Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (372K) Freely available
Right arrow Supplementary Material
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Related articles in Nucl. Acids Res.
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Chin, K.-H.
Right arrow Articles by Chou, S.-H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chin, K.-H.
Right arrow Articles by Chou, S.-H.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research, 2003, Vol. 31, No. 10 2622-2629
© 2003 Oxford University Press

Solution structure of the ActD–5'-CCGTT3GTGG-3' complex: drug interaction with tandem G·T mismatches and hairpin loop backbone

Ko-Hsin Chin1, Fu-Ming Chen2 and Shan-Ho Chou1,3

1 Institute of Biochemistry, National Chung-Hsing University, Taichung, 40227, Taiwan, 2 Department of Chemistry, Tennessee State University, Nashville, TN 37209-1561, USA and 3 Department of Life Science, National Central University, Jung-Li, 320, Taiwan

*To whom correspondence should be addressed at Institute of Biochemistry, National Chung-Hsing University, Taichung, 40227, Taiwan. Tel: +886 42 285 3486; Fax: +886 42 285 3487; Email: shchou{at}nchu.edu.tw

Received January 17, 2003; Revised February 27, 2003; Accepted March 12, 2003

PDB accession no. 1OBF.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Binding of actinomycin D (ActD) to the seemingly single-stranded DNA (ssDNA) oligomer 5'-CCGTT3 GTGG-3' has been studied in solution using high-resolution nuclear magnetic resonance (NMR) techniques. A strong binding constant (8 x 106 M–1) and high quality NMR spectra have allowed us to determine the initial DNA structure using distance geometry as well as the final ActD–5'-CCGTT3 GTGG-3' complex structure using constrained molecular dynamics calculations. The DNA oligomer 5'-CCGTT3GTGG-3' in the complex forms a hairpin structure with tandem G·T mismatches at the stem region next to a loop of three stacked thymine bases pointing toward the major groove. Bipartite T2O–GH1 and T2O–G2NH2 hydrogen bonds were detected for the G·T mismatches that further stabilize this unusual DNA hairpin. The phenoxazone chromophore of ActD intercalates nicely between the tandem G·T mismatches in essentially one major orientation. Additional hydrophobic interactions between the ActD quinoid amino acid residues with the loop T5–T6–T7 backbone protons were also observed. The hydrophobic G–phenoxazone–G interaction in the ActD–5'-CCGTT3GTGG-3' complex is more robust than that of the classical ActD– 5'-CCGCT3GCGG-3' complex, consistent with the roughly 2-fold stronger binding of ActD to the 5'-CCGTT3GTGG-3' sequence than to its 5'-CCG CT3GCGG-3' counterpart. Stabilization by ActD of a hairpin containing non-canonical stem base pairs further strengthens the notion that ActD or other related compounds may serve as a sequence- specific ssDNA-binding agent that inhibits human immunodeficiency virus (HIV) and other retroviruses replicating through ssDNA intermediates.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Actinomycin D (ActD) is a highly potent antitumor agent (Fig. 1) that has been used as a chemotherapy drug for treating certain types of cancer. Its action has long been attributed to its ability to bind to double-stranded DNA through intercalation of the planar phenoxazone ring into the double helix, hence inhibiting the DNA-directed RNA synthesis (1,2). Several nuclear magnetic resonance (NMR) and X-ray investigations of complexes with self-complementary oligonucleotides have been reported (3–8), which indicate that ActD intercalates at GpC steps with the cyclic pentapeptide lactone rings situated in the minor groove and spans 2 bp on either side of the phenoxazone intercalation site of the DNA duplex. The complex is stabilized by hydrogen bonds between the drug L-ThrHN/Thr(C=O) atoms and the DNA GN3/GNH2 atoms, and the stacking forces between the drug phenoxazone ring and the cross-strand DNA guanine bases to form a strong hydrophobic G–phenoxazone–G stack. Additional stabilization of the complex comes from hydrophobic interactions between the cyclic pentapeptides and the DNA minor groove surface atoms.



View larger version (32K):
[in this window]
[in a new window]
  		Figure 1. The ActD chemical structure and the investigated DNA sequence.

 
Although interaction of ActD with a classical GpC binding site has been well studied before, a great deal of interest has been triggered recently toward the investigation of unusual ActD–DNA binding, due to the unexpected finding that ActD is also capable of binding to DNA sequences that are devoid of classical (GpC)2 sequence (9–14) and/or to seemingly single-stranded DNA (ssDNA) sequences (11,15–19). Importantly, these complexes also exhibit stability and affinities comparable with those of the classical ActD–(GpC)2 complex. The recent discovery that sequence-specific ActD binding to ssDNA can inhibit human immunodeficiency virus (HIV) reverse transcriptase and other polymerases (16) further reinforces the need for structural studies of unusual ActD–DNA complexes, since these may assist the development of drugs against HIV and other retroviruses that replicate through ssDNA intermediates. Drugs highly selective for single-stranded forms of the HIV genome should cause minimal damage to the host genome DNA.

