Saturation transfer difference NMR reveals functionally essential kinetic differences for a sugar-binding repressor protein

Article abstract of DOI:10.1039/b913489a

The binding kinetics of disaccharides trehalose and trehalose-6-phosphate to repressor protein TreR have been determined using STD NMR and shed light on the contrasting biological roles of these two sugars.

Full text of DOI:10.1039/b913489a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Saturation transfer difference NMR reveals functionally essential kinetic
differences for a sugar-binding repressor proteinw
a
a
a
a
Ignacio Pe
´
James C. Errey, Lucia F. Primavesi, Matthew J. Paul, Timothy D. W. Claridge*
rez-Victoria,z Sebastian Kemper,z Mitul K. Patel, John M. Edwards,
a b b a
a
and Benjamin G. Davis*
Received (in Cambridge, UK) 7th July 2009, Accepted 3rd August 2009
First published as an Advance Article on the web 19th August 2009
DOI: 10.1039/b913489a
6
The binding kinetics of disaccharides trehalose and trehalose-6-
phosphate to repressor protein TreR have been determined using
STD NMR and shed light on the contrasting biological roles of
these two sugars.
confirmed by competition STD NMR experiments. Titration
of 2 into a solution of 1ꢀTreR complex (Fig. 1) confirmed that
both ligands bound competitively to the same binding site and
that trehalose-6-phosphate (2) is the stronger binder of the
two: at a ratio of 1 : 4 : 100 TreR : 2 : 1 (X = 4) the STD effect
for trehalose was essentially absent.
Saturation transfer difference NMR (STD NMR) is a power-
ful and popular tool to identify and characterize the binding of
The absence of STD effect for the phosphate derivative of
trehalose was intriguing since its dissociation constant
1
,2
ligands to protein receptors. The technique is extremely well
suited to the study of protein–carbohydrate interactions due to
D
(K = 10 mM) is well within the range of amenable binding
2
affinities for the technique; we reasoned that an unusually
slow off (dissociation) rate koff might explain this experimental
observation since in this case transfer of saturation to ligand
the broad range of binding affinities that may be studied
2
from 100 nM to 10 mM; K
(
K
D
D
s that are often typical of
3
such interactions).
2
We have employed STD NMR to study the binding of
trehalose (1) and trehalose-6-phosphate (2) to the trehalose
repressor protein (TreR) from Escherichia coli, the repressor
which regulates the pathway of trehalose utilization in
molecules into solution is not very efficient. This striking
kinetic contrast between 1 and 2 was probed further using
comparative computational analysis employing the program
7
,8
CORCEMA-ST
to calculate theoretical STD effects via
4
this bacteria. TreR binds with 1 and 2 with similar
complete relaxation and exchange rate matrices based
on the three-dimensional coordinates of TreR complexes,
concentrations, equilibrium binding constants, and association
and dissociation rate constants. Such an approach might allow
affinities (trehalose-6-phosphate, KD = 10 mM cf. trehalose,
KD = 280 mM), but it has long been known that only the
binding and engagement of 2 results in a biologically relevant
reduction of the repressor affinity for binding to the DNA
4
operator site. Despite this essential physiological functional
difference, the crystal structures of the complexes of TreR with
its inducer trehalose-6-phosphate (2) and the non-inducer
5
trehalose (1) were found to be virtually identical. This apparent
D
similarity in such ground state measurements (structure, K )
suggests that key mechanistic differences lie elsewhere.
Parent disaccharide trehalose 1 and its synthetic phosphate
derivative 2 (see ESIw for synthesis) were surveyed for binding
to TreR using a range of methods. Of these, the use of STD
NMR was particularly noteworthy; unusually, under identical
experimental conditions we observed a clear STD effect for 1
but none for 2. Absence of an STD signal can be due to a
number of mechanistic origins (including absence of binding).
However, 2 is in fact a more potent ligand than 1, critically
a
Department of Chemistry, University of Oxford, Chemistry Research
Laboratory, Mansfield Road, Oxford, UK OX1 3TA.
E-mail: Tim.Claridge@chem.ox.ac.uk, Ben.Davis@chem.ox.ac.uk;
Fax: +44 (0)1865 285002; Tel: +44 (0)1865 285001
Plant Science, Rothamsted Research, Harpenden, Herts,
UK AL5 2JQ
b
w Electronic supplementary information (ESI) available: Trehalose-6-
phosphate synthesis, STD NMR measurements, CORCEMA-ST
calculations and trehalose-6-phosphate conformation discussion.
Links to view 3D visualisations of structures using FirstGlance. See
DOI: 10.1039/b913489a
Fig.
trehalose-6-phosphate (2). All STD spectra were acquired with a ratio
: X : 100 TreR : 2 : 1 as indicated. The lowest reference spectrum is
for trehalose (1).
1 Competition STD NMR experiment with increasing
1
z Both of these authors contributed equally to this work.
5
862 | Chem. Commun., 2009, 5862–5864
This journal is ꢁc The Royal Society of Chemistry 2009

