Computational Studies Support the Role of the C7-Sibirosamine

Articles pubs.acs.org/acschemicalbiology

Computational Studies Support the Role of the C7-Sibirosamine Sugar of the Pyrrolobenzodiazepine (PBD) Sibiromycin in Transcription Factor Inhibition Paul J. M. Jackson,† Colin H. James,‡ Terence C. Jenkins,§ Khondaker M. Rahman,† and David E. Thurston*,† †

Institute of Pharmaceutical Science, King’s College London, Britannia House, 7 Trinity Street, London SE1 1DB, United Kingdom The School of Pharmacy, University College London, 29−39 Brunswick Square, London WC1N 1AX, United Kindgom § School of Science, University of Greenwich, Chatham Maritime, Kent ME4 4TB, United Kingdom ‡

S Supporting Information *

ABSTRACT: The pyrrolo[2,1-c][1,4]benzodiazepines (PBDs) are a group of sequence-selective, DNA minor-groove binding agents that covalently attach to guanine residues. Originally derived from Streptomyces species, a number of naturally occurring PBD monomers exist with varying A-Ring and C2-substituents. One such agent, sibiromycin, is unusual in having a glycosyl residue (sibirosamine) at its A-Ring C7-position. It is the most cytotoxic member of the naturally occurring PBD family and has the highest DNA-binding affinity. Recently, the analogue 9-deoxysibiromyin was produced biosynthetically by Yonemoto and co-workers.1 Differing only in the loss of the A-Ring C9-hydroxyl group, it was reported to have a significantly higher DNA-binding affinity than sibiromycin based on DNA thermal denaturation studies, although these data have since been retracted.2 As deletion of the C9-OH moiety, which points toward the DNA minor groove floor, might intuitively be expected to reduce DNA-binding affinity through the loss of hydrogen bonding, we carried out molecular dynamics simulations on the interaction of both molecules with DNA over a 10 ns time-course in explicit solvent. Our results suggest that the two molecules may differ in their sequence-selectivity and that 9-deoxysibiromycin should have a lower binding affinity for certain sequences of DNA compared to sibiromycin. Our molecular dynamics results indicate that the C7sibirosamine sugar does not form hydrogen bonding interactions with groups in the DNA minor-groove wall as previously reported, but instead points orthogonally out from the minor groove where it may inhibit the approach of DNA control proteins such as transcription factors. This was confirmed through a docking study involving sibiromycin and the GAL4 transcription factor, and these results could explain the significantly enhanced cytotoxicity of sibiromycin compared to other PBD family members without bulky C7-substituents.

down-regulate the expression of specific genes critical for the survival and proliferation of cancer cells through transcription factor inhibition (e.g., NFκB13 and NF-γ10). PBD monomers have been shown by footprinting,9,18 NMR,19,20 molecular modeling,21 and X-ray crystallography22 to span three base pairs with a reported preference for orientation of their A-Rings toward the 3′-end of the covalently modified strand (i.e., A-Ring-3′).15 In the past, PBD molecules have been reported to have a thermodynamic preference for 5′Pu-G-Pu-3′sequences,15,23 although more recent data suggest that they have a kinetic preference for 5′-Py-G-Py-3′ motifs24 (where Pu = purine, Py = pyrimidine, G = reacting guanine). Synthetic PBD monomers with noncovalent DNA-interactive components joined to the C8-position of their A-Rings have been reported,25−29 and examples such as GWL-7829 and

The pyrrolo[2,1-c][1,4]benzodiazepines (PBDs) are a group of sequence-selective DNA minor-groove binding agents originally discovered in Streptomyces species.3−7 Anthramycin (1, Figure 1) was the first PBD to be isolated and studied,8 although now more than 12 naturally occurring compounds are known.3 PBDs are characterized by an electrophilic N10−C11 imine group (or the hydrated equivalent) which forms a covalent bond with a C2-NH2 group of a guanine9 in the DNA minor groove. The molecules have (S)-chirality at their C11a-position, and this provides them with the appropriate isohelicity to fit perfectly into the DNA minor groove. PBD/DNA adduct formation has been shown to inhibit a number of biological processes, including the binding of transcription factors to DNA10−14 and the function of enzymes such as endonucleases15,16 and RNA polymerase.17 In particular, transcription factors bind to specific DNA consensus sequences within gene regulatory regions, and the critical points of contact between these proteins and DNA can be blocked by sequence-specific DNA-interactive molecules such as the PBDs. Thus, PBDs can © XXXX American Chemical Society

Received: March 22, 2014 Accepted: July 15, 2014

A

dx.doi.org/10.1021/cb5002203 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

claiming that it had a significantly greater DNA-binding affinity compared to sibiromycin based on the results of a DNA thermal denaturation (Tm) assay (i.e., induced ΔTm = 22 °C versus 16 °C, respectively, after incubation for 18 h at 37 °C for a DNA/PBD molar ratio of 1:20). As the DNA-binding affinity of the PBD family is usually rationalized in terms of their ability to form intermolecular contacts with functional groups within the walls and floor of the minor groove,41 the 6 °C enhancement in Tm compared to sibiromycin is counterintuitive, as loss of the C9-OH would be expected to lead to a lower binding affinity due to fewer opportunities for H-bonding and electrostatic interactions. In 2014, the ΔTm results originally published by Yonemoto and co-workers were retracted.2 Not having access to 9-deoxysibiromycin (4), this prompted us to carry out comparative molecular dynamics simulations on the interaction of both sibiromycin (3) and 9-deoxysibiromycin (4) with double-stranded DNA in an attempt to rationalize sequence selectivity and associated cytotoxicity data. As interaction of a PBD within the minor groove of DNA is thought to occur via a multistage process, with initial noncovalent interactions allowing the PBD to recognize its binding site followed by nucleophilic attack of the N10−C11 imine by the C2-amino group of a guanine base, two sets of simulations were undertaken for each adduct. The first involved simulation of the noncovalent interactions of sibiromycin (3) and 9-deoxysibiromycin (4) with DNA, and the second focused on their covalent interactions. This approach allowed assessment of the contributions of both the noncovalent and covalent components to the DNA-binding affinity of each PBD. It also allowed the calculation of free energy of binding values. In our analysis, we examined the interactions of sibiromycin (3) and 9-deoxysibiromycin (4) with duplex DNA sequences optimized for PBD binding according to the literature9 in order to provide information about both the orientation of the PBD in the minor groove (i.e., A-Ring-3′ or A-Ring-5′) and the likely tolerability of C11(R) or C11(S) stereochemistries for each adduct. Taken together, the results suggested that for both sibiromycin and 9-deoxysibiromycin, each molecule can form four possible adducts differing in C11-stereochemistry and ARing-3′/5′-orientation with respect to the covalently modified strand, a phenomenon first observed through high-field NMR studies of tomaymycin.18,41 We also used modeling approaches to examine potential differences in sequence selectivity for sibiromycin versus 9deoxysibiromycin, concluding that sibiromycin prefers a guanine residue adjacent to the covalently modified guanine on either the reacting or complementary strands. Conversely, 9deoxysibiromycin favors an adenine/thymine doublet adjacent to the reacting guanine. Finally, according to early SAR studies and models of PBD/ DNA adducts,42,43 the C6- and C7-substituents of a PBD should point out of the minor groove, with larger groups such as sugars thought to form hydrogen bonding interactions with functional groups in the minor groove wall. Using our more advanced MD simulations (where DNA and ligand were simulated without restraint), we have been able to show that a C7-substituent of a PBD is capable of orienting in two directions, either along the floor of the minor groove or protruding orthogonally from the helix axis. This led us to the conclusion that the bulky C7-sibirosamine sugar of sibiromycin protrudes orthogonally from the minor groove and may be responsible for inhibiting the interaction of transcription factors

Figure 1. Structures of the four pyrrolobenzodiazepine (PBD) monomers 1−4.

