Tetrepressorresiduesindirectlyrecognizing anhydrotetracycline
Peter Schubert*, Klaus Pfleiderer and Wolfgang Hillen
Lehrstuhl fu
¨
r Mikrobiologie, Institut fu
¨
r Mikrobiologie, Biochemie und Genetik, Friedrich-Alexander-Universita
¨
t Erlangen, Germany
Two tetracycline repressor (TetR) sequence variants sharing
63% identical amino acids were investigated in terms of their
recognition specificity for tetracycline and anhydrotetra-
cycline. Thermodynamic complex stabilities determined by
urea-dependent unfolding reveal that tetracycline stabilizes
both variants to a similar extent but that anhydrotetracycline
discriminates between them significantly. Isofunctional
TetR hybrid proteins of these sequence variants were
constructed and their denaturation profiles identified resi-
dues 57 and 61 as the complex stability determinant.
Association kinetics reveal different recognition of these
TetR variants by anhydrotetracycline, but the binding
constants indicate similar stabilization. The identified
residues connect to an internal water network, which
suggests that the discrepancy in the observed thermo-
dynamics may be caused by an entropy effect. Exchange of
these interacting residues between the two TetR variants
appears to influence the flexibility of this water organiza-
tion, demonstrating the importance of buried, structural
water molecules for ligand recognition and protein func-
tion. Therefore, this structural module seems to be a key
requisite for the plasticity of the multiple ligand binding
protein TetR.
Keywords: Tet repressor; gene regulation; protein stability;
ligand binding; antibiotic resistance.
The biological function of many proteins is triggered and
modulated by binding of effector molecules or a variety of
extrinsic cofactors that greatly expand the repertoire of
cellular processes executed by polypeptides, DNA or small
proteins [1]. Therefore, molecular recognition is a funda-
mental process in all living systems, regulating processes as
diverse as transcription, cell signalling and immunity [2–4].
Recognition mechanisms may be divided into two general
categories, named specific and related specificities [5]. For
example, mature immunoglobulins (Ig) are highly specific
while those in the germline bind a broad range of antigens
[6]. This effect is explained by a diversity of conformations
for the related specificity germline antibody, a pre-existing
subpopulation of antibody isomers based on increased
flexibility [7,8].
To understand the thermodynamic and kinetic principles
of protein ligand binding in more detail, concepts of energy
landscapes and folding funnels were used [9]. Characteriza-
tion of binding sites revealed that such regions are usually
depressions in the protein surface where a greater average
degree of exposure of hydrophobicity groups occurs [10].
However, for a detailed knowledge of protein specificity at
the molecular level it is essential to understand the
mechanisms of protein–ligand recognition by obtaining
information about the structure, energetics and dynamics of
the free and complexed species under a variety of condi-
tions. We used two sequence variants of the tetracycline
repressor (TetR) to investigate ligand recognition of two
different tetracycline (tc) derivatives, tc and anhydrotetra-
cycline (atc). TetR regulates resistance to the antibiotic tc in
Gram-negative bacteria by inducer binding [11]. This system
is successfully adapted for regulation of gene expression in
different organisms [12]. Based on sequence similarities of
isolates from various bacteria TetR variants were grouped
into nine classes called A to E, G, H, J and 30 [13]. The
proteins share between 38 and 88% sequence identity
and are presumably isostructural and isofunctional. Each
homodimeric protein consistes of an N-terminal DNA-
binding domain connected to a core domain harbouring the
dimerization motif and the effector binding site. Repression
of gene expression occurs by specific binding to tetO via a
helix–turn–helix motif. Binding of the effector molecule
leads to a conformational change resulting in the loss of
DNA binding and initiation of transcription [11,14]. The
crystal structure of TetR(D)[tc–Mg]
2
indicates the position-
ing of the inducer inside the protein core and reveals
interactions of the drug with both monomers [15,16].
Here we studied the thermodynamic complex stability
and binding affinity of the two naturally occuring TetR
variants B and D with tc and atc. The sequence identity
between these two TetR variants is only 63%, but it includes
most residues involved in operator and DNA binding. The
sequence identity for the helix–turn–helix domain is 94%
and for the residues contacting tc it is 68%. Urea-dependent
unfolding yields similar stabilization of the two sequence
variants by tc binding. However, atc shows a strong
discrimination between TetR(B) and TetR(D) in terms of
stabilization. Using TetR(B/D) hybrid proteins we have
Correspondence to W. Hillen, Lehrstuhl fu
¨
r Mikrobiologie, Institut fu
¨
r
Mikrobiologie, Biochemie und Genetik, Friedrich-Alexander-
Universita
¨
t Erlangen-Nu
¨
rnberg, Staudtstraße 5, 91058 Erlangen,
Germany. Fax: + 49 91318528082, Tel.: + 49 91318528081,
E-mail: whillen@biologie.uni-erlangen.de
Abbreviations: TetR, tetracycline repressor; tc, tetracycline;
atc, anhydrotetracycline.
*Present address: Biomedical Research Centre, University of British
Columbia, Vancouver, V6T 1Z3, Canada.
(Received 5 February 2004, revised 21 March 2004,
accepted 30 March 2004)
Eur. J. Biochem. 271, 2144–2152 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04130.x
narrowed the determinants for the different atc complex
stabilities to two residues localized at one side of the
tc-binding pocket, where they are involved in the last step of
the proposed induction mechanism of TetR by stabilizing
the induced complex by a water zipper [14]. Thermo-
dynamic and kinetic investigations reveal that replacement
of these residues in TetR(D) to the ones found in TetR(B)
reduces complex stability and recognition of atc. This
indicates that this water network is important for stability
and drug affinity, but not for induction.
