J Biol Inorg Chem (2007) 12:819–824 DOI 10.1007/s00775-007-0235-9
ORIGINAL PAPER
DNA affinity binding studies using a fluorescent dye displacement technique: the dichotomy of the binding site Caitriona B. Spillane Æ Jayden A. Smith Æ Joy L. Morgan Æ F. Richard Keene
Received: 7 November 2006 / Accepted: 30 March 2007 / Published online: 8 May 2007 SBIC 2007
Abstract We have observed a number of discrepancies and contradictions in the use of a fluorescent intercalator displacement assay in surveying the binding affinities of dinuclear polypyridyl ruthenium(II) complexes with DNA. By a modification of the assay using the fluorescent minorgroove binder 4¢,6-diamidino-2-phenylindole, rather than intercalating dyes (ethidium bromide or thiazole orange), results were obtained for all complexes studied which were consistent with relative affinities and stereoselectivities observed with other techniques, including NMR, affinity chromatography and equilibrium dialysis. It is believed that the difference in binding mode between the minor groovebinding Ru(II) complexes and the intercalating fluorescent dyes they are displacing may contribute to these discrepancies. Keywords Bulge DNA Dinuclear ruthenium Fluorescence assay 4¢,6-Diamidino-2-phenylindole Binding selectivity
Introduction Control over the processes regulating gene expression is a significant aim of interdisciplinary research encompassing
Electronic supplementary material The online version of this article (doi:10.1007/s00775-007-0235-9) contains supplementary material, which is available to authorized users. C. B. Spillane J. A. Smith J. L. Morgan F. R. Keene (&) School of Pharmacy and Molecular Sciences, James Cook University, Townsville, QLD 4811, Australia e-mail:
[email protected]
the fields of chemistry, molecular biology and medicine [1]. Perhaps the most promising approach to accomplishing this goal is the use of small, cell-permeable molecules that selectively bind DNA sequences and structures of interest, thus enhancing or inhibiting the activity of those proteins that modulate gene expression. Over the past decade, there has been extensive progress in the design of small molecules recognising the minor groove of B-DNA, and it is now apparent that the minor groove is an important carrier of DNA sequence information [2]. Understanding the features that contribute to recognition of DNA by small ligands or metal complexes is crucial for the development of drugs targeted at DNA. Our own area of interest is concerned with synthesising dinuclear ruthenium complexes: these nonintercalating positively-charged metal complexes associate within the minor groove of polyanionic DNA, with binding being further stabilised by a variety of intermolecular forces such as van der Waals, hydrophobic interactions and hydrogen bonding. Our studies have shown that while these bulky dinuclear ruthenium complexes associate with the minor groove they also demonstrate a high affinity for secondary DNA motifs such as bulge sites [3–6] or hairpins [7]. Techniques such as affinity cleavage and footprinting [8, 9] and NMR [10–12] are commonly used to establish the DNA-binding properties of small molecules. However, these are time-consuming and are not applicable to highthroughput screening. A recent method developed by Boger et al. [13] establishes the DNA-binding affinity and sequence selectivity of a library of compounds against an ensemble or complete library of sequences. This fluorescent intercalator displacement (FID) assay is nondestructive and relies on the fluorescence decrease derived from the displacement of DNA-bound ethidium bromide (EthBr) or thiazole orange (TO) by a DNA-binding compound
123
820
J Biol Inorg Chem (2007) 12:819–824
DAPI (95% or better purity, high-performance liquid chromatography grade) was obtained from Fluka. Oligonucleotides were obtained from GeneWorks.
