This journal is c The Royal Society of Chemistry 2013 Chem. Commun.

Cite this: DOI: 10.1039/c3cc44569k

Different thermodynamic signatures for DNA minorgroove binding with changes in salt concentrationand temperature†

Shuo Wang,a Arvind Kumar,a Karl Aston,b Binh Nguyen,za James K. Bashkin,b

David W. Boykina and W. David Wilson*a

The effects of salt concentration and temperature on the thermo-

dynamics of DNA minor groove binding have quite different

signatures: binding enthalpy is salt concentration independent

but temperature dependent. Conversely, binding free energy is

salt dependent but essentially temperature independent through

enthalpy–entropy compensation.

For DNA interactions, from protein binding to DNA strand associa-tion and similar RNA reactions such as hairpin kissing complexes,polyelectrolyte effects on the thermodynamics of the DNA com-plexes have been extensively studied.1 These results help to providea fundamental understanding of the driving force for binding toDNA and the molecular basis of DNA recognition.2 Privalov andcoworkers have conducted studies of the thermodynamics of proteinsand peptides with DNA sites as a function of salt concentration andthey observed that the enthalpy for binding is essentially salt inde-pendent while its change with temperature is generally significant.3

Given the quite different effects that proteins can have on DNA thisis an important observation for biomolecular reactions.

Minor groove targeting by small molecules occurs in the oppositegroove for most proteins, and is an important and quite differentsequence-specific mechanism for DNA recognition. Small moleculescan target a broad range of DNA sequences by different modes, suchas monomers, cooperative dimers and covalent hairpin structures,with high affinity and selectivity.4 Even with hundreds of paperspublished on DNA minor groove binding, there are still a number ofimportant unanswered questions in the fundamentals of DNA minorgroove recognition that are required for rational design of novel minorgroove agents: (i) is the minor groove binding enthalpy independent ofsalt concentration as with protein–DNA interactions; (ii) how does theenthalpy and its salt effects change with recognition sequences (Fig. 1)

and monomer or dimer complex formation; (iii) how do temperatureeffects on the binding enthalpy and energy compare to salt effects? Tobegin to fill in this essential missing information, and also to extendour understanding of the energetic basis of DNA molecular recogni-tion, salt concentration and temperature effects on the thermo-dynamics of five quite different DNA minor groove complexes havebeen evaluated in detail in this work. They were chosen because theyare representative of all the current minor groove binding modes forrecognition of AT or GC-rich sites as monomers, dimers or hairpincomplexes. The binding enthalpy was obtained through isothermaltitration calorimetry (ITC) and the binding Gibbs free energy wasdetermined by biosensor surface plasmon resonance (SPR).

Netropsin (Net, Fig. 1) is a natural heterocyclic dication having veryhigh specificity for monomer binding to DNA sites containing four ormore AT base pairs (bp).5 Its binding enthalpies (DHb) with two DNAhairpin duplex sequences, AAAA and ATAT (Fig. 1) which have verydistinct structural features, have been measured. AAAA has a verynarrow groove with a local bend while alternating AT has a muchwider groove and a more linear conformation.5b,6 The ITC titrationsunder several salt concentrations are shown in Fig. S1 (ESI†) and theDHb values are plotted in Fig. 2a. Strikingly, theDHb values of Net withboth sequences are clearly independent of salt concentration eventhough they are quite different for binding to the two DNAs. The DHb

with AAAA (�9.1 � 0.1 kcal mol�1) is less negative than with ATAT

Fig. 1 Compound structures and hairpin DNA sequences.

a Department of Chemistry, Georgia State University, Atlanta, GA 30303, USA.

E-mail: wdw@gsu.edu; Fax: +1 404-413-5505; Tel: +1 404-413-5503b Department of Chemistry & Biochemistry, Center for Nanoscience,

University of Missouri-St. Louis, St. Louis, MO 63121, USA

† Electronic supplementary information (ESI) available: Experimental details, ITCtitrations and steady state fits of SPR data. See DOI: 10.1039/c3cc44569k‡ Current address: Department of Biochemistry and Molecular Biophysics,Washington University School of Medicine, St. Louis, MO 63110, USA.

Received 18th June 2013,Accepted 7th August 2013

DOI: 10.1039/c3cc44569k

www.rsc.org/chemcomm

ChemComm

COMMUNICATION

Publ

ishe

d on

08

Aug

ust 2

013.

Dow

nloa

ded

by W

ashi

ngto

n U

nive

rsity

in S

t. L

ouis

on

26/0

8/20

13 1

3:12

:29.

