Báo cáo khoa học: Effect of flanking bases on quadruplex stability and Watson–Crick duplex competition pptx

In the cellular environment, G-rich sequences are flanked by other bases and are present with their complementary strands, leading to a dynamic equilib-rium between quadruplex and duplex

and Watson–Crick duplex competition

Amit Arora, Divya R Nair and Souvik Maiti

Proteomics and Structural Biology Unit, Institute of Genomics and Integrative Biology, Council for Scientific and Industrial Research (CSIR), Mall Road, Delhi, India

G-quadruplexes are unique secondary structures

formed by inter- or intramolecular association of

guan-ine-rich nucleic acid sequences in the presence of metal

ions [1–10] A genome-wide search showed that as

many as 376 000 potential quadruplexes could exist in

the functionally important regions of genes [11] The

biological significance of G-quadruplexes is further

highlighted by their presence in the promoter regions

of the c-myc [12–15], c-kit [16], k-ras [17] and Rb [18]

genes, the immunoglobin switch region [19], insulin

regulatory sequences [20], the fragile X gene [21], the

cystatin B promoter [22], the Hif-1a promoter [23] and

the proximal promoter of the VEGF gene [24] The

possible existence and roles of G-quadruplexes in vivo

have been corroborated by the detection of proteins

that bind specifically to G-quadruplexes and proteins

that have biological activities, such as helicases and nucleases, that are specific for G-quadruplexes [25]

In the cellular environment, G-rich sequences are flanked by other bases and are present with their complementary strands, leading to a dynamic equilib-rium between quadruplex and duplex structures [26] Depending on the cellular requirements, this equi-librium favors either quadruplex or Watson–Crick duplex formation for execution of their respective biological functions Studies have been performed to elucidate the role of various factors in guiding the direction of the equilibrium [27–37] Previous studies have mostly assessed the significance of changes in the intracellular environment in terms of pH, the presence

of cations, temperature and molecular crowding on the quadruplex to duplex transition It has been

Keywords

c-kit; equilibrium; flank length; quadruplex;

Watson–Crick duplex

Correspondence

S Maiti, Proteomics and Structural Biology

Unit, Institute of Genomics and Integrative

Biology, CSIR, Mall Road, Delhi 110 007,

India

Fax: +91 11 2766 7471

Tel: +91 11 2766 6156

E-mail: souvik@igib.res.in

(Received 8 February 2009, revised 5 April

2009, accepted 1 May 2009)

doi:10.1111/j.1742-4658.2009.07082.x

Guanine-rich DNA sequences have the ability to fold into four-stranded structures called G-quadruplexes, and are considered as promising antican-cer targets Although the G-quadruplex structure is composed of quartets and interspersed loops, in the genome it is also flanked on each side by numerous bases The effect of loop length and composition on quadruplex conformation and stability has been well investigated in the past, but the effect of flanking bases on quadruplex stability and Watson–Crick duplex competition has not been addressed We have studied in detail the effect of flanking bases on quadruplex stability and on duplex formation by the G-quadruplex in the presence of complementary strands using the quadru-plex-forming sequence located in the promoter region of the c-kit onco-gene The results obtained from CD, thermal difference spectrum and UV melting demonstrated the effect of flanking bases on quadruplex structure and stability With the increase in flank length, the increase in the more favorable DHvH is accompanied by a striking increase in the unfavorable

DSvH, which resulted in a decrease in the overall DGvH of quadruplex formation Furthermore, CD, fluorescence and isothermal titration calori-metry studies demonstrated that the propensity to attain quadruplex struc-ture decreases with increasing flank length

Abbreviations

ITC, isothermal titration calorimetry; LNA, locked nucleic acid; TDS, thermal difference spectrum.

demonstrated that molecular crowding agents such as

osmolytes significantly affect this transition, as living

cells are crowded with various biomolecules [38,39] It

is apparent that the composition of the base sequences

in the loops between the G-quartets, the loop length

and the base sequences flanking the quadruplex may

also affect the transition between quadruplex and

duplex in the natural environment of biological

sys-tems Recently, Kumar et al [40] demonstrated the

role of a locked nucleic acid (LNA) modified

comple-mentary strand in the quadruplex⁄ Watson–Crick

duplex equilibrium The study indicated that LNA

modifications in the complementary strand shift the

equilibrium toward the duplex state Moreover, it has

also been shown that an increase in loop length favors

duplex formation and competes out the quadruplex

[41] However, to obtain a greater insight into the

dynamics of the equilibrium between the folded motif

and the duplex form, the G-rich sequences must also

be considered within the genomic framework Previous

studies have focused on quadruplex sequences in

isola-tion, but the cellular environment is significantly

dif-ferent In the genome, these unique sequences are

flanked by other sequences that might influence the

stability of these folded motifs and their ability to

compete with the duplex form in the presence of their

complementary strands It thus seems logical to study

the influence of flanking regions on the existence of

quadruplexes, their stability and quadruplex⁄ duplex

competition in the presence of the complementary

strand

As the quadruplex-forming region does not occur

in isolation, and instead is flanked by other

sequences, it is imperative to analyze the effect of

these neighboring sequences on quadruplex stability

and on the duplex⁄ quadruplex equilibrium In the

current study, we have explored the influence of

flanking sequences in the quadruplex-forming region

of the c-kit proto-oncogene promoter [16]

