form a fatty alde-
hyde reductase complex that provides the luciferase substrate
(29). The gene encoding the cognate TetR-like transcription
factor is cloned into a second plasmid such that cells containing
the lux plasmid are spontaneously bioluminescent while those
containing both are not, due to repression of the synthetic
promoter by the repressor. We showed previously with biosen-
sors of this type based on TetR and ActR that bioluminescence
could be induced by tetracycline and by actinorhodin and some
of its biosynthetic intermediates, respectively (24).
In this study, we have explored the utility of this biosensor
mechanism in greater depth, focusing on ActR, and the un-
characterized TetR-like SCO7222 protein. Based on the tetR/
tetA paradigm, the target of the SCO7222 protein is predicted
to be the divergently transcribed gene SCO7223, which en-
codes a probable TetX-like monooxygenase. We show that the
SCO7222 protein interacts with palindromic sequences in the
SCO7222/7223 intergenic sequence. These sequences are re-
lated to those bound by ActR, and indeed the two proteins
exhibit considerable affinities for each other’s operators,
though we do not believe that these heterologous interactions
are biologically relevant. SCO7222 protein does not respond to
the ligands recognized by either ActR or TetR. Mutagenesis of
the binding sites for the SCO7222 protein and ActR revealed
that the two proteins recognize distinct nucleotides and, most
importantly, that there is a perfect correlation between the in
vivo and in vitro DNA-protein interactions. We suggest, there-
fore, that this biosensor mechanism is likely to be broadly
applicable to the TetR-like proteins.
MATERIALS AND METHODS
Bacterial strains, plasmids, culture conditions, and bioluminescence mea-
surements. Bacterial strains and plasmids used in this study are described in
Table 1. E. coli cultures were grown using Luria broth (LB) or LB agar medium
containing the appropriate antibiotics when required (22). Streptomyces cultures
were grown at 30°C in YEME broth or maintained on solid MS agar medium
(14). Isolated E. coli colonies were used to inoculate 1-ml amounts of reporter
cultures, which were grown for 16 to 20 h before measurement of luminescence
using a Lumat 9507 luminometer (Bertholt Technologies). In some cases, E. coli
reporter cultures were supplemented with spent Streptomyces culture superna-
tant (prepared as described in reference 24) or with purified tetracyclines. The
half-maximal concentrations of various tetracyclines required for induction were
determined using a standard sigmoidal dose-response regression equation:
y ? bottom ? (top ? bottom)/(1 ? 10(EC50? x) · hill slope), where bottom (ymin)
was set to 1, as data were normalized using signal/noise ratios, and where
EC50represents the 50% effective concentration.
TABLE 1. Bacterial strains and plasmids used in this work
Strain or plasmidDescription (selection marker) Background
F?dcm ompT hsdS(rB
?(lacIZYA-argF)U169 deoR ?80dlac (lacZ) M15?
recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac?F? proAB lacIqZ?M15
?) gal met ?(DE3)
?) glnV44 thi-1 recA1 gyrA (Nalr) relA1
S. coelicolor prototroph, SCP1?SCP2?
S. aureofaciens tetracycline-producing strain
Promoterless luxCDABE reporter (Kanr)
Reporter based on tetA operator (Kanr)
Reporter based on actII-ORF2 operator (Kanr)
Reporter based on SCO7223 operator (Kanr)
Reporter based on actII-ORF2 operator with random mutations (Kanr)
Reporter based on SCO7223 operator with random mutations (Kanr)
Plasmid required for repressor expression vector constructions; Cmr
gene from Tn9 and Tcrgene from pSC101 (CmrTcr)
tetR gene in pACYC184; TetR-expressing vector (Cmr)
actII-ORF1 (actR) gene in pACYC184; ActR-expressing vector (Cmr)
SCO7222 gene in pACYC184; SCO7222 protein-expressing vector (Cmr)
Plasmid required to provide tetracycline resistance for DH5?-based
strains; Tcrgene from pUA466 (TcrApr)
SCO7222 gene in pBluescript to provide repressor gene flanked by
BamHI and NdeI restriction sites for pTO7222 construction (Apr)
ActR-overexpressing vector for protein purification
SCO7222-overexpressing vector for protein purification (Kanr)
pBS7222pBluescript SK II?
aKanr, kanamycin resistance; Cmr, chloramphenicol resistance; Tcr, tetracycline resistance; Apr, ampicillin resistance.
6656AHN ET AL.J. BACTERIOL.