Recently, we have solved the first high-resolution complex structures of ActD binding to 5'-CGXCACCGYCG-3' oligomers devoid of classical GpC sequence (where X/Y is a G·C or A·T canonical base pair) (14), and found that such complexes exhibit some rather unusual structural features. As expected, the free DNA oligomers form hairpins consisting of a mini-ACC loop closed by a sheared AC base pair. Under a 1/1 drug/strand ratio and low salt conditions, however, the canonical X·Y base pair in the GXC/GYC stem region is totally different; they are disrupted and pushed out of the duplex (14). These disrupted base pairs are, however, not disordered, but are perpendicular to the base plane and form specific hydrogen bonds with the backbone phosphate and sugar oxygen (14). Such unusual ActD–DNA complex structures indicate that ActD is able to cause considerable perturbations to the partner DNA structure and further increases our understanding regarding drug–DNA or drug–protein interaction.

To study further other unique ActD–DNA binding modes, we have embarked on high-resolution NMR studies of ActD binding with the 5'-CCGTT3GTGG-3' sequence. This is because in our past thermodynamics studies, we have found that ActD is capable of binding to a series of seemingly single-stranded GT-rich sequences with strong binding constants in the micromolar range (19). In the present study, we therefore describe the detailed structural studies of the ActD–5'-CCGTT3GTGG-3' complex using NMR techniques. This ssDNA sequence is partially structured in low salt conditions (20 mM pH 6.8 sodium phosphate buffer), but forms a stable hairpin containing a 5'-GT/TG-5' mismatch motif in the stem region looped by TTT bases when titrated by ActD at a 1/1 ratio. The phenoxazone chromophore of ActD is found to intercalate nicely between the tandem G·T mismatches in essentially one major orientation. This results in greatly simplified NMR spectra that facilitate the complex structural analyses. Strong hydrophobic interactions between the ActD quinoid L-Pro, Sar and L-Me-Val residues with the loop TTT backbone protons are also observed, which enhance the ActD–hairpin complex formation. The strong binding of ActD to tandem G·T mismatches further enlarges the binding motifs available for this important chemotherapy drug and could provide valuable information regarding the action of ActD on DNA.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Sample preparation
All DNA samples were synthesized in a 3 µmol scale on an Applied Biosystems 380B DNA synthesizer with the final 5'-dimethoxytrityl groups attached. The samples were purified and prepared for NMR studies as described before (20). ActD was purchased from Sigma and used without further purification. The complex was generated by adding aliquots of ActD dissolved in methanol and was monitored by proton NMR. Addition of the drug was stopped when the original free DNA signals were completely absent. The sample was re-annealed during each addition. Free ActD is only sparingly soluble in H2O solution, and can thus be removed by centrifugation if necessary.

NMR experiments
All NMR experiments were carried out on a Varian Unity Inova 600 MHz spectrometer. One-dimensional imino proton spectra at 0°C were acquired using the jump–return pulse sequence (21). The spectral width was 16 000 Hz, with the carrier frequency set at the resonance of water. The maximum excitation was set at 12.5 p.p.m. For each experiment, 32 000 complex points were collected and 64 scans were averaged with a 2 s relaxation delay.

2D NOESY in 90% H2O/10% D2O was performed at 0°C in a pH 6.8 low salt (10 mM sodium phosphate, 20 mM NaCl) buffer with the following parameters; delay time 1 s, mixing time 0.12 s, spectra width 11 204 Hz, complex points 2048, number of transients 96, and number of increments 250.

NOESY experiments in D2O were carried out at 20°C in the hypercomplex mode with a spectral width of 5100 Hz. Spectra were collected using two mixing times of 100 and 300 ms with a relaxation delay of 1 s between each transient and with 2048 complex points in the t2 and 300 complex points in the t1 dimension. For each t1 increment, 40 scans were averaged.

A DQF-COSY spectrum was collected in the TPPI mode with a spectral width of 5100 Hz in both dimensions; 2048 complex points in the t2 dimension and 532 (real) points in the t1 dimension were collected with a relaxation delay of 1 s, and 40 scans were averaged for each t1 increment.

A proton-detected 31P–1H heteronuclear correlation spectrum (22) was collected in the TPPI mode with a spectral width of 5100 Hz in the 1H dimension and of 1000 Hz in the 31P dimension. A total of 2048 complex points in the t2 (1H) dimension and 100 complex points in the t1 (31P) dimension were collected. Protons were pre-saturated for 1.0 s, and 128 scans were accumulated for each t1 increment.

The acquired data were transferred to an IRIS 4D workstation and processed by the software FELIX (Accelrys Inc.) as described previously (23).