View Article OnlineView Journal

Chem. Commun. This journal is c The Royal Society of Chemistry 2013

(�12.0 � 0.2 kcal mol�1), as previously observed with other mono-mers.7 The narrow minor groove of A-tract sequences has an array oftightly structured water molecules which are highly ordered in a spineof hydration.8 These ordered water molecules have high entropy, andthus contribute a significant favourable entropy component when theyare displaced by ligand binding in A-tracts (Table S1, ESI†). Conversely,the binding in alternating AT sequences has a larger DHb along with asmaller binding entropy (DSb) in agreement with its wider minorgroove that releases less-ordered, less-tightly bound water molecules.

The heterocyclic diamidine DB293 (Fig. 1) targets the minorgroove of DNA sites that contain only AT bps as a monomer, but itbinds to a mixed sequence, ATGA, as a cooperative, stacked dimer.4b

The introduction of a GC bp widens the minor groove and favoursdimer formation in ATGA. In order to clarify the correlations amongsalt concentration, enthalpy and binding modes, ITC experimentswith DB293 were carried out with both the AAAA and ATGA sites(Fig. S2, ESI†). The plot of DHb versus salt concentration (Fig. 2a)shows that the DHb values for these two very different complexes arequite unlike, but both have a salt independent DHb. The dimerformation of DB293 on ATGA has a very negative enthalpy,�9.6� 0.2 kcal mol�1, and the 1 : 1 interaction between DB293 andAAAA shows a much less negative heat (�3.8 � 0.1 kcal mol�1). Aswith Net, these very different binding enthalpies are due to thedistinct minor groove structures, hydrations and binding modes inthe target sites. This comparison clearly shows that the presence ofa GC bp can strongly affect the binding mode of a minor groovebinder but the DHb remains independent of salt concentration.

To probe the effects of additional GC bps in the binding site onDHb, a synthetic hairpin polyamide (PA), KA1039, which is verydifferent in structure from Net and DB293, has been studied withits cognate binding sequence, TGGCTT. This PA molecule bindswith high affinity that is mainly driven by a favourable, negativeDHb.9 Surprisingly, when the standard compound concentration of50 mM is used, the DHb of KA1039 displayed significant decreaseswith increasing salt concentration (Fig. S3, ESI† and Fig. 2b). Toevaluate whether the salt effect on DHb is due to PA aggregation, aspreviously observed with larger PAs,9,10 more dilute KA1039 samples(37.5 and 20 mM) were used. The DHb values of 20 mM, 37.5 mMand 50 mM for KA1039 at 50 mM salt concentration are the same(�11.1 � 0.3 kcal mol�1), which indicates that all concentrationswork well in obtaining the DHb. The DHb values obtained at 20 mMof KA1039 are salt-independent as with other minor groovebinders in this work and this indicates that the salt dependent

DHb changes at 50 mM and 37.5 mM for KA1039 are due tocompound aggregation. The DHb of the monomer complex atGC-rich sites, even with this large molecule, is independent ofsalt concentration as with the simpler compounds.

In summary, regardless of monomer or dimer binding atAT-rich or GC-rich sites, or the size of the compound or DNA site,the enthalpy of DNA minor groove binding is salt concentrationindependent. This is similar to the results observed by Privalovand coworkers with protein–DNA interactions.3

To compare the effects of both salt concentration and temperatureon the thermodynamics for DNA minor groove binding, SPR experi-ments were conducted at 25 1C to obtain DGb under several saltconcentrations. The response unit (RU) values at each ligand concen-tration were determined in the steady state region, where the on andoff rates are equal and there is no mass transfer interference, andthey are plotted as a function of the free concentration (Cf) of KA1039in equilibrium with the complex in each flow solution (Fig. S4a, ESI†).The equilibrium binding constants obtained by steady state fits andthe rate constants determined from global kinetic fits (Fig. S4b, ESI†)of the sensorgrams are listed in Table 1. Based on the equilibriumbinding constants, DGb values at different salt concentrations werecalculated (DGb =�RT ln Keq, where R is 1.987 cal mol�1 K�1 and T is298 K) and are plotted with DHb versus salt concentration in Fig. 3a.It is clear that the DHb is salt independent while the DGb decreaseslinearly as the salt concentration increases. Similar changes in DHb

andDGb with salt concentration have been observed with proteins.3,11

The counterion condensation theory1a,12 predicts that thelogarithm of the equilibrium binding constants (Ka) of KA1039is a linear function of the logarithm of salt concentration and this isobserved in Fig. 3b. The slopes of the linear fits are around one and arequite consistent: �0.99 � 0.02 for kinetic and 0.93 � 0.03 for steady

Fig. 2 Binding enthalpies of (a) 50 mM Net with AAAA and ATAT, 50 mM DB293with AAAA and ATGA, (b) 50 mM, 37.5 mM and 20 mM KA1039 with TGGCTTmeasured by ITC at 25 1C.