Experimental procedures
Material and general methods
Anhydrotetracycline (atc) was purchased from Acros (Geel,
Belgium), all other chemicals were from Merck (Darmstadt,
Germany), Roth (Karlsruhe, Germany) or Sigma (Mu
¨
n-
chen, Germany) at the highest available purity. Enzymes for
DNA restriction and modification were from New England
Biolabs (Schwalbach, Germany), Boehringer (Mannheim,
Germany), Stratagene (Heidelberg, Germany) or Pharma-
cia (Freiburg, Germany). Oligonucleotides were from PE
Applied Biosystems (Weiterstadt, Germany). Sequencing
was carried out according to the protocol provided by
Perkin Elmer for cycle sequencing and sequence waas
analysed with an ABI PRISM
TM
310 Genetic Analyzer (PE
Applied Biosystems, Weiterstadt, Germany).
Bacterial strains and plasmids
All bacterial strains were derived from Escherichia coli K12.
Strain DH5a [hsdR17(r
K
m
K
+
), recA1, endA1, gyrA96, thi,
relA1, supE44, /80 dLacZ(M15, [(lacZYA-argF)U169] was
used for general transformation procedures. Strain WH207
(lacX74, gaK2, rpsL, recA13) [17] served as host strain
for b-galactosidase assays. The plasmids pWH806 and
pWH853(B) [17] and pWH853(D) [18] used in the in vivo
assay as well as pWH1950 [19] for overexpression have been
described before.
Construction of the chimeric tetR genes
All tetR variants were constructed by PCR according to the
three-primer method [20]. The products of the second PCR
reaction were purified and digested with XbaI/MluIor
MluI/NcoI and cloned in pWH853 to replace the respective
position of tetR. For overexpression these constructs were
digested with XbaIandNcoI and cloned into likewise
digested pWH1950. DNA of positive candidates was
analysed by sequencing of tetR.
b-Galactosidase assay
Repression and induction efficiencies of the TetR variants
were assayed in E.coliWH207ktet50 carrying the respective
pWH853 derivatives. The phage ktet50 contains a tetA–
lacZ transcriptional fusion [17] integrated as single copy into
the WH207 genome. Bacteria were grown at 28 °Cin
Luria–Bertani medium supplemented with the appropriate
antibiotics. Quantification of induction efficiencies was
carried out with 0.2 lgÆmL
)1
atc in overnight and log phase
cultures. b-Galactosidase activities were determined as
described by Miller [21]. Three independent cultures were
assayed for each strain and measurements were repeated at
least twice.
Purification of the TetR variants
pWH1950 derivatives of the different constructs were
transformed into E.coliRB791. Cells were grown in 3 L
of Lurian–Bertani medium at 28 °C in shaking flasks. TetR
expression was induced by adding isopropyl thio-b-
D
-
galactoside to a final concentration of 1 m
M
at D
600
¼
0.7–1.0. Cells were pelleted and resuspended in buffer A
containing 50 m
M
NaCl, 2 m
M
dithiothreitol, 20 m
M
sodium phosphate buffer pH 6.8 and broken by sonication.
TetR variants were purfied by cation exchange chromato-
graphy and gel filtration as described [19]. The amounts of
the proteins were obtained from the UV absorption at
280 nm [22] and their activities were determined by titration
with atc [23].
Fluorescence and CD spectroscopy
Fluorescence intensities were measured with a Spex Fluo-
rolog 1680 double spectrometer in 1 cm cells at protein
concentrations of 1 l
M
or 5 l
M
. Excitation was at 280 nm
and emission was recorded at the maximum of the
difference between the native and the denatured fluores-
cence spectrum. The bandwidth for excitation and emission
was 2.2 mm. CD measurements were carried out on a Jasco
J-715 spectropolarimeter in 0.5 cm cells at protein concen-
trations of 5 l
M
TetR monomer. The TetR[atc–Mg]
2
complex was formed by adding 10 l
M
atc.
Unfolding of the TetR complexes, thermodynamic
and kinetic constants
We used F-buffer containing 100 m
M
NaCl, 100 m
M
Tris/
HClpH7.5,5m
M
MgCl
2
,1m
M
EDTA, 1 m
M
dithiothre-
itol for all spectroscopic measurements. Urea was obtained
from ICN Biochemicals (Eschwege, Germany) and urea
solutions were prepared each day. Equilibrium denaturation
was performed by incubating protein samples overnight at
the indicated urea concentration. Renaturation reactions
were achieved by incubating the samples overnight at 8
M
urea and then diluting them 200-fold with F-buffer. All
reactions were performed at 22 °C and all TetR concentra-
tions relate to the monomer. Tetracycline or its derivative
(xtc) was used in a onefold molar excess over protein.
Thermodynamic calculations of the urea-induced denatur-
ation of the TetR[tc/atc–Mg] complex variants were
performed as described before [24,25] by extending the
calculation as it applies to the monomeric (¼ TetR) as
follows:
2TetR
N
þ 2xtc þ 2Mg
2þ
¼½TetRÀxtcÀMg
2
¼ 2TetR
U
þ 2xtc þ 2Mg
2þ
where N is native and U is unfolded.
The left side of the equation shows the association/
dissociation equilibrium and the right side the folding/
Ó FEBS 2004 TetR anhydrotetracycline recognition (Eur. J. Biochem. 271) 2145
unfolding equilibrium. For the unfolding process the
equilibium constant (K
U
)couldbegivenas:
K
U
¼½U
2
½Mg
2þ
2
½xtc
2
=½TetRÀxtcÀMg
2
The equilibrium constant for ligand-free systems is given as:
K
U
¼ 2P
t
f
2
U
=f
N
½24
where P
t
is the total protein concentration.