Structure determination
The final ActD–5'-CCGTT3GTGG-5' complex structures were accomplished by first applying distance geometry calculations (DGII from Accelrys Inc.) using constraints derived from NMR experiments to determine the initial DNA structures. Most distance constraints from NOESY/D2O were classified as very strong, strong, medium or weak, based on their relative intensities at 120 ms mixing time and given generous distance bounds of 2.0–3.0, 2.0–4.0, 3.0–5.0 or 4.0–6.0 Å, respectively. Canonical hydrogen bond distances with bounds of 1.8–2.2 Å were assigned to Watson–Crick and G·T wobble base pairs. Distance constraints involving exchangeable protons were derived from NOESY/H2O and given only two wide distance bounds of either 2.0–5.0 or 3.0–6.0 Å. ß and {gamma} torsion angle constraints were determined semi-quantitatively from the 31P–1H hetero-nuclear correlation data using the in-plane ‘W’ rule (24). Based on the absence of long-range 4JH2'-P coupling, all {epsilon} torsion angles were constrained to the trans domain (180 ± 30°) (25). The {zeta} and {alpha} dihedral angles were all left unconstrained. The initial DNA structures were generated by embedding the DNA-bound matrix using the DGII program (Accelrys Inc.). Half of the initial DNA structures (15 out of 30) encompass conformations suitable for subsequent docking and molecular dynamics studies. Since the majority of conformational parameters of the drug in the ActD–hairpin complex were similar to those of the free drug in the crystal (6,7,13,26), ActD coordinates derived from the X-ray diffraction method were employed to dock against the initial DNA structures. Energy minimization and constrained molecular dynamics were then applied using the DISCOVER program (Accelrys Inc.) with 39 intermolecular distance constraints (Supple mentary table S3, available at NAR Online), as well as 171 intramolecular constraints (including 103 intra-DNA and 68 intra-ActD, Table 1) to determine the final complex structures that best fit the experimental NMR data. Dynamics were initiated at 500 K with a 1 fs time step. After a total of 10 ps of molecular dynamics at 500 K, the system was slowly cooled to 300 K in 10 ps. The system was then equilibrated at 300 K for 5 ps. Well-converged final structures with pair-wise r.m.s.d. values of ~0.69 Å were obtained after the final molecular dynamics calculations. The intermolecular distance constraints applied and the distance measured from the determined ActD–hairpin complex structure agree well, with no major deviation larger than 0.1 Å (Supplementary table S3). The coordinates of the ActD–5'-CCGTT3GTGG-5' complex structures have been deposited in the Protein Data Bank under the ID 1OBF.


View this table:
[in this window]
[in a new window]
  		Table 1. Structural statistics for the ActD–5'-d(CCGTT3GTGG)-3' complex
 

   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Exchangeable DNA proton assignments
The ActD-d(CCGTT3GTGG) complex was generated by the addition of one equivalent of ActD to the d(CCGTT3GTGG) oligomer in low salt buffered solution at 0°C. Only one major set of exchangeable and non-exchangeable proton resonances was observed (Fig. 2), indicating the formation of a unique complex. From the 1D imino proton spectrum shown in the lower panel of Figure 2, it is clear that there is already some partial structural formation for the free DNA oligomer, as evidenced by the two broad G-imino proton signals at the paired region (~12.6 p.p.m.) and two clusters of imino proton resonances at ~10.6 and 11.1 p.p.m. Upon the addition of ActD at 1/1 ratio, the original DNA spectrum underwent considerable changes, as shown in the middle panel. The two broad peaks at ~12.6 p.p.m. have shifted downfield and become sharper, while two new peaks at 9.6 and 9.9 p.p.m. have appeared, along with four sharp peaks scattered at the ~8 p.p.m. region. These exchangeable protons were assigned successfully through the 2D-NOESY (120 ms mixing time) recorded in H2O at 0°C (top panel). Formation of two pairs of G·T mismatches in the complex was revealed by the medium to strong NOE cross-peaks between the GH1 and TH3 imino protons (labeled a and b), characteristic of formation of the wobble G·T mismatch (27). Assignments of the imino protons in the tandem G·T mismatches are made possible by the observation of NOEs between the imino protons in one of the G·T mismatches and the GH1 imino proton involved in canonical pairing (labeled c and d). Obviously the imino protons resonating at 10.55 and 9.6 p.p.m. stem from the G3·T9 mismatch, while the other set of imino protons resonating at 11.1 and 9.9 p.p.m. come from the G8·T4 mismatch, because it is the G3·T9 mismatch that is adjacent to a canonical G10·C2 base pair. All imino proton signals were therefore assigned concurrently. The C2NH2 and C1NH2 amino protons were also identified through the observation of strong GH1–CNH2 cross-peaks from the assigned G10H1 and G11H1 imino protons. The large peak at 10.5 p.p.m. must correspond to the resonances of three unpaired T loop imino protons. It is important to note that no NOE was observed between the tandem G·T mismatches (circled in the top figure) even in the NOESY with a long 200 ms mixing time (data not shown), indicating that the phenoxazone chromophore of ActD is probably intercalated between the tandem G·T mismatches.



View larger version (32K):
[in this window]
[in a new window]
  		Figure 2. Imino proton spectra of the 5'-CCGTT3GTGG-3' oligomer before (bottom) and after (middle) addition of ActD at 1/1 ratio. The 2D-NOESY in water showing correlations between the imino protons and the amide protons is expanded in the top panel. NOE cross-peaks between the paired G8H1–T4H3 protons and the paired G3H1–T9H3 protons of the G·T mismatches are labeled with lower case letters a and b, and those between the paired G3H1–T9H3 protons and their neighboring G10H1 proton with c and d, respectively. The absence of NOE between the G·T mismatches is marked with circles.

