Tet repressor mutants with altered effector binding and
allostery
Eva-Maria Henßler, Ralph Bertram, Stefanie Wisshak and Wolfgang Hillen
Lehrstuhl fu
¨
r Mikrobiologie, Institut fu
¨
r Biologie, Friedrich-Alexander Universita
¨
t Erlangen-Nu
¨
rnberg, Erlangen, Germany
Tetracycline (tc) resistance in Gram-negative bacteria
is often regulated by the Tet Repressor (TetR), a tc-
responsive allosterical DNA-binding protein. Due to
three very advantageous properties, namely the highly
specific binding of TetR to tet operator (tetO), the sen-
sitive induction by small amounts of tc, and the ability
of this drug to penetrate into most cells, the TetR
based regulation systems are widely used for condi-
tional gene expression [1]. Many biochemical studies
and crystal structures of TetR in all complexed forms
[2–4] have led to a detailed understanding of the regu-
latory mechanism.
TetR is an all a-helical, dimeric protein in which
tetO recognition is accomplished by a helix-turn-helix
motif consisting of helices a2 and a3 at the N-termi-
nus. The core domain (a5toa10) contains the tc bind-
ing pocket and the dimerization motif. Both domains
are connected by helix a4, and their interface is formed
by residues of helices a1, a4, and a6 (Fig. 1A). As

allostery; effector specificity; reverse TetR;
tetracycline derivatives; Tet repressor
Correspondence
W. Hillen, Lehrstuhl fu
¨
r Mikrobiologie,
Institut fu
¨
r Biologie, Friedrich-Alexander
Universita
¨
t Erlangen-Nu
¨
rnberg,
Staudtstraße 5, 91058 Erlangen, Germany
Fax: +49 9131 ⁄ 85 28082
Tel: +49 9131 ⁄ 85 28081
E-mail:
(Received 12 May 2005, revised 12 July
2005, accepted 15 July 2005)
doi:10.1111/j.1742-4658.2005.04868.x
To learn about the correlation between allostery and ligand binding of the
Tet repressor (TetR) we analyzed the effect of mutations in the DNA read-
ing head–core interface on the effector specific TetR
i2
variant. The same
mutations in these subdomains can lead to completely different activities,
e.g. the V99G exchange in the wild-type leads to corepression by 4-ddma-
atc without altering DNA binding. However, in TetR
i2

i2
variant
exhibiting 4-ddma-atc specificity to combine it with
degenerations of residues giving rise to revTetR
mutants and screen for the combined phenotype. The
results lead to insights about the allostery of TetR.
Results
Randomization of residues in helices a1, a4,
or a6 in TetR
i2
As the most efficient and mechanistically most interest-
ing mutations leading to revTetR occurred in helices
a1, a4, and a6 [8], we decided to combine randomiza-
tions in these helices with TetR
i2
(containing the muta-
tions H64K, S135L and S138I). We screened the
resulting candidates for TetR variants with 4-ddma-atc
specific corepression in Escherichia coli WH207 ⁄ ktet50
[10]. The specificity is scored against atc, the most
efficient effector of TetR known so far. The DNA
fragments containing the randomized codons 14–25
(C-terminal part of helix a1 and the following loop)
and 93–102 (helix a6 and the b-turns N-terminal and
C-terminal of a6) as described previously were intro-
duced into pWH1925-tetR
i2
. The randomized codons
50–63 in helix a4 generated by PCR mutagenesis using
a ‘doped’ oligonucleotide also encoding the H64K

C-terminus of helix a4 is connected to the [tc-Mg]
+
binding pocket
by interaction of His64 with tc. Leu52 and Leu56 form the hydro-
phobic region contacting Val99. Thr103 is located in the C-terminal
helical turn of a6 which is transformed into a type II b-turn upon
induction [20]. (C) Chemical structures of anhydrotetracycline (atc)
and 4-de-dimethylamino-anhydrotetracycline (4-ddma-atc).
Altered TetR effector binding and allostery E M. Henßler et al.
4488 FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS
WH207 ⁄ ktet50 containing a chromosomal tetA-lacZ
fusion as summarized in Table 1.
Three candidates with mutations in helix a1 exhibit
4-ddma-atc dependent repression, but show different
specificities as scored by the lack of atc-mediated repres-
sion. TetR
i2
-E15A L17V shows repression to 12.5%
with 4-ddma-atc and to 58% with atc, thus showing
only fivefold improved repression with 4-ddma-atc over
atc. A similar result is found for TetR
i2
-L17V V20F,
while the single exchange mutant TetR
i2
-N18Y shows
no distinction between atc and 4-ddma-atc.
Randomization of helix a4 also led to three candi-
dates. The amino-acid exchanges I59L L60M yield a
TetR