Conforma-tional analysis of preformed quadruplexes with flank

lengths from 0 to 12 was performed using CD and

thermal difference spectrum (TDS) Thermal

denatur-ation⁄ renaturation profiles using UV-visible

spectro-scopy were obtained in order to create a complete

thermodynamic profile for formation of quadruplexes

with different flank lengths Binding parameters and

the thermodynamic profile of preformed quadruplexes

in the presence of the complementary strand were

evaluated by fluorescence and isothermal titration

cal-orimetry (ITC) studies The data obtained in this

study highlight the influence of flanking bases on

quadruplex stability and structural competition

between the G-quadruplex and the duplex

Results and Discussion

To be able to assign a biologically relevant role to quadruplexes, they must be considered in the genomic context and natural cellular environment We have addressed this question in this study because of its wider implication on the practicality of using G-quad-ruplexes as therapeutic targets The telomeric quadru-plex has been well investigated and characterized

in terms of its structural and functional relevance [2,5,42–44] However, this quadruplex, formed by the 3¢ overhang of the telomere, lacks a complementary strand and hence does not suffer competition with the Watson–Crick duplex In addition to the telomeric quadruplex, the G-quadruplex present in the promoter region of the c-myc proto-oncogene has also been well characterized in terms of its structure and function [12–15,45] However, this quadruplex adopts multiple conformations that make structural⁄ biophysical inves-tigations difficult [45] Recently, quadruplex formation has been reported in the promoter region of the c-kit proto-oncogene (87 bp upstream of the transcription start site) [16] The solution structure of this quadru-plex is also well characterized [46–49], and it has also been investigated as an attractive therapeutic target [50] This has generated interest with respect to further biophysical and structural characterization of c-kit quadruplex However, to design effective drugs against quadruplex targets, it is essential to study the role of various factors affecting quadruplex stability and influ-encing the equilibrium between quadruplex and duplex formation Therefore, we have used the c-kit

GG-3¢) as the model sequence for our study (Fig 1)

To analyze the effect of flanking bases on quadruplex

Fig 1 Schematic representation of the 21-mer G-rich sequence located )87 bp upstream of the transcription start site of the c-kit gene The sequences shown to the left of )108 and to the right of )87 are the flanking sequences.

stability and quadruplex⁄ duplex transition, we used

four sequences with 4, 6, 8 or 12 bases on either side

of the 21-base naturally occurring c-kit

quadruplex-forming sequence (Table 1)

The structural topology of the c-kit quadruplex

sequences (c-kitG0, c-kitG4, c-kitG6, c-kitG8 and

c-kitG12) was characterized as parallel or anti-parallel

using CD in the presence of 100 mm KCl, although CD

only provides an indication of the presence of any

secondary structure rather than a confirmation

Figure 2 (black squares) shows the CD spectra obtained

for the various sequences We observed a positive band

at around 262 nm and a negative band near 240 nm,

suggesting the presence of a quadruplex signature

characteristic of the parallel conformation [51] in the

G-rich sequences c-kitG0, c-kitG4 and c-kitG6

(Fig 2A–C, black squares) This observation is in

agree-ment with a reported NMR study on the structural

conformation of the c-kit quadruplex [48] For c-kitG8,

two positive peaks at 265 and 286 nm and a negative

peak at 240 nm were observed (Fig 2D, black squares)