 
Non-exchangeable DNA proton assignments
The right part of Figure 3 shows the expanded contour plot of NOESY in H2O (120 ms mixing time) of the complex correlating the DNA base proton region with the DNA sugar H1'/H3' proton region. Sequential NOE connectivities between base protons and H1'/H3' protons in the complex are traced in the bottom and top boxes, respectively. In the top box, the G10H8–T9H3' cross-peak was missing (labeled with a x), but could be detected in a longer 200 ms mixing time spectrum (data not shown). However, the T9H6–G8H3' and T4H6–G3H3' cross-peaks were not detected even at this mixing time, consistent with the proposal that the ActD chromophore is intercalated between the tandem G·T mismatches. In the lower box, similar sequential NOE connectivities between the base protons and the H1' protons could be traced, but are complicated by the presence of two rather strong H5–H6 cross-peaks of the C1 and C2 residues (labeled with bold letters), as well as the cross-peaks exhibited by the ActD phenoxazone benzenoid H7/H8 protons (labeled with lower case letters g, h and i). Unfortunately, the chemical shift of the BH8 proton overlaps with that of the T9H6, so that the strong BH8–G8H1' (g) and BH8–T9H1' (h) cross-peaks cannot be distinguished from those of the T9H6–G8H1' and T9H6–T9H1' cross-peaks. This problem was solved by collecting the NOESY at a higher temperature of 10°C. Strong BH8–G8H1' (g) and BH8–T9H1' (h) cross-peaks were still detected, but not the T9H6–G8H1' cross-peak. These critical BH7- and BH8-related NOE cross-peaks further imply that the ActD chromophore intercalates between the tandem G3·T9 and T4·G8 mismatches in only one orientation, with the benzenoid ring stacked between the G8–T9 step.



View larger version (40K):
[in this window]
[in a new window]
  		Figure 3. Expanded NOESY contour plot of the aromatic protons to H1'/H5/H3' protons of the 5'-CCGTT3GTGG-3'–ActD complex in water. The left box shows the connections of L-ThrHN- and D-ValHN-related NOE cross-peaks, while the bottom right box shows the tracing of sequential H6/H8–H1' NOE connectivities, and the top right box the tracing of sequential H6/H8–H3' NOE connectivities. The C1NH2b/C1NH2n–C1H5 and C2NH2b/C2NH2n–C2H5 NOE cross-peaks are labeled with lower case letters a, b, and c, d, respectively. Critical inter-DNA–ActD cross-peaks between the L-ThrQHN–G3H1' protons and the L-ThrBHN–G8H1' protons are labeled with e and f, and those between the ActD H8–G8H1' protons, ActD H8–T9H1' protons and ActD H7–G8H8 protons with g, h and i, respectively. These critical NOEs indicate that the ActD phenoxazone ring intercalates into the G–p–T site in essentially one single orientation, with the benzenoid side intercalating into the G8–p–T9 step and the quinoid side into the G3–p–T4 step.

 
Additional NOESY contour plots correlating the DNA base protons with the DNA sugar H2'/H2'' protons, and the DNA H1' protons with the DNA H2'/H2'' protons in the complex are shown in the left and right side of supplementary figure S1, respectively.

ActD proton assignments
The ActD molecule consists of two cyclic pentapeptide lactone rings that are covalently connected to the benzenoid or quinoid ring of the phenoxazone chromophore (Fig. 1). These two cyclic pentapeptide lactone rings are designated as B and Q, based on their covalent links to the benzenoid and quinoid rings of the chromophore, respectively. Each cyclic pentapeptide lactone ring consists of a (L-Thr)-(D-Val)-(D-Pro)-(Sar)-(L-MeVal) sequence. The amino acid spin systems in the two cyclic peptapeptide lactone rings were assigned by the analysis of through-bond coupling cross-peaks in the COSY and TOCSY spectra of the complex, while the sequential connectivities within each lactone ring were derived from the through-space NOE cross-peaks in the NOESY spectrum of the complex, as shown in supplementary figure S2.

All assigned exchangeable and non-exchangeable protons of the DNA molecule and the ActD ligand are listed in supplementary table S1.

Intermolecular ActD/DNA NOEs
Many intermolecular NOEs between ActD and the d(CCG TT3GTGG) hairpin were detected, which are critical in constraining the ActD–hairpin complex structure formation during the molecular dynamics calculation. A complete set of intermolecular NOEs between the ActD and the d(CCG TT3GTGG) hairpin is listed in supplementary table S3. These NOEs originate chiefly from interaction between the ActD protons and the protons in the tandem G·T mismatches.

Structural features of the ActD–hairpin complex
Stereo views of the final 10 superimposed structures of the ActD–d(CCGTT3GTGG) complex in two different perspectives are shown in Figure 4. The presence of tandem G·T mismatches at the stem region and the T3 loop next to the intercalation site does not appear to affect the binding of ActD to DNA significantly. In fact, a rather strong binding constant of up to 8 x 106 M–1 was observed for this complex (19). As usual, phenoxazone ring stacks selectively between the G3/G8 bases, and strong interactions between the L-ThrQ and L-ThrB residues of the ActD cyclic pentapeptide lactone rings and the G3/G8 residues were observed. These are revealed by the strong NOEs detected between the G3H1'/G8H1' protons and the L-ThrQHN/L-ThrBHN amide protons (cross-peaks e and f in Fig. 3) as well as between the G3H1'/G8H1' protons and the L-ThrQCH3/L-ThrBCH3 protons (cross-peaks b and a in supplementary fig. S1), respectively. Furthermore, hydrogen bonds between the D-ValQHN amide proton and the D-ValB carbonyl oxygen atom as well as between the D-ValBHN amide proton and the D-ValQ carbonyl oxygen atom were also found. This is consistent with the very slow H–D exchange rate of the D-ValQHN and D-ValBHN amide protons, which were detectable even after the complex sample was exchanged into D2O buffer for a week (data not shown). These H-bonds help stabilize the relative positions of the two cyclic pentapeptide lactone rings.