100 mM NaCl, pH 7.7] and concentrations were determined spectrophotometrically using extinction coefficients provided by the supplier (GeneWorks). A DAPI to doublestranded oligonucleotide ratio of 1:1 was found to provide the best resolution between complex stereoisomers for a given oligonucleotide; however, higher ratios (approximately 3:1) better facilitate comparisons between oligonucleotides. Accordingly, in a typical experiment each well of a Costar black flat-bottomed 96-well plate was loaded with the fluorescent groove binder (88 lL of 32 lM DAPI in Tris buffer) and the oligonucleotide (17 lL of 165 lM DNA oligonucleotide concentration in Tris buffer). Ten readings were taken on the plate counter for each cell (DAPI kex = 358 nm, kem = 461 nm) at 25 C, giving the 100% emission reading. The wells were then loaded with the appropriate ruthenium complex solution (11.2 lL, 1 mM in H2O), with the same sample in four of the cells. Accordingly, before the addition of the metal complex the well occupancy was 105 lL [27 lM dye; 27 lM DNA (oligonucleotide)] and after addition of the metal complex the total volume was 116 lL (24 lM dye; 24 lM DNA; 96 lM Ru). A further ten emission readings were recorded for each cell, giving 40 data points for the complex being studied, which were then averaged to give the reported value. Approximately 10% discrepancy between wells was observed, possibly due to surface effects such as bubbles and dust. Readings were taken after incubation for 10 min and the difference in fluorescence for each well in the absence and presence of the metal complex was calculated and averaged for each sample, giving the relative fluorescence displacement as a percentage.
Synthesis of metal complexes
Binding-affinity chromatography [17]
The syntheses of the complexes discussed in this study and the isolation of the stereoisomers have been described elsewhere: [{Ru(Me2bpy)2}2(l-bpm)]4+ (Me2bpy is 4,4¢-dimethyl-2,2¢-bipyridine; bpm is 2,2¢-bipyrimidine), [4] [{Ru(phen)2}2(l-dppm)]4+ [phen is 1,10-phenanthroline; dppm is 4,6-bis(2-pyridyl)pyrimidine], [6] and [{Ru(phen)2}2(l-bb7)]4+ {bb7 is 1,7-bis[4(4¢-methyl-2,2¢bipyridyl)]heptane}. The bridging and terminal ligands are shown in Fig. 1.
A Tricorn column (GE Healthcare; 10 mm · 150 mm) was packed with Streptavidin Sepharose HP (approximately 12 mL) in 20% ethanol. The column was attached to a Gilson Minipuls 3 peristaltic pump and equilibrated with 10 mM Na3PO4/75 mM NaCl/pH 7.5 buffer (approximately 100 mL) solution at a rate of 5 mL min–1. Approximately 2.3 lmol of biotinylated A1-bulge [d(CCGAGAATTCCGG)2] was dissolved in the same buffer solution (5 mL), filtered through 0.45-lm syringe filters and loaded upon the column at a rate of 0.5 mL min–1. The eluate from the DNA-loading step was recycled around the column for approximately 1 h so as to maximise binding. The column was subsequently washed with additional clean buffer solution (100 mL). Once the DNA had been immobilised and washed, approximately 200 nmol of a racemic mixture of [{Ru(phen)2}2(l-bb7)]4+ was loaded upon the column in buffer solution (2 mL) and eluted with more of the same buffer solution at a rate of approximately 1 mL min–1.
[13, 14]. At present, FID has been used to successfully assess the DNA-binding abilities of several organic DNA binders such as distamycin A, netropsin, 4¢,6-diamidino-2phenylindole (DAPI), Hoechst 33258 and berenil [15, 16]. We investigated this procedure to evaluate the DNAbinding affinities of a wide range of dinuclear ruthenium complexes with various oligonucleotides. While our results were generally in agreement with those obtained by other means (NMR spectroscopy, DNA-based affinity chromatography [17] and equilibrium dialysis), there were some notable disagreements and contradictions. We have developed a modification of the fluorescence displacement technique—using a fluorescent groove binder (DAPI) rather than the intercalators (EthBr or TO)—which resolves these conflicts: we now report details of this method.
