Single-molecule approach to DNA minor-groove association dynamics

Article abstract of DOI:10.1002/anie.201201099

Getting in the groove: Fluorescence correlation spectroscopy reveals that the dynamics of the association process of the bisbenzamidine minor-groove binder BBA-OG (blue with green star, see scheme) to dsDNA is not controlled by diffusion, but by the inser

Full text of DOI:10.1002/anie.201201099

Angewandte
Chemie
DOI: 10.1002/anie.201201099
DNA Binding Dynamics
Single-Molecule Approach to DNA Minor-Groove Association
Dynamics**
Jorge Bordello, Mateo I. Sꢀnchez, M. Eugenio Vꢀzquez, Josꢁ L. MascareÇas, Wajih Al-Soufi,
and Mercedes Novo*
Chemists have long pursued the design and preparation of
small molecules that can recognize specific DNA sequences.
Deciphering the human genome and on-going efforts to
sequence the genomes of many other organisms have
provided a wealth of information about DNA targets of
both therapeutic and diagnostic interest, and therefore there
is a renewed interest in the development of smart DNA
minor-groove binders.[1,2] Despite the detailed structural and
thermodynamic information available regarding the interac-
tion of a number of minor-groove binders with DNA, kinetic
data are much more scarce. In addition, the very low
dissociation rates typical of these agents result in very slow
binding dynamics, which are hard to measure experimen-
tally.[3] Typically, stopped-flow methods are used, but this
technique can yield complex kinetics and artifacts arise from
the relatively high binder concentrations required. Moreover,
the dissociation rates are usually obtained with indirect SDS-
sequestering techniques that only give apparent rates.
Dynamic data of archetypical minor-groove binders such as
distamycin or Hoechst 33258 indicate that the association
process is very fast and nearly diffusion limited, which is
surprising given the severe geometric constraints imposed
upon the inclusion of a binder into the narrow minor
groove.[4–7]
bisbenzamidines[9] have encouraged the search for new
derivatives with improved efficacy, improved pharmacolog-
ical properties, and reduced adverse effects.[10] Bisbenzami-
dines bind short AT-rich DNA sequences by insertion in the
narrow minor groove. NMR spectroscopy and X-ray crystal-
lographic studies agree on the general model of the inter-
action, wherein the positively charged amidinium groups are
situated deep in the minor groove, making both direct and
indirect hydrogen bonds with the DNA bases and electro-
static contacts with the bottom of the groove.[11]
Recently, we reported several rapid and practical
approaches to synthesizing bisbenzamidine DNA binders,
some of which can be easily conjugated to functional groups
or fluorophores.[12] Using these methods, we made the
conjugate BBA-OG (Scheme 1), which features an Oregon
Cationic bisbenzamidines are able to target AT-rich
sequences preferentially over those containing GC pairs.
Prominent examples of this family of molecules, such as
pentamidine or furamidine, have found clinical applications
and are used for the treatment of several major tropical
diseases.[8] Nevertheless, the toxic side effects of these
Scheme 1. Synthesis of Oregon Green-labeled bisbenzamidine BBA-
OG. DMF=dimethylformamide; DIEA=N,N-diisopropylethylamine.