Procedures for DNA manipulation. Standard procedures were used for plas-
mid DNA isolation, manipulation, and analysis (22). Oligonucleotide primers
used in this study were obtained from the Institute for Molecular Biology and
Biotechnology (MOBIX) facility at McMaster University or from Sigma. PCRs
were carried out using Vent DNA polymerase (New England Biolabs) with the
following program: 95°C for 5 min followed by 30 cycles of 95°C and 55°C (30 s
each), followed by 72°C (30 or 60 s, depending on the length of a template), with
a final extension step at 72°C for 7 min. DNA sequencing was carried out by the
MOBIX facility to select/isolate the appropriate PCR products.
Construction of lux-based reporter plasmids and expression vectors for the
TetR, ActR, and SCO7222 protein biosensors. pOtetlux and pOactlux were con-
structed previously (24). Similarly, a DNA fragment containing a synthetic pro-
moter—consisting of SCO7223 operator (O7223) flanked by the ?35 and ?10
regions from the Tn10 tetA promoter—was prepared by annealing T7223-1 and
T7223-2 oligonucleotides (Fig. 1B and Table 2). These oligonucleotides were
introduced into the BamHI-XhoI sites of pCS26-Pac to give reporter plasmid
pO7223lux, in which the expression of the lux operon was under the control of the
SCO7222 protein (Table 1).
Two expression vectors, pTetR and pActR (encoding TetR and ActR, respec-
tively) were previously prepared (24). In parallel, the primer pair 7222-1/7222-2
(Table 2) along with S. coelicolor M145 chromosomal DNA (as template) was
used to PCR amplify DNA fragments containing SCO7222. Gel-purified PCR
product was then digested with BamHI and ligated to BamHI-EcoRV-treated
pACYC184 to give pSCO7222, which served as the source of the SCO7222
protein in the biosensors.
Operator mutagenesis. Double-stranded DNA products, obtained by anneal-
ing the imperfectly matched primers TA3 and TA2 and T7223-3 and T7223-2,
were inserted into the BamHI-EcoRV sites of pCS26-Pac to give pMAlux (TA3-
TA2) and pMTlux (T7223-3-T7223-2), respectively. TA3 and T7223-3 are iden-
tical to TA1 and T7223-1, respectively, except that the operator sequences had
been doped during synthesis as follows. At each operator position, while the
concentration of the correct nucleotide was as usual during synthesis, the other
three nucleotides were also present, as defined by x/3n, where x represents the
concentration of principle nucleotide and n represents the length of the operator
sequence. The result of this is that each mutagenic oligonucleotide would be
expected to possess at least one nucleotide sequence change in the operator
region, embedded in an otherwise “wild-type” synthetic promoter. In both cases,
we mutagenized the top strands shown in Fig. 1B. The plasmids pMAlux and
pMTlux were introduced into E. coli strains containing pActR or pSCO7222,
respectively, to isolate colonies that produced luminescence, indicating that the
respective repressors could not bind to the mutagenized operators. Plasmid
DNA was isolated from these strains and reintroduced into E. coli to isolate
kanamycin-resistant and chloramphenicol-sensitive colonies, which had lost
pActR or pSCO7222 but still contained pMAlux or pMTlux. The DNA se-
FIG. 1. Putative binding sites for ActR and the SCO7222 protein. (A) Repeated elements in the actR/actA (modified from reference 24) and
SCO7222/SCO7223 intergenic regions are shown. Between actR and actA, there are three weakly palindromic sequences (underlined) that exhibit
a low degree of conservation. Between SCO7222 and SCO7223 there are three 15-bp repeats that are perfectly palindromic and highly conserved.
Arrows indicate repeated sequence (actR/actA) and palindromes (SCO7222/SCO7223). (B) Sequence of the synthetic promoter used in the
pO7223lux. The DNA fragment contained ?35 and ?10 regions from the Tn10 tetA promoter, flanking O7223. Arrows indicate the palindromic
nucleotides in O7223and Otet.
VOL. 189, 2007REGULATORY ACTIVITY OF TetR-LIKE TRANSCRIPTION FACTORS 6657
quences of the operators cloned in pMAlux and pMTlux were determined for
Expression and purification of His6-ActR and His6-SCO7222 protein. Previ-
ously prepared pET28a-ActR was used to express and purify N-terminal six-His-
tagged ActR (His6-ActR) in E. coli (24). The primers TOEV-1 and TOEV-2
(Table 2), along with pSCO7222 plasmid DNA (as template), were used to PCR
amplify a 721-bp fragment containing SCO7222, which was introduced into the
EcoRV site of pBlueScript SK II?to give pBS7222. After sequencing SCO7222,
the DNA fragment encoding SCO7222 protein was isolated as an NdeI-BamHI
fragment from pBS7222 and was ligated to pET28a, giving pTO7222 (Table 1).