TetR variants were found for the residues V99 and
L17 [8] and V99G did reverse the TetR
i2
activity [9] we
analyzed the role of substitutions at these positions for
both effectors in the wild-type and TetR
i2
sequence
backgrounds.
TetR
i2
variants with mutations at valine 99
We revisited the 19 possible exchanges at positions 99
in TetR [8] and introduced them into TetR
i2
. The
in vivo repression and induction for these 20 TetR vari-
ants is shown in Fig. 2A,B. Fig. 2A shows the activit-
ies of V99 exchanges in TetR for the effectors atc and
4-ddma-atc and without effector. 11 out of 19 substitu-
tions at V99 do not lead to large changes of the phe-
notype. However, 11 out of the 19 exchanges show
slightly enhanced repression with 4-ddma-atc com-
pared to without thus making it a corepressor for these
variants. Interestingly, the mutations V99I, V99M,
V99P, V99G and V99F turn 4-ddma-atc into a core-
pressor while atc is still an inducer. This is remarkable
because a residue at position 99 is not in contact with
the effector [2]. Six out of these 11 exchanges exhibit a
reverse phenotype with 4-ddma-atc in the TetR

TetR background. The V99G exchange in TetR only
marginally influences induction with atc but leads to
Table 1. In vivo repression and induction of TetR
i2
variants. The
expression of 100% b-galactosidase corresponds to 6300 ± 1050
units.
TetR variant
b-Gal activity (%)
Induction with
4-ddma-atc
(0.4 l
M)
atc
(0.4 lM)
TetR
i2
1.6 ± 0.1 57 ± 5 3.1 ± 0.5
TetR
i2
-N18Y 60 ± 1.3 20 ± 1 14 ± 1.2
TetR
i2
-L17V V20F 80 ± 6 11 ± 0.5 34 ± 0.7
TetR
i2
-E15A L17V 85 ± 2 12.5 ± 0.4 58 ± 8
TetR
i2
-I59S 87 ± 1 19 ± 0.6 46 ± 1

TetR
i2
-R94H D95N G96R 60 ± 1.4 2 ± 0.1 4 ± 0.4
E M. Henßler et al. Altered TetR effector binding and allostery
FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS 4489
improved repression with 4-ddma-atc. The same
exchange leads to one of the best 4-ddma-atc specific
reverse phenotypes in TetR
i2
-V99G. Another very spe-
cific reverse phenotype is found for TetR
i2
-V99N with
a 34-fold better repression in the presence of 4-ddma-
atc compared to atc. Again, this exchange has only a
slight effect on induction with atc when introduced in
the TetR sequence background.
Western blot analyses of the V99A, V99G, V99S,
V99T and V99N exchanges in wild-type and TetR
i2
revealed only small differences of the intracellular pro-
tein amounts (Fig. 2C), which may correlate to the
alterations seen in the repressed expression levels of
the respective mutants. The observed specificity chan-
ges are clearly not influenced by protein amounts.
TetR
i2
variants with mutations at leucine 17
The effects of all possible exchanges of leucine at
position 17 in the wild-type or TetR

showing repression to 3% with 4-ddma-atc
and only to 59% with atc. In contrast, TetR-L17A is
noninducible with atc or 4-ddma-atc, while TetR-
L17G shows the best atc dependent reverse pheno-
type. TetR
i2
-L17G, on the other hand, is inactive.
Taken together, it is surprising that single residue
exchanges cause quite different effects in these two
sequence backgrounds. Moreover, contrary activities
are caused by very small differences in side chains.
The determination of the intracellular protein
amounts (Fig. 3C) excludes contributions to these spe-
cificity changes.
Specificity determining residues in TetR
i2
-V99G
TetR
i2
-V99G is one of the best revTetR
i2
variants with
4-ddma-atc specific repression, almost completely lack-
ing repression with atc [9]. As the V99G exchange in
the wild-type sequence background displays no reverse
phenotype and slightly increased repression in the
presence of 4-ddma-atc, we decided to determine the
contribution of each amino-acid exchange in TetR
i2
-