Green (OG) fluorophore attached to the DNA binder,
bisbenzamidine (BBA). We envisioned that the presence of
a negatively charged carboxylate in the fluorophore would
increase the DNA dissociation rate, and thereby allow us to
extract kinetic information using fluorescence correlation
spectroscopy (FCS). This technique provides a sensitive
single-molecule approach to study the fast kinetics of
biomolecular interactions,[13] yielding accurate values for the
binding equilibrium constant, the association and dissociation
rate constants, and the diffusion coefficients,[14,15] from which
the diffusion-controlled rate constants can be estimated.[16] To
the best of our knowledge, this technique has not been
previously used to analyze the dynamics of minor-groove
binders. The wide dynamic range of FCS makes it possible to
identify the rate-limiting step in the association event by
comparison of the dynamics of the association to DNA with
different sequences and chain lengths, and thus to evaluate
processes such as sliding or two-dimensional diffusion, which
[*] J. Bordello, Prof. Dr. W. Al-Soufi, Prof. Dr. M. Novo
Departamento de Quꢀmica Fꢀsica, Universidade de Santiago de
Compostela, Facultade de Ciencias
27001 Lugo (Spain)
E-mail: m.novo@usc.es
M. I. Sꢁnchez, Prof. Dr. M. E. Vꢁzquez, Prof. Dr. J. L. MascareÇas
Departamento de Quꢀmica Orgꢁnica y Centro Singular de Inves-
tigaciꢂn en Quꢀmica Biolꢂgica y Materiales Moleculares, Unidad
Asociada al CSIC, Universidade de Santiago de Compostela
15782 Santiago de Compostela (Spain)
[**] We thank the Ministerio de Ciencia e Innovaciꢂn (CTQ2010-21369,
SAF2007-61015, SAF2010-20822-C02, CTQ2009-14431/BQU, Con-
solider Ingenio 2010 CSD2007-00006) and the Xunta de Galicia
(INCITE09262304PR, INCITE09E2R209064ES, IN845B-2010/094,
INCITE09 209 084PR, GRC2010/12, PGIDIT08CSA-047209PR) for
financial support. J.B. and M.I.S thank the Ministerio de Ciencia
e Innovaciꢂn for their research scholarships.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201201099.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 1: Association (k+) and dissociation (kꢁ) rate constants and
binding equilibrium constant (K) of BBA-OG with dsDNA. The values of
k+ are calculated from the fit parameters kꢁ and K (K=k+/kꢁ).
have been proposed as likely mechanisms for the binding of
proteins to DNA.[16,17]
FCS titrations of BBA-OG with three short hairpin (hp)
oligonucleotides (12 bp + loop) containing as key target
sequences AAATTT, AATTT, and a non-target control
GGCCC and with a longer dsDNA (50 bp) containing one
AAATTT site, were performed. For the AAATTT-hp, an
additional FCS titration at a higher salt concentration was
measured (see Supporting Information for full DNA sequen-
ces). Figure 1 shows normalized fluorescence correlation
DNA[a]
k+ [ꢃ108 mꢁ1 sꢁ1
]
kꢁ [ꢃ104 sꢁ1
]
K [ꢃ103 mꢁ1
]
AAATTT-hp[a]
AATTT-hp[a]
GGCCC-hp[a]
AAATTT-ds[a]
AAATTT-hp[b]
1.36ꢂ0.01
0.95ꢂ0.01
0.083ꢂ0.003
1.56ꢂ0.02
0.65ꢂ0.03
3.19ꢂ0.02
4.89ꢂ0.05
14.0ꢂ0.4
4.27ꢂ0.03
1.95ꢂ0.02
0.059ꢂ0.001
5.47ꢂ0.05
1.41ꢂ0.03
2.84ꢂ0.02
4.63ꢂ0.01
[a] [NaCl]=0.1m, [b] [NaCl]=1.0m. See the Supporting Information for
complete oligonucleotide sequences.
Figure 2. Association (gray, left scale) and dissociation (white, right
scale) rate constants of BBA-OG with DNA, calculated from FCS
measurements. Estimates of the diffusion-controlled association rate
contants kd and kd’, defined in Scheme 2 (see Supporting Information).
Figure 2 shows that the association rate constant
decreases by more than one order of magnitude from the
AT-rich to the GC-containing sequences, whereas the disso-
ciation rate constant increases three to five times. Both
association and dissociation contribute to the specificity for
the AT-rich tracts, but the association has a higher effect. This
is in agreement with reported results for the binding of the
antibiotic distamycin to sequences with different numbers and
positions of mismatched sites.[4] Moreover, the absolute
values of k+ for the AT sequences coincide well with those
reported for other minor-groove binders[4–7] and show the
small decrease with increasing Na+ ion concentration
expected for groove binders,[20] suggesting they might have
analogous association mechanisms. Nevertheless, the dissoci-
ation rate constants are about four orders of magnitude
higher than those of typical minor-groove binders, resulting in
much lower binding constants.
The above results indicate much weaker specific inter-
actions between BBA-OG and the DNA minor groove, which
is probably owing to the electrostatic repulsion between the
negatively charged OG moiety and the DNA phosphate
backbone. The higher binding constants of unlabeled bisbenz-
amidines supports this explanation.[12] On this basis, we can
consider BBA-OG as a valid model for the study of the
association process of small molecules binding to the minor
groove of DNA.