KMR-28−3913 can span up to six or seven DNA base pairs. Furthermore, two PBD units have been joined through their C730 or C831 positions to form PBD dimers, which can form monoadducts, and inter- and intrastrand DNA cross-links. One example, SJG-136, is presently in phase II clinical trials as an anticancer agent.32 Both PBD monomers and dimers have in common the feature that their covalent interaction in the minor groove significantly enhances the stability of the DNA double helix to thermal denaturation,15,29,33 and this stabilizing property correlates well with their in vitro cytotoxicity.29,31,33 PBDs are thought to interact with DNA by first locating a lowenergy binding site (e.g., a 5′-Pu-G-Pu-3′ triplet) through van der Waals (vdW), H-bonding, and electrostatic interactions.29 Once in place, nucleophilic attack by the exocyclic C2-NH2 of the central guanine then occurs to form a covalent adduct.34 Once bound, the PBD remains anchored in the DNA host minor groove, avoiding DNA repair through negligible distortion of the helix.22,35,36 Sibanomycin37 and sibiromycin38 (2 and 3, Figure 1) differ from most other naturally occurring PBDs by the presence of a glycosyl residue (a C7-sibirosamine sugar) at their C7positions, along with a unique C2-substitution pattern.28 There is little biological (i.e., P388 mouse model) and no biophysical data available for sibanomycin, although sibiromycin remains the most cytotoxic PBD monomer to be identified to date,28 a feature assumed until now to be related to its C7sugar substituent.28 It also remains one of the few PBD monomers yet to be synthesized due to the complexity of the C7-sugar substituent. Sibiromycin has been shown to be cardiotoxic in animal models, and this has been attributed to possible free radical formation by the C9-OH group in a similar manner to anthramycin.39 The 9-methoxy analogue of anthramycin has been synthesized but is biologically inactive,40 suggesting that a C9-hydroxy is tolerated in the minor groove, but larger substituents at this position may not be (a point confirmed by modeling studies in this laboratory, unpublished data). The enhanced cytoxicity of the glycosylated PBD monomer sibiromycin over non-glycosylated members of the family prompted us to investigate its DNA-interactive behavior through molecular dynamics techniques. During these studies, in 2012 Yonemoto and co-workers reported the isolation of 9deoxysibiromycin (4, Figure 1)1 via a biosynthetic route, B



Articles

9-deoxysibiromycin docked both noncovalently and covalently (with either C11-R or C11-S stereochemistry) in both orientations in the relevant sequences (Table 1). DNA-Binding Orientation and C11 Stereochemistry. Simulations with each of Seq 1−4 (Table 1) showed that both sibiromycin and 9-deoxysibiromycin were snugly accommodated in the minor groove with either R or S stereochemistry at C11. For example, simulations of sibiromycin covalently bound to G6 of Seq 3 indicated that, as predicted, the C9-OH of the PBD formed a constant H-bond with the exocyclic C2-NH2 group of the adjacent G7 for both the C11(R) and C11(S) adducts. A second H-bond was also observed between C9−OH of the PBD and the ring N3 nitrogen of G7 (Figure 3, upper panel) (see H-bond lifetime plots in S1 of Supporting Information). When covalently bound in the C11(R) configuration, the H-bond interactions between C9-OH and G7 were more transient in nature throughout the simulation (see H-bond lifetime plots in S2, Supporting Information) compared to those for the C11(S) adduct, suggesting a preference for the S configuration with R tolerated. In the case of C11(R)-stereochemistry for Seq 4, 9deoxysibiromycin formed a single transient H-bond between N10−H of the PBD and the N3 ring nitrogen of A7 (S3, Supporting Information), with similar less-transient H-bonds forming with C17 in the S adduct (S4, Supporting Information). These results suggested that either stereochemistry at C11 could be accommodated with the molecule covalently bound in the A-Ring-3′ orientation. Simulations of 9-deoxysibiromycin bound to Seq 1 and Seq 3 in the A-Ring-5′ orientation with either C11(R) or C11(S) stereochemistry suggested that this arrangement is also well accommodated with negligible DNA distortion, but no Hbonding interactions were observed in any of these adducts. Overall, these observations suggested that the A-Ring-3′ orientation is preferred, with the A-Ring-5′ orientation tolerated. Simulations of sibiromycin bound in the A-Ring-5′ orientation in Seq 1 and Seq 2 indicated that both S and R configurations should be possible. When covalently bound in the S configuration, the C9-OH did not interact with G5 as predicted, and did not form any contacts with other bases in the minor groove. For the R stereochemistry a similar situation was observed, with the C9-OH or N10−H of the PBD not producing any H-bonding interactions with the DNA groove, but the C9-OH of sibiromycin laying closer to the groove floor (S5 and S6, Supporting Information). This suggested that both

(e.g., GAL4) with DNA as reported by Khokhlov and coworkers.44 This steric inhibition of binding of transcription factors may explain why sibiromycin, and potentially other glycosylated PBDs, are significantly more cytotoxic than PBDs that lack a bulky C7-substituent.

■

RESULTS AND DISCUSSION As PBDs such as sibiromycin are known to have a thermodynamic preference for Pu-G-Pu motifs,24,34,45 modeling studies were initially restricted to DNA sequences containing a Pu-G-Pu PBD-binding triplet so that the sequence selectivity of sibiromycin and 9-deoxysibiromycin could be evaluated and compared in both the A-Ring-3′ and A-Ring-5′ directions, and with different C11-stereochemistries (Figure 2). It was initially

Figure 2. Two possible binding orientations (top panel, A-Ring-3′; bottom panel, A-Ring-5′) for PBDs (3 and 4, Figure 1) covalently bound to the DNA duplexes shown in Table 1. R9 is either OH (sibiromycin, 3) or H (9-deoxysibiromycin, 4), R7 is the sibirosamine sugar, Z is any base (A, G, T, or C), X is A, G, or T, and Y is the complementary base to X (either T, C or A). Z−Z pairings are any combination of A−T, T−A, G−C or C−G.

postulated that, due to the C9-OH group of sibiromycin, binding should favor a triplet with a guanine base situated below the A-Ring of the molecule adjacent to the covalently modified guanine in a similar manner to anthramycin.22 As 9deoxysibiromycin lacks this C9 functionality, it was initially assumed to have a lower degree of selectivity for 5′-AGG-3′ or 5′-GGG-3′ versus 5′-AGA-3′ in the A-Ring-3′ direction and for 5′-GGA-3′ versus 5′-AGA-3′ in the A-Ring-5′ direction. Therefore, simulations were undertaken with sibiromycin and

Table 1. Free Energy of Binding Values (kcal/mol) Calculated for Sibiromycin (3) and 9-Deoxysibiromycin (4) Noncovalently Bound to DNA Duplex Sequences Seq 1 to Seq 8a sibiromycin (3) (kcal/mol) label Seq Seq Seq Seq Seq Seq Seq Seq