-
V99G activity profile. The mutants TetR-V99G S138I,
TetR-H64K V99G S135L, and TetR-H64K V99G
clearly have reverse activities, albeit with different effi-
ciencies, but do not show the effector specificity.
TetR-V99G S135L S138I is inactive. These results
demonstrate that both the H64K and S138I mutations
are necessary in combination with V99G to produce
revTetR variants with 4-ddma-atc specificity. It is also
remakable that the S138I mutation does not always
lead to 4-ddma-atc specificity as seen in TetR-V99G
S138I, however, this mutant is only slightly reverse
with both effectors.
Thermodynamic analysis of atc and 4-ddma-atc
binding to TetR
i2
-L17A and TetR
i2
-V99N
For overexpression of the proteins the respective genes
were introduced into pWH610 and the resulting plas-
mids were transformed in E. coli RB791 [11]. TetR
i2
-
L17A was purified to homogeneity employing the
protocol described for wild-type TetR [11]. The purifi-
cation protocol for TetR
i2
-V99N had to be modified
as described in experimental procedures and resulted

wild-type TetR sequence background and (B) for the TetR
i2
sequence background. (C) Steady-state levels of selected TetR vari-
ants with mutations of L17. The first lane (TetR) contains 50 ng of
purified wild-type TetR and the other lanes 50 lg of a soluble pro-
tein extract from E. coli WH207 ⁄ ktet50.
E M. Henßler et al. Altered TetR effector binding and allostery
FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS 4491
[atc-Mg]
+
or [4-ddma-atc-Mg]
+
complex to TetR can
be directly monitored. Atc or 4-ddma-atc fluorescence
were employed to observe complex formation. The fits
of the data for all TetR variants indicated positive
cooperativity. As described previously for tetracycline
and atc [12,13], we observed only weak cooperativity
for binding of 4-ddma-atc to the wild-type TetR and
of atc to TetR
i2
. In contrast, 4-ddma-atc binding to
the revTetR
i2
variants showed large cooperativity.
Scatchard analysis confirmed positive cooperativity for
TetR
i2
-L17A and -V99N (Fig. 5). The resulting equi-
librium binding constants are summarized in Table 2.

the affinities are lower than the respective ones to
TetR
i2
.
Fig. 5. Binding curves and Scatchard plots
of binding of TetR to [4-ddma-atc-Mg]
+
.
Fluorescence titrations were carried out at
0.1 l
M, 0.01 lM and 0.005 lM [Atc-Mg]
+
.
m is the average number of 4-ddma-atc mole-
cules bound to one TetR monomer. The
circles show the data, and the lines indicate
the fit according to the binding function. The
nonlinear curve progression shows the pres-
ence of positive cooperativity for the two
4-ddma-atc binding sites. (A) Fluorescence
titration, Langmuir fit and Scatchard plot of
the titration of 4-ddma-atc with TetR
i2
-L17A.
(B) Fluorescence titration, Langmuir fit and
Scatchard plot of the titration of 4-ddma-atc
with TetR
i2
-V99N.
Table 2. Atc and 4-ddma-atc binding constants of TetR variants. All constants have been determined by direct titration of 0.1 lM, 0.01 lM or

TetR
i2
-V99N
TetR
i2
-L17A TetR TetR
i2
TetR
i2
-V99N
TetR
i2
-L17A
TetR
M
+ [tc-Mg]
+
?TetR
M
[tc-Mg]
+
0.3
b
132
b
119600
b
1.7
b
< 0.02