Figure 1. FCS curves of the titration of BBA-OG with AAATTT-ds and
results of a global fit of the dynamic-equilibrium model (Supporting
Information, Equation S1–S4). Insets: variation of the mean diffusion
ꢀ
time tD, the relaxation time tR, and the relaxation amplitude AR with
the [DNA] calculated from the fit parameters. Lower panel: unweighted
fit residuals. The small deviations at high DNA concentrations are
from minor fluorescent impurities in the DNA samples. Above 100 ms
a factor of 10 was applied.
curves of the binding of BBA-OG to the longer dsDNA.
Without DNA, only a translational diffusion term of the
binder with correlation time tD is observed. Addition of DNA
increases tD, reflecting the slower diffusion of the dye when
bound to DNA. Additionally, a new term appears with
a shorter correlation time tR, which decreases as the concen-
tration of DNA is increased (see insets in Figure 1). tR is the
relaxation time of the reversible binding process:
tR ¼ ðkþ½DNAꢀ þ kꢁÞꢁ1 with k+ and kꢁ being the association
and dissociation rate constants, respectively. The large
amplitude AR of this term is because of the large increase in
the brightness of bound BBA-OG.[18] The diffusion and
relaxation terms show the expected concentration depend-
ence and are independent of the irradiance.[18,19] To reduce
parameter correlation, the FCS titration series were analyzed
by global target fits[14,15] with a common dynamic-equilibrium
model of the diffusion, triplet, and relaxation terms (Figure 1
and Figure S1–S4). The parameters that best fit the whole
data set are given in Table 1 and Figure 2.
To determine the rate-limiting step for the association
process, we compared the dynamics of binding to the short
DNA AAATTT-hp with that to the much longer AAATTT-
ds, which has been designed to avoid additional AT-rich sites.
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Curiously, we did not find significant differences in the
binding dynamics of these two duplexes (Table 1, Figure 2).
The similar dissociation (kꢁ) is readily explained in terms of
similar specific interactions with the AAATTT sequence.
However, if diffusion plays an important role, the association
dynamics should be affected by the increase of the DNA
length.
For a more detailed analysis, we propose a two-step
mechanism like that previously used to describe other
supramolecular binding processes (Scheme 2).[14,15] The first
pre-equilibrium with the separated species so that the overall
rate of association between BBA-OG and DNA is not
diffusion limited but determined by the next step of the
process, namely the unimolecular inclusion into the minor
groove. This step is probably controlled by geometric and
orientational requirements, which depend on the critical
dimensions of both the binder and the minor groove. The
presence of an exocyclic 2-amino group in the minor groove
of GC tracts[21] may explain the much lower value of k+
obtained for GGCCC-hp.
We can also conclude that processes that increase the rate
of formation of localized encounter complexes, such as sliding
or two-dimensional diffusion along the DNA strand,[16,17] do
not play a key role in the association rate of BBA-OG, and
probably also do not for other minor-groove binders.
In summary, our results indicate that the association
process of BBA-OG to dsDNA is not controlled by diffusion,
but by the rate-limiting insertion of the binder into the minor-
groove. Moreover, we find that the association process has an
important effect on the specificity to AT-rich sites, whereas
the differences in the binding affinity are mainly determined
by the dissociation rate. This mechanism might constitute
a general mechanism for small minor-groove binders, but this
should be confirmed with further studies on more typical
minor-groove agents. This information should be useful for
the design of new DNA binders with optimized properties
and for the future understanding of the behavior of these
molecules in more complex cellular environments. This work
has also shown the potential of FCS for the study DNA
binding dynamics of minor-groove binders using labeled
derivatives.
Scheme 2. Two-step mechanism proposed for the binding of BBA-OG
to DNA involving the formation of an encounter complex, with the
binder localized at the DNA reactive site (L) or unlocalized (U).
step is the formation of an encounter complex between DNA
and binder, with a diffusion-controlled association rate which
can be estimated based on the geometry of the DNA and
binder. The second step is the insertion of the binder into the
minor-groove with a rate constant kr, a process which conveys
structural rearrangements or the breakdown and reconstitu-
tion of the network of water molecules surrounding the Experimental Section
Oligonucleotides were purchased from Thermo Fisher Scientific Inc.
interacting species. The experimentally determined k+ corre-
sponds to the overall reaction described by a single relaxation
time and is determined by the rate-limiting step. This
mechanism is therefore different to the sequential model
proposed for the binding of Hoechst 33258 to DNA[7] which
involves two reactions, with their corresponding two observed
reaction times.