Moreover, unlike the G-rich sequences c-kitG0, c-kitG4

and c-kitG6, a broad positive CD signal ranging from

250 to 290 nm and a negative peak at 233 nm were

observed in the CD spectrum of c-kitG12 (Fig 2E,

black squares) Thus, the CD spectra of c-kitG8 and

c-kitG12 showed the presence of secondary structures

other than quadruplex

TDS complement CD as a tool for the structural

characterization of nucleic acids in solution TDS

pro-vide a simple, inexpensive and rapid method to obtain

structural insight into nucleic acid structures, and may

be used for both DNA and RNA from short oligomers

to polynucleotides [52] Figure 3 shows TDS for

c-kitG0, c-kitG4, c-kitG6, c-kitG8 and c-kitG12

seq-uences The TDS for c-kitG0, c-kitG4, c-kitG6 showed

two positive maxima at 245 and 270 nm, a shoulder at

255 nm, and a negative minimum at 295 nm, thus exhibiting the presence of quadruplex structure How-ever, the TDS for c-kitG8 and c-kitG12 sequences showed loss of both the positive peak at 245 nm and the negative peak at 295 nm that are characteristic of G-quadruplex structure The presence of a positive peak at 270 nm in the TDS of the c-kitG8 and c-kitG12 sequences indicated the presence of a Watson–Crick duplex-like structure as shown in Fig 3 The TDS data presented here are in agreement with previously reported TDS data for G-quadruplexes and GC-rich duplexes [52] The TDS analysis thus supports the exis-tence of G-quartets in c-kitG0, c-kitG4 and c-kitG6 and the absence of Hoogsteen-bonded G-quartets in c-kitG8 and c-kitG12 The presence of non-Hoogsteen-bonded multiple structures in c-kitG8 and c-kitG12 as shown by TDS prompted us to perform UV melting of c-kitG0, c-kitG4, c-kitG6, c-kitG8 and c-kitG12 sequences at 260 nm in 100 mm KCl (Fig S1) The c-kitG8 and c-kitG12 sequences showed considerable hyperchromic effects in the range 10–12%, while c-kitG0, c-kitG4 and c-kitG6 showed only 2–6% hyper-chromicity at 260 nm The presence of 10–12% of hyperchromicity at 260 nm for c-kitG8 and c-kitG12 also resulted from disruption of Watson–Crick base pairing in the secondary structure (Fig S1) mFOLD analysis [53] also indicated the presence of stem-loop structures with Watson–Crick base pairing in the stem region in c-kitG8 and c-kitG12, and thus supported the absence of Hoogsteen-bonded G-quartets (Fig S2) Together, these data clearly indicate that the G-rich c-kit sequence with 8 and 12 flanking bases can adopt a Watson–Crick duplex-like ‘stem-loop’ structure, and thus lose the ability to form prominent quadruplex structure, unlike the c-kitG0, c-kitG4 and c-kitG6 sequences Figure S2 shows the topology of the parallel G-quadruplex formation for c-kitG0, c-kitG4 and

Table 1 Quadruplexes with various flank lengths and their respective complementary strand sequences used in this study G0 and C0 are the core c-kit quadruplex-forming sequence and its complementary strand sequence, respectively G4–G12 and C4–C12 indicate the number

of bases 5¢ and 3¢ to the core c-kit quadruplex-forming sequences and their respective complementary strand sequences.

Oligo name Oligonucleotide sequence (5¢- to 3¢)

Number of flanking bases

c-kitG6 and predicted secondary structures for c-kitG8

and c-kitG12 We wish to highlight that the c-kitG8

sequence can adopt a Watson–Crick duplex-like

‘stem-loop’ structure together with G-quadruplex structure

in KCl buffer The contribution of two different

secondary structure populations is quite evident from

the CD spectrum (Fig 2) and the hypochromic (Fig 4)

and hyperchromic transitions (Fig S1) obtained from

UV melting at 295 and 260 nm, respectively

Our next aim was to determine the effect of

increasing the flank length on the thermodynamic

stability of formation of secondary structures We

have used a spectroscopic method to obtain thermal

denaturation⁄ renaturation profiles to detect G-quartet formation [54] The thermal denaturation⁄ renaturation profiles for c-kitG0, c-kitG4 and c-kitG6 were charac-terized by a clear and reversible transition, such that melting and annealing curves were super-imposable (Fig 4A–C) For c-kitG8, the melting and annealing curves showed considerable hysteresis (Fig 4D), and c-kitG12 showed no clear transition at 295 nm, sug-gesting the absence of stable G-quartets (Fig 4E) The Tm values for the c-kitG0, c-kitG4 and c-kitG6 sequences were calculated as shown in Table 2 The midpoints of the melting transition (Tmelt) and the annealing transition (Tanneal) were also calculated for

A B

E

Fig 2 CD spectra of preformed

quadru-plexes with various flank lengths in the

absence (black squares) and presence

(white squares) of equimolar concentrations

of corresponding complementary strands for

(A) c-kitG0, (B) c-kitG4, (C) c-kitG6, (D)

c-kitG8 and (E) c-kitG12 in 10 m M sodium

cacodylate buffer with 100 m M KCl, pH 7.0,

at 25 C.