E. coli BL21(DE3) cultures containing pTO7222 were grown at 37°C to an
optical density at 600 nm of ?0.4 to 0.6 and were then induced with 1 mM
isopropyl ?-D-thiogalactopyranoside for 20 h at 30°C. From this point, the same
procedures were taken to purify His6-ActR and His6-SCO7222 (24).
EMSAs. The primers TET-EMSA-F and TET-EMSA-R (24), along with
pOactlux, pO7223lux, or the reporter plasmids with point-mutagenized operators
(as templates), were used in PCRs to isolate double stranded DNA fragments
containing operator regions, which served as probes for electrophoretic mobility
shift assays (EMSAs). The DNA fragments were end labeled using [?-32P]ATP
and T4 polynucleotide kinase (22).
Labeled probe (12.7 fmol), 1.5 to 3,200 fmol of purified protein and 90 ng of
salmon sperm DNA were used in 15-?l reactions containing 1? reaction buffer
(10 mM Tris-Cl [pH 7.8], 150 mM NaCl, 2 mM dithiothreitol, 10% glycerol).
Reaction mixtures were incubated at 30°C for 10 min and were fractionated on
12% nondenaturing polyacrylamide gels containing 1.5% glycerol. The gels were
exposed using a phosphor screen (Amersham), and bands were detected using a
PhosphorImager (Molecular Dynamics).
Determination of KD. ImageQuant software (Molecular Dynamics) was used
to analyze EMSA results to determine the percentages of shifted and unshifted
probes, which represent bound and unbound substrate, respectively. Saturation
curves (% probe bound against [protein]) were drawn with SigmaPlot 2000 to
determine the dissociation constant (KD). Binding cooperativity was determined
by Hill’s equation: log (Y/1 ? Y) ? h log[S] ? log KD, where Y ? protein-DNA
complex/total DNA, S is the protein of interest, and KDis the dissociation
constant of protein binding to DNA (9). The slope of a straight line passing the
point where 50% DNA binding occurs, is the Hill’s coefficient (h) and indicates
the binding cooperativity.
RESULTS AND DISCUSSION
Construction of biosensors using TetR, ActR, and the
SCO7222 protein. Given the TetR paradigm, we predicted that
the likely target of the SCO7222 protein would be SCO7223.
Examination of the 110 bp that separates these genes revealed
the sequences N1, N2, and N3 (Fig. 1A) corresponding to
likely binding sites for the SCO7222 protein. All of these se-
quences exhibit significant similarity to the nearly perfect pal-
indrome having the consensus sequence C/TTGGAACGNCG
TTCCAG/C. This is similar to the intergenic DNA of actR and
actA that includes the putative ActR-binding sequences P1, P2
(Oact), and P3. We carried out gel mobility shift assays using
both intergenic fragments and purified His6-SCO7222 and
His6-ActR. As shown in Fig. 2, both fragments formed three
exceptionally tight complexes with their cognate repressors,
consistent with the presence of three repressor binding sites in
each. We determined approximate KDs for each of the
SCO7222 protein complexes and found them to be 0.4 nM, 3.4
nM, and 17.3 nM for the L, M, and H complexes, respectively.
A Hill plot of this data demonstrates a Hill’s coefficient of 1.6,
consistent with positive cooperative binding of the SCO7222
protein (Hill’s coefficient of ?1) to its three binding sites. The
interaction of ActR with the actR/actA intergenic sequence was
also characterized by strong binding (KDs of 0.1 nM, 0.3 nM,
and 5.8 nM), but with a Hill’s coefficient of ?1 (data not
shown), there was no evidence of cooperativity.
As previously, we designed synthetic promoters in which
each of the N1, N2, and N3 sequences was introduced between
functional E. coli ?10 and ?35 promoter elements and intro-
duced them upstream of luxCDABE operon in the vector
pCS26-Pac (2). The resulting plasmids pN1lux, pN2lux, and
pO7223lux were introduced into E. coli along with the previ-
ously constructed pOtetlux and pOactlux, and all conferred sig-
nificant bioluminescence compared to promoterless reporter
vector pCS26-Pac (Table 3). Similarly, we constructed an
SCO7222 protein-producing plasmid by amplifying the gene
with oligonucleotides that included a Shine-Dalgarno se-
quence upstream of the initiator codon and a BamHI site at
the downstream end. This fragment was ligated downstream of
a tetA promoter in the vector pACYC184 to create the plasmid
pSCO7222. We introduced pTetR, pActR, pSCO7222, and the
control pACYC184 into the E. coli strains bearing pOtetlux,
TABLE 2. Primers used in this study
Name Sequence (5?33?)a
T7223-1TCGAGTTGACACTGGAACGCCGTTCCAGTTATTTTACCAPositive-strand oligonucleotide for preparing
Negative-strand oligonucleotide for preparing the
Forward primer for amplifying SCO7222 gene for
Reverse primer for amplifying SCO7222 gene for
Forward primer for amplifying SCO7222 gene
from pSCO7222 for pBS7222 construction
Reverse primer for amplifying SCO7222 gene
from pSCO7222 for pBS7222 construction
Forward primer for amplifying intergenic region
between SCO7222 and SCO7223
Reverse primer for amplifying intergenic region
between SCO7222 and SCO7223
Forward primer for amplifying intergenic region
between actR and actA
Reverse primer for amplifying intergenic region
between actR and actA
aCompatible cohesive ends and restriction endonuclease recognition sequences introduced by these oligonucleotides are underlined.