Assuming a rod-like structure of the DNAwith a localized
reactive site, we estimate lower bounds (see Supporting
Information) of the diffusion-controlled association rate
constants kd and kd’ for the formation of unlocalized (U)
and localized (L) encounter complexes, respectively, where
the binder is either located arbitrarily at any position of the
DNA (U) or already localized near to the reactive site (L)
(Scheme 2).[16]
The estimated values of kd’ for the formation of localized
complexes L are of course significantly lower than those of kd
corresponding to the formation of U, especially in the case of
the long dsDNA (Figure 2). Nevertheless, even for the most
specific sequences, both kd and kd’ are at least one order of
magnitude higher than the observed overall rate constant k+.
This shows that the diffusion-controlled formation of the
encounter complex is not the rate-limiting step, even if the
binder is required to be located near the reactive sequence
(L complex). The encounter complexes U and L are in rapid
Nucleotide sequences and the procedure for reconstitution and
annealing are provided in the Supporting Information. Details of the
synthesis of the BBA-OG can also be found in the Supporting
Information. Experimental setup and conditions, as well as the details
of the data analysis for the FCS measurements have been published
elsewhere[18,22] and are specified for these experiments in the
Supporting Information. The uncertainties given in Table 1 represent
one standard deviation as obtained by the fits.
Received: February 9, 2012
Revised: March 26, 2012
Published online: && &&, &&&&
Keywords: binding dynamics · DNA ·
.
fluorescence correlation spectroscopy · minor-groove binders ·
molecular recognition
[1] a) D. R. Boer, A. Canals, M. Coll, J. Chem. Soc. Dalton Trans.
2009, 3, 399 – 414; b) W. C. Tsee, D. L. Boger, Chem. Biol. 2004,
11, 1607 – 1617; c) E. Pazos, J. Mosquera, M. E. Vꢀzquez, J. L.
MascareÇas, ChemBioChem 2011, 12, 1958 – 1973; d) M. E.
Vꢀzquez, A. M. CaamaÇo, J. L. MascareÇas, Chem. Soc. Rev.
2003, 32, 338 – 349.
[2] a) M. E. Vꢀzquez, A. M. CaamaÇo, J. Martꢁnez-Costas, L.
Castedo, J. L. MascareÇas, Angew. Chem. 2001, 113, 4859 –
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
4861; Angew. Chem. Int. Ed. 2001, 40, 4723 – 4725; b) O.
Vꢀzquez, M. E. Vꢀzquez, J. B. Blanco, L. Castedo, J. L. Mascar-
eÇas, Angew. Chem. 2007, 119, 7010 – 7014; Angew. Chem. Int.
Ed. 2007, 46, 6886 – 6890.
[12] a) O. Vꢀzquez, M. I. Sꢀnchez, J. Martꢁnez-Costas, M. E.
Vꢀzquez, J. L. MascareÇas, Org. Lett. 2010, 12, 216 – 219; b) O.
Vꢀzquez, M. I. Sꢀnchez, J. L. MascareÇas, M. E. Vꢀzquez, Chem.
Commun. 2010, 46, 5518 – 5520; c) M. I. Sꢀnchez, O. Vꢀzquez, J.
Martꢁnez-Costas, M. E. Vꢀzquez, J. L. MascareÇas, Chem. Sci.
2012, 3, 2383 – 2387; d) M. I. Sꢀnchez, O. Vꢀzquez, M. E.
Vꢀzquez, J. L. MascareÇas, Chem. Commun. 2011, 47, 11107—
11109.
[13] a) A. van Oijen, Curr. Opin. Biotechnol. 2011, 22, 75 – 80; b) D.
Magde, E. L. Elson, W. W. Webb, Biopolymers 1974, 13, 29 – 61.
[14] W. Al-Soufi, B. Reija, M. Novo, S. Felekyan, R. Kꢃhnemuth,
C. A. M. Seidel, J. Am. Chem. Soc. 2005, 127, 8775 – 8784.
[15] W. Al-Soufi, B. Reija, S. Felekyan, C. A. Seidel, M. Novo,
ChemPhysChem 2008, 9, 1819 – 1827.