Materials and methods Instrumentation Fluorescence displacement assays were performed using a PerkinElmer Wallac Victor3 V multilabel plate counter. Circular dichroism (CD) spectra were recorded using a JASCO J-715 spectropolarimeter. Materials
Fluorescence displacement assays The FID assays were performed using techniques adapted from those proposed by Boger et al. [13, 14] and described elsewhere [5–7]. The method was further adapted for use with the fluorescent groove binder, DAPI. The oligonucleotide used for minor-groove displacement studies was rehydrated with buffer [100 mM tris(hydroxymethyl)aminomethane (Tris),
123
J Biol Inorg Chem (2007) 12:819–824
821
bb7 N
N N
N
N
N
N
N
N N
N
N
bpm
N
dppm
N
N
phen
N
Me2bpy
Fig. 1 Structures of bridging and terminal ligands
Separation into two distinct yellow bands was observed within 2–3 cm of travel (taking several hours). Complete elution of the bands from the column took approximately 70 h. The identities of the species eluted were confirmed using CD spectroscopy. Equilibrium dialysis Equilibrium dialysis was conducted at room temperature (22 C) with d(CCGAGAATTCCGG)2 (0.1 mM; 2.5 mL) sealed in dialysis tubing (Sigma) and placed in a solution containing the racemic complex (50 lM; 10 mL). After 24 h the CD spectrum of the dialysate was measured.
Results and discussion In our studies of the relative affinities and stereoselectivities observed in the binding of dinuclear metal complexes with a wide range of oligonucleotides, we have found that the results of FID surveys have generally been consistent with those of other techniques—such as NMR, affinity
chromatography and equilibrium dialysis. However, we have encountered a small number of instances in which the results are not in agreement. A summary of these cases is provided in Table 1. Since the complexes are minor-groove binders rather than intercalators, we believed it possible that the displacement method might benefit from the use of a dye having a similar binding mode to the metal complexes themselves. In our systematic evaluation of the DNA-binding affinities of a wide range of dinuclear ruthenium complexes with various oligonucleotides, the first and most notable inconsistency that became apparent amongst the FID assay results concerned the binding of the enantiomers of the complex [{Ru(Me2bpy)2}2(l-bpm)]4+ to a tridecanucleotide featuring an unpaired adenine base or ‘‘bulge site’’, d(CCGAGAATTCCGG)2 (Fig. 2). In the FID assay, both the DD and the LL enantiomers induced a fluorescence decrease of 29% in the bound EthBr, suggesting that the two possess equivalent binding affinities to this particular oligonucleotide (Fig. 3) [7]. This contradicts the findings of an NMR study in which the stereoisomers of this complex were found to exhibit total enantioselectivity (DD > DL > LL) for this same bulge sequence [4]. The discrepancy in the FID results was confirmed through the use of DNA-assisted affinity chromatography: DD-[{Ru(Me2bpy)2}2(l-bpm)]4+ was found to have a significantly higher affinity to a column coated with the bulge tridecanucleotide than did the analogous LL enantiomer [17]. A second significant anomaly in the FID results was in the binding of meso-[{Ru(phen)2}2(l-dppm)]4+ and DD-[{Ru(phen)2}2(l-dppm)]4+ to the same bulge oligonucleotide. The assay found no discernable difference between the diastereoisomers, both causing a 52% reduction in fluorescence (Fig. 3) [6]. This result contrasted with DNA-based affinity chromatography observations in which the meso isomer was eluted from the column before the DD isomer (Fig. 4), which reflects the relative binding affinities of these diastereoisomers with the oligonucleotide (DD > DL). Accordingly, there is a discrepancy between the apparent binding affinities of the [{Ru(phen)2}2-
Table 1 Comparison of DNA-binding enantioselectivities in the binding of dinuclear ruthenium complexes to a bulge-containing tridecanucleotide as observed using different techniques Complex
NMR
Dialysis
Chromatography
FID (EthBr or TO)
FD (DAPI)
[{Ru(Me2bpy)2}2(l-bpm)]4+
DD > LL [4]
–
DD > LL [17]
DD LL [17]
DD > LL
[{Ru(phen)2}2(l-dppm)]4+
–
–
DD > meso [17]
DD meso [6]
DD > meso
–
DD > LL
DD > LL [17]
LL > DD
DD > LL
[{Ru(phen)2}2(l-bb7)]
4+
See ‘‘Synthesis of metal complexes’’ for an explanation of the complexes DAPI 4¢,6-diamidino-2-phenylindole, EthBr ethidium bromide, FD fluorescence displacement, FID fluorescent intercalator displacement, TO thiazole orange
123
822
J Biol Inorg Chem (2007) 12:819–824 C1 -- C2 -- G3 -- A4 -- G5 -- A6 -- A7 -- T8 -- T9 -- C10 - - - - - C11 - G12 - G13 G13 - G12 - C11 - - - - - C10 --T9 -- T8 -- A7 -- A6 -- G5 -- A4 -- G3 -- C2 -- C1