View Article Online
5
Fig. 2 Repressor protein TreR (PDB entry: 1BYK) showing the
ligand trehalose-6-phosphate (2, blue) and the residues (green)
included within the CORCEMA-ST cutoff distance for calculations
in both surface (a) and ribbon (b) representations generated using
Fig. 3 (a) Dependence of predicted STD factor at 5 s saturation time
(
for H-2 in 2) with off rate. Dashed line represents the calculated
12
saturation for the unbound state. (b) Dependence of predicted STD
factor at 5 s saturation time (for H-2 in both ligands) with the on rate
PyMOL.
(
solid line K 1, dashed line ’ 2).
the calculation of STD effects as a function of the dissociation
9
rate (koff).
In this way the thresholds for the rates of binding of
trehalose-6-phosphate to TreR could be probed. Using the
crystal structure of trehalose-6-phosphate in complex with
associated conformational change. Such large changes in
10
shape are often the mode of action of sugar-triggered repressors
and is here apparently caused by a subtle difference in ligand
identity, the phosphate at C-6 of trehalose. Nonetheless, other
mechanistic modes, including those utilizing through-protein
5
˚
TreR (PDB entry: 1BYK), residues within 15 A of the bound
ligand in the binding pocket (Fig. 2) were explicitly used to
predict STD factors (using representative ligand protons) as a
function of the dissociation rate (koff) (Fig. 3). These revealed
1
1
electrostatic effects (such as is the case for ‘‘electric-genetic’’
switches that utilize effector molecules like 2 that can contain
only a single charged group), cannot be excluded.
ꢂ1
that a low STD factor is obtained for off rates r0.1 s . The
residual calculated STD factor is caused nearly exclusively by
the ligand in its bound state; because of the slow kinetics only
a negligible part will be released to the solution and the
calculated unbound ligand saturation converges to zero at
low off rates (Fig. 3a). The bound ligand saturation is hardly
detectable since the ligand signals in its bound state have a
similar linewidth to that of the protein and thus will merge
with the protein background; therefore the unbound ligand
saturation is more representative and allows the establishment
Consistent with the difference in biological effect, use of the
same novel ‘kinetic STD’ approach revealed key differences in
on rate. Previous analysis of the STD effects of trehalose using
a novel approach of group epitope mapping considering the
1
3
relaxation of the ligand (GEM-CRL)
combined with
CORCEMA-ST calculations has suggested that the reported
5
structure by Hars and co-workers is valid for the solution
structure of the TreR–trehalose complex. Use of the
GEM-CRL approach here revealed that the binding kinetics
of trehalose (1) are in striking contrast to those of 2; an off rate
ꢂ1
of an upper limit value for koff(2) = 0.1 s , which implies a
4
ꢂ1 ꢂ1
ꢂ1
kon(2) r 10 M
is typically observed in situations where there is a large
s
(KD = koff/k ). Such a low on rate
of 4100 s was required to properly reflect calculated values.
Moreover, CORCEMA-ST calculations with different on
on
6
rates resulted in a lower limit for the kon(1) of 10 M
ꢂ1 ꢂ1
conformational rearrangement of receptor or ligand upon
2
binding. Strong shifts in the frequency of the fluorescence
4
emission maximum of TreR during binding of 2 but an
s
ꢂ1
(koff Z 280 s ). This threshold value was determined by
comparing the experimental STD build-up curves with those
7
calculated in NOE R-factor determination. The minimum
essential absence of difference in the conformation of bound
or unbound ligand 2 (see ESIw) suggests that this is a protein
R value, which allows identification of the ‘‘true’’ on rate, was
Chem. Commun., 2009, 5862–5864 | 5863
This journal is ꢁc The Royal Society of Chemistry 2009

