Interaction of DAPI with individual strands of trinucleotide repeats Effects on replication in vitro of the AATÆATT triplet Edoardo Trotta 1 , Nicoletta Del Grosso 1 , Maura Erba 2 , Sonia Melino 2 , Daniel Cicero 2,3,4 and Maurizio Paci 2,3 1 Istituto di Neurobiologia e Medicina Molecolare, Consiglio Nazionale delle Ricerche, Roma, Italy; 2 Dipartimento di Scienze e Tecnologie Chimiche, Universita ` di Roma ‘Tor Vergata’, Roma, Italy; 3 INFM sez. B, Roma, Italy; 4 SISSA, Trieste, Italy The structural changes produced by the minor-groove binding ligand DAPI (4¢,6-diamidine-2-phenylindole) on individual strands of trinucleotide repeat sequences were detected by electrophoretic band-shift analysis and related to their effects on DNA replication in vitro. Among the 20 possible single-stranded trinucleotide repeats, only the T-rich strand of the AATÆATT triplet exhibits an observable fluorescence band and a change in electrophoretic mobility due to the drug binding. This is attributable to the property of DAPI that favours folding of the random coil ATT strand into a fast-migrating hairpin structure by a minor-groove binding mechanism. Electrophoretic characteristics of AAT, ACT, AGT, ATG and ATC are unchanged by DAPI, suggesting the crucial role of TÆT with respect to AÆA, CÆC and GÆG mismatch, in favouring the binding properties and the structural features of the ATT–DAPI complexes. Primer extension experiments, using the Klenow fragment of DNA polymerase I, demonstrate that such a selective structural change at ATT targets presents a marked property to stall DNA replication in vitro in comparison with the comple- mentary AAT and a random GC-rich sequence. The results suggest a novel molecular mechanism of action of the DNA minor-groove binding ligand DAPI. Keywords: DAPI; trinucleotide repeats; Klenow; DNA structure; hairpin. DAPI (4¢,6-diamidino-2-phenylindole) is a DNA minor- groove binding ligand used largely as a fluorescent dye for DNA and chromosomes [1]. It interferes with the activity of DNA processing enzymes involved in regulatory and structural functions such as DNA polymerase I [2], RNA polymerases [2–4], topoisomerases [2,5–7], DNA ligase [2], exonuclease III [2], DNAase I [8] and restriction endonuc- leases [8], showing varying levels of inhibitory effects [2]. DAPI preferentially binds into the minor-groove of at least two consecutive AÆT base-pairs [9,10] or TÆTmismatches flanked by AÆT base-pairs [11]. Quite different binding mechanisms and a lower affinity with DNA sequences that contain no adjacent AÆT base-pairs have been reported: intercalation [12,13], major-groove binding [14] and p,p- stacking interactions with double-helix ends [15]. It has also been shown that DAPI favours folding into hairpin structures of the T-rich strand of AATÆATT trinucleotide repeat sequences [16]. In this work we investigate the effects of DAPI on replication in vitro of AATÆATT trinucleotide repeats in relation to the structural changes induced by the drug on individual strands of trinucleotide repeats. The AATÆATT triplet presents a number of biological features that require further investigation to understand the possible role of this frequent class of trinucleotide repeat. This triplet has been reported as the most abundant and polymorphic trinucleotide repeat in the human genome [17]. Distribution of the AATÆATT trinucleotide repeat does not appear to be random in relation to different types of genomic sequences, but it is frequent in introns and uncommon in exons, suggesting a possible role in regulating transcription [18–20]. In contrast to the CG-rich triplet repeats associated with human diseases, AATÆATT did not show any pausing of DNA polymerases in the primer extension assay [21] and has not yet been found associated with human diseases, although its expansion was observed during DNA replica- tion in vitro [22,23]. Also unlike other trinucleotide repeats, the AATÆATT repeat in plasmids shows the unusual propensity to adopt nonhydrogen-bonded structures, sug- gesting that it may play a different role in gene regulation [21]. The AATÆATT repeat also constitutes a binding site for at least one nuclear protein [24]. The present study demonstrates that DAPI causes a sequence- and strand-dependent stalling of DNA poly- merase at AATÆATT trinucleotide repeat regions. This effect is associated with the strongly selective property of DAPI to bind and favour the hairpin structure of the T-rich strand of the AATÆATT trinucleotide repeat. Our results suggest a novel mechanism of action of DAPI and show the unusual replication feature of a common class of genomic DNA that have unknown functions. Materials and methods Materials DNA 18-mers: (AAC) 6 ,(AAG) 6 ,(AAT) 6 , (ACC) 6 , (ACG) 6 ,(ACT) 6 ,(AGC) 6 , (AGG) 6 ,(AGT) 6 ,(ATC) 6 , (ATG) 6 , (ATT) 6 , (CCG) 6 ,(CCT) 6 ,(CGG) 6 ,(CGT) 6 , Correspondence to E. Trotta, Istituto di Neurobiologia e Medicina Molecolare, Consiglio Nazionale delle Ricerche, Via del Fosso del Cavaliere 100, 00133 Roma, Italy. Fax: + 39 064 993 4257, Tel.: + 39 064 993 4567, E-mail: edoardo.trotta@ims.rm.cnr.it Abbreviation:DAPI,4¢,6-diamidino-2-phenylindole. (Received 5 September 2003, revised 10 October 2003, accepted 14 October 2003) Eur. J. Biochem. 