Fig. 2 A structural representation of the tridecanucleotide containing an adenine bulge (A4)
Fig. 3 Comparison of fluorescent DNA-binding agents with [{Ru(Me2bpy)2}2(l-bpm)]Cl4: (a), [{Ru(phen)2}2(l-dppm)]Cl4 (b), and [{Ru(phen)2}2(l-bb7)]Cl4 (c), The intercalating dye is ethidium bromide for a and b, and is thiazole orange for c. DAPI 4¢,6diamidino-2-phenylindole
(l-dppm)]4+ stereoisomers as seen in the FID and chromatographic experiments. The third and final disagreement within our FID assay was found in the binding of LL-[{Ru(phen)2}2(l-bb7)]4+/ DD-[{Ru(phen)2}2(l-bb7)]4+ and the same bulge oligonucleotide. The FID assay indicated that the LL enantiomer showed the higher binding affinity, causing a 66% reduction in fluorescence in comparison with a 46% decrease recorded for the DD enantiomer (Fig. 3). In contrast, DNAbased affinity chromatography efficiently separated the enantiomers and the CD spectra indicated that the LL enantiomer was eluted first followed by the DD enantiomer (Fig. 5). Equilibrium dialysis was also used to examine the enantioselectivity in the binding of the metal complex to DNA and CD spectroscopy showed an optical enrichment of the less-favoured isomer in the dialysate which was later assigned to the LL-[{Ru(phen)2}2(l-bb7)]4+ (Fig. 6). This result suggests that the DD enantiomer preferentially binds to the bulge oligonucleotide, which again disagrees with the FID assay results. As cautioned by Boger et al. [13], the FID data can only be semiquantitative. The mechanism of displacement of the known intercalators EthBr and TO by the metal complexes used in this study—which are themselves minor-groove binders—is uncertain. A direct competitive displacement would arise when the ruthenium complex binds (though not necessarily by intercalation) at the same site as the fluorescent dye. Although a direct expulsion of the dye by the complex is unlikely, the resultant dynamic equilibria and concentration differences would effectively exclude the dye from the binding site. Such a mechanism could lead to a direct comparison of the relative difference of the two
123
Fig. 4 Circular dichroism (CD) spectra of the first (blue line meso) and second (red line DD) bands collected from a mixture of meso[{Ru(phen)2}2(l-dppm)]4+ and DD-[{Ru(phen)2}2(l-dppm)]4+ separated on a DNA-affinity column loaded with the bulge oligonucleotide, d(CCGAGAATTCCGG)2
association constants. On the other hand, an indirect displacement could occur whereby the ruthenium complex induces a structural change in the deoxyoligonucleotide and so decreases the binding affinity for EthBr or TO. This would lead to a smaller detected change in the FID assay, but does not necessarily indicate that the metal complex interacts with the DNA sequence to a lesser extent; an apparently weaker displacement ratio could arise since a large proportion of the indicator dye remains intercalated in the process despite the additional association with the ruthenium(II) complex. The effect of a bulge sequence in the oligonucleotide is a further uncertainty in these cases —clearly the complexes generally have a greater affinity for the bulge region than for the normal duplex [3, 4]. This selectivity was highlighted by experiments comparing the relative fluorescence decreases between DAPI-bound bulged DNA and the analogous canonical duplex sequence. At higher DAPI-to-DNA ratios (approximately 3:1, to ensure occupancy of the bulge site by the dye) the percentage decrease in fluorescence upon addition of the known bulgebinding complex [{Ru(Me2bpy)2}2(l-bpm)]4+ was found to be significantly greater in the bulged DNA than in the duplex DNA (Fig. S1). In the comparable experiment conducted with EthBr, the fluorescence decreases of the bulged and duplex sequences were reversed (at least in the case of the DD enantiomer) and not nearly as dramatic. While the possibility of multiple potential binding sites and/or modes exists for both DAPI and EthBr when encountering a noncanonical DNA structure, and may
J Biol Inorg Chem (2007) 12:819–824
Fig. 5 CD spectra of the first (blue line LL) and second (red line DD) bands collected from a mixture of DD-[{Ru(phen)2}2(l-bb7)]4+ and LL-[{Ru(phen)2}2(l-bb7)]4+ separated on DNA-affinity column loaded with the bulge oligonucleotide, d(CCGAGAATTCCGG)2
Fig. 6 CD spectra of the racemic mixture (blue line) of DD[{Ru(phen)2}2(l-bb7)]4+ and LL-[{Ru(phen)2}2(l-bb7)]4+ and the LL-enriched dialysate (red line) after dialysis with the A1-bulge oligonucleotide for 24 h