Mass balance yields:
½Mg
2þ
t
¼½Mg
2þ
þ½C
N
½xtc
t
¼½xtcþ½C
N
where [C
N
] is the concentration of the native complex,
resulting in:
K
U
¼ 2P
t
f
U
ð½Mg
2þ
t
À P
t
f
N
Þ
2
ð½xtc
t
À P
t
f
N
Þ
2
=f
N
where f
N
and f
U
are the fraction native or unfolded,
respectively. This equation was used to calculate DGinthe
different states of the unfolding pathway.
Mg
2+
-independent and -dependent atc equilibrium
association constants were determined as published [26].
The association rate constants were determined at 28 °C
[27] with equimolar concentrations of TetR monomer and
atc in F-buffer as mentioned above. All experiments were
repeated at least twice.
Fig. 1. TetR crystal structure and amino acid composition. (A) Structure of TetR[tc–Mg]
2
. Monomers are shown in blue and red, helices are
indicated by numbers in the blue monomer and Tc is shown in green. (B) Alignment of TetR(B) and TetR(D). Conserved residues are shown in
reverse type. Tc binding residues (black filled point) and residues involved in coordinating the water zipper (›) are indicated [14,30].
2146 P. Schubert et al. (Eur. J. Biochem. 271) Ó FEBS 2004
Results
Unfolding of TetR complexed with tc or atc
We used urea-dependent denaturation as described before
[24,25] to determine the stabilities of TetR[atc–Mg]
2
com-
plexes in comparison with TetR[tc–Mg]
2
. As shown by the
crystal structure of the latter complex binding of tc occurs
inside the TetR dimer (Fig. 1A) [15]. The two TetR B and D
sequence variants share 63% amino acid identity (Fig. 1B)
and fold into the same quaternary structure in the free and
tc-complexed forms [15]. Structural changes during unfold-
ing of the complexes were observed by fluorescence and CD.
As the fluorescence and CD properties of the two proteins in
their complexed forms are analogous, we show only the
data for the TetR(D)[atc–Mg]
2
complex. The main fluor-
escence of the TetR[atc–Mg]
2
complex originates from atc
and four naturally occuring trp residues. The emission
spectrum excited at 280 nm shows two maxima at 362 nm
and 515 nm (Fig. 2A), resulting from trp emission [22]
and energy transfer from trp to atc [28], respectively. In the
denatured state at 8
M
urea the trp emission maximum shifts
to at 354 nm and is slightly increased in intensity, but the
515-nm band is absent, indicating the loss of ligand binding
(Fig. 2A). Therefore, the intensity change of the 515-nm
band was used to follow denaturation. The TetR[atc–Mg]
2
complexexcitedat455nmshowsalsoanemissionmaxi-
mum at 515 nm which is also apparently absent at 8
M
urea
(Fig. 2B). This fluorescence change of atc was also used to
monitor complex denaturation to compare with energy
transfer. For the respective tc-complexes unfolding was
followed by the change of the energy transfer band at
508 nm and the tc fluorescence at 508 nm excited at 370 nm
(data not shown).
The TetR[atc–Mg]
2
complex shows CD minima at
208 nm and 222 nm (data not shown) reflecting the high
content of a-helical structure [15]. The CD spectra in the
presence or absence of the inducer are the same in that curve
segment. Since both absorption bands are absent at 8
M
urea the ellipticity at 222 nm was also used to quantify
TetR[atc–Mg]
2
unfolding. Each of the three probes yields
identical results when used to observe denaturation, show-
ing monophasic, sigmoidal curves indicating the absence of
stable folding intermediates (Fig. 2C). This demonstrates
that denaturation of the protein fold as observed by CD,
and release of the ligand as observed by fluorescence, occur
simultaneously. As expected for a bimolecular reaction, the
midpoint of the unfolding transition depends on the protein
concentration (Table 1). The transition midpoints shift
from 6.2
M
to about 6.5
M
, when the protein concentration
is increased fivefold. To analyse the efficiency of the
refolding reaction the TetR(D)[atc–Mg]
2
complex was
treated with 8
M
urea and subsequently diluted 200-fold
with urea-free buffer. The fluorescence emission spectrum
and CD of these renatured complexes were identical with
those obtained from the native form (Fig. 2A,B). We
conclude that denaturation under these conditions is a
completely reversible, single-step reaction for the TetR[atc–
Mg]
2
and the TetR[tc–Mg]
2
complexes. This allows quan-
tification using the two-state model [29] in which only folded
complexes, unfolded monomers and free ligand exist at
equilibrium in significant concentrations.
Thermodynamic stabilities of TetR[tc–Mg]
2
and
TetR[atc–Mg]
2
complexes
Extrapolation of the urea-induced unfolding curves to 0
M
urea (Fig. 2C, inset) to calculate the Gibbs free energy of
unfolding DG°
U
(H
2
O) gave the results shown in Table 1.
The DG°
U
(H
2
O) of 123 kJÆmol
)1
for TetR(D)[atc–Mg]
2
Fig. 2. Fluorescence probes to determine complex stability. (A) Fluor-
escence emission spectra of TetR(D)[atc–Mg]
2
; native (top curve), re-
natured (dotted line below) and denatured (dashed line). (B) Emission
spectra of atc; native (top curve), renatured (dotted line, below) and
denatured (dashed line). (C) Urea-induced unfolding curve of TetR(-
D)[atc–Mg]
2
. Unfolding was followed by different probes (s,energy
transfer; d, atc fluorescence; m, CD) and extrapolated for determin-
ation of complex stability (inset).