c-kitG8, and are also shown in Table 2 The Tm

val-ues for c-kitG0, c-kitG4 and c-kitG6 were 60, 55 and

51C, respectively The Tmelt and Tanneal values for

c-kitG8 were 50 and 47C, respectively, and were

accompanied by considerable hysteresis No

melt-ing⁄ annealing transition was observed for the

c-kitG12 sequence at 295 nm This observation

sug-gested that increasing the flank length led to a

decrease in the Tmvalue that reflects the reduced

ther-mal stability (Table 2) The various thermodynamic

parameters are summarized in Table 2 The

thermody-namic parameters DGvH, DHvH and DSvH were not

determined for c-kitG8 and c-kitG12 due to the

hys-teresis in c-kitG8 and the lack of a clear transition at

295 nm for c-kitG12 DHvH increases with the increase

in flank length from 0 to 6 bases This increase in the

enthalpy change (DHvH) may be due to the increase

in the base-stacking interaction among the flanking

bases A striking observation was the high increase in

the unfavorable negative entropy change DSvH, which

resulted in a decrease in the overall free energy

change (DGvH) The decrease in entropy upon increase

in flank length arises due to the intra-residue stacking

interaction in the flank bases Furthermore, we also

performed concentration-dependent melting of all the

sequences to deduce the molecularity of the structures

The sequences formed intramolecular structures as

suggested by the concentration-independent thermal

stability (Tm) (data not shown) Overall, the results

indicated that quadruplex formation becomes less

favorable with the increase in flank length on each

side of the core c-kit quadruplex sequence (Table 2)

Fig 3 Thermal difference spectrum of c-kitG0 (black squares),

c-kitG4 (open squares), c-kitG6 (open circles), c-kitG8 (open

triangles) and c-kitG12 (open diamonds) resulting from subtraction

of the spectrum obtained at 25 C from that obtained at 90 C in

10 m M sodium caodylate buffer with 100 m M KCl, pH 7.0.

A

B

C

D

E

Fig 4 UV melting (open triangles) and annealing (open diamonds) profiles of preformed c-kit quadruplex with various flank lengths: (A) c-kitG0, (B) c-kitG4, (C) c-kitG6, (D) c-kitG8 and (E) c-kitG12 in

10 m M sodium cacodylate buffer, pH 7.0, with 100 m M KCl.

As the quadruplex-forming region does not occur in

isolation, but instead is flanked by other sequences

and is present together with its complementary strand,

it is imperative to analyze the effect of these

neighbor-ing sequences on the quadruplex⁄ duplex equilibrium

also

In order to assess the influence of flanking bases on

the quadruplex⁄ duplex equilibrium, we investigated the

CD spectral changes on addition of corresponding

complementary strand to preformed quadruplex in

KCl buffer The CD spectra are shown in Fig 2 (white

squares) For c-kitG0, c-kitG4 and c-kitG6, CD

spec-tra recorded for equimolar concenspec-trations of

quadru-plex and its respective complementary strand showed a

positive peak at 270 nm coupled with an increase in

the intensity of bands at 240 nm (Fig 2A–C, white

squares) These spectral features are characteristic of

the B-DNA form, and suggest the formation of duplex

when complementary strands are added to a

quadru-plex However, they do not confirm complete duplex

formation for equimolar mixtures of preformed

quad-ruplex and its complementary strand in 100 mm KCl

buffer For the c-kitG8⁄ C8 system, a small positive

peak at 270 nm together with a more intense positive

band at 286 nm coupled with an increase in the

inten-sity of the negative band at 240 nm was observed

(Fig 2D, white squares) A broad positive CD band at

286 nm together with negative CD band at 240 nm for

c-kitG12⁄ C12 indicates that there is no change in the

CD spectrum of c-kitG12 when incubated with its

complementary strand (c-kitC12) in 1 : 1 ratio

(Fig 2E, white squares) These observations for

c-kitG8⁄ C8 and c-kitG12 ⁄ C12 can be ascribed to the

presence of multiple secondary structures in both the

G-rich as well as in the C-rich complementary strands

as discussed below

Next, to assess the competition between the

quadru-plex and duquadru-plex forms under the influence of

increas-ing flank lengths, fluorescence bindincreas-ing experiments

were performed using a donor⁄ quencher pair of 5¢-fluorescein (donor) and 3¢-dabsyl chloride (quencher) This technique was chosen because it offers the advantage of working in the nanomolar range, which is not possible with UV or CD FRET-based studies have also been used effectively to understand quadruplex structures [37,38,55–57] Complementary strands of respective flank lengths were used for hybridization with fluorophore-labeled sequences We investigated the binding affinity of complementary strands to preformed quadruplexes with various flank lengths of 0–12 bases in KCl We observed enhance-ment of fluorescence intensity on increasing the con-centration of the complementary strand, indicative of a greater extent of quadruplex opening The normalized relative changes in fluorescence intensity were plotted against the complementary strand concentration (Fig 5), and the binding affinities were calculated by fitting the plots using Eqn (8), as described in Experi-mental procedures The estimated binding affinities are summarized in Table 3 The binding affinity value for the complementary strands towards the preformed G-quadruplexes with various flank lengths increased with the increase in the flank length from c-kitG0 to c-kitG6 (Table 3) The KA value obtained for c-kitG8⁄ C8 was same as that for c-kitG6 ⁄ C6, and was decreased for c-kitG12⁄ C12 (Table 3)

