Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Recent published observations from a large number of molecular biology laboratories demonstrate that DNA-binding drugs, including distamycin (Broggini et al., 1989; Gambari et al., 1991; Dorn et al., 1992; Feriotto et al., 1994), chromomycin (Bianchi et al., 1996), mithramycin (Snyder et al., 1991), actinomycin D (Vaquero and Portugal, 1998), elsamicin A (Vaquero and Portugal, 1998), and CC-1065 (Welch et al., 1994), interfere in a sequence-dependent manner with biological functions such as restriction enzyme cleavage of DNA, topoisomerase activity, protein-DNA interactions, in vitro and in vivo transcription (Van Dyke and Dervan, 1983; Broggini et al., 1989; Gambari et al., 1991; Dorn et al., 1992; Feriotto et al., 1994, Gambari and Nastruzzi, 1994). For instance, the G+C selective DNA binding drugs mithramycin and chromomycin (Van Dyke and Dervan, 1983) suppress molecular interactions between transcription factor Sp1 and viral and eukaryotic promoters, leading to an alteration of the expression of Sp1-regulated transcriptional units, including the oncogene Ha-ras, the collagen-1(I) gene and HIV-1 (Snyder et al., 1991; Bianchi et al., 1996; Vaquero and Portugal, 1998). By contrast, the A+T selective DNA binding drug distamycin does not interfere with the interaction between Sp1 and target DNA sequences, being able on the contrary to fully suppress binding of TF-IID/TBP to the TATA-box element of the HIV-1 long terminal repeat (LTR) (Feriotto et al., 1994). These and similar effects could have important practical applications in the experimental therapy of many human pathologies, including neoplastic diseases (Vaquero and Portugal, 1998) and AIDS (Gambari and Nastruzzi, 1994). Following this hypothesis, a large number of investigations on the design, synthesis, and characterization of the potential sequence-selectivity of DNA binding drugs was recently published using DNase I footprinting, gel retardation studies, polymerase chain reaction, and in vitro transcription (Portugal and Waring, 1987; Broggini et al., 1989; White and Phillips, 1989; Churchill et al., 1990; Fox et al., 1990; Laughton et al., 1990; Gambari et al., 1991; Montecucco et al., 1991; Snyder et al., 1991; Dorn et al., 1992; Chiang et al., 1994; Feriotto et al., 1994; Passadore et al., 1994, 1995; Welch et al., 1994; Bellorini et al., 1995; Bianchi et al., 1996; Vaquero et al., 1998). It should be stressed that most of these techniques are steady-state methodologies and therefore are not suitable for an easy determination of 1) the binding efficiency of DNA-binding drugs to target DNA elements and 2) the stability of drugs-DNA complexes. Of course, these parameters are of great interest, because they could be intimately associated with the biological activity of DNA-binding drugs.

The recent development of biosensor technologies for biospecific interaction analysis (BIA) (Jonsson et al., 1991; Vadgama and Crump, 1992; Malmqvist, 1993) enables monitoring of a variety of molecular reactions in real-time by surface plasmon resonance (SPR). This optical technique detects and quantifies changes in refractive index in the vicinity of the surface of sensor chips to which ligands are immobilized (Vadgama and Crump, 1992; Malmqvist, 1993). Because the changes in the refractive index are proportional to the changes in the adsorbed mass (Malmqvist, 1993), the SPR technology allows detection of biomolecules (analytes) interacting with the ligand immobilized on the sensor chip. For instance, if the ligand is a biotinylated single-stranded DNA, SPR technology could monitor DNA-DNA hybridization in the same time as it occurs (Wood, 1993; Nilsson et al., 1995). SPR technology was also applied to studies on protein-protein interactions, protein-DNA recognition, antibody-antigen binding, as well as to a large variety of biomolecular applications involving high-molecular weight molecules (Malmqvist, 1993). In sharp contrast, few reports were published on BIA analysis on low-molecular weight compounds, and to our knowledge, no data are available in the literature on the possible use of SPR technology to detect differential interactions between DNA-binding drugs and target DNA sequences. It is well known that detection of molecular interactions of low-molecular weight analytes (<2000 Da) is expected to be difficult, especially if the analyte is not pure. This is largely due to the low expected increase of resonance units (RU) following binding of a low-molecular weight (200-1000 Da) analyte to its immobilized ligand (Karlsson and Stahlberg, 1995). For these reasons, the demonstration of the possible use of BIA analysis for studies on DNA-binding drugs could be of interest.