Ó FEBS 2004 TetR anhydrotetracycline recognition (Eur. J. Biochem. 271) 2147
represents a stabilization of 69 kJÆmol
)1
compared to the
free protein [25]. In contrast, the stabilization of TetR(B) by
complex formation with [atc–Mg]
+
is only 26 kJÆmol
)1
.
Although free TetR(D) is less stable than TetR(B) [25] it is
stabilized more by binding of [atc–Mg]
+
. The complexes of
the two sequence variants with [tc–Mg]
+
are stabilized by
13 kJÆmol
)1
and 12 kJÆmol
)1
, respectively. This indicates
that the TetR(D) sequence variant must undergo more
favourable interactions with atc than does TetR(B), despite
of the fact that 63% of the amino acids (Fig. 1B) and
the folds of the polypeptide chains are identical. These
interactions must also be specific for atc as no difference is
observed for the tc complexes.
Interaction of TetR(B/D) chimera with [atc–Mg]
+
We used some TetR(B/D) hybrid proteins from previous
work [18,25] and constructed five new chimeras based on
the assumption that residues contacting tc in the
TetR(D)[tc–Mg]
2
structure (Fig. 3) should be involved in
binding of atc as well. Residues contacting [tc–Mg]
+
are
marked with open circles in Fig. 1B. The chimeric proteins
analysed in this work for binding and affinity to [tc–Mg]
+
and [atc–Mg]
+
areshowninFig.4.
To determine both their in vivo repression efficiency and
inducibility all chimeric genes were cloned into pWH853
[17] and transformed into E.coliWH207ktet50 containing a
chromosomal tetA–lacZ-fusion. The b-galactosidase activ-
ities at 28 °C were determined in the presence and absence
of 0.2 lgÆmL
)1
atc. As shown in Table 2 none of the
mutants shows a significantly reduced inducibility com-
pared to TetR(D). The repression efficiencies are nearly
identical to TetR(B), only TetR(D) is a less efficient
repressor and is also less inducible as observed before [18].
The TetR(B/D) chimeras were overexpressed and purified
to homogenity as described previously [19]. All purified
TetR variants show similar spectral properties in the free
and complexed forms (data not shown). The unfolding of
the free and complexed proteins was carried out by urea-
induced denaturation as mentioned above showing identical
denaturation pathways (data not shown) indicated by
monophasic denaturation curves. The thermodynamic
stabilities given as transition midpoints are summarized in
Table 2. Tc binding results in a similar stabilization for all
sequence variants. For atc only TetR(B/D)51–208 exhibits
TetR(D)[atc–Mg]
2
-like stability, all other chimera show
TetR(B)[atc–Mg]
2
-like stability. These data indicate that
residues between positions 51 and 63 are responsible for the
differences in [atc–Mg]
+
binding.
Identifying single residues for atc recognition
Three residues of the segment 51–62 are different between
TetR(B) and TetR(D) (Fig. 1B). The mutants TetR(B)57/
Table 1. Thermodynamic stability of the TetR(B)[tc/atc–Mg]
2
and TetR(D)[tc/atc–Mg]
2
complexes.
Fluorescence of energy transfer
a
CD
b
1 l
M
5 l
M
5 l
M
DG°
U
[kJÆmol
)1
]
Urea
1/2
[
M
]
DG°
U
[kJÆmol
)1
]
Urea
1/2
[
M
]
DG°
U
[kJÆmol
)1
]
Urea
1/2
[
M
]
TetR(B)[tc–Mg]
2
83 ± 7 4.7 83 ± 4 5.0 84 ± 4 5.1
TetR(B)[atc–Mg]
2
96 ± 4 5.0 94 ± 3 5.4 93 ± 6 5.5
TetR(D)[tc–Mg]
2
66 ± 3 4.1 67 ± 4 4.3 69 ± 3 4.6
TetR(D)[atc–Mg]
2
123 ± 9 6.2 122 ± 6 6.5 125 ± 4 6.7
a
Urea-dependent unfolding was followed by the change of the energy transfer signal at the wavelength with the maximal difference of
fluorescence between the native and the denatured forms, at protein concentrations of 1 l
M
(excitation at 280 nm) and 5 l
M
(excitation at
295 nm); tc and atc in onefold excess at 28 °C.
b
The change of CD was observed at 222 nm.
Fig. 3. TetR–tc interactions. Binding to the
different monomers is shown by blue and
green symbols, respectively, and involved
water molecules are depicted as red spheres.
For comparison the chemical structure of the
atc molecule is shown to the right. The two
molecules have similar chemical structures,
differing only in that the hydroxyl group at
position 6 in tc and the neighbouring hydro-
gen bond are eliminated in atc, resulting in an
aromatic ring C.
2148 P. Schubert et al. (Eur. J. Biochem. 271) Ó FEBS 2004
59/61D and TetR(D)57/59/61B contain replacements of
these three amino acids by the respective other residues. The
denaturation results shown in Table 2 demonstrate that
these three residues determine the stability difference of the
TetR–inducer complexes. When they originate from the
TetR(B) sequence the atc-mediated stabilization is small,
whereas the TetR(D) residues lead to higher stabilization,
hence better recognition. We then constructed all possible
double and single TetR variants for these positions in the
TetR(D) sequence background. They show the same in vivo
repression and inducibility as TetR(D) (Table 2). The
denaturation data reveal that complex stability is strongly
affected when the residue at position 57 is exchanged, but
the alterations at positions 57 and 61 are necessary to yield
the fully reduced TetR(B)[atc–Mg]
2
-like stability. Thus, they
are the determinant for the improved recognition of atc by
TetR(D).