1 2 3 4 5 6 7 8

9-deoxysibiromycin (4) (kcal/mol)

duplex sequence

A-Ring-3′

A-Ring-5′

A-Ring-3′

A-Ring-5′

5′-TATAGGGTATA-3′ 5′-TATAGGATATA-3′ 5′-TATAAGGTATA-3′ 5′-TATAAGATATA-3′ 5′-TATAAGTTATA-3′ 5′-TATAAGCTATA-3′ 5′-TATATGGTATA-3′ 5′-TATATGATATA-3′

−26.07 −23.17 −25.12 −24.87 −23.52 −24.73 −25.05 −24.85

−23.92 −25.73 −23.69 −25.01 −20.32 −23.45 −20.52 −21.76

−24.93 −23.87 −22.35 −25.08 −25.81 −21.44 −24.67 −26.37

−23.37 −21.89 −23.74 −26.77 −25.79 −23.52 −21.75 −25.74

a

Only one strand is shown for each sequence, and the triplet PBD docking sites are underlined. The oligonucleotides were designed to have varying base pairs adjacent to the central covalently-reacting guanine. Base pairs are numbered from the 5′-end of each oligonucleotide shown. C



Articles

preferred form of the adduct,22,46−48 R is also likely in many cases.19,46−48 Therefore, the formation of a C11(R)- or C11(S)PBD adduct is likely to depend on both the individual structural characteristics of the PBD, and the DNA sequence to which the molecule is bound. Electronic calculations were also conducted on free sibiromycin and 9-deoxysibiromycin to ascertain whether the presence or absence of a C9-OH group could affect the electrophilicity of the N10−C11 imine functionality, and thus the rate of reaction with DNA. These calculations were considered important, as a PBD molecule with full C-ring unsaturation was previously shown to have negligible DNAbinding affinity due, in part, to electron redistribution effects causing a low electrophilicity at the C11-position.49 Conversely, PBDs with C7 and C8 electron-donating groups (for example sibiromycin) are known to be more electrophilic at the C11position than those without them, and are thus more DNAinteractive.48 However, negligible differences in electronic charge distributions were observed between sibiromycin (3) and 9-deoxysibiromycin (4) suggesting that differences in DNA binding affinity between them are most likely due to structural rather than electronic features. In their 2012 publication,1 Yonemoto and co-workers reported that, in the calf thymus DNA melting assay routinely used to evaluate the DNA binding affinity of PBDs,13 9deoxysibiromycin (4) showed a significantly greater bindinginduced stabilization (ΔTm = 22 °C) of DNA toward melting than either anthramycin (1) or sibiromycin (3) (ΔTm = 13 and 16 °C, respectively). On this basis, they concluded that removal of the 9-OH group increased rather than reduced DNA-binding affinity. However, on the basis of our MD simulations, we found no evidence to support the notion that 9-deoxysibiromycin induces a higher ΔTm value compared to sibiromycin. In contrast, our results suggest that deletion of the 9-OH group from sibiromycin should lead to a loss of favorable H-bonding, particularly to guanine bases adjacent to the covalently modified guanine, thus leading to decreased stabilization of the host duplex. It is noteworthy that of all the naturally occurring PBD monomers, sibiromycin and anthramycin have the highest reported ΔTm values (16.3 and 13.0 °C, respectively),15 and both have C9-hydroxy groups. The next highest is tomaymycin (2.6 °C) which lacks a C9-OH, and all other PBD monomers without C9-substituents have Tm values lower than this. This prompted us to look carefully at the methodology used by Yonemoto and co-workers1 to carry out their ΔTm measurements. The reported ΔTm experiments were carried out with a final concentration molar ratio of DNA/PBD of 1:20 (i.e., with the PBD in excess), and measurements were taken after 18 h incubation at 37 °C. However, these results are inconsistent with literature reports,15,28,50 as anthramycin and sibiromycin are known to induce ΔTm shifts of 13.0 and 16.3 °C, (18 h/37 °C incubation), respectively, but at a molar DNA/PBD ratio of 5:1 (i.e., with the DNA in excess). Based on the 100-fold greater molar concentration of PBD relative to DNA used in the experiments by Yonemoto and co-workers,1 it is unclear why the ΔTm shift values for the two PBDs match the literature values despite the contrast in molar ratios. Yonemoto and coworkers withdrew the ΔTm component of their 2012 publication in 2014.2 Our simulations predicted that deletion of the C9-OH group of sibiromycin should lead to fewer favorable H-bonded interactions in the minor groove and a lower DNA sequence

Figure 3. Upper Panel: A low energy snapshot from an MD simulation of sibiromycin covalently bound in the C11(S) A-ring-3′ configuration to G6 (magenta) of 5′-TATAAGGTATA-3′ (Seq 3), showing the ligand perfectly accommodated in the DNA minor groove. Two Hbonds (yellow) were observed between the C9-OH and N3 of G6, and the C9−O and exocyclic C2-NH2 of guanine G7 (green) adjacent to the covalently modified guanine. Lower Panel: A low energy snapshot from an MD simulation of 9-deoxysibiromycin covalently bound in the C11(S) A-ring-3′ configuration to G6 (magenta) of 5′-TATAAGATATA-3′ (Seq 4), showing the ligand perfectly accommodated in the minor groove. An H-bond (yellow) was observed between the N10−H of the PBD and ring nitrogen of A7 (orange). In both models, the C7sibirosamine sugar (black box) is positioned above the plane of the minor groove (gray ribbon) at a 90° angle to the duplex axis.

R and S stereochemistries may be tolerated for the A-Ring-5′ orientation, with R preferred. In summary, the results obtained from these simulations suggested that both orientations and C11 stereochemistries should provide stable adducts of both sibiromycin and 9deoxysibiromycin with Seq 1−4. This is in accord with previous studies on PBD monomer adducts which have shown that although NMR, molecular mechanics calculations, and crystallographic analyses indicate that S is a potentially D



Articles

orientation due to stabilizing vdW interactions between C9−H and T6/A18. Taken together, these results suggest that, when docked ARing-3′, sibiromycin prefers a guanine residue on the 3′-end of the triplet on either strand of DNA, and thus has a preference for 5′-X-G-G/C-3, where X is any base. Conversely, 9deoxysibiromycin docked in the same direction prefers an adenine group on the 3′-end of the reacting triplet on either DNA strand, thus preferring 5′-X-G-A/T-3′. In the A-Ring-5′ orientation, sibiromycin prefers the sequence 5′-G/C-G-X-3′, with 9-deoxysibiromycin showing selectivity for 5′-A/T-G-X-3′. Overall, simulations and free energy of binding calculations suggest that the presence of a G/C or C/G base-pair below the A-Ring is preferred in both orientations for sibiromycin, with 9deoxysibiromycin preferring an A/T or T/A base-pair under its A-Ring in both directions. Furthermore, for every case except Seq 1 and Seq 4, free energy of binding calculations suggest that interaction of sibiromycin in the A-Ring-3′ orientation is preferred to A-Ring-5′, supporting previous literature reports on PBDs.15 Orientation of the C7-Sugar Substituent. According to early studies based on CPK models, the C7-sibirosamine substituent of sibiromycin was thought to protrude from the minor groove but still form H-bonds with charged phosphate oxygen atoms of the DNA backbone,9,52 thereby stabilizing the DNA/PBD adduct. Later studies on C7-aryl PBDs reinforced this concept.43 To investigate this further, MD simulations were undertaken at increased temperature (343 K) to simulate the environment used in DNA thermal denaturation studies.15 Interestingly, during 10 ns explicit solvent simulations, the C7sibirosamine sugar group formed transient electrostatic interactions with the DNA minor-groove wall enabled by significant movement (including rotation) of the sibirosamine sugar. These electrostatic interactions, which were absent in baseline simulations at 310 K, may explain the greater helix stabilization observed for sibiromycin in thermal denaturation studies (i.e., 16.3 °C for sibiromycin) compared to the PBD with the next lowest value (i.e., ΔTm = 13.0 °C for anthramycin). We also investigated the possibility of water-mediated Hbonding interactions between the sibirosamine sugar and minor groove wall. Water molecules close to the ligand were determined using a 3 Å distance cutoff. Although, as expected, water molecules were observed in the region of the ligand/ DNA complex, these solvent molecules did not, at any point, form water-mediated stabilizing H-bonding interactions between the sugar moiety and minor groove wall. Importantly, in the more advanced MD simulations reported here, we observed the potential for C7-substituents to orient in two directions by pivoting on the C7−O-dihedral without forming stabilizing H-bonds. To further explore this, we conducted simulations of a DNA adduct of a PBD monomer containing a C7-oxypropanoic acid substituent (i.e., −OCH2CH2CH2COOH) starting with two potential orientations (i.e., toward or distant from the DNA minor groove floor). In these simulations both adducts were found to have similar energies suggesting that both could form. However, in similar simulations of a DNA-sibiromycin adduct starting with the C7-sibirosamine substituent pointing toward the minor groove floor, the C7-sugar rotated away from the groove floor pivoting on its C7−O-dihedral until it was pointing out of the minor groove at an approximate angle of 90° to the helix axis. This rotation of the bulky C7-substituent out of the minor