View larger version (64K):
[in this window]
[in a new window]
  		Figure 4. Stereo views of the final 10 structures of the ActD–d(CCG TT3GTGG) complex shown along the phosphate backbone with the major groove exposed on the left (top). ActD protons are drawn in yellow, and DNA hairpin protons in blue, except for the three loop thymine bases, which are in pink. A stereo view looking into the minor groove of the DNA hairpin is shown in the lower panel.

 
However, there are also some unique features of this ActD–hairpin complex structure. First, the TTT loop conformation of the DNA hairpin in the complex is very different from that of the free DNA hairpin. In a free DNA hairpin containing either a TTT or TCC loop, the first thymidine has the potential to loop into the minor groove and interact with the stem base pairs (28–30). In the current complex, however, the DNA minor groove is occupied by the ActD cyclic pentapeptide lactone rings, and hence is not available for interacting with the looped thymidine. The three loop thymidine have, instead, adopted an alternative stacked conformation, with the bases pointing toward the major groove (in pink in Fig. 4), and the loop backbone toward the minor groove to interact with the L-Pro, Sar and N-MeVal residues of the quinoid cyclic pentapeptide rings of ActD (lower panel of Fig. 4). These hydrophobic interactions may enhance the stability of ActD–hairpin binding and help explain the observation that the hairpin loop interacts significantly with the cyclic pentapeptide rings during the drug binding (31). The detailed overlapping pattern of the T5–T6–T7 loop bases with adjacent T4·G8 mismatch is shown at the top of Figure 5. From the figure it is clear that, while it is the base of the T5 residue that exhibits a good overlap with the T4 base, it is, on the other hand, the deoxyribose of the T7 residue that exhibits some partial stacking with the G8 base. Such a conformation is also consistent with the observation of unusual upfield shifting of the T7H4' proton to the ~3 p.p.m. region, because the T7H4' proton is now located directly above the G8 base and experiences an upfield shifting of ~0.5 p.p.m. due to the ring current-shifting effect of the guanine base (32,33).



View larger version (55K):
[in this window]
[in a new window]
  		Figure 5. Stereo views of the overlap of loop T5–T6–T7 bases with the T4·G8 mismatch (in pink) at the stem–loop junction (top) and between the phenoxazone chromophore of ActD (in red) and the flanking G3·T9 (in pink) and T4·G8 (in blue) mismatches. Good T4–T5–T6–T7 base stacking is observed, with the T7C4' (marked by a brown arrow) and its connected T7H4' situated directly above the G8 base. This is consistent with a ring current shifting of ~0.5 p.p.m. experienced by the T7H4' proton. The bottom part shows the excellent hydrophobic interaction of the G3–phenoxazone– G8 bases. The six-membered rings of G8, phenoxazone and G3 are stacked almost on top of each other. The T4 and T9 bases of the T4·G8 and G3·T9 mismatches are, on the other hand, completely de-stacked from the hydrophobic core. This is different from the observed stacking pattern in the crystal structure of tandem GT mismatches (34), in which excellent intra-strand G/T stacking is detected.

 
Secondly, intercalation of the phenoxazone ring into the tandem G·T mismatches has caused considerable changes in their stacking pattern as compared with that in a free B-DNA environment (34). In a pure DNA structure containing tandem G·T mismatches, excellent intra-strand guanine/thymine base stacking was observed. However, it is not the case for the present ActD–(GT)2 complex structure, in which the six-membered rings of G8, phenoxazone and G3 were found to stack almost on top of each other, while the bases of their respective binding partners T4 and T9 are completely de-stacked from the hydrophobic core (bottom of Fig. 5). There is thus considerable relative shifting of guanine bases toward each other to form a strong G–phenoxazone–G stack at the cost of de-stacking the original intra-strand G–T stacking when the ActD chromophore is inserted between the tandem G·T mismatches.

Structural comparison between the ActD–(GC)2 and ActD–(GT)2 complex motifs
Since the ActD–(GC)2 complex structure has been well studied before (3), it is thus of interest to compare the present ActD–(GT)2 complex structure with that of the ActD–(GC)2 complex. Figure 6 shows the superimposition of one of the G·C and G·T base pairs flanking the ActD phenoxazone chromophore of the ActD–(GC)2 and ActD–(GT)2 complex structures. It is clear that due to the different hydrogen bonding schemes adopted by the G·C and G·T base pairs, the thymidine residue has been displaced considerably toward the major groove in the ActD–(GT)2 complex when the guanine bases are superimposed. The thymine 2O atom is also found to form a bipartite H-bond with the guanine H1 and 2NH2 atoms in the final molecular dynamics refined complex structure. These may help stabilize the G·T mismatches in the complex. Such displacements have changed the binding surface of the G–chromophore–G hydrophobic stack to some extent, as shown in Figure 7 in space-filling plots. Although the ActD–(GC)2 and ActD–(GT)2 complex structures are similar overall, both with a thymidine or cytosine tilted toward the G–chromophore–G hydrophobic core, the stacking surface in the ActD–(GT)2 complex is more robust, with the six-membered rings of the phenoxazone and the flanking guanine bases almost stacking on top of each other (see also the lower half of Fig. 5). This may explain the stronger binding of ActD to the 5'-d(CCGTTTTGTGG) sequence than to the 5'-d(CCGCTTTGCGG) sequence (Ka of 8.2 ± 1.5 x 106 M–1 and 3.1 ± 0.5 x 106 M–1, respectively).