[3] a) L. M. Wilhelmsson, P. Lincoln, B. Nordꢂn in Sequence-
Specific DNA Binding Agents (Ed.: M. Waring), The Royal
Society of Chemistry, London, 2006, pp. 69 – 95; b) T. C. S. Pace,
C. Bohne, Adv. Phys. Org. Chem. 2007, 42, 167 – 223.
[4] R. Baliga, D. M. Crothers, J. Am. Chem. Soc. 2000, 122, 11751 –
11752.
[5] R. Baliga, D. M. Crothers, Proc. Natl. Acad. Sci. USA 2000, 97,
7814 – 7818.
[6] S. Y. Breusegem, S. Sadat-Ebrahimi, K. T. Douglas, R. M. Clegg,
F. G. Loontiens, J. Mol. Biol. 2001, 308, 649 – 663.
[7] S. Y. Breusegem, R. M. Clegg, F. G. Loontiens, J. Mol. Biol. 2002,
315, 1049 – 1061.
[16] O. G. Berg, P. H. von Hippel, Annu. Rev. Biophys. Biophys.
Chem. 1985, 14, 131 – 158.
[8] a) R. BalaÇa-Fouce, M. C. Redondo, Y. Pꢂrez-Pertejo, R. Dꢁaz-
Gonzꢀlez, M. R. Reguera, Drug Discovery Today 2006, 11, 733 –
740; b) I. Midgley, K. Fitzpatrick, L. M. Taylor, T. L. Houchen,
S. J. Henderson, S. J. Wright, Z. R. Cybulski, B. A. John, A.
McBurney, D. W. Boykin, K. L. Trendler, Drug Metab. Dispos.
2007, 35, 955 – 967.
[9] a) W. D. Wilson, B. Nguyen, F. A. Tanious, A. Mathis, J. E. Hall,
C. E. Stephens, D. W. Boykin, Curr. Med. Chem. Anti-Cancer
Agents 2005, 5, 389 – 408; b) A. H. Fairlamb, Trends Parasitol.
2003, 19, 488 – 494.
[17] D. Vuzman, A. Azia, Y. Levy, J. Mol. Biol. 2010, 396, 674 – 684.
[18] J. Bordello, M. Novo, W. Al-Soufi, J. Colloid Interface Sci. 2010,
345, 369 – 376.
[19] M. Novo, D. Granadero, J. Bordello, W. Al-Soufi, J. Inclusion
Phenom. Macrocyclic Chem. 2011, 69, 1 – 10.
[20] a) W. D. Wilson, F. A. Tanious, H. J. Barton, R. L. Jones, K. Fox,
R. L. Wydra, L. Strekowski, Biochemistry 1990, 29, 8452 – 8461;
b) F. A. Tanious, J. M. Veal, H. Buczak, L. S. Ratmeyer, W. D.
Wilson, Biochemistry 1992, 31, 3103 – 3112; c) T. M. Lohman,
CRC Crit. Rev. Biochem. 1986, 19, 191 – 245.
[10] R. R. Tidwell, D. W. Boykin in DNA and RNA Binders: From
Small Molecules to Drugs (Eds.: M. Demeunynck, C. Bailly,
W. D. Wilson), Wiley-VCH, Weinheim, 2003, pp. 414 – 460.
[11] C. M. Nunn, S. Neidle, J. Med. Chem. 1995, 38, 2317 – 2325.
[21] S. Neidle, Nat. Prod. Rep. 2001, 18, 291 – 309.
[22] D. Granadero, J. Bordello, M. J. Pꢂrez-Alvite, M. Novo, W. Al-
Soufi, Int. J. Mol. Sci. 2010, 11, 173 – 188.
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
DNA Binding Dynamics
J. Bordello, M. I. Sꢁnchez, M. E. Vꢁzquez,
J. L. MascareÇas, W. Al-Soufi,
M. Novo*
&&&& — &&&&
Single-Molecule Approach to DNA
Minor-Groove Association Dynamics
Getting in the groove: Fluorescence cor-
relation spectroscopy reveals that the
dynamics of the association process of
the bisbenzamidine minor-groove binder
BBA-OG (blue with green star, see
scheme) to dsDNA is not controlled by
diffusion, but by the insertion of the
binder into the groove at the specific site
(red), as shown by the rate constants for
each step of the binding event.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!