View Article Online
6
ꢂ1 ꢂ1
s
reached at kon Z 10 M
(see ESIw). At such high rates it
Work in this communication was supported by
BBSRC grant BB/D006112/1, the European Community
(PIEF-GA-2008-221066, I. P-V.), the Deutscher Akademischer
cannot be discounted that binding of 1 to TreR is rapid,
essentially diffusion controlled, and causes no functional
change in TreR.
¨
Austausch Dienst (S. K.) and the Kolner Gymnasial- und
Taken together these results suggest that for TreR, although
the ratio K (trehalose)/K (trehalose-6-phosphate) is small
Stiftungsfonds (S. K.). Rothamsted Research receives
grant-aided support from the Biotechnological and Biological
Research Council (BBSRC) of the United Kingdom. We are
grateful to Prof. Winfried Boos, University of Konstanz,
Germany, for the gift of plasmid pCYTEXPtreR and to Prof.
Rama Krishna, University of Alabama, USA, for making
available to us the CORCEMA-ST program.
D
D
(
only 28), the corresponding ratio of dissociation rates
off(trehalose)/koff(trehalose-6-phosphate) is Z 2800 and the
k
2
ratio of association (kon) rates Z 10 (Fig. 3). This suggests
that the different biological function of both sugars (trehalose-
6
-phosphate active, trehalose inactive) can be explained by
considering kinetic ratios; these in turn suggest a large
conformational rearrangement of TreR upon binding to
trehalose-6-phosphate (the active inducer) or a strong electro-
Notes and references
1
1
static mechanism which triggers an inability of the repressor
to bind the operator site in the DNA sequence, thus allowing
gene expression. Such a rearrangement does not occur upon
trehalose binding, allowing the union of the repressor to the
operator inhibiting gene expression.
1 M. Mayer and B. Meyer, Angew. Chem., Int. Ed., 1999, 38,
1
784–1788.
2
3
B. Meyer and T. Peters, Angew. Chem., Int. Ed., 2003, 42, 864–890.
J. Angulo, C. Rademacher, T. Biet, A. J. Benie, A. Blume,
H. Peters, M. Palcic, F. Parra, T. Peters and F. Minoru, Methods
Enzymol., 2006, 416, 12–30.
4
5
R. Horlacher and W. Boos, J. Biol. Chem., 1997, 272,
13026–13032.
U. Hars, R. Horlacher, W. Boos, W. Welte and K. Diederichs,
In summary, taking advantage of the availability of
structural protein data for TreR, calculations have allowed
us to determine key kinetic data for an effector (2) and a
non-effector ligand (1) (threshold dissociation and association
rate values) that was probed experimentally using STD NMR.
In this way insights into binding kinetics of both ligands with
the repressor can be used to explain their very different
biological roles and effects. To our knowledge, this is the first
use of such STD kinetic analysis in the discovery of a
functionally important kinetic difference in ligand–protein
interactions. Such solution phase kinetic analysis we suggest
might be generally useful to complement other biophysical
Protein Sci., 1998, 7, 2511–2521.
6 Y.-S. Wang, D. Liu and D. F. Wyss, Magn. Reson. Chem., 2004,
2, 485–489.
V. Jayalakshmi and N. R. Krishna, J. Magn. Reson., 2002, 155,
06–118.
4
7
1
8 N. R. Krishna and V. Jayalakshmi, Prog. Nucl. Magn. Reson.
Spectrosc., 2006, 49, 1–25.
9
J. Angulo, B. Langpap, A. Blume, T. Biet, B. Meyer,
N. R. Krishna, H. Peters, M. M. Palcic and T. Peters, J. Am.
Chem. Soc., 2006, 128, 13529–13538.
10 M. Taraban, H. Zhan, A. E. Whitten, D. B. Langley,
K. S. Matthews, L. Swint-Kruse and J. Trewhella, J. Mol. Biol.,
2
008, 376, 466–481.
1 K. Phillips and S. E. V. Phillips, Structure (London), 1994, 2,
09–316.
1
4
15
techniques such as SPR or QCM. These other methods can
also allow determination of kinetic values but require one
component (typically ligand) to be bound to a solid phase
1
3
12 W. L. DeLano, The PyMOL Molecular Graphics System, DeLano
Scientific, Palo Alto, CA, USA, 2002.
(
SPR chip or microbalance), which can bring with it a bias that
1
3 S. Kemper, M. K. Patel, J. C. Errey, B. G. Davis, J. A. Jones and
T. D. W. Claridge, submitted.
may be less relevant to a solution phase ligand or interaction.
The kinetic STD method therefore might valuably allow
more general and relevant determinations free from such
constraints.
1
4 J. M. McDonnell, Curr. Opin. Chem. Biol., 2001, 5, 572–577.
5 T. Rudd, J. T. Gallagher, D. Ron, R. J. Nichols and D. G. Fernig,
Biochem. Soc. Trans., 2003, 31, 349–351.
1
5
864 | Chem. Commun., 2009, 5862–5864
This journal is ꢁc The Royal Society of Chemistry 2009