Association rate constants for inducer binding
to the TetR(D/B) variants
From the thermodynamic point of view the complex
stability should reflect the equilibrium binding constant
(K
A
) of atc to TetR. Therefore we determined K
A
for
TetR(B), TetR(D) and TetR(D/B) single, double and triple
mutants using an improved assay [26] which takes into
account the effect that atc binds Mg
2+
independently of
TetR to a small extent. The Mg
2+
independent atc
equilibrium constants (K
T
) of the TetR(D/B) constructs
were determined by titration with inducer in the presence of
EDTA following complex formation by atc fluorescence
emission at 545 nm excited at 455 nm. The K
T
values were
calculated from these titration curves as described [26] and
are listed in Table 3. The Mg
2+
-dependent binding
constants were obtained from titrations of TetR complexed
with atc in the presence of Mg
2+
[26]. The resulting
association constants K
A
arealsolistedinTable3.The
value for TetR(D) is identical to that published before [26].
The single exchanges of the residues at positions 57 and 61
show a partial alteration of affinity from the TetR(D) to the
TetR(B) sequence variant, whereas the double mutation
shows the full reduction. Single and combined alterations
involving position 59 show either no or smaller effects.
Association rate constant of TetR[atc–Mg]
2
complex
formation
The different equilibrium binding constants of atc for both
TetR sequence variants could be caused by different
association or dissociation rate constants, or both. The
time-dependent association rate constants of atc for both
sequence variants were determined by measuring the
increase in the atc fluorescence upon addition of TetR as
described [27]. They were fitted using second-order kinetics
for a bimolecular reaction and the results are also presented
in Table 3 showing a k
ass
value of 1 · 10
)6
M
)1
Æs
)1
for
TetR(B) and a sevenfold higher rate constant for TetR(D).
Since this accounts for the total difference seen in equilib-
rium constants the different stability of TetR[atc–Mg]
2
is
based only on different association rates and therefore on
molecular recognition.
Discussion
In this study we used urea-induced unfolding to determine
the thermodynamic stabilities of two TetR sequence vari-
ants B and D complexed with the inducer atc or tc. The
change of the spectral probes used show identical results
reflecting the coordinated destruction of the tertiary struc-
ture, the loss of the ligand binding and the break down of
the secondary structure. The monophasic, sigmoidal curves
for urea-dependent unfolding allow the use of a two-state
model for calculating thermodynamic complex stabilities.
The values for the Gibbs free energy DG°
U
(H
2
O) for the two
TetR sequence variants B and D complexed with tc were
determined to 83 kJÆmol
)1
and 66 kJÆmol
)1
, which com-
pared to their free forms [25] reveals that tc stabilizes both
sequence variants to similar extents of 13 kJÆmol
)1
and
12 kJÆmol
)1
, respectively. However, atc binding results in
complex stabilities of 96 kJÆmol
)1
for the B and 123 kJÆ
mol
)1
for the D variant, which leads to stabilizations of
26 kJÆmol
)1
and 69 kJÆmol
)1
, respectively.
This surprising stabilization difference of atc with the
TetR variants B and D is reflected in the respective
association constants. The increased affinity for TetR(D)
compared to TetR(B) is accounted for by different associ-
ation rate constants, thus indicating the identical overall
structures of TetR(D)[tc–Mg]
2
and TetR(B)[tc–Mg]
2
[15]
that is used to identify the determinant for this difference in
Fig. 4. Overview of the chimeric TetR(B/D) constructs. The respective
wild-type proteins are shown by filled (TetR(D)) and open (TetR(B))
bars. The hybrid proteins are shown with their designation given on
the right. The top panel indicates the location of the TetR residues
interacting with tc (Fig. 1B).
Ó FEBS 2004 TetR anhydrotetracycline recognition (Eur. J. Biochem. 271) 2149
ligand stabilization. The sequence alignment of B and D
shows that the 16 residues that interact with the inducer
(Fig. 3) are highly conserved (Fig. 1B). The fact that both
sequence variants form identical primary contacts to the
inducer supports the hypothesis that the different affinities
must be due to an indirect effect. Functional TetR(B/D)
hybrid proteins enabled us to narrow the determinant for
the different stabilities to the residues 57 and 61, located at
the C terminus of the hinge helix a4 connecting the DNA
binding domain with the protein core (Fig. 5). Although
located close to residue His64 making contact to tc in the
crystal structure [15], these residues are too far away to
exert a direct influence. The replacement of the solvent-
exposed Val by the chemically similar Ile residue at position
57 should cause just a small effect, but this exchange
contributes the most to destabilization. This may be
explained by a special feature observed in the crystal
structure of the TetR(D)–tetO complex [14].
The comparison of the crystal structures in the tc-induced
with the DNA-bound forms leads to a proposed induction
mechanism [14]. After [tc–Mg]
+
insertion into the binding
tunnel, ring A of the tc molecule is anchored by hydrogen
bonds to different residues including His64, Asn82, Phe86
and Gln116 (Fig. 3). His64 is involved in the conforma-
tional change of TetR associated with induction [14] and its
interaction with tc fixes the C terminus of helix a4. This state
is stabilized by a network of cooperative hydrogen bonds
including a chain of eight water molecules (Fig. 5) that is
not found in the free form of TetR. Val57 participates in
this so-called Ôwater zipperÕ [30], representing the only
Table 2. In viv o data and thermodynamic stabilities of different TetR variants.