270, 4755–4761 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03877.x (CTG) 6 , (CTT) 6 ,(GGT) 6 and (GTT) 6 were used for evaluating electrophoretic band-shifts induced by DAPI on trinucleotide repeat sequences. Three different 70-base templates for primer extension experiments were synthesized with a common 40-base sequence at the 3¢-end constituted by a random region with no adjacent AÆTbases:5¢-GGTCCCAGTCCACTCTGT CGCCGCACCCTGCGGACCTGCT-3¢ (Fig. 1A,B). The remaining 30-base sequence at the 5¢-end was distinctive for each template: (ATT) 10 ,(AAT) 10 and 5¢-CAGGTGGA GCTGTGTCAGTGCCACTGACCC-3¢ for ATT-, AAT- and random-template, respectively (Fig. 1B). A 5¢-fluoresc- ein labeled 15-base primer was synthesized that was complementary to the 41–55 template region. DAPI (Fig. 1C) and Klenow fragment [3¢fi5¢ exonuc- lease-deficient (exo-) of Escherichia coli DNA polymerase I] were purchased from Sigma (St. Louis, MO, USA) and New England Biolabs Inc. (Beverly, MA, USA), respectively. Electrophoretic mobility-shift analysis DAPI-induced electrophoretic mobility-shift analyses were performed on 12% (w/v) native polyacrylamide gel at 4 °C. Before gel inoculation, DNA samples were denatured at 90 °C and slowly annealed to room temperature. The DAPI–DNA complexes were prepared by adding the drug after DNA annealing. Polyacrylamide gels were stained with 0.1% (v/v) methylene blue in 0.5 M sodium acetate. Samples for comparing binding effects of DAPI on the 18-mer trinucleotide repeat sequences were 150 l M DAPI and 35 l M DNA strand in Tris/borate/EDTA buffer. Primer–template DNA samples were 50 l M DAPI and 8 l M primer–template in Tris/borate/EDTA buffer. Fluorescence Fluorescence spectra were recorded on a SPEX FluoroMax photon counting spectrofluorometer equipped with a ther- mostatic cell holder (SPEX Industries Inc., Edison, NJ, USA). Excitation (350 nm) and emission (448 nm) with 1.5 nm band-pass were used and spectra were corrected for background signal. The absorbance at the excitation wave- length was less than 0.03 ([DAPI] ¼ 1 l M )makingthe inner filter effects negligible. The experiments were per- formed at 37 °C, in 100 m M NaCl and 10 m M phosphate buffer at pH 7.00, by adding an increasing amount of DNA oligomers to the drug solutions up to 2 : 1 (DAPI : tem- plate) molar ratio. UV melting experiments Optical melting experiments were performed on a PC controlled PerkinElmer Lambda Bio20 double beam spec- trophotometer equipped with a programmable Peltier temperature-controlled cell holder. Quartz cuvettes of 1 cm path length were used and closed with a Teflon cap incorporating the temperature sensor. A layer of paraffin was also placed on top of the sample solutions to prevent solvent evaporation. Absorbance of each DNA sample was  0.25 per mL, in 10 m M sodium phosphate (pH 7.00) and 100 m M NaCl. Complexes were prepared at a base to drug molar ratio of 3.5. Before thermal melting analysis, samples were heated to 80 °C for 5 min and then slowly cooled to the starting temperature of 3 °C. Melting profiles were acquired by heating samples from 3 °Cto85°Catarateof 1 °CÆmin )1 and measuring absorbance corresponding to the maximum around 260 nm at 25 °C. To avoid condensation of water vapor at low temperatures, the cuvette-holding chamber was dried by flushing with N 2 gas. The melting temperature (Tm) was estimated from the maximum of the first-derivative curve of the absorbance vs. temperature. Primer extension experiments Chain extension reactions were catalyzed by the Klenow fragment of Escherichia coli DNA polymerase I (exo-) in a volume of 20–40 lL containing 8 l M template, 8 l M primer, 200 l M dNTP, 100 m M NaCl and EcoPol buffer (New England Biolabs). After the addition of primer to the template solutions, samples were denatured at 94 °Cfor 4 min and slowly annealed to 37 °C. DAPI was added after primer–template annealing, the solution reactions were incubated with 2.5 U of Klenow fragment at 37 °Cfor1h andstoppedwithEDTAtoafinalconcentrationof10m M . Reaction products were desalted by successive dilutions and filtrations using Centricon YM3 (Millipore) (3000 MW Fig. 1. DNA sequences and DAPI chemical structure. Schematic rep- resentation (A) and sequences (B) of the templates and primer used for chain extension reaction in vitro, catalyzed by Klenow fragment (3¢fi5¢ exo-) of E. coli DNA polymerase I. The underlined region of common sequence in (B) represents the primer-complementary region of the template. (C) Chemical structure of DAPI. 4756 E. Trotta et al. (Eur. J. Biochem. 