View larger version (17K):
[in this window]
[in a new window]
  		Figure 6. Superimposition of the G·C (in blue) and G·T (in pink) base pairs flanking the ActD phenoxazone chromophore in the ActD–(GC)2 and ActD–(GT)2 complexes. The coordinates of the ActD–(GC)2 complex were taken from the literature (3). Hydrogen bonds between the base pairs are connected by blue and pink dotted lines, respectively. It is clear that the thymidine residue is displaced toward the major groove to a larger extent compared with the cytidine residue when the guanine bases are superimposed, due to the different hydrogen bonding schemes adopted by the G·C and G·T base pairs. A bipartite H-bond between the thymine 2O atom and the guanine H1 and 2NH2 atoms possibly forms in the G·T mismatch of the final ActD–(GT)2 complex.

 


View larger version (91K):
[in this window]
[in a new window]
  		Figure 7. Stereo views in space filling of the ActD–(GT)2 (top) and ActD–(GC)2 (bottom) complexes looking into the major groove. The coordinates of the ActD–(GC)2 complex were taken from the literature (3). Guanosine residues are drawn in red, thymidines in blue, cytosines in yellow, ActD carbons in gray and ActD hydrogens in white. These two motifs are quite similar, with a thymidine or cytosine residue tilted toward the G–chromophore–G stacking core. However, the stacking surface in the ActD(GT)2 complex is more intense than that in the ActD–(GC)2 complex (see also Fig. 5).

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
SUPPLEMENTARY MATERIAL
 REFERENCES
 
A major ActD-binding mode
The ActD–DNA complex is the prototype of drug–DNA binding and has therefore been well studied before. Due to the asymmetry of the drug chromophore (Fig. 1), however, NMR studies of these ActD–DNA complexes have been difficult until now. Even with self-complementary oligonucleotides, multiple bindings of ActD with DNA are still possible, which complicates detailed NMR spectral analysis (3–5). Very few well-defined NMR structures are therefore available (3,8,35). Recently, we have studied ActD binding to the 5'-GTC ACCGAC-3' and 5'-GGCACCGCC-3' hairpins closed by an ACC mini-loop, and found that only one major set of NMR signals was detected (14). In the present studies of ActD binding to the 5'-CCGTT3GTGG-3' hairpin closed by a TTT mini-loop, again only one major set of NMR signals was observed. However, it is interesting to note that two major sets of NMR signals were observed when ActD is bound to a DNA hairpin closed by a TTTT loop that is 2 bp away from the intercalation site, albeit with a different sequence context (35). Therefore, a hairpin mini-triloop that is only 1 bp away from the intercalation site seems to favor an orientation with the quinoid side of phenoxazone chromophore intercalated between the 5'-G/X step prior to the loop, and the benzenoid side intercalated between the 5'-G/X step after the loop. The reasons for this orientation specificity are, however, still unclear. More studies about ActD binding to a series of DNA hairpins of different stem and loop sequences are necessary to clarify this situation.

Diversified ActD-binding sites
There is renewed interest in studying ActD–DNA binding due to the finding that ActD is able to bind to ssDNA and inhibit replication of retroviruses, including HIV (16). Such ssDNAs usually form hairpins with non-canonical base pairs at the stem region that exhibit specific interaction with ActD (15,18). Since these drug–hairpin complexes can be quite stable, with a Kd in the micromolar range, it has been suggested recently that ActD acts by binding to the hairpin structures formed from the template DNA strands and hence restricts polymerase in traveling along the template DNA strands during DNA replication or RNA transcription (16,18). Since ActD is the first small molecule drug capable of binding ssDNA with both high affinity and sequence specificity, there is hope that detailed studies of such unusual ActD–hairpin complexes may open up the possibility of designing novel drugs unique to viral infections (16,18).

We recently have determined the novel three-dimensional structures of ActD–hairpin complexes containing a (GXC)2 binding site adjacent to an ACC mini-loop (14). In the present report, we have shown further that ActD is able to interact strongly with a ssDNA hairpin containing a (GT)2 binding site and an adjacent T3 loop. These findings, together with the recent NMR and spectroscopic observations that ActD is capable of binding to the 5'-(GC)2 sites with adjacent A·A mismatch (36) or to the T·T mismatches involved in the (GCA)n or (GCT)n triplet repeats (26,36,37), should provide valuable information regarding unusual ActD–DNA action. Other special ActD–DNA binding modes are currently under investigation and will be reported at a later date.


   SUPPLEMENTARY MATERIAL
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY MATERIAL
REFERENCES
 
Supplementary Material is available at NAR Online.


   ACKNOWLEDGEMENTS
 
We thank the National Science Council and the Chung-Zhen Agricultural Foundation Society of Taiwan, ROC for the instrumentation grants. S.H.C. is the recipient of the outstanding research award of NSC and the outstanding research scholarship of the Chung-Shan Foundation. This work was supported by NSC grants 91-2113-M-005-014 to S.H.C. and a subproject of MBRS grant S06GM0892 to F.M.C.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY MATERIAL
 REFERENCES
 

  1. Waring,M.J. (1981) DNA modification and cancer. Annu. Rev. Biochem., 50, 159–192.[CrossRef][Web of Science][Medline]

  2. Chen,F.-M. (1992) Binding specificities of actinomycin D to non-self-complementary -XGCY- tetranucleotide sequences. Biochemistry, 31, 6223–6228.[CrossRef][Medline]

  3. Liu,X., Chen,H. and Patel,D.J. (1991) Solution structure of actinoymcin–DNA complexes: drug intercalation at isolated G–C sites. J. Biomol. NMR, 1, 323–347.[CrossRef][Medline]