Repression Induction
a
Free +tc
b
Difference +atc
2
Difference
b-gal [%] b-gal [%] Urea
1/2
[
M
] Urea
1/2
[
M
] DUrea
1/2
[
M
] Urea
1/2
[
M
] DUrea
1/2
[
M
]
TetR(D) 8.6 ± 1.0 100 ± 1.1 3.7 4.3 0.5 6.3 2.5
TetR(B)51–208D 0.9 ± 0.1 94 ± 1.5 3.7 4.2 0.4 6.1 2.3
TetR(B)63–208D 0.8 ± 0.0 83 ± 1.1 3.5 4.0 0.5 4.7 1.2
TetR(B)75–208D 0.8 ± 0.0 97 ± 1.4 3.4 3.8 0.4 5.0 1.7
TetR(B)84–208D 0.7 ± 0.1 90 ± 1.2 3.3 3.7 0.4 4.8 1.5
TetR(B)92–208D 0.8 ± 0.0 99 ± 2.3 3.5 3.9 0.4 5.1 1.6
TetR(B)110–208D 0.8 ± 0.0 94 ± 4.1 3.7 4.1 0.4 5.2 1.5
TetR(B)123–208D 0.7 ± 0.0 83 ± 2.2 3.9 4.2 0.3 5.0 1.1
TetR(B)168–208D 1.3 ± 0.1 100 ± 2.3 3.9 4.4 0.5 5.1 1.2
TetR(B)179–208D 1.1 ± 0.1 105 ± 2.0 4.0 4.5 0.5 5.2 1.2
TetR(B) 1.2 ± 0.0 96 ± 2.2 4.2 4.7 0.3 5.2 0.9
TetR(B)57,59,61D 1.8 ± 0.0 90 ± 2.7 4.0 n.d. – 6.2 2.2
TetR(D)57,59,61B 7.8 ± 0.5 92 ± 6.1 3.9 n.d. – 5.0 1.1
TetR(D)57,59B 12 ± 0.7 76 ± 1.8 4.1 n.d. – 5.6 1.5
TetR(D)57,61B 8.2 ± 0.4 92 ± 5.7 3.4 n.d. – 5.1 1.7
TetR(D)59,61B 9.7 ± 0.2 89 ± 1.5 3.5 n.d. – 5.8 2.3
TetR(D)57B 7.6 ± 0.2 79 ± 2.0 3.7 n.d. – 5.4 1.7
TetR(D)59B 9.0 ± 1.2 88 ± 6.1 3.7 n.d. – 6.0 2.3
TetR(D)61B 8.4 ± 0.8 83 ± 1.1 3.4 n.d. – 5.7 2.3
a
Induction was determined at 0.2 lgÆmL
)1
atc. The 100% expression of b-galactosidase corresponds to 10920 ± 1451 units.
b
Urea
1/2
-
values of chimeric TetR variants were calculated in the absence of the inducer atc by the change of the fluorescence at 330 nm or in the
presence of atc from the change of the energy transfer signal at 515 nm for tc and 545 nm for atc excited at 280 nm detected at 28 °C.
Table 3. Mg
2+
-independent (K
T
) and -dependent (K
A
) binding constants, association and calculated dissociation rate constants of atc to TetR variants.
K
T
[· 10
7
M
)1
]
K
A
[· 10
11
M
)1
]
k
ass
[· 10
6
M
)1
Æs
)1
]
k
diss
a
[· 10
)6
s
)1
]
TetR(B) 1.95 ± 0.11 2.30 ± 0.21 1.0 ± 0.2 5.0 ± 0.5
TetR(B)57,59,61D 7.10 ± 0.12 7.62 ± 0.24
TetR(D)57,59,61B 2.20 ± 0.17 2. 94 ± 0.17
TetR(D) 10.0 ± 0.31 9.11 ± 0.30 7.0 ± 0.9 7.0 ± 0.4
TetR(D)57,59B 4.90 ± 0.15 4.70 ± 0.16
TetR(D)57,61B 1.95 ± 0.14 2.12 ± 0.18
TetR(D)59,61B 1.25 ± 0.09 6.78 ± 0.22
TetR(D)57B 3.00 ± 0.16 3.34 ± 0.22
TetR(D)59B 7.85 ± 0.20 7.72 ± 0.24
TetR(D)61B 5.45 ± 0.17 4.74 ± 0.19
a
Calculated as k
diss
¼ k
ass
/K
A
.
2150 P. Schubert et al. (Eur. J. Biochem. 271) Ó FEBS 2004
nonconserved residue of the contacting amino acids (in
Fig. 1B marked by ›). The assumption that the exchange to
Ile could lead to a distortion of the arrangement in this
water network connecting the helixes a4 and the loop
between helices a6anda7 is in agreement with the strong
decreased complex stability of this single mutation alone.
Although Ala61 is not directly involved in interacting
with the Ôwater zipperÕ the exchange to Asp could sterically
influence residue 59 coordinating water W6 due to the larger
size and the introduction of a charged residue.