subsequently affect the observed fluorescence response, the use of DAPI yields results that better highlight the known bulge selectivity of these metal complexes. It seems that a possible reason behind the discrepancies described here is the difference between the binding mode of the fluorescent dye (EthBr/TO) and that adopted by the ruthenium complexes. EthBr and TO both intercalate between the base pairs in double-stranded nucleic acids and consequently elongate the duplex, or between bases in single-stranded nucleic acids. In contrast, our metal complexes are minor-groove binders like the dye DAPI (see Fig. S2 for a comparative model of the binding modes of EthBr and DAPI). The minor-groove-binding geometries of the complexes DD-[{Ru(Me2bpy)2}2(l-bpm)]4+ and DD[{Ru(phen)2}2(l-bb7)]4+ with the A1-bulge oligonucleotide are depicted in molecular models (based upon NMR experiment derived constraints) in [3] and in unpublished
823
work (J.L. Morgan, C.B. Spillane, J.A. Smith, D.P. Buck, J.G. Collins and F.R. Keene), respectively. As the analogous NMR experiments have not been performed with DD[{Ru(phen)2}2(l-dppm)]4+ a corresponding model of the complex bound to the A1-bulge is not available; however, the same complex is modelled binding to the minor groove of a three-base bulge in [6]. The relationship between the DNA-binding affinities of these metal complexes and their ability to induce a fluorescence decrease in DNA-bound EthBr is therefore less straightforward than previously imagined. For this reason we decided to revise the FID assay. We hypothesised that by using a fluorescent minor-groove binder instead of an intercalator a more efficient and accurate assay could be developed. DAPI (Fig. 7) is a thoroughly studied example of such a minor-groove-binding dye that preferentially binds at AT-rich regions of B-DNA, a feature commonly observed in metal complexes. Upon binding to DNA, this ligand exhibits an increase in fluorescence intensity, which enables its widespread use as a fluorescent marker for DNA in various contexts [18, 19]. Although the increase in fluorescence by DAPI upon binding to DNA is relatively modest (i.e. 20-fold), an attractive feature of minor-groove binders is their selectivity for double-stranded compared with single-stranded DNA and the fact that they do not cause elongation of the duplex. When DAPI is bound to double-stranded DNA, the excitation and emission maxima of DAPI (kex = 358 nm, kem = 461 nm) [20] are distinct from those of EthBr (kex = 545 nm, kem = 595 nm) [13]. This is advantageous with respect to ruthenium complexes that have competing emissions with EthBr, a number of which we have encountered in past FID assays. These spectroscopic properties, in addition to DAPI’s minorgroove-binding nature, have been exploited in several studies intended to identify the site of interaction—major or minor groove—of DNA-binding polypyridyl ruthenium(II) complexes [20–23]. While it is noted that even with a groove-binding rather than an intercalating dye, there can still be some uncertainty as to the direct or indirect nature of its displacement by the complexes, the results obtained with the minor groove displacement assay proved to be successful, highlighting the expected differences in the binding affinities
Fig. 7 Structure of DAPI
123
824
between the stereoisomers of complexes [{Ru(Me2bpy)2}2(l-bpm)]4+, [{Ru(phen)2}2(l-dppm)]4+ and [{Ru(phen)2}2(l-bb7)]4+. These results are summarised in Fig. 3, which emphasises how the binding mode of the fluorescent dye influences the results obtained with metal complexes and shows the comparisons with the FID technique. When DAPI was used a 15, 33 and 24% difference in emission between the stereoisomers of [{Ru(Me2bpy)2}2(l-bpm)]4+, [{Ru(phen)2}2(l-dppm)]4+ and [{Ru(phen)2}2(l-bb7)]4+, respectively, was recorded. These results coincide with the various methods previously described above. Our groove binder displacement results resolve the discrepancies and contradictions initially observed with the FID assay. We are currently using this technique to conduct a comprehensive survey into the interactions between various metal complexes and oligonucleotides. We believe that these modifications to the FID method improve the accuracy of the technique in assessing the binding affinities and selectivities of minor-groove-binding metal complexes to nucleic acids over the analogous intercalator-based method, owing to the similar binding mode of DAPI. This modification has been found to rectify any discrepancies, improving on an already efficient technique for the appraisal of the complex–DNA interactions of numerous systems in a relatively short period of time. Acknowledgement We gratefully acknowledge the financial support of the Australian Research Council.