6658AHN ET AL.J. BACTERIOL.
pOactlux, pN1lux, pN2lux, and pO7223lux and tested each strain
for bioluminescence. As previously reported, the pTetR and
pActR eliminated pOtetlux- and pOactlux-dependent biolumi-
nescence, respectively, while the control plasmid pACYC184
had no effect (24). Similarly, pSCO7222 eliminated the biolu-
minescence produced by pN1lux, pN2lux, and pO7223lux (Ta-
ble 3), confirming direct binding as observed in vitro. The in
vivo function of these three sites agrees with the mobility shift
demonstrating three complexes between the SCO7222 protein
and the SCO7222/SCO7223 intergenic sequence. This sug-
gested that the SCO7222 protein is a repressor of SCO7223,
and we focused the rest of our work on N3, which we refer to
O7223. These results underscore the fact that in those cases
(which are common) in which a tetR-like gene is paired with a
putative target gene in a manner similar to the well-character-
ized tetR/tetA configuration, repeated palindromic sites in the
intergenic sequences can be assumed to be candidate binding
sites for the TetR-like protein.
We noted that the putative operators recognized by ActR
(ACGCGACCACCGTTCCAT) and the SCO7222 protein (C
TGGAACGACGTTCCAG) were similar, particularly in their
right half-sites. To determine whether either repressor could
bind the other’s operator, we combined pActR with pO7223lux
biosensor and pSCO7222 with pOactlux biosensor. We ob-
FIG. 2. Interactions of the SCO7222 protein and ActR with intergenic sequences. (A) A32P-labeled probe containing the entire SCO7222/
SCO7223 or actR/actA intergenic DNA (IR) sequences was incubated with the indicated concentrations of the SCO7222 protein or ActR at 30°C
for 10 min and then separated in 12% nondenaturing polyacrylamide gels. Three distinguishable shifts, L (lower shift), M (middle shift), and H
(higher shift), were observed in both cases. (B to D) The KDvalue of each SCO7222 protein shift was determined by plotting the percentage of
bound DNA against the concentration of the SCO7222 protein. (E) Hill’s plot was drawn to measure binding cooperativity by the SCO7222 protein.
VOL. 189, 2007 REGULATORY ACTIVITY OF TetR-LIKE TRANSCRIPTION FACTORS6659
served strong repression of pO7223lux-dependent biolumines-
cence by pActR (to background) and partial repression of
pOactlux-dependent bioluminescence (?1.5-fold) by pSCO7222
Effects of various ligands on TetR, ActR, and the SCO7222
protein. To determine whether tetracycline could relieve re-
pression by SCO7222 protein, we introduced plasmid pUAO1,
bearing the tetO gene from Campylobacter jejuni (28), into each
biosensor strain to protect them against the antibiotic. The
TetO gene product is a ribosomal protection protein that al-
lows relatively high cytoplasmic concentrations of the antibi-
otic to be tolerated without substantial damage to the cells.
The addition of pUAO1 had no major impact, although it
slowed bacterial growth to ?1 doubling per hour.
To determine whether the resulting strains could detect tet-
racycline, we cultured each of them in the presence of various
concentrations of the drug (Fig. 3A). As discussed previously,
we observed the dose-responsive induction of bioluminescence
in the TetR biosensor that peaked when 4 ?g/ml tetracycline
was used (24). Higher concentrations of drug inhibited cell
growth, in spite of the presence of the tetO gene, but even at
levels where there was significant inhibition of growth, we
could still detect bioluminescence in the TetR biosensor. Con-
sistent with previous results, tetracycline did not induce biolu-
minescence in the ActR biosensor (24). The presence of tet-
racycline also did not relieve repression by SCO7222 protein.