Table 1 SPR analysis of kinetic rate constants and equilibrium affinities forKA1039 binding to its cognate site TGGCTTa

[NaCl](mM)

ka

(� 106 M�1 s�1)kd

(� 10�3 s�1)

Kd (nM)

Kinetic fit Steady state

50 7.1 � 0.8 26 � 2 3.7 � 0.4 4.0 � 0.2100 4.5 � 0.9 29 � 3 6.5 � 0.6 6.3 � 0.6200 2.4 � 0.2 32 � 6 13 � 1.1 13 � 0.5300 1.5 � 0.2 33 � 4 22 � 0.3 21 � 1.0

a Kinetic analysis was performed by global fitting of 1 : 1 binding model:ka = (d[AB]/dt)/([A][B]), kd = (�d[AB]/dt)/[AB] and Kd = kd/ka, where [A], [B]and [AB] are the concentrations of the immobilized DNA, PA, andcomplex, respectively. Errors listed are the standard errors for the global fit.

Fig. 3 (a) Plot of DGb and DHb versus salt concentrations for KA1039 at 25 1C.(b) Salt dependence of Ka for KA1039 binding as determined by SPR. TheKa values were obtained by both global kinetic and steady state fits.

Communication ChemComm

Publ

ishe

d on

08

Aug

ust 2

013.

Dow

nloa

ded

by W

ashi

ngto

n U

nive

rsity

in S

t. L

ouis

on

26/0

8/20

13 1

3:12

:29.

View Article Online

This journal is c The Royal Society of Chemistry 2013 Chem. Commun.

state fits. Therefore, approximately one cation has been displaced forthe complex formation. The number of phosphate contacts (Z)between KA1039 and DNA can be determined by the slope/C, whereC is the fraction of phosphate shielded by condensed counterionsand is 0.88 for double stranded B-DNA: Z = 0.96/0.88 = 1.09.1 Thusthere is about one phosphate contact between KA1039 and DNAwhich is reasonable since this PA has a single positive charge (Fig. 1)that can have electrostatic interactions with DNA phosphate groups.The rate constants also depend on salt concentration. As the saltconcentrations increase, the association rate (ka) of KA1039 becomesremarkably slower while the dissociation rate (kd) is slightly fasterand thus the Kd decreases, as expected.1 The salt concentrationdependency of both kinetic constants are calculated and shown inFig. S5 (ESI†). The linear change of the logarithm values [slopes =+0.14 for log(kd) and �0.86 for log(ka)] are as predicted by thecounterion condensation theory for one charge interaction.12

ITC and SPR experiments for Net with AAAA, DB293 with ATGA13

and KA1039 with TGGCTT9 have been conducted as a function oftemperature (Table S1, ESI†), and the thermodynamic profiles areshown in Fig. 4. Minor groove complex formation typically has anegative heat capacity14 and, as expected and in contrast to saltindependency, a significant temperature dependence of DHb

was observed for all three complexes. The linear fits of the DHb

values yield the heat capacities for binding of Net with AAAA(DCp = �168 � 5 cal M�1 K�1), DB293 with ATGA (DCp =�180 � 10 cal M�1 K�1 (ref. 13)) and KA1039 with TGGCTT(DCp = �287 � 10 cal M�1 K�1 (ref. 9)). The DCp values for thesecompounds correlate with their buried surface area as observed withother similar minor groove binders,13,15: Net (monomer) o DB293(stacked dimer) o KA1039 (covalent hairpin dimer). Moreover,a number of heterocyclic diamidines have been investigatedwith AATT binding sites.15 In all cases a negative heat capacity wasobtained with enthalpy–entropy compensation to give a smallchange in DGb. Interestingly, complex formation of DB293-ATGAand KA1039-TGGCTT is strongly driven by enthalpy which isopposite to most of the minor groove binders which have beeninvestigated,16 and their enthalpy–entropy compensation profiles aredistinct from those of A-tracts complexes as generalized by Chaires.16

The presence of GC bps in the binding site alters the hydration

character of the minor groove, and the stacking of heterocycles andnumerous H-bonds between ligands and DNA make minor groovebinding thermodynamics at GC sites more similar to that normallyobserved for intercalators. This is a very important extension of thethermodynamic signatures of DNA minor groove binding.

In conclusion, the fundamental energetic features of DNA minorgroove targeting small molecules as functions of minor groovegeometries, binding modes, DNA sequences, salt concentrationsand temperatures have been determined. The results clearly illus-trate that despite substantial differences in binding sequences andstructures of the minor groove complexes, the DHb for bindingchanges by a negligible amount but is accompanied by large changesin DGb and DSb as a function of salt concentration. Interestingly andconversely, these reactions have significant increases in DHb withtemperature and a negative heat capacity for binding. The tempera-ture-dependent DHb and DSb compensate, however, to give arelatively small change in DGb for binding versus temperature, unlikethe large changes with salt concentration. The effects of salt concen-tration and temperature thus have quite different signatures for DNAminor groove binding and they provide important insights intosmall molecule–DNA complexes and significantly expand our under-standing of the molecular basis of DNA recognition.