6.3 –
d
22 31 –
d
3.7 < 0.02
c
< 0.02
c
a
The standard deviations typically range from 10% to 40%.
b
See [5].
c
The affinity was too low to be quantified.
d
The affinity was too high
for quantification by direct titrations.
Altered TetR effector binding and allostery E M. Henßler et al.
4492 FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS
TetO binding in the presence of both effectors was
qualitatively analyzed by EMSA for the revTetR
i2
vari-
ants. We used an at least 55-fold excess of 4-ddma-atc
or atc over TetR to ensure complete complex forma-
tion. The results are shown in Fig. 6. Both proteins
exhibit residual binding to tetO without effector and
with atc. The strongest tetO binding is observed for
TetR
i2

i2
mutant, however, the same
exchange leads to loss of DNA binding in the absence
of effector and 4-ddma-atc dependent corepression.
Moreover, analysis of the contributions of each
exchange in TetR
i2
-V99G to effector specificity
revealed a similar role for V99G and S135L. It was
shown for S135L previously that it confers relaxed
effector specificity to TetR [5,6]. S135 belongs to the
secondary shell of the effector binding pocket which
does not directly contact tetracycline in the crystal
structure [2] but is located next to tc contacting resi-
dues. V99 is not in contact with S135 (Fig. 1B). Thus,
the effect of V99G on effector binding must be trans-
ferred to the effector binding pocket. V99A also leads
to similar properties as it has no effect on the wild-type
but shows corepression with atc and 4-ddma-atc in
TetR
i2
. V99S has unaltered properties in the wild-type
TetR, but leads to loss of DNA binding and 4-ddma-
atc dependent corepression in TetR
i2
. V99T, on the
other hand, shows partial DNA binding, corepression
by atc and induction by 4-ddma-atc when introduced
in TetR
i2

i2
-V99N and TetR
i2
-L17A.
The EMSA was performed without effector and in the presence of
0.1 m
M atc or 4-ddma-atc. Hybridized oligonucleotides (0.3 lM) car-
rying tetO or a nonpalindromic sequence (usp. DNA) were incuba-
ted for 15 min with 0.3 l
M,0.9lM or 1.8 lM of the respective
TetR variant, electrophoresed on an 8% polyacylamide gel and
stained with ethidium bromide. The contents of the mixtures ana-
lyzed are indicated below the respective slots.
E M. Henßler et al. Altered TetR effector binding and allostery
FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS 4493
L205S may be due to changes in the positioning of the
effector contacting residue H100 [7]. This assumption
could apply to the properties of TetR
i2
-V99N, but the
exchanged residue in TetR
i2
-L17A is not in proximity to
H100 or to any other residue contacting the effector [2].
In conclusion, we propose that revTetR mutations
do not only lead to the previously proposed reposition-
ing of the DNA reading heads with respect to the core
domain [7,8] but also to altered effector binding via
structural changes in the effector binding pocket. It
was proposed on the basis of the TetR crystal struc-

taining randomized codons for helices a1 and a6 from
pWH1925 [8] were introduced in pWH1925-tetR
i2
(enco-
ding the mutations H64K S135L S138I) via XbaI ⁄ ApaI and
ApaI ⁄ FspI, respectively. Randomization of codons 50–63
in helix a4 was performed by PCR mutagenesis with the
primers a4deg_H64K (5¢-aataagcgggcccta
ctggatgcgctggcggt
ggagatcttggcgcgtcataaggattat-3¢; the underlined positions
contain 89% wild-type and 11% of the three non-wt bases
resulting in a predominant frequency of three to four muta-
tions) and 1925gh (5¢-gcaaaccgcctctcgccgc-3¢) using tetR
i2
as template. The resulting fragment was introduced in
pWH1925 via ApaI ⁄ NcoI for constitutive expression. All
other TetR variants were constructed using single restric-
tion enzyme sites in pWH1925.
E. coli screening system
E. coli WH207 ⁄ ktet50 [10,22] was transformed with the
mutant pools. It contains a chromosomal tetA-lacZ fusion
under tetR control. The cells were plated on MacConkey
Agar Base (Becton Dickinson, San Jose, CA, USA) con-
taining 14 gÆL
)1
lactose, 0.0042% (w ⁄ v) neutral red and
0.0014% (w ⁄ v) crystal violet. The colonies were screened
for their ability to repress b-galactosidase in the presence of
0.4 lm 4-ddma-atc and to express b-galactosidase on plates
containing 0.4 lm atc.