Similar analysis was conducted using various concentrations of
demeclocycline, doxycycline, oxytetracycline, chlortetracycline,
TABLE 3. Effects of TetR/ActR/SCO7222 on various
Error (?1 SD)
pN1lux ? pACYC184
pN1lux ? pSCO7222
pN2lux ? pACYC184
pN2lux ? pSCO7222
pO7223lux ? pACYC184
pO7223lux ? pSCO7222
pO7223lux ? pTetR
pO7223lux ? pActR
pOactlux ? pACYC184
pOactlux ? pSCO7222
2.19 ? 105
1.21 ? 105
1.09 ? 104
1.25 ? 104
8.87 ? 104
5.70 ? 104
8.22 ? 104
3.01 ? 104
1.86 ? 104
3.9 ? 104
3.4 ? 104
2.0 ? 103
1.0 ? 103
2.7 ? 103
5.8 ? 103
1.1 ? 104
2.4 ? 103
2.4 ? 103
aAnalysis was carried out using E. coli XL1-Blue as the host.
bAll values are in relative light units (RLU) and represent the average of at
least three independent readings.
FIG. 3. Effect of purified tetracycline (A) or S. coelicolor supernatant (B) on TetR-, ActR-, and SCO7222 protein-mediated repression. (A) One
to 5 ?g of pure tetracycline was added to E. coli biosensors harboring pUOA1 and either pOtetlux, pOactlux, or pO7223lux along with the vector
expressing the cognate repressors, respectively. (B) Ten to 40% (vol/vol) S. coelicolor supernatant was added to the same biosensor strains
described above. All values were measured in relative light units (RLU), and error bars indicate ?1 standard deviation of values obtained from
three independent readings.
6660 AHN ET AL. J. BACTERIOL.
and methatetracycline, all antibiotics related in structure and
mechanism to tetracycline. While the concentrations required
for half-maximal induction varied (Table 4), all of the tetracy-
cline derivatives analyzed were good inducers of the TetR
biosensor (data not shown). Again, none of these molecules
had any effect on the ActR-based (24) or the SCO7222 pro-
tein-based strain (data not shown) biosensors. We then applied
culture supernatants from Streptomyces aureofaciens (a tetra-
cycline producer) and S. coelicolor to all three biosensors to
determine whether either streptomycete produced secondary
metabolites that could interact with the SCO7222 protein. As
demonstrated previously, the S. aureofaciens and S. coelicolor
supernatants were able to activate bioluminescence in the
TetR- and ActR-based biosensors, respectively. Neither, how-
ever, had any effect on the SCO7222 protein-based biosensor
(Fig. 3B). Under the growth conditions we have employed in
this work, therefore, neither strain produces an SCO7222 pro-
tein ligand, although it is possible that there was some inducing
molecule in the S. coelicolor supernatant that could not cross
the E. coli envelope.
Mutagenesis of operator sequences. Unlike pCS26-Pac,
which is a low-copy-number plasmid, pACYC184 is propa-
gated at a moderately high copy number in E. coli: ?20 copies
per cell (4). As a result, the TetR-like transcription factors in
our biosensors are probably in excess of their targets, poten-
tially impairing small-molecule induction and explaining the
observed ActR-SCO7222 protein cross talk. To provide a basis
for comparing the in vivo interactions of the repressors and
operators with their biochemical affinities, we constructed mu-
tants of Oactand O7223using two strategies. In the first, we
subjected each operator sequence to randomization and
screened for nonfunctional operators that exhibited biolumi-
nescence in the presence of their cognate repressors (see Ma-
terials and Methods). We screened ?1,600 colonies (?800 for
each repressor) and isolated 16 Oactsequences impaired in
their interaction with ActR and 18 O7223sequences impaired in
their interaction with the SCO7222 protein (Table 5).
Interestingly, all but one (pMTlux10) of the repressor-resis-
tant alleles exhibited multiple mutations, suggesting either that
the high in vivo concentration of the repressor resulted in
DNA binding that was relatively resistant to mutation or that
the affinity of each repressor for its cognate operator was
simply very high due to a large number of specific repressor-
operator interactions. In Oact, many of the alleles were altered
TABLE 5. Sequences of O7223and Oactmutants not recognized by
the SCO7223 protein and ActRa
aOnly nucleotides found to be different from those present at the same
position in the wild-type operator sequences are shown. ?, nucleotide deleted; N,
not applicable as either this nucleotide or the previous one has been deleted in
the same operator sequence analyzed.