References 1. Quivy V, Calomme C, Dekoninck A, Demonte D, Bex F, Lamsoul I, Vanhulle C, Burny A, Lint CV, (2004) Cloning Stem Cells 6:140–149 2. Geierstanger BH, Wemmer DE (1995) Annu Rev Biophys Biomol Struct 24:463–493
123
J Biol Inorg Chem (2007) 12:819–824 3. Patterson BP, Collins JG, Foley FM, Keene FR (2002) J Chem Soc Dalton Trans 4343–4350 4. Smith JA, Collins JG, Patterson BT, Keene FR (2004) Dalton Trans 1277–1283 5. Morgan JL, Buck DP, Turley AG, Collins JG, Keene FR (2005) Inorg Chim Acta 359:888–898 6. Morgan JL, Buck DP, Turley AG, Collins JG, Keene FR (2006) J Biol Inorg Chem 11:824–834 7. Smith JA, Morgan JL, Turley AG, Collins JG, Keene FR (2006) Dalton Trans 3179–3187 8. Taylor JS, Schultz PG, Dervan PB (1984) Tetrahedron 40:457– 465 9. Trauger JW, Dervan PB (2001) Methods Enzymol 340:450–466 10. Lukin M, Santos C (2006) Chem Rev 106:607–686 11. Lee J, Guelev V, Sorey S, Hoffman DW, Iverson BL (2004) J Am Chem Soc 126:14036–14042 12. Wheate NJ, Evison BJ, Herlt AJ, Phillips DR, Collins JG (2003) Dalton Trans 3486–3492 13. Boger DL, Fink BE, Brunette SR, Tse WC, Hedrick MP (2001) J Am Chem Soc 123:5878–5891 14. Boger DL, Tse WC (2001) Bioorg Med Chem 9:2511–2518 15. Lewis MA, Long EC (2006) Bioorg Med Chem 14:3481–3490 16. Goodwin KD, Lewis MA, Tanious FA, Tidwell RR, Wilson WD, Georgiadis MM, Long EC (2006) J Am Chem Soc 128:7846– 7854 17. Smith JA, Keene FR (2006) Chem Commun 2583–2585 18. Kapuscinski J (1995) Biotech Histochem 70:220–233 19. Wilson WD, Tanious FA, Barton HJ, Jones RL, Fox K, Wydra RL, Strekowski L (1990) Biochemistry 29:8452–8461 20. Yun BH, Kim JH, Lee BW, Lincoln P, Norden B, Kim J, Kim SK (2003) J Phys Chem 107:9858–9864 21. Youn MR, Moon SJ, Lee BW, Lee D, Kim J, Kim SK, Lee C (2005) Bull Korean Chem Soc 26:537–542 22. Lee BW, Moon SJ, Youn MR, Kim JH, Jang GH, Kim SK (2003) Biophys J 85:3865–3871 23. Westerlund F, Wilhelmsson LM, Norde´n B, Lincoln P (2005) J Phys Chem B 109:21140–21144