270) Ó FEBS 2003 cut-off). Finally, samples were lyophilized, suspended in 10 lL of formamide and run at 37 °C and 290 V on a 12% (w/v) denaturing polyacrylamide gel containing 40% (v/v) formamide and 7 M urea. Gels were stained with 0.1% (w/v) methylenebluein0.5 M sodium acetate. Gels containing 5¢-fluorescein labeled DNA were analysed by UV transillu- mination. Results and discussion Electrophoretic mobility-shift induced by DAPI on trinucleotide repeats Electrophoretic analysis, comparing gel mobility-shift induced by DAPI on the 20 different individual strands of trinucleotide repeats, shows an observable increase in electrophoretic mobility induced by the drug for the triplet sequence (ATT) 6 only (Fig. 2). As shown previously, such increased migration is due to the reported ability of the drug to fold the (ATT) 6 into a hairpin-like structure by a minor- groove binding mechanism which stabilizes base pairing of DNA regions containing TÆT mismatches flanked by AÆT base-pairs [K a ¼ 3.4 · 10 6 M )1 considering two binding sites for (ATT) 6 ] [16]. The DAPI–(ATT) 6 complex is also the only fluorescent band detected in the electrophoretic gel. It has been reported that binding of DAPI into the minor- groove of DNA is characterized by a distinctive large increase of the drug fluorescence quantum yield [25,26] and by changes in DNA electrophoretic mobility [16,27]. As the electrophoretic bands of (AAT) 6 ,(ACT) 6 ,(AGT) 6 ,(ATG) 6 and (ATC) 6 do not exhibit such characteristic properties caused by the drug interaction, our results indicate that binding of DAPI to two consecutive AÆT base-pairs flanked by AÆA, GÆGorCÆC mismatches leads to weaker minor- groove complexes than two consecutive AÆTbase-pairs flanked by TÆT mismatches. Unexpectedly, even though intercalation mechanisms have been reported for binding of DAPI to GC-rich or mixed sequences [12,13], gel mobility of the GC-rich triplet is unchanged by DAPI and no fluorescent bands are detected for their DAPI-containing samples, including those triplets that spontaneously migrate faster due to their inherent propensity for self-annealing into hairpin structures: ACG, AGC, CCG, CGG, CGT and CTG (Fig. 2) [28]. In particular, CGT and CTG triplets, which self-anneal spontaneously by forming two consecu- tive Watson–Crick GÆC base-pairs flanked by TÆTmismat- ches, do not lead to stable complexes with DAPI as in the case of the ATT triplet. This indicates the crucial role played by AÆT base-pairs in addition to TÆT mismatches. Therefore, although our electrophoretic results represent the limited case of six triplet sequences, the specificity of the observed effects confirms the higher selectivity of DAPI for the minor-groove of the ATT-triplet with respect to all the other individual strands of trinucleotide repeats. Both AÆTand TÆT base-pairs appear to be necessary conditions for determining ATT-regions as the most favourable targets among all the individual strands of trinucleotide repeats. Binding and conformational properties of DAPI–template complexes The binding effects of DAPI on the replication processes of ATT-triplet sequences were evaluated by comparing prod- ucts of primer extension reactions by three distinct 70-base templates containing, respectively (ATT) 10 ,(AAT) 10 or a 30-base long GC-rich random sequence at the 5¢ end, as described in Materials and methods and illustrated in Fig. 1A,B. The conformational changes induced by DAPI on the different templates were evaluated to exclude unexpected binding and structural effects due to the presence of common primer-complementary and random sequences. As shown in Fig. 3, DAPI (50 l M ) does not induce appreciable changes in the electrophoretic mobility of AAT- (lane 6) and GC-rich random-templates (lane 2). Conversely, in the presence of DAPI, the ATT-template migrates faster than its drug-free form (lanes 4 and 3, respectively). This increased gel-mobility induced by DAPI is not merely attributable to the drug interaction per se, as binding of DAPI normally leads to a reduction in DNA electrophoretic mobility [16,27]. Moreover, the relatively low mobility of 1 : 1 AAT-/ATT-template mixture in the presence of DAPI (Fig. 3, lane 7) safely excludes the possibility that fast migrating DAPI–ATT-template com- plex could be constituted by dimeric homoduplex DNA structures. In contrast, the electrophoretic