To complement the fluorescence studies, ITC experi-ments were performed to obtain the complete thermo-dynamic profile for quadruplex hybridization to its complementary strand The hybridization event was dependent on nearest-neighbor Watson–Crick base pairing Figure 6 shows characteristic sigmoidal curves obtained for heat of injection for hybridization of preformed quadruplex to its complementary strand Table 4 summarizes the thermodynamic parameters for the duplex formation obtained from ITC experiments The heat of injection profile for duplex formation is exothermic The magnitude of negative DHITC reflects

Table 2 Thermodynamic parameters obtained from UV experiments performed in 10 m M sodium cacodylate buffer, 100 m M KCl at pH 7.0 and

25 C T m is the melting temperature DH vH is the enthalpy change and DS vH is the entropy change for G-quadruplex formation DG vH is the free energy change for G-quadruplex formation All parameters were calculated as described in Experimental procedures All the parameters obtained were within 10% error Tmvalues differed by ± 1.0 C ND indicates that values were not determined for c-kitG8 and c-kitG12.

Quadruplex

Tm(C)

DH vH (kcalÆmol)1)

DS vH (calÆmol)1ÆK)1)

DG vH (kcalÆmol)1)

the binding enthalpy for duplex formation, which

increased with the increase in the number of flanking

bases from 0 to 6, and deviation from the increasing

DHITC trend was observed for c-kitG8 and c-kitG12,

as shown in Table 4 The DH values obtained from

ITC experiments were much lower than the expected

value for duplex formation in all cases, i.e )166.30 kcalÆmol)1for c-kitG0⁄ C0,)232.70 kcalÆmol)1

for c-kitG4⁄ C4, )268.50 kcalÆmol)1 for c-kitG6⁄ C6, )309.40 kcalÆmol)1 for c-kitG8⁄ C8 and )382.20 kcalÆ mol)1 for c-kitG12⁄ C12, obtained using the nearest-neighbor method [58] The enthalpy change in this process involves endothermic and exothermic contribu-tions from opening up of the preformed quadruplex and hybridization between G- and C-rich strands, respectively The overall enthalpy change is the sum of the contribution from each process, leading to a lower

DHITCvalue than calculated using the nearest-neighbor method The DSITC for duplex formation decreased with the increase in the number of flanking bases from

0 to 6 However, there was deviation from the decreas-ing DSITCvalues for c-kitG8 and c-kitG12 as shown in Table 4 A detailed inspection of Table 4 reveals that the DHITC values as well as the DGITC values for c-kitG8⁄ C8 and c-kitG12 ⁄ C12 deviate from the increas-ing trend as observed for c-kitG0⁄ C0, c-kitG4 ⁄ C4 and c-kitG6⁄ C6 As shown by TDS, UV hyperchromic tran-sition at 260 nm and mFOLD analysis, the G-rich strands of c-kitG8 and c-kitG12 can adopt intra-molecular stem-loop structure with Watson–Crick base pairing in the stem region Likewise, the C-rich comple-mentary strand can also form such stem-loop structures Moreover, the C-rich complementary strand also has the potential to form a secondary structure called an i-motif in all the sequences ranging from c-kitC0 to c-kitC12 at near physiological pH 7.0 [59], although these structures would be less stable, at physiological

pH as compared to acidic pH, as it has been shown that intercalated hemiprotonated C:C+base pairs are stable

at acidic pH [60,61,62] To understand the structure

Table 3 Binding affinities (KA) of quadruplex with complementary

strands obtained from fluorescence studies in 10 m M sodium

caco-dylate buffer with 100 m M KCl, pH 7.0 at 25 C The quadruplex

concentration used was 50 n M and the respective complementary

strand concentration varied from 0 to 1 l M The values obtained

were within 10% error.

Fig 6 ITC binding profile of equimolar mixtures of c-kitG0 (black square), c-kitG4 (open square), c-kitG6 (open circle), c-kitG8 (open triangle) and c-kitG12 (open diamond) preformed quadruplex sequences with corresponding complementary strands in 10 m M

sodium caodylate buffer with 100 m M KCl, pH 7.0, at 25 C.