natant by cation exchange and size exclusion chromatogra-
phy as described previously [11].
The protein concentrations were determined by UV
spectroscopy and their activity was assessed by saturating
titration with 4-ddma-atc observing the change of fluores-
cence.
Fluorescence measurements
The fluorescence measurements were performed in a Spex
Fluorolog 3 with two double monochromators. To observe
Altered TetR effector binding and allostery E M. Henßler et al.
4494 FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS
4-ddma-atc fluorescence we excited at 420 nm and observed
emission at 540 nm. Excitation of atc fluorescence was per-
formed at 454 nm and emission was observed at 545 nm.
The equilibrium binding constants were obtained from
fluorescence titrations under equilibrium conditions. The
titrations were carried out in buffer containing 100 mm
Tris ⁄ HCl, pH 8.0, 100 mm NaCl and 20 mm MgCl
2
.
0.1 lm, 0.01 lm or 0.005 lm atc or 4-ddma-atc were titra-
ted with TetR concentrations from 2 · 10
)10
m to
1 · 10
)5
m. All measurements were repeated at least twice.
The binding constants were calculated by fitting of a hyper-
bolic binding function and including cooperative binding.
Electrophoretic mobility shift assay

References
1 Berens C & Hillen W (2003) Gene regulation by tetra-
cyclines. Constraints of resistance regulation in bacteria
shape TetR for application in eukaryotes. Eur J Bio-
chem 270, 3109–3121.
2 Kisker C, Hinrichs W, Tovar K, Hillen W & Saenger WB
(1995) The complex formed between Tet repressor and
tetracycline-Mg2+ reveals mechanism of antibiotic
resistance. J Mol Biol 247, 260–280.
3 Orth P, Alings C, Schnappinger D, Saenger W & Hin-
richs W (1998) Crystallization and preliminary X-ray
analysis of the Tet-repressor ⁄ operator complex. Acta
Crystallogr D54, 99–100.
4 Orth P, Schnappinger D, Sum PE, Ellestad GA,
Hillen W, Saenger W & Hinrichs W (1999) Crystal
structure of the Tet repressor in complex with a novel
tetracycline, 9-(N,N-dimethylglycylamido)-6-demethyl-
6-deoxy-tetracycline. J Mol Biol 285, 455–461.
5 Henssler EM, Scholz O, Lochner S, Gmeiner P & Hillen
W (2004) Structure-based design of tet repressor to opti-
mize a new inducer specificity. Biochemistry 43, 9512–
9518.
6 Scholz O, Ko
¨
stner M, Reich M, Gastiger S & Hillen
WB (2003) Teaching TetR to recognize a new inducer.
J Mol Biol 329, 217–227.
7 Kamionka A, Bogdanska-Urbaniak J, Scholz O &
Hillen W (2004) Two mutations in the tetracycline
repressor change the inducer anhydrotetracycline to a

tetracycline complex by fluorescence measurements.
Evidence for divalent metal ion requirement and energy
transfer. J Mol Biol 187, 341–348.
13 Scholz O, Kintrup M, Reich M & Hillen WB (2001)
Mechanism of Tet repressor induction by tetracyclines:
length compensates for sequence in the alpha8-alpha9
loop. J Mol Biol 310, 979–986.
14 Helbl V & Hillen W (1998) Stepwise selection of TetR
variants recognizing tet operator 4C with high affinity
and specificity. J Mol Biol 276, 313–318.
15 Helbl V, Tiebel B & Hillen W (1998) Stepwise selec-
tion of TetR variants recognizing tet operator 6C
E M. Henßler et al. Altered TetR effector binding and allostery
FEBS Journal 272 (2005) 4487–4496 ª 2005 FEBS 4495
with high affinity and specificity. J Mol Biol 276,
319–324.
16 O’Gorman RB, Rosenberg JM, Kallai OB, Dickerson
RE, Itakura K, Riggs AD & Matthews KS (1980)
Equilibrium binding of inducer to lac repressor.
operator DNA complex. J Biol Chem 255, 10107–
10114.
17 Lederer T, Takahashi M & Hillen W (1995) Thermo-
dynamic analysis of tetracycline-mediated induction of
Tet repressor by a quantitative methylation protection
assay. Anal Biochem 232, 190–196.
18 Friedman BE, Olson JS & Matthews KS (1977) Inter-
action of lac repressor with inducer, kinetic and equili-
brium measurements. J Mol Biol 111, 27–39.
19 Hinrichs W, Kisker C, Duvel M, Mu
¨