From the thermodynamic point of view, this different
stabilization of TetR(B) and TetR(D) should be reflected in
their equilibrium association constants K
A
. However, cal-
culating DG from the determined binding constants K
A
by
DG ¼ ) RT*lnK
A
reveals 65.5 kJÆmol
)1
for TetR(B) and
68.9 kJÆmol
)1
for TetR(D). Taking into account the stability
of the free proteins [25] leads to complex stabilities of
140 kJÆmol
)1
for TetR(B)[atc–Mg]
2
and 123 kJÆmol
)1
TetR(D)[atc–Mg]
2
, respectively. Only the value for
TetR(D)[Mg-atc]
2
resembles the result from the denatura-
tion experiment. This discrepancy for TetR(B) could be
explained by an enthalpy–entropy compensation effect
taking into account that amino acid replacements alter
both enthalpy and entropy contributions to ligand binding.
As shown for the rat intestinal fatty acid-binding protein the
changes in molecular interactions may not necessarily
correlate with changes in affinity [31,32]. The two identified
residues responsible for the different complex stability of
TetR(B) and TetR(D) with atc belong to an internal water
network which could be partially destroyed by the replace-
ment to the respective residues of TetR(B). This fact might
lead to the consequence that the local conformational
flexibility of the ligand recognition site is increased due to
the observation that binding of buried structural water
molecules increase flexibility [33]. This increased flexibility
leads to an increase of entropy, which is probably not
thecaseinTetR(B)duetoaÔdisorderedÕ water network
organization. This compensation explains the reduced
DG°
U
(H
2
O) value for TetR(B)[atc–Mg]
2
deduced from
urea-induced denaturation.
The water zipper could be part of a functional epitope.
Taking into account that interactions between biological
molecules cannot be reduced to the description of static
molecular structures the function of a protein depends also
on the distribution and the populations of its conforma-
tional states [34]. Such a mechanism provides multiple
pathways and allows a single molecular surface to
interact with numerous structurally distinct binding part-
ners, accommodate mutations through shifts in the
dynamic energy landscape and as such is evolutionarily
advantageous [9].
Although we are just beginning to understand the
properties that makes these consensus binding sites
unique, the role of conformational changes induced upon
binding at the protein interface has emerged as a factor of
key importance. Because this separation/discrepancy is not
seen for tc it is most likely that the water zipper plays an
important role for stabilization and ligand affinity but not
for induction. This arrangement therefore mediates an
indirect recognition mode of TetR. This points out that
internal bound water molecules increase protein flexibility
which is responsible for specificity of ligand binding.
These findings contribute to the basic knowledge for drug
design necessary to improve specificity of the TetR
system.
Acknowledgements
We thank Prof F. X. Schmid (University Bayreuth) for helpful and
stimulating discussions and Dr Oliver Scholz for help in calculating the
binding constants. This work was supported by the Deutsche
Forschungsgemeinschaft through SFB 473 and the Fonds der
Chemischen Industrie.
References
1. Zhang, J. & Matthews, C.R. (1998) The role of ligand binding in
the kinetic folding mechanism of human p21 (H-ras) protein.
Biochemistry 37, 14891–14899.
2. Garvie, C.W. & Wolberger, C. (2001) Recognition of specific
DNA sequences. Mol. Cell 8, 937–946.
3. Christophe, D., Christophe-Hobertus, C. & Pichon, B. (2000)
Nuclear targeting of proteins: how many different signals? Cell
Signal. 12, 337–341.
4. Kleanthous, C. & Walker, D. (2001) Immunity proteins: enzyme
inhibitors that avoid the active site. Trends Biochem. Sci. 26,
624–631.
5. Ma, B., Shatsky, M., Wolfson, H.J. & Nussinov, R. (2002) Mul-
tiple diverse ligands binding at a single protein site: a matter of pre-
existing populations. Protein Sci. 11, 184–197.
Fig. 5. Detailed view of TetR(D)[tc–Mg]
2
. The water-network con-
necting helices a4anda7 and involved residues V57 (red) and Q109
(orange), and the neighbouring residue A61 (green) are shown.
Ó FEBS 2004 TetR anhydrotetracycline recognition (Eur. J. Biochem. 271) 2151
6. Ma, B., Kumar, S., Tsai, C.J. & Nussinov, R. (1999) Folding
funnels and binding mechanisms. Protein Eng. 12, 713–720.
7. Foote, J. & Milstein, C. (1994) Conformational isomerism and the
diversity of antibodies. Proc. Natl Acad. Sci. USA 91, 10370–
10374.
8. Wedemeyer, W.J., Ashton, R.W. & Scheraga, H.A. (1997)
Kinetics of competitive binding with application to thrombin
complexes. Anal Biochem. 248, 130–140.
9. Kumar, S., Ma, B., Tsai, C.J., Sinha, N. & Nussinov, R. (2000)
Folding and binding cascades: dynamic landscapes and popula-
tion shifts. Protein Sci. 9, 10–19.
10. Ringe, D. (1995) Structure-aided drug design: crystallography and
computational approaches. J. Nucl. Med. 36, 28S–30S.
11. Hillen, W. & Berens, C. (1994) Mechanisms underlying expression
of Tn10 encoded tetracycline resistance. Annu. Rev. Microbiol. 48,
345–369.
12. Berens, C. & Hillen, W. (2003) Gene regulation by tetracyclines.
Constraints of resistance regulation in bacteria shape TetR for
application in eukaryotes. Eur. J. Biochem. 270, 3109–3121.
13. Levy, S.B., McMurry, L.M., Barbosa, T.M., Burdett, V.,
Courvalin, P., Hillen, W., Roberts, M.C., Rood, J.I. & Taylor,
D.E. (1999) Nomenclature for new tetracycline resistance
determinants. Antimicrob. Agents Chemother. 43, 1523–1524.