Fig 5 Plots of normalized relative fluorescence emission intensity

(DF, as described in the text) of quadruplex (30 n M ) at 520 nm

ver-sus complementary strand concentration in 10 m M sodium

caody-late buffer with 100 m M KCl, pH 7.0, at 25 C The complementary

strands used were c-kitC0 (black square), c-kitC4 (open square),

c-kitC6 (open circle), c-kitC8 (open triangle) and c-kitC12 (open

diamond).

adopted by c-kit C-rich strands, we performed CD

stud-ies (Fig S3) and UV melting studstud-ies at 287 and 265 nm

(Fig S4) at both pH 6.0 and 7.0 Both CD and UV

melting studies showed that the C-rich strands c-kitC0,

c-kitC4 and c-kitC6 adopt i-motif structure at pH 6.0,

but no such structure formation takes place at pH 7.0

Moreover, c-kitC8 and c-kitC12 did not show the

i-motif structure signature; instead, the UV melting

profiles at 265 nm for c-kitC8 and c-kitC12 showed

 10% hyperchromicity, thus indicating the presence of

an intramolecular stem–loop structure at both pH 6.0

and 7.0 (Fig S4) The sum of the CD spectra for both

the G- and C-rich individual strands (Fig S5) was also

found to be similar to the CD spectra of mixtures of

both the strands at pH 7.0 as shown in Fig 2

Furthermore, to understand the contribution of

i-motif structures in quadruplex⁄ Watson–Crick duplex

competition, we also performed ITC titrations at pH 6.0

(Fig S6), and data are presented in Table S1 ITC

experiments at pH 6.0 showed that the binding affinity

of c-kitC4 and c-kitC6 strands to their respective G-rich

strands decreases almost by one order of magnitude

(Fig S6 and Table S1) compared to the affinity at pH

7.0 (Table 4), but G0 sequences remained unopened in

the presence of respective complementary c-kitC0

strands (Fig S6) However, ITC titration data for the

c-kitG8⁄ C8 and c-kitG12 ⁄ C12 system remain unaffected

and similar at both pH 6.0 and 7.0 (Fig S6 and

Table S1) This rules out the possibility of a significant

contribution of i-motif structures in c-kit C-rich strands

to quadruplex⁄ Watson–Crick duplex competition at

physiological pH 7.0

Together, the results obtained from CD, fluorescence

and ITC titration experiments in the c-kitG8⁄ C8 and

c-kitG12⁄ C12 systems demonstrate that structural

com-petition is imposed by the intramolecular stem–loop

structure in C-rich complementary strand, thus affecting

the quadruplex⁄ Watson–Crick duplex equilibria at both

pH 7.0 and the near physiological pH of 6.0 The pres-ence of intramolecular stem–loop structures in the C-rich complementary strand leads to the existence of competition and thus hinders opening of the secondary structures in c-kitG8 and c-kitG12 G-rich sequences These observations indicate that the increase in flank length from 0 to 6 on each side of the c-kit core quadru-plex-forming sequence drives efficient invasion and bet-ter conversion of quadruplex to duplex, and competes out quadruplex in this structural competitive equilib-rium However, this is not the case for flank lengths

of 8 and 12 due to the presence of intramolecular stem–loop structures in both the G-rich and C-rich strands

The parameter that highlights the predominance of either of population (duplex or quadruplex) is the rela-tive free energy difference, DDG25C, between duplex and quadruplex structures In this study, we have obtained thermodynamic profiles of quadruplexes by

UV melting experiments However, it was difficult to obtain the thermodynamic parameters involved in duplex formation from the same sequences and their respective complementary strands by UV melting stud-ies, as this includes contributions from both duplex and quadruplex Therefore, we obtained the thermo-dynamic profile for duplexes from ITC experiments (Table 4) The relative free energy difference, DDG25C, between duplex and quadruplex structure increase from)3.3 to )5.6 kcalÆmol)1upon an increase in flank length from 0 to 6 It is noteworthy that DDG25C val-ues are reasonably negative in all cases, indicating that duplex is the predominant structure We also observed

an increase in duplex stability upon an increase in flank length (Table 4) The greater the negative magni-tude of DDG25C, the higher is the predominance of duplex at equilibrium

Table 4 Thermodynamic parameters obtained from ITC experiments performed in 10 m M sodium cacodylate buffer, pH 7.0, 100 m M KCl at

25 C Thermodynamic parameters were obtained for complementary strand binding to the preformed quadruplexes at 25 C The quadru-plex concentration in the cell was 5–10 l M and the complementary strand concentration in the syringe was 100–250 l M N is the stoichiom-etry of complementary strand binding to preformed quadruplex DHITCis the enthalpy change and DSITCis the entropy change for duplex formation DG ITC is the free energy change for duplex formation and was determined using the relationship DG = )RT ln K A , where R is the universal gas constant, T is the temperature in Kelvin (K), and KAis the binding affinity for duplex formation All the parameters obtained were within 10% error.