results account for a more compact structure of DAPI-bound ATT- template with respect to its free form, which is consistent with the previously reported lower mobility of the free monomeric random coil structure of (ATT) 6 in comparison with its hairpin conformer bound to DAPI [16]. Binding mechanisms of DAPI and their effects on the stability of template intramolecular base-pairing were Fig. 2. Effect of DAPI on the electrophoretic mobility of trinucleotide repeat DNA sequences. (A) Nondenaturing 12% polyacrylamide gel migration of the 20 possible triplet repeat strands and their complexes with DAPI at 4 : 1 drug/template molar ratio. (B) Before gel inocu- lation, samples were denatured at 90 °C and slowly annealed to room temperature. Drug–DNA complexes were prepared by adding DAPI following DNA annealing. Gels were stained with 0.1% (v/v) methy- lene blue in 0.5 M sodium acetate. Samples were 35 l M DNA and 150 l M DAPI in Tris/borate/EDTA buffer. Ó FEBS 2003 DAPI effects on replication of ATT-triplet (Eur. J. Biochem. 270) 4757 evaluated by fluorescence and UV thermal melting studies. The addition of ATT-template to DAPI solution determines drug fluorescence changes that are distinctive of the drug binding into the minor-groove of the double-helix structure of AT-rich sequences [25]: the drug fluorescence signal is strongly increased (8.3-fold at a drug/template molar ratio of 2 : 1) and the maximum emission is shifted to lower wavelengths (20 nm) (Fig. 4). The observation of such distinctive fluorescence properties of the minor-groove binding mechanism implies the presence in the DAPI– ATT-template complex of a double-helix structure, in agreement with the hairpin suggested by the increased gel mobility of the DAPI–ATT-template complex. In contrast, addition of AAT- and GC-rich random-template oligomers to the drug solution causes changes in DAPI fluorescent properties which are indicative of a weak minor-groove interaction, GÆC-like intercalation or uncharacterized bind- ing mechanisms: the shift in the maximum emission at a 2 : 1 drug/template molar ratio was about 12 and 15 nm, and the increase in the fluorescence intensity was only 2.2- and 1.6-fold for AAT- and random-template, respect- ively (Fig. 4). UV thermal melting studies are also consistent with a double-helix structure of ATT-template complexed with the drug. Binding of DAPI strongly increases the melting temperature of the ATT-template from 17 °Cto45°Cas measured by the shift in sigmoidal melting profiles illustra- ted in Fig. 5. Consequently, at the Klenow reaction temperature of 37 °C, the drug-free ATT-template is almost completely in random coil conformation (Tm ¼ 17 °C) whereas its DAPI-bound form is prevalently base-paired (Tm ¼ 45 °C). Moreover, along with a single electropho- reticbandproducedbytheATT-templateinthepresenceof DAPI (Fig. 3, lane 4), the observation of a monophasic thermal transition is consistent with the prevalent presence of a single DNA structure in the complex, which excludes the observable presence of dimeric homoduplex structures Fig. 3. Effect of DAPI on the electrophoretic mobility of DNA tem- plates. Nondenaturing 12% (w/v) polyacrylamide gel migration of GC-rich random-template (lanes 1 and 2), ATT-template (lanes 3 and 4), AAT-template (lanes 5 and 6) and 1 : 1 mixture of ATT- and AAT- template. Samples with (+) and without (–) DAPI are indicated. Before gel inoculation, samples were denatured at 90 °Candslowly annealed to room temperature. Drug–DNA complexes were prepared by adding DAPI following DNA annealing. Gels were stained with 0.1% (w/v) methylene blue in 0.5 M sodium acetate. Samples were 8 l M template–primer and 50 l M DAPI in Tris/borate/EDTA buffer. Fig. 4. Effect of DNA templates on the fluorescence spectrum of DAPI. DAPI fluorescence spectra in (a) the free form, in the presence of (b) random-template, (c) AAT-template and (d) ATT-template at 2 : 1 drug/template molar ratio. The experiments were performed at 37 °C, in 100 m M NaCl and 10 m M phosphate buffer at pH 7.00. Excitation (350 nm) and emission (448 nm) with 1.5 nm band-pass were used and spectra were corrected for background signal. Fig. 5. Effect of DAPI on the melting curve of ATT-template. Melting curves of ATT-template in free (- - -) and DAPI-bound forms (––) at a base/drug molar ratio of 3.5. The DNA solution absorbance was 0.25 per mL in 10 m M sodium phosphate (pH 7.00) and 100 m M NaCl, and quartz cuvettes of 1 cm path length were used. Samples were heated at arateof1°CÆmin )1 and absorbance was measured corresponding to the maximum around 260 nm, at 25 °C. Hyperchromicity (%) at the temperature T °C was given by [A (T °C) – A (3 °C) ]/A (3 °C) ,whereA (T °C) and A (3 °C) represent the A 260 at T °Cand3°C, respectively. 