14. Orth, P., Schnappinger, D., Hillen, W., Saenger, W. & Hinrichs,
W. (2000) Structural basis of gene regulation by the tetracycline
inducible Tet repressor-operator system. Nat. Struct. Biol. 7,
215–219.
15. Hinrichs, W., Kisker, C., Duvel, M., Muller, A., Tovar, K., Hillen,
W. & Saenger, W. (1994) Structure of the Tet repressor-tetra-
cycline complex and regulation of antibiotic resistance. Science
264, 418–420.
16. Kisker, C., Hinrichs, W., Tovar, K., Hillen, W. & Saenger, W.
(1995) The complex formed between Tetrepressor and tetra-
cycline-Mg2+ reveals mechanism of antibiotic resistance. J. Mol.
Biol. 247, 260–280.
17. Wissmann, A., Wray, L.V. Jr, Somaggio, U., Baumeister, R.,
Geissendorfer, M. & Hillen, W. (1991) Selection for Tn10 tet
repressor binding to tet operator in Escherichia coli:isolationof
temperature-sensitive mutants and combinatorial mutagenesis in
the DNA binding motif. Genetics 128, 225–232.
18. Schnappinger, D., Schubert, P., Pfleiderer, K. & Hillen, W. (1998)
Determinants of protein-protein recognition by four helix bundles:
changing the dimerization specificity of Tet repressor. EMBO J.
17, 535–543.
19. Ettner,N.,Muller,G.,Berens,C.,Backes,H.,Schnappinger,D.,
Schreppel, T., Pfleiderer, K. & Hillen, W. (1996) Fast large-scale
purification of tetracycline repressor variants from overproducing
Escherichia coli strains. J Chromatogr. A. 742, 95–105.
20. Landt, O., Grunert, H.P. & Hahn, U. (1990) A general method for
rapid site-directed mutagenesis using the polymerase chain reac-
tion. Gene 96, 125–128.
21. Miller, J.H. (1992) A Short Course in Bacterial Genetics. A
Laboratory Manual and Handbook for Escherichia Coli and
Related Bacteria. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor.
22. Hansen, D., Altschmied, L. & Hillen, W. (1987) Engineered Tet
repressor mutants with single tryptophan residues as fluorescent
probes. Solvent accessibilities of DNA and inducer binding sites
and interaction with tetracycline. J. Biol. Chem. 262, 14030–14035.
23. Takahashi, M., Degenkolb, J. & Hillen, W. (1991) Determination
of the equilibrium association constant between Tetrepressor and
tetracycline at limiting Mg
2+
concentrations: a generally applic-
able method for effector-dependent high-affinity complexes. Anal.
Biochem. 199, 197–202.
24. Backes, H., Berens, C., Helbl, V., Walter, S., Schmid, F.X. &
Hillen, W. (1997) Combinations of the alpha-helix-turn-alpha-
helix motif of TetR with respective residues from LacI or 434Cro:
DNA recognition, inducer binding, and urea-dependent dena-
turation. Biochemistry 36, 5311–5322.
25. Schubert, P., Schnappinger, D., Pfleiderer, K. & Hillen, W. (2001)
Identification of a stability determinant on the edge of the Tet
repressor four-helix bundle dimerization motif. Biochemistry 40,
3257–3263.
26. Scholz, O., Schubert, P., Kintrup, M. & Hillen, W. (2000) Tet
repressor induction without Mg
2+
. Biochemistry 39, 10914–10920.
27. Takahashi, M., Altschmied, L. & Hillen, W. (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.
28. Kintrup, M., Schubert, P., Kunz, M., Chabbert, M., Alberti, P.,
Bombarda, E., Schneider, S. & Hillen, W. (2000) Trp scanning
analysis of Tetrepressor reveals conformational changes asso-
ciated with operator and anhydrotetracycline binding. Eur. J.
Biochem. 267, 821–829.
29. Pace, C.N. (1986) Determination and analysis of urea and
guanidine hydrochloride denaturation curves. Methods Enzymol.
131, 266–280.
30. Orth, P., Cordes, F., Schnappinger, D., Hillen, W., Saenger, W. &
Hinrichs, W. (1998) Conformational changes of the Tet repressor
induced by tetracycline trapping. J. Mol. Biol. 279, 439–447.
31. Richieri, G.V., Low, P.J., Ogata, R.T. & Kleinfeld, A.M. (1997)
Mutants of rat intestinal fatty acid-binding protein illustrate the
critical role played by enthalpy-entropy compensation in ligand
binding. J. Biol. Chem. 272, 16737–16740.
32. Dill, K.A. (1997) Additivity principles in biochemistry. J. Biol.
Chem. 272, 701–704.
33. Fischer, S. & Verma, C.S. (1999) Binding of buried structural
water increases the flexibility of proteins. Proc. Natl Acad. Sci.
USA 96, 9613–9615.
34. Van Regenmortel, M.H. (1999) Molecular recognition in the post-
reductionist era. J. Mol. Recognit. 12,1–2.
2152 P. Schubert et al. (Eur. J. Biochem. 271) Ó FEBS 2004
. Tet repressor residues indirectly recognizing anhydrotetracycline Peter Schubert*, Klaus Pfleiderer and Wolfgang Hillen Lehrstuhl. sequence variants of the tetracycline repressor (TetR) to investigate ligand recognition of two different tetracycline (tc) derivatives, tc and anhydrotetra- cycline (atc). TetR regulates resistance. 91318528081, E-mail: whillen@biologie.uni-erlangen.de Abbreviations: TetR, tetracycline repressor; tc, tetracycline; atc, anhydrotetracycline. *Present address: Biomedical Research Centre, University