  4. Scott,E.V., Zon,G., Marzilli,L.G. and Wilson,W.D. (1988) 2D NMR investigation of the binding of the anticancer drug actinomycin D to duplexed dATGCGCAT: conformational features of the unique 2:1 adduct. Biochemistry, 27, 7940–7951.[CrossRef][Medline]

  5. Scott,E.V., Jones,R.L., Banville,D.L., Zon,G., Marzilli,L.G. and Wilson,W.D. (1988) 1H and 31P NMR investigations of actinomycin D binding specificity with oligonucleotides containing multiple adjacent d(GC) sites. Biochemistry, 27, 915–923.[CrossRef][Medline]

  6. Kamitori,S. and Takusagawa,F. (1992) Crystal structure of the 2:1 complex between d(GAAGCTTC) and the anticancer drug actinomycin D. J. Mol. Biol., 225, 445–456.[CrossRef][Web of Science][Medline]

  7. Kamitori,S. and Takusagawa,F. (1994) Multiple binding modes of anti-cancer drug actinomycin D: X-ray, molecular modeling and spectroscopic studies of d(GAAGCTTC)2–actinomycin D complexes and its host DNA. J. Am. Chem. Soc., 116, 4154–4165.

  8. Chen,H., Liu,X. and Patel,D.J. (1996) DNA bending and unwinding associated with actinoymcin D antibiotics bound to partially overlapping sites on DNA. J. Mol. Biol., 258, 457–479.[CrossRef][Web of Science][Medline]

  9. Snyder,J.G., Hartman,N.G., D’Estantott,B.L., Kennard,O., Remeta,D.P. and Breslauer,K.J. (1989) Binding of actinomycin D to DNA: evidence for a nonclassical high-affinity binding mode that does not require GpC sites. Proc. Natl Acad. Sci. USA, 86, 3968–3972.[Abstract/Free Full Text]

  10. Chen,F.-M. (1998) Binding of actinomycin D to DNA oligomers of CXG trinucleotide repeats. Biochemistry, 37, 3955–3964.[CrossRef][Medline]

  11. Chen,F.-M. and Sha,F. (2001) Actinomycin D binds strongly to d(TGTCATTG), a single-stranded DNA devoid of GpC sites. Biochemistry, 40, 5218–5225.[Medline]

  12. Sha,F. and Chen,F.-M. (2000) Actinomycin D binds strongly to d(CGACGACG) and d(CGTCGTCG). Biophys. J., 79, 2095–2104.[Medline]

  13. Robinson,H., Gao,Y.-G., Yang,X.-L., Sanishvili,R., Joachimiak,A. and Wang,A.H.-J. (2001) Crystallographic analysis of a novel complex of actinomycin D bound to the DNA decamer CGATCGATCG. Biochemistry, 40, 5587–5592.[CrossRef][Medline]

  14. Chou,S.-H., Chin,K.-H. and Chen,F.-M. (2002) Looped out and perpendicular: deformation of Watson–Crick base pair associated with actinomycin D binding. Proc. Natl Acad. Sci. USA, 99, 6625–6630.[Abstract/Free Full Text]

  15. Bailey,S.A., Graves,D.E. and Rill,R. (1994) Binding of actinomycin D to the T(G)nT motif of double-stranded DNA: determination of the guanine requirement in nonclassical, non-GpC binding sites. Biochemistry, 33, 11493–11500.[CrossRef][Medline]

  16. Rill,R.L. and Hecker,K.H. (1996) Sequence-specific actinomycin D binding to single-stranded DNA inhibits HIV reverse transcriptase and other polymerases. Biochemistry, 35, 3525–3533.[CrossRef][Medline]

  17. Wadkins,R.M., Jares-Erijman,E.A., Klement,R., Rudiger,A. and Jovin,T.M. (1996) Actinomycin D binding to single-stranded DNA: sequence specificity and hemi-intercalation model from fluorescence and 1H NMR spectroscopy. J. Mol. Biol., 262, 53–68.[CrossRef][Web of Science][Medline]

  18. Wadkins,R.M., Vladu,B. and Tung,C.-S. (1998) Actinomycin D binding to metastable hairpins in single-stranded DNA. Biochemistry, 37, 11915–11923.[CrossRef][Medline]

  19. Chen,F.-M., Sha,F., Chin,K.-H. and Chou,S.-H. (2003) Binding of actinomycin D to single-stranded DNA of sequence motifs d(TGTCTnG) and d(TGTnGTCT). Biophys. J., 84, 432–439.[Medline]

  20. Tseng,Y.-Y. and Chou,S.-H. (1999) Systematic NMR assignment pathways for DNA exchangeable protons. J. Chin. Chem. Soc., 46, 699–706.