4758 E. Trotta et al. (Eur. J. Biochem. 270) Ó FEBS 2003 in addition to the monomeric hairpin. In conclusion, the spectroscopic and electrophoretic results shows that the common random GC-rich sequences of the ATT-template as well as the increased length of its triplet sequence, do not prevent its (ATT) 10 region from adopting a hairpin-like structure the equivalent of that induced by DAPI in the (ATT) 6 oligomer [16]. Primer extension experiments DNA polymerization studies for evaluating the biological effects of the binding of DAPI to ATT-, AAT- and random- template sequences were performed by the Klenow frag- ment of DNA polymerase I using a 15-base primer complementary to the 41–55 base-pair region of the 70-base templates reported in Fig. 1. The fully extended product of Klenow synthesis was 15 bases shorter than the templates due to the 15 unpaired bases at the 3¢-end of the templates annealed to the primer (Fig. 1). As shown by electrophoresis (Fig. 6), at the 16 : 1 DAPI/template molar ratio, the expected fully extended 55-base product of Klenow synthesis is present in the lane containing the reaction products of AAT-template. In contrast at the 16 : 1 DAPI/template molar ratio, DAPI inhibits synthesis of the 55-base product in the ATT-template directed reaction. Such a sequence-dependent inhibition was effective when the DAPI/template molar ratio is equal to 8 (DAPI/ (ATT) 2 ¼ 1.6), while full-length synthesis products are observed at a drug/template molar ratio of 4 (DAPI/ (ATT) 2 ¼ 0.8), although DAPI/template values of both 4 and 8 present visible products of partial synthesis. The experiments in Fig. 6 were performed at a constant template concentration of 8 l M and drug concentration ranging from 8 to 128 l M . Such a tight binding condition (K a ¼ 3.4 · 10 6 M )1 considering two binding sites for (ATT) 6 [16]) strongly shifts the binding equilibrium towards the bound species. This indicates that DAPI strongly increases its inhibitory activity when complexes are formed at a stoichiometric molar ratio of around 1 : 1 with (ATT) 2 target sites. The observed dependence on template sequence excludes the possibility that the inhibitory property of DAPI could only be attributed to a direct interaction of the drug with the Klenow fragment. A DNA-mediated step correlating with the different structural and binding characteristics of DAPI–template complexes is involved. The lanes of Fig. 6 relating to synthesis products using the AAT-template (8 l M ), show a weak inhibitory activity of DAPI at high concentration (128 l M ), which is more evident at the highest drug concentration (256 l M ) (Fig. 7). Such a weak inhib- itory activity on the AAT-triplet is also sequence-dependent. At the highest concentration (256 l M ), DAPI has a very weak effect on the random-template synthesis appearing to be almost ineffective on GC-rich sequences (Fig. 7). This finding is also supported by the observation that the first 10 bases of the GC-rich random sequence of the AAT- and ATT-templates are copied without any pausing in relation to the drug-free syntheses. This provides evidence for a relatively weak sequence-independent inhibitory effect of DAPI on Klenow activity at our experimental conditions. Therefore, inhibitory activity of DAPI appears to be significantly effective only for the AT-rich sequences of the AAT- and ATT-templates, even though with different characteristics, as it appears weaker and more homogeneous along AAT- than ATT-triplet sequences. The weak inhibi- tion at AAT-triplet sites is attributable to a binding mechanism that presents spectroscopic characteristics that are different from those reported for intercalation and minor-groove binding, indicating minor involvement of DNA bases [16]. However, the absence of a comparable inhibitory effect at GC-rich random sequences indicates a significant direct or indirect involvement of AAT bases in Fig. 6. Effects of different DAPI/template molar ratios on Klenow synthesis. Denaturing gel analysis showing the effect of DAPI on Klenow-catalyzed 5¢-fluorescein labeled primer elongation by ATT- and AAT-template. Reactions were performed at the constant tem- plate concentration of 8 l M with different DAPI/template molar ratios (R). Ladder is constituted by 5¢-fluorescein-labeled DNA oligomers with sequences similar to those expected for the reaction products of the ATT-template. Fig. 7. Effect of high DAPI concentration on template-directed Klenow synthesis. Denaturing gel analysis showing the effect of high DAPI concentration on the elongation of 5¢-fluorescein labeled primer by ATT-, AAT- and random-template-directed Klenow synthesis. Sam- ples were 8 l M primer–template and 256 l M DAPI. Presence (+) and absence (–) of DAPI in the reaction solution is indicated. Ladder is constituted by 5¢-fluorescein labeled DNA oligomers with sequences similar to those expected for the reaction products of the ATT- template. Ó FEBS 2003 DAPI effects on replication of ATT-triplet (Eur. J. Biochem. 