selectivity compared to sibiromycin. In this context the different proportions of A/T (59.3%) and G/C base pairs (40.7%) in the human genome51 may provide a greater number of potential 9deoxysibiromycin binding sites compared to sibiromycin, and it is this promiscuity which may contribute to the small but significant increase in cytotoxicity for 9-deoxysibiromycin versus sibiromycin (i.e., average GI50 [MID] of 24.5 nM versus 114 nM, respectively, across all cells lines in the NCI 60 Panel).1 Alternatively, this 4.6-fold difference could be due to a number of other factors including a small improvement (i.e., Log P = +0.39) in lipophilicity for 9-deoxysibiromycin compared to sibiromycin (i.e., Log P values = 1.26 versus 0.87, respectively, calculated using ChemBioOffice 13) with potentially improved cell penetration properties. Sequence-Selectivity. Simulations were initially undertaken with selected Pu-G-Pu (Pu = Purine) triplets, as these are considered the most thermodynamically favored sites for PBD binding.18 Free energy of binding calculations for sibiromycin when noncovalently bound in the A-Ring-3′ orientation suggested that a guanine base is favored on the 3′-side of the reacting triplet (i.e., 5′-Pu-G-G-3′) (see Table 1) due to Hbonding contacts. This effect is evident in Seq 1 where there is a −1.14 kcal/mol difference in binding affinity favoring sibiromycin compared to 9-deoxysibiromycin. This is further evident in calculations for Seq 3, where sibiromycin is favored by −2.77 kcal/mol. Conversely, calculations for 9-deoxysibiromycin in the ARing-3′ orientation showed a slight preference for the sequence 5′-Pu-G-A-3′ compared to sibiromycin, with differences of −0.70 and −0.21 kcal/mol for Seq 2 and Seq 4, respectively. This may be explained by the absence of the C9-OH, allowing greater vdW interactions between the A-Ring of the PBD and an adenine/thymine base-pair (i.e., A7/T16) below it. This suggests that 9-deoxysibiromycin may show a preference for 5′Pu-G-A-3′ sequences when noncovalently bound in the A-Ring3′ orientation. In order to study further the significance of a guanine or adenine base under the A-Ring of sibiromycin or 9deoxysibiromycin, simulations were carried out on DNA fragments 11 base-pairs in length containing a single PBD binding site. These sequences consisted of two TATA flanking regions and a central binding triplet: 5′-AGT-3′ (Seq 5), 5′AGC-3′ (Seq 6), 5′-TGG-3′ (Seq 7) and 5′-TGA-3′ (Seq 8). For sibiromycin binding to Seq 6 with A-Ring-3′ orientation, a preference of −3.29 kcal/mol was observed (compared to 9deoxysibiromycin), with a H-bond forming between the C9OH and the C2-NH2 group of G16 (S7, Supporting Information). Similarly, calculations suggested that 9-deoxysibiromycin should favor Seq 5 compared to sibiromycin by −2.33 kcal/mol in the A-Ring-3′ orientation due to favorable vdW contacts between the A-Ring and the T7/A16 base-pair. Simulations of the interaction of 3 and 4 at the thyminecontaining triplet binding sites within Seq 7 and Seq 8 further supported the observations above, with free energy of binding differences suggesting that sibiromycin prefers to interact in the A-Ring-3′ direction in the case of Seq 7 (by −0.38 kcal/mol) due to the formation of a sequence-selective H-bond between C9-OH of the PBD and G7. In the A-Ring-5′ orientation, 9deoxysibiromycin is preferred when interacting with Seq 7 due to the formation of stabilizing nonbonded interactions between C9−H of the PBD and the T6/A18 doublet. Identical interactions were observed for Seq 8, with a −1.52 kcal/mol difference in favor of 9-deoxysibiromycin in the A-Ring-5′ E



Articles

this observation suggests that it may prove possible to increase the inhibitory activity of PBDs as a class by adding bulky substituents to the C7-position. Conclusion. We report the first molecular dynamics simulation of the interaction of sibiromycin and its 9-deoxy analogue with double-stranded DNA. The results predict that the C7-sibirosamine substituents of both sibiromycin and 9deoxysibiromycin protrude from the minor groove at an angle of ∼90° and have no significant nonbonded contacts with the floor or walls of the groove either directly or through watermediated H-bonds. This suggests that the unusually high cytotoxicity of sibiromycin compared to the other naturally occurring PBD monomers could be due to the positioning of its C7-sugar outside of the minor groove, thus causing steric interaction with approaching DNA control proteins such as transcription factors. In this context, we have modeled the reported44 inhibition of the DNA interaction of the GAL4 transcription factor by sibiromycin, demonstrating that the C7subsituent may block or impede approach of the protein by protruding out of the minor groove. While it is possible that the C7-sugar of sibiromycin may also contribute to its potency by enhancing cellular penetration, we are presently pursuing further physical evidence for the possible steric inhibition of approaching proteins. Furthermore, through rotation of the C7−O-dihedral, smaller C7-substituents such as methylene chains can be accommodated within the minor groove which explains the reported cross-linking activity of the C7-linked PBD dimers.30 This contrasts with early studies based on CPK models which suggested that C7-substituents can only point out of the groove, with the C7-sugar of sibiromycin participating in Hbonding and other interactions to stabilize the adduct. The results reported here also highlight the likely role of the C9-OH group of sibiromycin in influencing its DNA binding affinity and sequence-selectivity, with a preference for sequences with a guanine base immediately under the A-Ring to which the C9-OH group can H-bond (i.e., 5′-X-G-G/C-3′ or 5′-G/C-G-X-3′ when bound in the A-Ring-3′ or A-Ring-5′ orientations, respectively). Our simulations predict that removal of the C9-hydroxy substituent (i.e., 9-deoxysibiromycin) should reduce overall DNA-binding affinity and modify sequence-selectivity, with vdW interactions between C9−H and the minor groove floor leading to a preference for sequences with an adenine-thymine base-pair immediately below the ARing. In this context, our simulations support the retraction2 of the ΔTm data originally reported by Yonemoto and coworkers.1 The improved understanding of the interaction of PBD molecules with DNA, as reported here, may help in the design of future generations of PBD-based transcription factor inhibitors.