  21. Plateau,P. and Gueron,M. (1982) Exchangeable proton NMR without base-line distortion, using new strong-pulse sequences. J. Am. Chem. Soc., 104, 7310–7311.[CrossRef]

  22. Sklenar,V., Miyashiro,H., Zon,G., Miles,H.T. and Bax,A. (1986) Assignment of the 31P and 1H resonances in oligonucleotides by two-dimensional NMR spectroscopy. FEBS Lett., 208, 94–98.[CrossRef][Web of Science][Medline]

  23. Chou,S.-H., Cheng,J.-W., Fedoroff,O. and Reid,B.R. (1994) DNA sequence GCGAATGAGC containing the human centromere core sequence GAAT forms a self-complementary duplex with sheared G:A pairs in solution. J. Mol. Biol., 241, 467–479.[CrossRef][Web of Science][Medline]

  24. Sarma,R.H., Mynott,R.J., Wood,D.J. and Hruska,F.E. (1973) Determination of the preferred conformations constrained along the C4'–C5' and C5'–O5' bonds of ß-5'-nucleotide in solution. Four-bond 31P–1H coupling. J. Am. Chem. Soc., 95, 6457–6459.[Medline]

  25. Altona,C. and Sundaralingam,M. (1972) Conformational analysis of the sugar ring in nucleosides and nucleotides. A new description using the concept of pseudorotation. J. Am. Chem. Soc., 94, 8205–8212.[CrossRef][Web of Science][Medline]

  26. Hou,M.-H., Robinson,H., Gao,Y.-G. and Wang,A.H.-J. (2002) Crystal structure of actinomycin D bound to the CTG triplet repeat sequences linked to neurological diseases. Nucleic Acids Res., 30, 4910–4917.[Abstract/Free Full Text]

  27. Hare,D., Shapiro,L. and Patel,D.J. (1986) Wobble dGdT pairing in right-handed DNA: solution conformation of the d(CGTGAATTCGCG) duplex deduced from distance geometry analysis of nuclear Overhauser effect spectra. Biochemistry, 25, 7445–7456.[CrossRef][Medline]

  28. Chou,S.-H., Tseng,Y.-Y. and Chu,B.-Y. (1999) Stable formation of a pyrimidine-rich loop hairpin in a cruciform promoter. J. Mol. Biol., 292, 309–320.[CrossRef][Web of Science][Medline]

  29. Chou,S.-H., Tseng,Y.-Y. and Chu,B.-Y. (2000) Natural abundance heteronuclear NMR studies of the T3 mini-loop hairpin in the terminal repeat of the adenoassociated virus 2. J. Biomol. NMR, 17, 1–16.[CrossRef][Web of Science][Medline]

  30. Chou,S.-H., Chin,K.-H. and Chen,C.W. (2001) Enhanced DNA folding induced by thymine-CH3 and perpendicular guanine–thymine interaction. J. Biomol. NMR, 19, 33–48.[CrossRef][Web of Science][Medline]

  31. Wadkins,R.M., Tung,C.-S., Vallone,P.M. and Benight,A.S. (2000) The role of the loop in binding of an actinoymcin D analog to hairpins formed by single-stranded DNA. Arch. Biochem. Biophys., 384, 199–203.[CrossRef][Web of Science][Medline]

  32. Chou,S.-H., Zhu,L., Gao,Z., Cheng,J.-W. and Reid,B.R. (1996) Hairpin loops consisting of single adenine residues closed by sheared A:A and G:G pairs formed by the DNA triplets AAA and GAG: solution structure of the d(GTACAAAGTAC) hairpin. J. Mol. Biol., 264, 981–1001.[CrossRef][Web of Science][Medline]

  33. Chou,S.-H. and Chin,K.-H. (2001) Quadruple-intercalated G-6 stacks: a possible motif in the fold-back structure of Drosophila centromeric dodeca-satellite? J. Mol. Biol., 314, 139–152.[CrossRef][Web of Science][Medline]

  34. Kneal,G., Brown,T., Kennard,O. and Rabinovich,D. (1985) GT base-pairs in a DNA helix: the crystal structure of d(G-G-G-G-T-C-C-C). J. Mol. Biol., 186, 805–814.[CrossRef][Web of Science][Medline]

  35. Brown,D.R., Kurz,M., Kearns,D.R. and Hsu,V.L. (1994) Formation of multiple complexes between actinomycin D and a DNA hairpin: structural characterization by multinulcear NMR. Biochemistry, 33, 651–664.[CrossRef][Medline]

  36. Lian,C., Robinson,H. and Wang,A.H.-J. (1996) Structure of actinomycin D bound with (GAAGCTTC)2 and (GATGCTTC)2 and its binding to the (CAG)n:(CTG)n triplet sequence as determined by NMR analysis. J. Am. Chem. Soc., 118, 8791–8801.

  37. Liu,C. and Chen,F.-M. (1996) Actinomycin D binds strongly and dissociates slowly at the dGpdC site with flanking T/T mismatches. Biochemistry, 35, 16346–16353.[CrossRef][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?

Related articles in Nucl. Acids Res.:

Unique actinomycin D binding to self-complementary d(CXYGGCCY'X'G) sequences: duplex disruption and binding to a nominally base-paired hairpin
Fu-Ming Chen, Feng Sha, Ko-Hsin Chin, and Shan-Ho Chou
Nucl. Acids Res. 2003 31: 4238-4246. [Abstract] [FREE Full Text]  



This article has been cited by other articles:


Home page
 Nucleic Acids ResHome page
F.-M. Chen, F. Sha, K.-H. Chin, and S.-H. Chou
The nature of actinomycin D binding to d(AACCAXYG) sequence motifs
Nucleic Acids Res., January 9, 2004; 32(1): 271 - 277.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (372K) Freely available
Right arrow Supplementary Material
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Related articles in Nucl. Acids Res.
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Chin, K.-H.
Right arrow Articles by Chou, S.-H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chin, K.-H.
Right arrow Articles by Chou, S.-H.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?