270) 4759 the interaction with DAPI. This suggests an efficient mechanism of single-stranded sequence recognition charac- terized by small changes in NMR resonances of DNA bases [16]. It is also relevant that the intercalation mechanism reported for the binding of DAPI at AÆU[25]andGÆC[12] sites does not appear to significantly affect Klenow activity catalyzed on the AAT- and GC-rich random-templates. The synthesis was stalled by DAPI on the ATT-bearing template giving rise to a prevalent product of about 31–33 bases (Fig. 7) indicating that, in addition to the initial 10-base random sequence, at least 6–8 bases of ATT-region are copied by the Klenow fragment before it stalls. In other words, the ATT motif itself does not appear to be sufficient to block DNA synthesis by DAPI, suggesting an effect by different structural features along the (ATT) 10 sequence caused by the drug interaction. As shown above, DAPI induces the formation of hairpin structures on the ATT- but not on the AAT- and GC-rich random-templates. In addition, it has been reported previously that long tracts of CTGÆCAG repeats, which have the propensity to fold into stable hairpin structures [29–31], block the replication fork in E. coli [32] and cause a pause in DNA synthesis in vitro [33]. These findings suggest that the stall of Klenow progression observed in the present study could be attrib- uted to the (ATT) 10 region (located at the 5¢-end of the ATT-template) folding into a hairpin-like structure. The double helix of the hairpin stem, stabilized by the binding of DAPI into the minor-groove of AÆTandTÆTbase-pairs, appears to be responsible for stalling the Klenow fragment. As AÆAandTÆT mismatches cause similar destabilizing effects on the double helix structure [34–36], the dissimilar influence of DAPI on the Klenow catalyzed replication of AAT- and ATT-containing sequences can only be attrib- uted to the more favourable characteristic of the TÆT minor- groove than the AÆA minor-groove, in the binding of DAPI [16]. In conclusion, the main cause of the strong inhibitory effect of DAPI on DNA replication in vitro reported in this work appears to be linked to the propensity of the drug to stabilize the double helix structure by favourable electro- static, H-bonds and van der Waals contacts with the DNA minor-groove surface [11,16]. Although it should be men- tioned that the drug could affect other cellular processes at the conditions necessary for stalling DNA polymerase progression, the results nevertheless suggest a novel struc- tural model for the molecular mechanism of the action of minor-groove binding ligands. Conclusion This study reports the effect of the classical minor-groove binding ligand DAPI on AATÆATT trinucleotide repeat replication in vitro, catalyzed by the Klenow fragment of DNA polymerase I. The high affinity and structural changes selectively produced by DAPI on the ATT-strand of the AATÆATT trinucleotide repeat are associated with the stalling of Klenow progression along the ATT-template sequence. This draws attention to the biological relevance of high affinity drugs with structural effects on single- stranded DNA and RNA. The main genetic processes such as replication, transcription and translation involve indi- vidual strands of DNA and RNA, emphasizing the importance of studies that attempt to clarify the relation- ship among binding, structural and biological aspects relating to this conformational class of nucleic acids. Particularly in the case of trinucleotide repeats, the secondary structures of individual strands are considered to be the main cause of their expansion, and are associated with a number of human genetic diseases [28]. In particular, the formation of stable intramolecular hairpins appears to be the most probable cause of CAGÆCTG and CGGÆCCG expansion, by favouring slippage of DNA strands during DNA replication. We have found previ- ously that the spontaneous hairpins reported for the T-rich strand of CAGÆCTG triplets are similar to those induced byDAPIandHoechstontheT-richstrandofAATÆATT trinucleotide repeats [16]. The present work shows that such a structural affinity is associated with a comparable ability to influence DNA polymerase activity, given that tracts of CAGÆCTG triplets can affect DNA replication in vivo [32] and can be sites of pausing of DNA synthesis in vitro [33]. The CAGÆCTG triplet repeats present additional