KA (106M )1)

DHITC (kcalÆmol)1)

DSITC (calÆmol)1ÆK)1)

DGITC (kcalÆmol)1)

Based on a search algorithm designed by Huppert

and Balasubramanian [11], it has been predicted that,

in principle, as many as 376 000 quadruplexes could

exist in the human genome However, our study

suggests a lower likelihood of quadruplex formation

at all these sites, as the presence of flanking bases

on each side of the c-kit core quadruplex sequence

destabilizes the quadruplex structure Furthermore,

an increase in the number of flanking bases leads to

the existence of alternative structures other than

quadruplexes In the genome, these sites will have

many more bases flanking them than used in our

studies, casting doubt on the global presence of

quadruplex structures with a general role in the

bio-logical system On the other hand, several studies

have indicated the existence of quadruplexes in vivo

[63] and their ability to regulate gene expression [12–

14] Their significant role in the telomeric region has

also been well established [2,5,42–44] Further

indica-tions of the presence of quadruplexes in the living

system come from the fact that cells contain factors

that actively cleave and unwind G4 DNA [25] It

thus seems apparent that cells may have some

mech-anism(s) that favors formation of either quadruplexes

or duplexes according to their biological relevance,

thus suggesting the importance of the

quadru-plex⁄ duplex equilibrium in modulating biological

activities

Conclusion

In the present study, we have explored the effect of

flanking sequences on quadruplex stability and

quadru-plex⁄ duplex competition in order to understand the

likely scenario in the cell, where quadruplex sites have

additional sequences at both their ends The study

shows that the presence of flanking bases affects the

thermodynamic stability of the G-quadruplex With an

increase in the flank length, the increase in the more

favorable negative enthalpy change (DHvH) is

accom-panied by an increase in the unfavorable negative

entropy change (DSvH), resulting in a decrease in the

overall free energy change (DGvH) The study also

shows that, with the increase in the number of flanking

bases, there is an increased propensity for the existence

of other alternative structures that may compete with

G-quadruplex formation Our work shows that the

presence of flanks destabilizes the G-quadruplex

struc-ture and drives the equilibrium towards duplex

forma-tion If this is indeed the case, the probability of the

existence of these structures as global regulatory motif

in the genome, prima facie, appears to be

context-dependent

Experimental procedures

Materials Oligonucleotides were obtained from Microsynth (Balgach, Switzerland) The sequences of the oligonucleotides used

in these studies are given in Table 1 c-kitG0, c-kitG4, c-kitG6, c-kitG8 and c-kitG12 represent the c-kit quadru-plex sequence with various flank lengths, and c-kitC0, c-kitC4, c-kitC6, c-kitC8 and c-kitC12 represent their respective complementary strands (Table 1) All the sequences containing core quadruplex-forming motifs with varying flank lengths used in our study were labeled using the fluorophores 5¢-fluorescein and 3¢-dabsyl chloride The concentrations of unlabeled oligonucleotide solutions were determined based on the absorbance at 260 nm and 80C using molar extinction coefficients of 213, 290, 321, 354 and 419 mm)1Æcm)1 for c-kitG0, c-kitG4, c-kitG6, c-kitG8 and c-kitG12, respectively, and 164, 234, 275, 308 and

382 mm)1Æcm)1 for c-kitC0, c-kitC4, c-kitC6, c-kitC8 and c-kitC12, respectively These values were calculated by extrapolation of tabulated values for the dimers and monomer bases at 25–80C using procedures described previously [64,65] Concentrations of the labeled oligonu-cleotide were determined by measuring the absorbance of the attached fluorescein moiety at 496 nm using a molar extinction coefficient of 4.1· 104m)1Æcm)1 [66] In all studies, we used preformed quadruplexes obtained by heating solutions containing G-rich sequences in 100 mm KCl to 100C for 5 min and gradually cooling to room temperature at the rate of 0.2CÆmin)1, and then kept for

7 days at 4C prior to experimentation

Circular dichroism spectroscopy

CD spectra were measured using a Jasco model J-715 spec-tropolarimeter (Jasco, Tokyo, Japan) equipped with a ther-moelectrically controlled cell holder and a cuvette with a path length of 1 cm Scans were performed over a range of 200–350 nm in 10 mm sodium cacodylate buffer (pH 7.0) with 100 mm KCl at 25C Preformed G-quadruplexes with various flank lengths were incubated with equimolar concen-trations of respective flank length at 25C for 24 h prior to

CD experiments The spectra of preformed quadruplexes at a concentration of 7.5 lm in the absence and presence of equi-molar concentrations of the complementary strand were obtained A buffer baseline spectrum was obtained using the same cuvette and subtracted from sample spectra