groove was prompted by steric interactions between the sugar and both the C8-methyl substituent of the PBD and functional groups within the floor of the minor groove. This model is consistent with the DNA cross-linking activity of the C7-linked PBD dimers reported by Suggs and co-workers30 in which the C7-polymethylene linker joining the two PBD units would have to run parallel to the minor groove for interstrand cross-linking to occur. It follows that C8-linked PBD dimers are more-potent cross-linking agents31,53 as their C8/C8′-linkers can follow the floor of the minor groove more closely. The increased cytotoxic potency of sibiromycin compared to other members of the PBD family has been previously attributed to possible enhancement of cellular penetration by the C7-sugar moiety.15 However, Khokhlov and co-workers44 have shown that sibiromycin can displace a peptide from a 19bp oligonucleotide representing the consensus recognition sequence of the transcription factor GAL4.44 This prompted us to question whether the C7-sugar of sibiromycin may protrude from the minor groove sufficiently to inhibit the approach of a DNA control protein such as a transcription factor, which could also account for its superior cytotoxicity compared to other natural PBD monomers. Thus, we undertook a docking study based on the crystal structure of the GAL4 homodimer bound to its consensus sequence (5′-ACCGGAGGACAGTCCGG3′), with sibiromycin covalently bound at each conceivable guanine docking point (i.e., 5′-CCG-3′, 5′-CGG-3′, 5′-AGG-3′, 5′-GGA-3′, 5′-ACA-3′, 5′-AGG-3′, 5′-GGA-3′, 5′-AGT-3′, 5′TCC-3′, 5′-CCG-3′ and 5′-CGG-3′). The drug/DNA adducts were energy minimized using AMBER 11, and then reintegrated into the crystal structure. In this model, the C7sibirosamine substituent of sibiromycin when docked at the 5′AGT-3′ site directly interacted with Arg 46 of homodimer B (red surface, Figure 4) which could prevent binding of the protein and explain the potent cytotoxicity observed for sibiromycin. There is growing interest in the use of PBD-type molecules as selective transcription factor inhibitors,10,54−56 so

■

METHODS

Molecular dynamics simulations were conducted using AMBER (v11)57 software. Each DNA sequence was constructed using nab, and antechamber was used to convert the structures to mol2 files with the application of Gasteiger charges. The parameters for each ligand were derived using the antechamber module of AMBER. Antechamber addresses the following issues during the MM calculations: recognizing the atom type, recognizing bond type, judging the atomic equivalence, and missing parameter generation. Further missing parameters were then generated using parmchk, where the gaff.dat force-field was used to facilitate this process, and nonbonded parameters were assigned using parm99.dat.

Figure 4. Molecular model of the GAL4 transcription factor (green) interacting with its consensus DNA sequence (gray), showing the potential steric interaction between the Arg 46 residue (red) of the protein and the C7-sibirosamine sugar (black box) of sibiromycin (blue) which protrudes at 90° to the helix axis. The drug/DNA adduct was energy-minimized using the AMBER software.57 F



Articles

The gaff and DNA-optimized parm99bsc058 force-fields were loaded for DNA, and xleap was used to manually position ligands into each sequence individually by creating the covalent bond between the exocyclic C2-NH2 of the reacting guanine and the C11-positions of sibiromycin (3, Figure 1) and 9-deoxysibiromycin (4, Figure 1). This positioning was based upon the known orientation of the PBD in the minor groove, as ascertained through crystallographic studies.19 Parm99bsc0 was used as it is a refined version of parm99 in which α/γ rotation of nucleic acids is considered.58 Parameters for the covalent attachment of the ligand to DNA were created using parameters derived previously through molecular mechanics calculations,59 and both sibiromycin (3) and 9-deoxysibiromycin (4) were placed within the minor groove to within 2 Å of the reacting guanine (based on the length of the covalent attachment derived through crystal studies22). Na+ ions were placed along the DNA backbone using xleap (to neutralize the DNA), and adducts were solvated using a truncated octahedron TIP3P water box of maximum dimension 10 Å. Next, each adduct was minimized in a gradient manner by initially placing the DNA under a high force constraint (i.e., high net force, in this instance a force constant of 500 kcal mol−1 Angstrom−2), which was reduced in stages to zero to enable the ligand to find its local energy minimum, followed by reduction in force in a periodic manner with a relaxation of restraints. Production Simulations. Once the full system was minimized, it was heated slowly to 300 K over 20 ps using the SHAKE algorithm to restrict the vibrations of C−H bonds,60 followed by an unrestrained equilibration step of 100 ps to relax the density of water. Once in equilibrium, production simulations were run for a period of 10 ns, and atomic coordinates were saved at 1 ps intervals. RMSD calculations of each DNA/ligand adduct indicated when simulations had reached an equilibrium state (selected adducts presented in Supporting Information S8−S11). Further simulations of 10 ns were conducted using an identical protocol, with sibiromycin and 9-deoxysibiromycin docked in the minor groove of each sequence in a noncovalent manner. In this instance, ligands were docked in the minor groove with the imines positioned to within 2 Å of the exocyclic amine of the reacting guanine. This allowed the formation of a hydrogen bond between the N10−C11 of the PBD and the exocyclic amine, followed by either the formation of noncovalent interactions or the maintenance of the Hbond, which in turn resulted in both sibiromycin and 9deoxysibiromycin remaining in the minor groove for the duration of each simulation. This was reaffirmed through measurements between the ligand and DNA (selected adducts presented in S12−S15, Supporting Information). Simulations of 10 ns duration were undertaken, and free energies calculated using the MM-PBSA method in AMBER 11, a procedure shown to be most accurate in free energy estimation in explicit solvent simulations61 considering the computational power at hand. Constant temperature (selected adducts presented in S16 and S17, Supporting Information) and kinetic, potential and total energies (selected adducts presented in S18 and S19, Supporting Information) were maintained throughout each simulation. Binding Free Energy Calculations. Binding free energy was calculated as follows:

ΔG°bind = ΔEMM + ΔG°solv − T ΔS It is possible to calculate the entropic contribution using normalmode analysis. However, this was impractical as normal-mode analysis calculations would have introduced significant error into final values and are computationally expensive. As states of similar entropy were being assessed (i.e., highly similar drug/DNA adducts), and were therefore comparable, these calculations were not undertaken. One hundred snapshots of the MD simulations were taken at equal intervals over the 10 ns duration, and molecular mechanics (MM) calculations were performed using pbsa.57 Molecular models were generated based on lowest energy conformations of sibiromycin and 9deoxysibiromycin interacting with DNA, derived through ptraj calculations. Docking studies on GAL4 were undertaken using the GAL4 (PDB ID 1D66) crystal structure.63 The DNA consensus sequence was removed, sibiromycin and 9-deoxysibiromycin were docked at a potential reacting guanine, and the complex energy minimized in AMBER using the protocol described above. This process was undertaken for every potential PBD binding site within the GAL4 transcription factor consensus sequence, and each model was realigned with the original crystal DNA sequence using Pymol.64 Potential steric interactions between ligands and DNA were then identified using VMD.65 All molecular models were created using Chimera,66 and electronic calculations were carried out using the Calculator Plugins within Marvin 6.1.0 (2013; www.chemaxon.com).