biological characteristics that have been attrib- uted to its propensity to fold into hairpin structures, suggesting further potential biological properties of DAPI- induced hairpins: regions of CAGÆCTG triplets are tran- scribed significantly slower than random sequences [37] and can elude cellular mechanisms designed to repair DNA [38]. Moreover, because it has emerged that specific DNA sequences within introns can regulate transcription [19,20], the prevalent location of ATT triplet repeats within introns in the human genome [18] suggests strongly that binding of DAPI to ATT-bearing genomic targets may influence the transcriptional processes. In conclusion, the results reported in this work suggest a novel molecular mechanism of action of the classical minor- groove binding ligand DAPI. This property is associated with the ability of the drug to bind and stabilize folded structures in ATT-triplet sequences, by favourable and selective interaction with the minor-groove of AÆTandTÆT base-pairs. The molecular mechanism proposed in this work may help in designing new minor-groove binding ligands with sequence-selective activity on DNA replication. Acknowledgements We thank Fabio Bertocchi for his technical assistance. This work is supportedinpartbytargetprojectÔBiotecnologiaÕ of Italian CNR. References 1. Kapuscinski, J. (1995) DAPI: a DNA-specific fluorescent probe. Biotech Histochem 70, 220–233. 2. Parolin, C., Montecucco, A., Chiarocchi, G., Pedrali-Noy, G., Valisena, S., Palumbo, M. & Palu` ,G.(1990)Theeffectofthe minor groove binding agent DAPI (4,6-diamidino-2-phenyl- indole) on DNA-directed enzymes: an attempt to explain inhibi- tion of plasmid expression in Escherichia coli. FEMS Microbiol. Lett. 68, 341–346. 3. Straney, D.C. & Crothers, D.M. (1987) Effect of drug–DNA interactions upon transcription initiation at the lac promoter. Biochemistry 26, 1987–1995. 4. Chiang, S Y., Welch, J., Rauscher, F.J. & Beerman, T.A. (1994) Effects of minor groove binding drugs on the interaction of TATA box binding protein and TFIIA with DNA. Biochemistry 33, 7033–7040. 4760 E. Trotta et al. (Eur. J. Biochem. 270) Ó FEBS 2003 5. Storl, K., Storl, J., Zimmer, C.H. & Lown, J.W. (1993) Minor- groove binders are inhibitors of the catalytic activity of DNA gyrases. FEBS Lett. 317, 157–162. 6. Woynarowski, J.M., McHugh, M., Sigmund, R.D. & Beerman, T.A. (1989) Modulation of topoisomerase II catalytic activity by DNA minor groove binding agents distamycin, Hoechst 33258, and 4¢,6-diamidine-2-phenylindole. Mol. Pharmacol. 35, 177–182. 7. McHugh, M.M., Woynarowski, R.D. & Beerman, T.A. (1989) Effect of minor groove binding drugs on mammalian topo- isomerase I activity. Biochem. Pharmacol. 38, 2323–2328. 8. Parolin, C., Zanotti, G. & Palu` , G. (1995) A model for the sequence-dependent DNA binding of 4¢,6-diamidino-2-phenyl- indole (DAPI). Biochem. Biophys. Res. Commun. 208, 332–338. 9. Wilson, W.D., Tanious, F.A., Barton, H.J., Jones, R.L., Fox, K., Wydra, R.L. & Strekowski, L. (1990) DNA sequence dependent binding modes of 4¢6-diamidino-2-phenylindole (DAPI). Bio- chemistry 29, 8452–8461. 10. Trotta, E., D’Ambrosio, E., Del Grosso, N., Ravagnan, G., Cirilli, M. & Paci, M. (1993) 1 H NMR study of [d(GCGATCGC)] 2 and its interaction with minor groove binding 4¢,6-diamidino-2-pheny- lindole. J. Biol. Chem. 268, 3944–3951. 11. Trotta, E. & Paci, M. (1998) Solution structure of DAPI selectively bound in the minor groove of a DNA TÆT mismatch-containing site: NMR and molecular dynamics studies. Nucleic Acids Res. 26, 4706–4713. 12. Wilson, W.D., Tanious, F.A., Barton, H.J., Strekowski, L. & Boykin, D.W. (1989) Binding of 4¢,6-diamidino-2-phenylindole (DAPI) to GC and mixed sequences in DNA: intercalation of a classical groove-binding molecule. J. Am. Chem. Soc. 111, 5008– 5010. 13. Trotta, E., D’Ambrosio, E., Ravagnan, G. & Paci, M. (1995) Evidence for DAPI intercalation in CG sites of DNA oligomer [d(CGACGTCG)] 2 :a 1 HNMRstudy.Nucleic Acids Res. 23, 1333–1340. 14.Kim,S.K.,Eriksson,S.,Kubista,M.&Norde ´ n, B. (1993) Interaction of 4¢,6-diamidino-2-phenylindole (DAPI) with poly[d(G-C) 2 ] and poly[d(G-m 5 C) 2 ]: evidence for major groove binding of a DNA probe. J. Am. Chem. Soc. 115, 3441–3447. 15. Trotta, E., D’Ambrosio, E., Ravagnan, G. & Paci, M. (1996) Simultaneous and different binding mechanisms of 4¢,6-diami- dino-2-phenylindole to DNA hexamer (d(CGATCG))2. A 1 H NMR Study. J. Biol. Chem. 271, 27608–27614. 16. Trotta,E.,DelGrosso,N.,Erba,M.&Paci,M.(2000)TheATT strand of AATÆATT trinucleotide repeats adopts stable hairpin structures induced by minor-groove-binding ligands. Biochemistry 39, 6799–6808. 