TABLE 6. Summary of in vivo and in vitro assay results with the
aFor point mutants, only the specific nucleotide changes made are shown.
bY, strong repression; N, either no or weak repression (less than fivefold
cNot applicable as no shift was observed even at the highest concentration of
TABLE 4. Induction of TetR-controlled gene expression
by various tetracyclinesa
Concn (?g/ml) for half-
aPlasmids were cultivated in E. coli strain DH5?.
bArbitrary light units were calculated from at least three independent read-
VOL. 189, 2007 REGULATORY ACTIVITY OF TetR-LIKE TRANSCRIPTION FACTORS6661
at position 8 or 11, while in O7223, many alleles had mutations
at position 5, 14, or 15 (positions 14 and 15 were two nucleo-
tides more frequently involved in O7223double mutants [Table
5]). In our second strategy, therefore, we subjected these base
pairs to site-directed mutagenesis. We generated the alleles
OactA8G, OactG11C, O7223A5G, O7223C14G, and O7223C15A
and assessed each for repression by their cognate repressors.
Position 5 of Oact(OactA5G) was also tested since it is one of
the palindromic nucleotides in the operator. As shown in Table
6, while the point mutation at position 8 of Oacthad no effect,
those at positions 5 and 11 eliminated in vivo repression by
ActR. None of the point mutations in O7223altered repression
of bioluminescence in vivo.
We also assessed the effects of each point mutation on cross-
repression by ActR and the SCO7222 protein and found that
the results were quite different. For example, O7223A5G and
O7223C15A, which had no effect on repression by the SCO7222
protein, eliminated cross-repression by ActR. Clearly, there-
FIG. 4. Characterization of interactions between ActR or SCO7222 protein and Oactor O7223. A 12.7-fmol amount of32P-labeled probes was
incubated with 1.5 to 210 nM protein at 30°C for 10 min and then separated in 12% nondenaturing polyacrylamide gels. Interactions shown on
the gels are between the SCO7222 protein and O7223/Oact(A), ActR and O7223/Oact(B), and SCO7222 protein/ActR and O7223A5G (C). CP,
protein-DNA complex. KDs were determined by drawing the saturation curves and obtaining [SCO7222 protein] at half-maximal saturation.
6662AHN ET AL. J. BACTERIOL.
fore, while the operator sequences are similar, the manners in
which they are recognized by the two proteins in the context of
our biosensors are subtly different.
Correlation of binding strengths with in vivo repression. To
assess the significance of these results, we measured the affin-
ities of the SCO7222 protein and ActR for O7223and Oactand
for the relevant point mutants described above. Using purified
His6-SCO7222 and His6-ActR, we carried out gel mobility shift
experiments with radioactively labeled probes corresponding
to each operator sequence. The interaction of both proteins
with their cognate operators was exceptionally strong (Fig. 4).
Figure 4A shows gel mobility shift results and saturation curves
for the SCO7222 protein. Consistent with the fact that the
SCO7222 protein caused complete inhibition of transcription
on O7223but not on Oactin vivo, KDs for these interactions
were ?2 nM for O7223and ?30 nM for Oact. Even though the
affinity of the SCO7222 protein for Oactwas much lower than
that for O7223, binding at higher concentrations explains the
consistent weak repression this protein brought about on the
Oact-regulated promoter in vivo (Table 3). Figure 4B shows
mobility shift results for ActR. ActR could bind to O7223as
efficiently as Oact, in agreement with the capacity of this protein
to repress both Oact- and O7223-regulated transcription in vivo
(Table 3). Unlike the SCO7222 protein, ActR produced two
shifted bands with both Oactand O7223, raising the possibility
that two ActR dimers interact with each site. Protein titration
and Hill plot analysis did not indicate significant cooperativity
in the assembly of these complexes (data not shown).
We went on to assess the effect of each point mutation in
Oactand O7223on the affinities of these DNAs for ActR and
the SCO7222 protein and found a nearly perfect correlation
between in vitro binding and in vivo repression (Fig. 4C and
Table 6). For example, in E. coli, the SCO7222 protein re-
pressed bioluminescence from the O7223A5G operator, but
ActR did not. Consistent with this, the SCO7222 protein
bound to O7223A5G with a similar affinity to O7223, while much
higher concentrations of ActR were required to form a com-
plex with O7223A5G compared to either Oactor O7223(Fig. 4C).
In summary, the KDs of the SCO7222 protein for the three
mutant operators it was able to repress in vivo (O7223A5G,
O7223C14G, and O7223C15A) were all in the 2 to 3 nM range,
whereas its KDs for those operators it could not repress in vivo
were at least five times higher, and in one case, OactG11C, we
observed no interaction at all.
This work raises interesting questions about possible cross
talk between these two repressors. In our view, it is unlikely
that the interaction of these proteins with heterologous oper-
ators is biologically meaningful. SCO7222 and -7223 are lo-
cated far outside the act gene cluster, which contains most, if
not all, of the genes necessary for actinorhodin biosynthesis
and self-resistance. How might cross talk be avoided between
these two and perhaps other members of this very large gene
family in vivo? One possibility is that coupled transcription and
translation, which is a characteristic of all bacteria, results in
the direct delivery of a TetR protein from the ribosome on
which it is synthesized to its cognate operator. This would be
consistent with the fact that the repressor-encoding genes tend
to be closely linked to their target genes. If, as is the case for
TetR, ActR and the SCO7222 protein regulate their own pro-
duction as well as that of their target resistance genes, they
would limit their own intracellular accumulation and hence
minimize or eliminate cross talk. Temporal and spatial regu-
lations might also be responsible for restricting the cross talk.