■

ASSOCIATED CONTENT

* Supporting Information S

Plots of H-bond lifetimes, distances between ligand and minor groove, RMSD calculations, graphs of simulation temperature and kinetics, potential and total energies. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Email: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS Drs B. Gerratana and I. Yonemoto are thanked for helpful discussions. ABBREVIATIONS PBD, pyrrolobenzodiazepine; CPK model, Corey-PaulingKoltun model; MM-PBSA, Molecular Mechanics Poisson− Boltzmann Surface Area; vdW, van der Waals; bp, base-pair; MM, molecular mechanics; MD, molecular dynamics

■

ΔG°bind = ΔG° bind vacuum + ΔG°complex

REFERENCES

(1) Yonemoto, I. T., Li, W., Khullar, A., Reixach, N., and Gerratana, B. (2012) Mutasynthesis of a potent anticancer sibiromycin analogue. ACS Chem. Biol. 7, 973−977. (2) Yonemoto, I. T., Li, W., Khullar, A., Reixach, N., and Gerratana, B. (2014) Correction to mutasynthesis of a potent anticancer sibiromycin analogue. ACS Chem. Biol. 9, 1214−1214. (3) Antonow, D., and Thurston, D. E. (2011) Synthesis of DNAinteractive pyrrolo[2,1-c][1,4]benzodiazepines (PBDs). Chem. Rev. 111, 2815−2864. (4) Cipolla, L., Araujo, A. C., Airoldi, C., and Bini, D. (2009) Pyrrolo[2,1-c][1,4]benzodiazepine as a scaffold for the design and synthesis of anti-tumour drugs. Anticancer Agents Med. Chem. 9, 1−31. (5) Gerratana, B. (2012) Biosynthesis, synthesis, and biological activities of pyrrolobenzodiazepines. Med. Res. Rev. 32, 254−293.

− (ΔG°ligand + ΔG°receptor ) where ΔG°bind was determined by solving the linearized Poisson− Boltzmann equation,62 and ΔG°complex is represented by the following:

ΔG°complex = EMM + Gpolar solvation energy + Gnonpolar solvation energy − TS Solvation energies (both polar and nonpolar) were considered, with EMM corresponding to internal, electrostatic and vdW interactions, and S to solute entropy. The final binding energy was represented as G



Articles

diazepine dimers: Stability, stereochemistry and drug orientation. Org. Biomol Chem. 10, 6850−6860. (21) Smellie, M., Bose, D. S., Thompson, A. S., Jenkins, T. C., Hartley, J. A., and Thurston, D. E. (2003) Sequence-selective recognition of duplex DNA Through covalent interstrand crosslinking: Kinetic and molecular modeling studies with pyrrolobenzodiazepine dimers. Biochemistry 42, 8232−8239. (22) Kopka, M. L., Goodsell, D. S., Baikalov, I., Grzeskowiak, K., Cascio, D., and Dickerson, R. E. (1994) Crystal structure of a covalent DNA-drug adduct: Anthramycin bound to C-C-A-A-C-G-T-T-G-G and a molecular explanation of specificity. Biochemistry 33, 13593− 13610. (23) Thurston, D. E. (2006) Chemistry and Pharmacology of Anticancer Drugs; Vol. 1, p. 281, CRC Press (Taylor & Francis), Boca Raton, FL. (24) Rahman, K. M., Vassoler, H., James, C. H., and Thurston, D. E. (2010) DNA sequence preference and adduct orientation of pyrrolo[2,1-c][1,4]benzodiazepine antitumor agents. ACS Med. Chem. Lett. 1, 427−432. (25) Kamal, A., Balakishan, G., Ramakrishna, G., Shaik, T. B., Sreekanth, K., Balakrishna, M., Rajender, Dastagiri, D., and Kalivendi, S. V. (2010) Synthesis and biological evaluation of cinnamido linked pyrrolo [2,1-c][1,4]benzodiazepines as antimitotic agents. Eur. J. Med. Chem. 45, 3870−3884. (26) Kamal, A., Devaiah, V., Reddy, K. L., and Kumar, M. S. (2005) Synthesis and biological activity of fluoroquinolone-pyrrolo[2,1c][1,4]benzodiazepine conjugates. Bioorg. Med. Chem. 13, 2021−2029. (27) Kamal, A., Khan, M. N. A., Srikanth, Y. V. V, Reddy, K. S., Juvekar, A., Sen, S., Kurian, N., and Zingde, S. (2008) Synthesis, DNAbinding ability and evaluation of antitumour activity of triazolo[1,2,4] benzothiadiazine linked pyrrolo[2,1-c][1,4] benzodiazepine conjugates. Bioorg. Med. Chem. 16, 7804−7810. (28) Baraldi, P. G., Balboni, G., Cacciari, B., Guiotto, A., Manfredini, S., Romagnoli, R., Spalluto, G., Thurston, D. E., Howard, P. W., Bianchi, N., Rutigliano, C., Mischiati, C., and Gambari, R. (1999) Synthesis, in vitro antiproliferative activity, and DNA-binding properties of hybrid molecules containing pyrrolo[2,1-c][1,4]benzodiazepine and minor-groove-binding oligopyrrole carriers. J. Med. Chem. 42, 5131−5141. (29) Wells, G., Martin, C. R. H., Howard, P. W., Sands, Z. A., Laughton, C. A., Tiberghien, A., Woo, C. K., Masterson, L. A., Stephenson, M. J., Hartley, J. A., Jenkins, T. C., Shnyder, S. D., Loadman, P. M., Waring, M. J., and Thurston, D. E. (2006) Design, synthesis, and biophysical and biological evaluation of a series of pyrrolobenzodiazepine-poly(N-methylpyrrole) conjugates. J. Med. Chem. 49, 5442−5461. (30) Farmer, J. D., Rudnicki, S. M., and Suggs, J. W. (1988) Synthesis and DNA crosslinking ability of a dimeric anthramycin analog. Tetrahedron Lett. 29, 5105−5108. (31) Gregson, S. J., Howard, P. W., Hartley, J. A., Brooks, N. A., Adams, L. J., Jenkins, T. C., Kelland, L. R., and Thurston, D. E. (2001) Design, synthesis, and evaluation of a novel pyrrolobenzodiazepine DNA-interactive agent with highly efficient cross-linking ability and potent cytotoxicity. J. Med. Chem. 44, 737−748. (32) Puzanov, I., Lee, W., Chen, A. P., Calcutt, M. W., Hachey, D. L., Vermeulen, W. L., Spanswick, V. J., Liao, C. Y., Hartley, J. A., Berlin, J. D., and Rothenberg, M. L. (2011) Phase I pharmacokinetic and pharmacodynamic study of SJG-136, a novel DNA sequence selective minor groove cross-linking agent, in advanced solid tumors. Clin. Cancer Res. 17, 3794−3802. (33) Kamal, A., Rajender, Reddy, D. R., Reddy, M. K., Balakishan, G., Shaik, T. B., Chourasia, M., and Sastry, G. N. (2009) Remarkable enhancement in the DNA-binding ability of C2-fluoro substituted pyrrolo[2,1-c][1,4]benzodiazepines and their anticancer potential. Bioorg. Med. Chem. 17, 1557−1572. (34) Thurston, D. E. (1993) Advances in the study of pyrrolo[2,1c][1,4]benzodiazepine (PBD) antitumour antibiotics. In Molecular Aspects of Anticancer Drug-DNA Interactions (Neidle, S., and Waring, M. J., Eds.) pp 54−88, The Macmillan Press Ltd., London, U.K.