17. Gastier, J.M., Pulido, J.C., Sunden, S., Brody, T., Buetow, K.H., Murray, J.C., Weber, J.L., Hudson, T.J., Sheffield, V.C. & Duyk, G.M. (1995) Survey of trinucleotide repeats in the human genome: assessment of their utility as genetic markers. Hum. Mol. Genet. 4, 1829–1836. 18. Stallings, R.L. (1994) Distribution of trinucleotide microsatellites in different categories of mammalian genomic sequence: implica- tions for human genetic diseases. Genomics 21, 116–121. 19. Lozano, G. & Levine, A.J. (1991) Tissue-specific expression of p53 in transgenic mice is regulated by intron sequences. Mol. Carci- nogen. 4, 3–9. 20. Brinster, R.L., Allen, J.M., Behringer, R.R., Gelinas, R.E. & Palmiter, R.D. (1988) Introns increase transcriptional efficiency in transgenic mice. Proc. Natl Acad. Sci. USA 85, 836–840. 21. Ohshima, K., Kang, S., Larson, J.E. & Wells, R.D. (1996) TTA.TAA triplet repeats in plasmids form a non-H bonded structure. J. Biol. Chem. 271, 16784–16791. 22. Lyons-Darden, T. & Topal, M.D. (1999) Effects of temperature, Mg 2+ concentration and mismatches on triplet-repeat expansion during DNA replication in vitro. Nucleic Acids Res. 27, 2235– 2240. 23. Lyons-Darden, T. & Topal, M.D. (1999) Abasic sites induce tri- plet-repeat expansion during DNA replication in vitro. J. Biol. Chem. 274, 25975–25978. 24. Richards,R.I.&Holman,K.,Yu,S.&Sutherland,G.R.(1993) Fragile X syndrome unstable element, p(CCG)n, and other simple tandem repeat sequences are binding sites for specific nuclear proteins. Hum. Mol. Genet. 2, 1429–1435. 25. Tanious, F.A., Veal, J.M., Buczak, H., Ratmeyer, L.S. & Wilson, W.D. (1992) DAPI (4¢,6-diamidino-2-phenylindole) binds differ- ently to DNA and RNA: minor-groove binding at AT sites and intercalation at AU Sites. Biochemistry 31, 3103–3122. 26. Eriksson, S., Kim, S.K., Chiarocchi, G., Kubista, M. & Norden, B. (1993) Binding of 4¢,6-diamidino-2-phenylindole (DAPI) to AT regions of DNA: evidence for an allosteric conformational change. Biochemistry 32, 2987–2998. 27. Larsson, A., Akerman, B. & Jonsson, M. (1996) DAPI staining of DNA: effect of change in charge, flexibility, and contour length on orientational dynamics and mobility of the DNA during agarose gel electrophoresis. J. Phys. Chem. 100, 3252–3263. 28. Mitas, M. (1997) Trinucleotide repeats associated with human disease. Nucleic Acids Res. 25, 2245–2254. 29. Gacy, A.M., Goellner, G., Juranic, N., Macura, S. & McMurray, C.T. (1995) Trinucleotide repeats that expand in human disease form hairpin structures in vitro. Cell 81, 533–540. 30.Mitas,M.,Yu,A.,Dill,J.,Kamp,T.J.,Chamber,E.J.& Haworth, I.S. (1995) Hairpin properties of single-stranded DNA containing a GC-rich triplet repeat: (CTG) 15 . Nucleic Acids Res. 23, 1050–1059. 31. Chen, X., Mariappan, S.V.S., Catasti, P., Ratliff, R., Moyzis, R.K., Laayoun, A., Smith, S.S., Bradbury, E.M. & Gupta, G. (1995) Hairpins are formed by the single DNA strands of the fragile X triplet repeats: structure and biological implications. Proc.NatlAcad.Sci.USA92, 5199–5203. 32. Samadashwily, G.M., Raca, G. & Mirkin, S.M. (1997) Tri- nucleotide repeats affect DNA replication in vivo. Nat. Genet. 17, 298–304. 33. Kang, S., Ohshima, K., Shimizu, M., Amirhaeri, S. & Wells, R.D. (1995) Pausing of DNA synthesis in vitro at specific loci in CTG and CGG triplet repeats from human hereditary disease genes. J. Biol. Chem. 270, 27014–27021. 34. Aboul-ela, F., Koh, D., Tinoco, I. & Martin, F.H. (1985) Base- base mismatches. Thermodynamics of double helix formation for dCA3XA3G + dCT3YT3G (X, Y ¼ A,C,G,T). Nucleic Acids Res. 13, 4811–4824. 35. Werntges, H., Steger, G., Riesner, D. & Fritz, H.J. (1986) Mismatches in DNA double strands: thermodynamic parameters and their correlation to repair efficiencies. Nucleic Acids Res. 14, 3773–3790. 36. Ikuta, S., Takagi, K., Wallace, R.B. & Itakura, K. (1987) Dis- sociation kinetics of 19 base paired oligonucleotide-DNA duplexes containing different single mismatched base-pairs. Nucleic Acids Res. 15, 797–811. 37. Parsons, M.A., Sinden, R.R. & Izban, M.G. (1998) Transcrip- tional properties of RNA polymerase II within triplet repeat- containing DNA from the human myotonic dystrophy and fragile Xloci.J. Biol. Chem. 273, 26998–27008. 38. Moore,H.,Greenwell,P.W.,Liu,C P.,Arnheim,N.&Petes, T.D. (1999) Triplet repeats form secondary structures that escape DNA repair in yeast. Proc. Natl Acad. Sci. USA 96, 1504–1509. Ó FEBS 2003 DAPI effects on replication of ATT-triplet (Eur. J. Biochem. 270) 4761 . Interaction of DAPI with individual strands of trinucleotide repeats Effects on replication in vitro of the AATÆATT triplet Edoardo Trotta 1 ,. strand of AATÆATT trinucleotide repeat sequences [16]. In this work we investigate the effects of DAPI on replication in vitro of AATÆATT trinucleotide repeats