More importantly for our immediate purposes, the high de-
gree of correlation between in vitro binding and in vivo repres-
sion and the fact that we were able to specifically detect several
tetracyclines as pure molecules and from a natural source
suggest that this biosensor mechanism is a valuable tool for
investigating TetR-like transcription factors.
We thank Michael Surette and Gerard Wright for supplying pCS26-
Pac and for sharing various tetracyclines, respectively. We also thank
Eric Brown, John Capone, and Gerard Wright for use of their labo-
K.T. was supported by a postdoctoral fellowship from the Natural
Science and Engineering Research Council. This work was funded by
grant MOP-57684 from the Canadian Institutes for Health Research.
1. Bannam, T. L., and J. I. Rood. 1999. Identification of structural and func-
tional domains of the tetracycline efflux protein TetA(P) from Clostridium
perfringens. Microbiology 145:2947–2955.
2. Bjarnason, J., C. M. Southward, and M. G. Surette. 2003. Genomic profiling
of iron-responsive genes in Salmonella enterica serovar Typhimurium by
high-throughput screening of a random promoter library. J. Bacteriol. 185:
3. Caballero, J. L., F. Malpartida, and D. A. Hopwood. 1991. Transcriptional
organization and regulation of an antibiotic export complex in the producing
Streptomyces culture. Mol. Gen. Genet. 228:372–380.
4. Chang, A. C. Y., and S. N. Cohen. 1978. Construction and characterization of
amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic
miniplasmid. J. Bacteriol. 134:1141–1156.
5. Cho, Y.-H., E.-J. Kim, H.-J. Chung, J.-H. Choi, K. F. Chater, B.-E. Ahn, J.-H.
Shin, and J.-H. Roe. 2003. The pqrAB operon is responsible for paraquat
resistance in Streptomyces coelicolor. J. Bacteriol. 185:6756–6763.
6. Christen, S., A. Srinivas, P. Bahler, A. Zeller, D. Pridmore, C. Bieniossek, U.
Baumann, and B. Erni. 2006. Regulation of the Dha operon of Lactococcus
lactis: a deviation from the rule followed by the TetR family of transcription
regulators. J. Biol. Chem. 281:23129–23137.
7. Connell, S. R., C. A. Trieber, G. P. Dinos, E. Einfeldt, D. E. Taylor, and K. H.
Nierhaus. 2003. Mechanism of Tet(O)-mediated tetracycline resistance.
EMBO J. 22:945–953.
8. Fernandez-Moreno, M. A., J. L. Caballero, D. A. Hopwood, and F. Malpar-
tida. 1991. The act cluster contains regulatory and antibiotic export genes,
direct targets for translational control by the bldA tRNA gene of Streptomy-
ces. Cell 66:769–780.
9. Fersht, A. 1985. Enzyme structure and mechanism, 2nd ed. W. H. Freeman
and Company, New York, NY.
10. Folcher, M., R. P. Morris, G. Dale, K. Salah-Bey-Hocini, P. H. Viollier, and
C. J. Thompson. 2001. A transcriptional regulator of a pristinamycin resis-
tance gene in Streptomyces coelicolor. J. Biol. Chem. 276:1479–1485.
11. Girard, G., S. Barends, S. Rigali, E. Tjeerd van Rij, B. J. J. Lugtenberg, and
G. V. Bloemberg. 2006. Pip, a novel activator of phenazine biosynthesis in
Pseudomonas chlororaphis PCL1391. J. Bacteriol. 188:8283–8293.
12. Grkovic, S., M. H. Brown, N. J. Roberts, I. T. Paulsen, and R. A. Skurray.
1998. QacR is a repressor protein that regulates expression of the Staphylo-
coccus aureus multidrug efflux pump QacA. J. Biol. Chem. 273:18665–18673.
13. Hillen, W., and C. Berens. 1994. Mechanisms underlying expression of Tn10
encoded tetracycline resistance. Annu. Rev. Microbiol. 48:345–369.
14. Kieser, T., M. J. Bibb, M. J. Buttner, K. F. Chater, and D. A. Hopwood. 2000.
Practical Streptomyces genetics. The John Innes Foundation, Norwich, En-
15. Kisker, C., W. Hinrichs, K. Tovar, W. Hillen, and W. Saenger. 1995. The
complex formed between Tet repressor and tetracycline-Mg2?reveals mech-
anism of antibiotic resistance. J. Mol. Biol. 247:240–280.