(6) Hartley, J. A. (2011) The development of pyrrolobenzodiazepines as antitumour agents. Expert Opin. Invest. Drugs 20, 733−744. (7) Kamal, A., Reddy, K. L., Devaiah, V., Shankaraiah, N., and Reddy, D. R. (2006) Recent advances in the solid-phase combinatorial synthetic strategies for the benzodiazepine based privileged structures. Mini-Rev. Med. Chem. 6, 53−69. (8) Leimgruber, W., Stefanovic, V., Schenker, F., Karr, A., and Berger, J. (1965) Isolation and characterization of anthramycin, a new antitumor antibiotic. J. Am. Chem. Soc. 87, 5791−5793. (9) Hurley, L. H., Reck, T., Thurston, D. E., Langley, D. R., Holden, K. G., Hertzberg, R. P., Hoover, J. R., Gallagher, G., Jr., Faucette, L. F., Mong, S. M., et al. (1988) Pyrrolo[1,4]benzodiazepine antitumor antibiotics: Relationship of DNA alkylation and sequence specificity to the biological activity of natural and synthetic compounds. Chem. Res. Toxicol. 1, 258−268. (10) Kotecha, M., Kluza, J., Wells, G., O’Hare, C. C., Forni, C., Mantovani, R., Howard, P. W., Morris, P., Thurston, D. E., Hartley, J. A., and Hochhauser, D. (2008) Inhibition of DNA binding of the NFY transcription factor by the pyrrolobenzodiazepine−polyamide conjugate GWL-78. Mol. Cancer Ther. 7, 1319−1328. (11) Wells, G., Howard, P. W., Martin, C., Sands, Z. A., Laughton, C. A., Tiberghien, A., Woo, C. K., Masterson, L. A., John, A. I., Jenkins, T. C., Shnyder, S. D., Loadman, P. M., and Thurston, D. E. (2005) Design, synthesis, biophysical, and biological evaluation of a series of methyl ester-terminated pyrrolobenzodiazepine (PBD)-poly(N-methylpyrrole) conjugates. Clin. Cancer Res. 11, 9015s−9015s. (12) Brucoli, F., Hawkins, R. M., James, C. H., Jackson, P. J., Wells, G., Jenkins, T. C., Ellis, T., Kotecha, M., Hochhauser, D., Hartley, J. A., Howard, P. W., and Thurston, D. E. (2013) An extended pyrrolobenzodiazepine-polyamide conjugate with selectivity for a DNA sequence containing the ICB2 transcription factor binding site. J. Med. Chem. 56, 6339−6351. (13) Rahman, K. M., Jackson, P. J. M., James, C. H., Basu, B. P., Hartley, J. A., de la Fuente, M., Schatzlein, A., Robson, M., Pedley, R. B., Pepper, C., Fox, K. R., Howard, P. W., and Thurston, D. E. (2013) GC-targeted C8-linked pyrrolobenzodiazepine−biaryl conjugates with femtomolar in vitro cytotoxicity and in vivo antitumor activity in mouse models. J. Med. Chem. 56, 2911−2935. (14) Mantaj, J., Jackson, P. J. M., Rahman, K. M., and Thurston, D. E. (2013) Interaction of SJG-136 with cognate sequences of oncogenic transcription factors. In American Association of Cancer Research Annual Meeting, Washington, DC. (15) Puvvada, M. S., Hartley, J. A., Jenkins, T. C., and Thurston, D. E. (1993) A quantitative assay to measure the relative DNA-binding affinity of pyrrolo[2,1-c][1,4]benzodiazepine (PBD) antitumour antibiotics based on the inhibition of restriction endonuclease BamHI. Nucleic Acids Res. 21, 3671−3675. (16) Clingen, P. H., De Silva, I. U., McHugh, P. J., Ghadessy, F. J., Tilby, M. J., Thurston, D. E., and Hartley, J. A. (2005) The XPFERCC1 endonuclease and homologous recombination contribute to the repair of minor groove DNA interstrand crosslinks in mammalian cells produced by the pyrrolo[2,1-c][1,4]benzodiazepine dimer SJG136. Nucleic Acids Res. 33, 3283−3291. (17) Puvvada, M. S., Forrow, S. A., Hartley, J. A., Stephenson, P., Gibson, I., Jenkins, T. C., and Thurston, D. E. (1997) Inhibition of bacteriophage T7 RNA polymerase in vitro transcription by DNABinding pyrrolo[2,1-c][1,4]benzodiazepines. Biochemistry 36, 2478− 2484. (18) Hertzberg, R. P., Hecht, S. M., Reynolds, V. L., Molineux, I. J., and Hurley, L. H. (1986) DNA sequence specificity of the pyrrolo[1,4]benzodiazepine antitumor antibiotics. MethidiumpropylEDTA-iron(II) footprinting analysis of DNA binding sites for anthramycin and related drugs. Biochemistry 25, 1249−1258. (19) Barkley, M. D., Cheatham, S., Thurston, D. E., and Hurley, L. H. (1986) Pyrrolo[1,4]benzodiazepine antitumor antibiotics: Evidence for two forms of tomaymycin bound to DNA. Biochemistry 25, 3021− 3031. (20) Seifert, J., Pezeshki, S., Kamal, A., and Weisz, K. (2012) Interand intrastrand DNA crosslinks by 2-fluoro-substituted pyrrolobenzoH