The authors thank NIH NIAID AI064200 to W.D.W and D.W.B,NIH NIAID AI083803 to J.K.B for support and the Center forDiagnostics and Therapeutics for a fellowship to S.W. The authorsthank Carol Wilson for manuscript proofreading.

Notes and references1 (a) M. T. Record Jr., C. F. Anderson and T. M. Lohman, Q. Rev. Biophys.,

1978, 11, 103; (b) L. A. Marky, K. S. Blumenfeld, S. Kozlowski andK. J. Breslauer, Biopolymers, 1983, 22, 1247; (c) N. Salim, R. Lamichhane,R. Zhao, T. Banerjee, J. Philip, D. Rueda and A. L. Feig, Biophys. J., 2012,102, 1097.

2 (a) J. B. Chaires, Anti-Cancer Drug Des., 1996, 11, 569; (b) P. L. Privalov,A. I. Dragan, C. Crane-Robinson, K. J. Breslauer, D. P. Remeta andC. A. Minetti, J. Mol. Biol., 2007, 365, 1.

3 P. L. Privalov, A. I. Dragan and C. Crane-Robinson, Nucleic Acids Res.,2011, 39, 2483.

4 (a) P. B. Dervan and R. W. Burli, Curr. Opin. Chem. Biol., 1999, 3, 688;(b) L. Wang, C. Bailly, A. Kumar, D. Ding, M. Bajic, D. W. Boykin andW. D. Wilson, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 12.

5 (a) M. L. Kopka, C. Yoon, D. Goodsell, P. Pjura and R. E. Dickerson,Proc. Natl. Acad. Sci. U. S. A., 1985, 82, 1376; (b) S. Wang, M. Munde,S. Wang and W. D. Wilson, Biochemistry, 2011, 50, 7674.

6 H. S. Koo, H. M. Wu and D. M. Crothers, Nature, 1986, 320, 501.7 Y. Liu, C. J. Collar, A. Kumar, C. E. Stephens, D. W. Boykin and

W. D. Wilson, J. Phys. Chem. B, 2008, 112, 11809.8 (a) H. R. Drew and R. E. Dickerson, J. Mol. Biol., 1981, 151, 535; (b) X. Shui,

L. McFail-Isom, G. G. Hu and L. D. Williams, Biochemistry, 1998, 37, 8541.9 S. Wang, R. Nanjunda, K. Aston, J. K. Bashkin and W. D. Wilson,

Biochemistry, 2012, 51, 9796.10 A. E. Hargrove, J. A. Raskatov, J. L. Meier, D. C. Montgomery and

P. B. Dervan, J. Med. Chem., 2012, 55, 5425.11 A. I. Dragan, C. M. Read, E. N. Makeyeva, E. I. Milgotina, M. E. A. Churchill,

C. Crane-Robinson and P. L. Privalov, J. Mol. Biol., 2004, 343, 371.12 (a) P. L. deHaseth, T. M. Lohman and M. T. Record Jr., Biochemistry,

1977, 16, 4783; (b) W. D. Wilson, C. R. Krishnamoorthy, Y. H. Wangand J. C. Smith, Biopolymers, 1985, 24, 1941; (c) T. M. Lohman,P. L. deHaseth and M. T. Record Jr., Biophys. Chem., 1978, 8, 281.

13 L. Wang, A. Kumar, D. W. Boykin, C. Bailly and W. D. Wilson,J. Mol. Biol., 2002, 317, 361.

14 K. Gallagher and K. Sharp, Biophys. J., 1998, 75, 769.15 (a) B. Nguyen, J. Stanek and W. D. Wilson, Biophys. J., 2006, 90, 1319;

(b) S. Mazur, F. A. Tanious, D. Ding, A. Kumar, D. W. Boykin, I. J.Simpson, S. Neidle and W. D. Wilson, J. Mol. Biol., 2000, 300, 321.

16 J. B. Chaires, Arch. Biochem. Biophys., 2006, 453, 26.

Fig. 4 Thermodynamic profiles of Net-AAAA (black), DB293-ATGA (blue) and KA1039-TGGCTT (red) as a function of temperature. The values of DHb are in triangles, DGb are incircles and TDSb are in diamonds. The thermodynamics of DB293-ATGA are from ref. 13.

ChemComm Communication

Publ

ishe

d on

08

Aug

ust 2

013.

Dow

nloa

ded

by W

ashi

ngto

n U

nive

rsity

in S

t. L

ouis

on

26/0

8/20

13 1

3:12

:29.

View Article Online