Thermal difference spectrum For each oligonucleotide sample, an UV spectrum was recorded above and below its melting temperature (Tm) The difference between the UV spectrum at high temperature

(95C) and the UV spectrum at low temperature (25 C) is

defined as the TDS, and represents the spectral difference

between the unfolded and the folded form The TDS were

normalized, using a value of +1 for the highest positive

peak

Thermal denaturation/renaturation using

UV-visible spectroscopy

Oligonucleotides were dissolved in 10 mm sodium cacodylate

buffer pH 7.0 with 100 mm KCl at final concentrations

rang-ing from 2 to 10 lm, dependrang-ing on the oligonucleotide

length Samples (1 ml) were placed in a stoppered quartz

cuvette of 1 cm path length, and then thermal

denatur-ation⁄ renaturation was performed using a Cary 100 UV ⁄

vis-ible spectrophotometer (Varian, Walnut Creek, CA, USA)

equipped with a Peltier effect heated cuvette holder A

tem-perature range of 25–95C was used to monitor the

absor-bance at 295 nm at a heating⁄ cooling rate of 0.2 CÆmin)1

The absorbance profiles recorded at 295 nm were analyzed

using a non-linear least-squares curve-fitting method This

method involved contributions from pre- and post-transition

baselines, and thermodynamic data were obtained using

equations described previously [67,68] The analysis was

performed using mathematica 5.1 (Wolfram Research,

Champaign, IL, USA) and origin 7.0 (Microcal Inc.,

Northampton, MA, USA)

The following equations were used to calculate the

thermodynamic data:

Au¼ buþ mð u TÞ ð1Þ

Al¼ blþ mð l TÞ ð2Þ

Keq¼ð1  aÞ

AðTÞ ¼ aðAu AlÞ þ Al ð4Þ

Keq¼ exp DG

o

RT

¼ exp DH

o

RT þ

DSo

R

ð5Þ

Equations 1 and 2 are linear equations where Au and Al

are terms describing upper and lower baselines, respectively,

bu and bl are fitted parameters for the intercepts for the

upper and lower baseline, and muand mlare the respective

slopes Keq is the equilibrium constant for the

unstruc-tured⁄ structured transition for an intramolecular system,

and a is the folded fraction A(T) is the dependent variable

and is the experimentally determined absorbance at each

temperature (T) Using these equations, the van’t Hoff

enthalpy (DHvH) and entropy (DSvH) were calculated, and

Tm was calculated from the peak value of the first deriva-tive of the fitted curve Tm values differed by ± 1.0C The Gibbs free energy (DGvH) was calculated at 25C using the equation DGvH= DHvH) TDSvH, assuming DCp = 0 Fluorescence studies

A FLUOstar OPTIMA fluorescence plate reader (BMG Lab technologies, Melbourne, Australia) was used to deter-mine the binding affinities of fluorophore-labeled c-kitG0, c-kitG4, c-kitG6, c-kitG8 and c-kitG12 to their respective complementary strands (sequences given in Table 1) in the presence of 100 mm KCl The plate reader makes it possi-ble to work on systems that suffer from thermodynamic and kinetic inertia, thus requiring prolonged incubation, and enables study of many samples at dilute concentration [38] The experiments were performed in 384-well plates, using 480 nm excitation and 520 nm emission filters The wells were loaded with solutions of a fixed concentration of preformed quadruplex (50 nm) and increasing concentra-tions of complementary strand (0–1 lm) Sample mixtures were incubated for 24 h at 25C, and the plate was read

at 520 nm For analysis of data, the observed fluorescence intensity was considered as the sum of the weighted contri-butions from folded G-quadruplex strand and extended G-strand in the duplex form:

F¼ 1  að bÞF0þ abFb ð6Þ where F is the observed fluorescence intensity at each titrant concentration, F0and Fbare the fluorescence intensi-ties of the initial and final states of titration, respectively, and ab is the mole fraction of quadruplex in duplex form Assuming 1 : 1 stoichiometry for the interaction involving complementary strand binding, it can be shown that:

Q

½ 0a2b Q½ 0þ C½  þ 1=KA

abþ C½  ¼ 0 ð7Þ where KA is the association constant, [Q]0 is the total G-strand concentration, and [C] is the complementary strand concentration

From Eqns (6) and (7), it can be shown that:

DF¼ DFð max=2 Q½ 0Þ ½ Q0þ C½  þ 1=KA





ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Q

½ 0þ C½  þ 1=KA

4 Q½ 0½ C

where DF = F) F0and DFmax= Fmax) F0 Isothermal titration calorimetry experiment The ITC experiment was performed using a Microcal VP-ITC titration calorimeter The 300 ll syringe was filled with 146 lm of complementary strand Titration was per-formed by injecting 10 ll aliquots of complementary strand