Articles

(35) Jenkins, T. C., Hurley, L. H., Neidle, S., and Thurston, D. E. (1994) Structure of a covalent DNA minor-groove adduct with a pyrrolobenzodiazepine dimerEvidence for sequence-specific interstrand cross-linking. J. Med. Chem. 37, 4529−4537. (36) Hopton, S. R., and Thompson, A. S. (2011) Nuclear magnetic resonance solution structures of inter- and intrastrand adducts of DNA cross-linker SJG-136. Biochemistry 50, 4720−4732. (37) Itoh, J., Watabe, H., Ishii, S., Gomi, S., Nagasawa, M., Yamamoto, H., Shomura, T., Sezaki, M., and Kondo, S. (1988) Sibanomicin, a new pyrrolo[1,4]benzodiazepine antitumor antibiotic produced by a Micromonospora sp. J. Antibiot (Tokyo) 41, 1281−1284. (38) Brazhnikova, M. G., Konstantinova, N. V., and Mesentsev, A. S. (1972) Sibiromycin: Isolation and characterization. J. Antibiot (Tokyo) 25, 668−673. (39) Lubawy, W. C., Dallam, R. A., and Hurley, L. H. (1980) Protection against anthramycin-induced toxicity in mice by coenzyme Q10. J. Natl. Cancer Inst. 64, 105−109. (40) Borkovec, A. B., Chang, S. C., and Horwitz, S. B. (1971) Chemosterilization of house flies with anthramycin methyl ether. J. Econ. Entomol. 64, 983−984. (41) Petrusek, R. L., Anderson, G. L., Garner, T. F., Fannin, Q. L., Kaplan, D. J., Zimmer, S. G., and Hurley, L. H. (1981) Pyrrol[1,4]benzodiazepine antibiotics. Proposed structures and characteristics of the in vitro deoxyribonucleic acid adducts of anthramycin, tomaymycin, sibiromycin, and neothramycins A and B. Biochemistry 20, 1111−1119. (42) Hurley, L. H., and Thurston, D. E. (1984) Pyrrolo(1,4)benzodiazepine antitumor antibioticsChemistry, interaction with DNA, and biological implications. Pharm. Res., 52−59. (43) Guiotto, A., Howard, P. W., Baraldi, P. G., and Thurston, D. E. (1998) Synthesis of novel C7-aryl substituted pyrrolo[2,1-c][1,4]benzodiazepines (PBDs) via pro-N10-Troc protection and Suzuki coupling. Bioorg. Med. Chem. Lett. 8, 3017−3018. (44) Khokhlov, D. N., Brusov, R. V., Grokhovskii, S. L., Nikolaev, V. A., Pis’menskii, V. F., Zhuze, A. L., and Gurskii, G. V. (1995) [A synthetic zinc chelating peptide competes for DNA binding sites with antibiotics, adsorbed in a minor DNA groove]. Mol. Biol. (Mosk) 29, 354−364. (45) Kizu, R., Draves, P. H., and Hurley, L. H. (1993) Correlation of DNA sequence specificity of anthramycin and tomaymycin with reaction kinetics and bending of DNA. Biochemistry 32, 8712−8722. (46) Boyd, F. L., Stewart, D., Remers, W. A., Barkley, M. D., and Hurley, L. H. (1990) Characterization of a unique tomaymycind(CICGAATTCICG)2 adduct containing two drug molecules per duplex by NMR, fluorescence, and molecular modeling studies. Biochemistry 29, 2387−2403. (47) Cheatham, S., Kook, A., Hurley, L. H., Barkley, M. D., and Remers, W. (1988) One- and two-dimensional 1H NMR, fluorescence, and molecular modeling studies on the tomaymycind(ATGCAT)2 adduct. Evidence for two covalent adducts with opposite orientations and stereochemistries at the covalent linkage site. J. Med. Chem. 31, 583−590. (48) Adams, L. J., Jenkins, T. C., Banting, L., and Thurston, D. E. (1999) Molecular modelling of a sequence-specific DNA-binding agent based on the pyrrolo[2,1-c][1,4]benzodiazepines. Pharm. Pharmacol Commun. 5, 555−560. (49) Malhotra, R. K., Ostrander, J. M., Hurley, L. H., McInnes, A. G., Smith, D. G., Walter, J. A., and Wright, J. L. (1981) Chemical conversion of anthramycin 11-methyl ether to didehydroanhydroanthramycin and its utilization in studies of the biosynthesis and mechanism of action of anthramycin. J. Nat. Prod. 44, 38−44. (50) Antonow, D., Barata, T., Jenkins, T. C., Parkinson, G. N., Howard, P. W., Thurston, D. E., and Zloh, M. (2008) Solution structure of a 2:1 C2-(2-naphthyl) pyrrolo[2,1-c][1,4]benzodiazepine DNA adduct: Molecular basis for unexpectedly high DNA helix stabilization. Biochemistry 47, 11818−11829. (51) Marmur, J., and Doty, P. (1962) Determination of the base composition of deoxyribonucleic acid from its thermal denaturation temperature. J. Mol. Biol. 5, 109−118.

(52) Hurley, L. H., Gairola, C., and Zmijewski, M. (1977) Pyrrolo(1,4)benzodiazepine antitumor antibiotics. In vitro interaction of anthramycin, sibiromycin, and tomaymycin with DNA using specifically radiolabelled molecules. Biochim. Biophys. Acta 475, 521− 535. (53) Jenkins, T. C., Hurley, L. H., Neidle, S., and Thurston, D. E. (1994) Structure of a covalent DNA minor groove adduct with a pyrrolobenzodiazepine dimer: Evidence for sequence-specific interstrand cross-linking. J. Med. Chem. 37, 4529−4537. (54) Baraldi, P. G., Cacciari, B., Guiotto, A., Romagnoli, R., Spalluto, G., Leoni, A., Bianchi, N., Feriotto, G., Rutigliano, C., Mischiati, C., and Gambari, R. (2000) [2,1-c][1,4]benzodiazepine (PBD)-distamycin hybrid inhibits DNA binding to transcription factor Sp1. Nucleosides Nucleotides Nucleic Acids 19, 1219−1229. (55) Brucoli, F., Hawkins, R. M., James, C. H., Wells, G., Jenkins, T. C., Ellis, T., Hartley, J. A., Howard, P. W., and Thurston, D. E. (2011) Novel C8-linked pyrrolobenzodiazepine (PBD)-heterocycle conjugates that recognize DNA sequences containing an inverted CCAAT box. Bioorg. Med. Chem. Lett. 21, 3780−3783. (56) Brucoli, F., Hawkins, R. M., Wells, G., Jenkins, T. C., Ellis, T., Kotecha, M., Hochhauser, D., Hartley, J. A., Howard, P. W., and Thurston, D. E. (2010) A potent PBD-heterocyclic polyamide conjugate targeting an ICB2 transcription factor binding site. EJC Suppl. 8, 168−168. (57) Case, D. A., Darden, T. A., Cheatham III, T. E., Simmerling, C. L., Wang, J., Duke, R. E., Luo, R., Walker, R. C., Zhang, W., Merz, K. M., Roberts, B., Wang, B., Hayik, S., Roitberg, A., Seabra, G., Kolossváry, I., Wong, K. F., Paesani, F., Vanicek, J., Liu, J., Wu, X., Brozell, S. R., Steinbrecher, T., Gohlke, H., Cai, Q., Ye, X., Wang, J., Hsieh, M.-J., Cui, G., Roe, D. R., Mathews, D. H., Seetin, M. G., Sagui, C., Babin, V., Luchko, T., Gusarov, S., Kovalenko, A., Kollman, P. A. (2010) AMBER 11, University of California, San Francisco, 2010. (58) Perez, A., Marchan, I., Svozil, D., Sponer, J., Cheatham, T. E., 3rd, Laughton, C. A., and Orozco, M. (2007) Refinement of the AMBER force field for nucleic acids: Improving the description of α/γ conformers. Biophys. J. 92, 3817−3829. (59) Rao, S. N., Singh, U. C., and Kollman, P. A. (1986) Molecular mechanics simulations on covalent complexes between anthramycin and B DNA. J. Med. Chem. 29, 2484−2492. (60) Ryckaert, J.-P., Ciccotti, G., and Berendsen, H. J. C. (1977) Numerical integration of the cartesian equations of motion of a system with constraints: Molecular dynamics of n-alkanes. J. Comput. Phys. 23, 327−341. (61) Wang, H., and Laughton, C. A. (2009) Evaluation of molecular modelling methods to predict the sequence-selectivity of DNA minor groove binding ligands. Phys. Chem. Chem. Phys. 11, 10722−10728. (62) Fogolari, F., Zuccato, P., Esposito, G., and Viglino, P. (1999) Biomolecular electrostatics with the linearized Poisson−Boltzmann equation. Biophys. J. 76, 1−16. (63) Marmorstein, R., Carey, M., Ptashne, M., and Harrison, S. C. (1992) DNA recognition by GAL4: Structure of a protein−DNA complex. Nature 356, 408−414. (64) Schrodinger, L. L. C. (2010) The PyMOL Molecular Graphics System, Version 1.3r1. (65) Humphrey, W., Dalke, A., and Schulten, K. (1996) VMD: Visual Molecular Dynamics. J. Mol. Graph 14 (33−38), 27−38. (66) Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF ChimeraA visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605−1612.

I


Computational Studies Support the Role of the C7-Sibirosamine

Recommend Documents