16. McMurry, L., R. E. Petrucci, Jr., and S. B. Levy. 1980. Active efflux of
tetracycline encoded by four genetically different tetracycline resistance de-
terminants in Escherichia coli. Proc. Natl. Acad. Sci. USA 77:3974–3977.
17. Natsume, R., Y. Ohnishi, T. Senda, and S. Horinouchi. 2004. Crystal struc-
ture of a gamma-butyrolactone autoregulator receptor protein in Streptomy-
ces coelicolor A3(2). J. Mol. Biol. 336:409–419.
18. Onaka, H., T. Nakagawa, and S. Horinouchi. 1998. Involvement of two
A-factor receptor homologues in Streptomyces coelicolor A3(2) in the regu-
lation of secondary metabolism and morphogenesis. Mol. Microbiol. 28:743–
VOL. 189, 2007REGULATORY ACTIVITY OF TetR-LIKE TRANSCRIPTION FACTORS 6663
19. Orth, P., W. Saenger, and W. Hinrichs. 1999. Tetracycline-chelated Mg2? Download full-text
ion initiates helix unwinding in Tet repressor induction. Biochemistry 38:
20. Ramos, J. L., M. Martinez-Bueno, A. J. Molina-Henares, W. Tera ´n, K.
Watanabe, X. Zhang, M. T. Gallegos, R. Brennan, and R. Tobes. 2005. The
TetR family of transcriptional repressors. Microbiol. Mol. Biol. Rev. 69:326–
21. Rodriguez-Garcia, A., P. Combes, R. Perez-Redondo, M. C. Smith, and
M. C. Smith. 2005. Natural and synthetic tetracycline-inducible promoters
for use in the antibiotic-producing bacteria Streptomyces. Nucleic Acids Res.
22. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a
laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY.
23. Sougakoff, W., B. Papadopoulou, P. Nordmann, and P. Courvalin. 1987.
Nucleotide sequence and distribution of the gene tetO encoding tetracycline
resistance in Campylobacter coli. FEMS Microbiol. Lett. 44:153–159.
24. Tahlan, K., S. K. Ahn, A. Sing, T. D. Bodnaruk, A. R. Willems, A. R.
Davidson, and J. R. Nodwell. 2007. Initiation of actinorhodin export in
Streptomyces coelicolor. Mol. Microbiol. 63:951–961.
25. Takahashi, M., L. Altschmied, and W. Hillen. 1986. Kinetic and equilibrium
characterization of the Tet repressor-tetracycline complex by fluorescence
measurements. Evidence for divalent metal ion requirement and energy
transfer. J. Mol. Biol. 187:341–348.
26. Takahashi, M., J. Degenkolb, and W. Hillen. 1991. Determination of the
equilibrium association constant between Tet repressor and tetracycline at
limiting Mg2?concentrations: a generally applicable method for effector-
dependent high-affinity complexes. Anal. Biochem. 199:197–202.
27. Takano, E., R. Chakraburtty, T. Nihira, Y. Yamada, and M. J. Bibb. 2001. A
complex role for the gamma-butyrolactone SCB1 in regulating antibiotic
production in Streptomyces coelicolor A3(2). Mol. Microbiol. 41:1015–1028.
28. Taylor, D. E., K. Hiratsuka, H. Ray, and E. K. Manavathu. 1987. Charac-
terization and expression of a cloned tetracycline resistance determinant
from Campylobacter jejuni plasmid pUA466. J. Bacteriol. 169:2984–2989.
29. Winson, M. K., S. Swift, P. J. Hill, C. M. Sims, G. Griesmayr, B. W. Bycroft,
P. Williams, and G. S. Stewart. 1998. Engineering the luxCDABE genes from
Photorhabdus luminescens to provide a bioluminescent reporter for consti-
tutive and promoter probe plasmids and mini-Tn5 constructs. FEMS Micro-
biol. Lett. 163:193–202.
30. Yamaguchi, A., T. Udagawa, and T. Sawai. 1990. Transport of divalent
cations with tetracycline as mediated by the transposon Tn10-encoded tet-
racycline resistance protein. J. Biol. Chem. 265:4809–4813.
31. Yang, W., I. F. Moore, K. P. Koteva, D. C. Bareich, D. W. Hughes, and G. D.
Wright. 2004. TetX is a flavin-dependent monooxygenase conferring resis-
tance to tetracycline antibiotics. J. Biol. Chem. 279:52346–52352.
6664 AHN ET AL.J. BACTERIOL.