In this paper we used the biosensor BIAcore (Pharmacia Biosensors, Uppsala, Sweden) to detect, by BIA and SPR technology, the binding activity of two G+C selective minor groove ligands, mithramycin and chromomycin, and two A+T selective binders, distamycin and MEN 10567. As a model system, we selected three oligonucleotides mimicking the binding sites of transcription factors Sp1, NF-kB and TFII-D, present within the LTR of HIV-1. These oligonucleotides were immobilized on streptavidin-coated sensor chips and used as ligands. The results obtained show that SPR technology is a suitable one-step, nonradioactive approach to demonstrating DNA-binding activity of low molecular minor groove ligands and stability of drugs-DNA complexes.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

DNA-Binding Drugs and Synthetic Oligonucleotides. Chromomycin (1183.3 Da), mithramycin (1085.2 Da), and distamycin A (518 Da) were purchased from Sigma Chemical Co. (St. Louis, MO). MEN 10567 (639 Da) was obtained from Menarini Research (Florence, Italy). Chemical structures of these compounds are shown in Fig. 1. The Sp1 (7507 Da), NFkB (8537 Da) and TATA (5503 Da) synthetic oligonucleotides are shown in Table 1 and were purchased from Pharmacia (Uppsala, Sweden). All stock solutions of DNA-binding drugs (1.2 mM) were stored at 20°C in the dark and diluted immediately before use in CH/MT buffer (100 mM NaCl, 10 mM Tris-HCl, pH 7.4, 20 mM MgCl₂).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Chemical structures of mithramycin, chromomycin, distamycin and MEN10567.

BIA Using Biosensor Technology. The BIAcore-1000 instrument (Pharmacia Biosensors) was used in all experiments (Jonsson et al., 1991; Vadgama and Crump, 1992; Malmqvist, 1993). Sensor chip SA5 (research grade), precoated with streptavidin was purchased from Pharmacia Biosensors. The experiments were conducted at 25°C temperature and at 4 µl/min flow rate (Nilsson et al., 1995). Rapid capture of about 700 to 900 RU of biotinylated Sp1, NF-kB, and TATA single-stranded DNA (0.5 µg/50 µl of HBS buffer, 30-µl injection) (HBS buffer = 10 mM Hepes, pH 7.4, 0.15 M NaCl, 3.4 mM EDTA, and 0.05% Surfactant P2; Pharmacia Biosensors) was reproducibly obtained within 4 to 6 min (data not shown). Double-stranded Sp1, NF-kB, and TATA mers were generated by a 20- to 30-µl injection of the complementary oligonucleotides (Sp1c, NF-kBc, and TATAc mers, 0.5 µg/50 µl of HBS). The obtained double-stranded DNAs were found to be stably immobilized on SA5 sensor chips, as 30-µl injections of running buffer did not cause a decrease of the SPR signals (data not shown). The binding of minor groove ligands to double-stranded target DNA sequences was monitored after a 20- to 30-µl injection of 0.125 to 5 µM compounds. Running buffers were HBS buffer and CH/MT buffer (100 mM NaCl, 50 mM Tris, pH 7.4, 20 mM MgCl₂). Buffers were filtered and degassed before use. Generation of double-stranded target DNA sequences was performed in HBS buffer. Analysis of the obtained sensorgrams was performed using the BIA-evaluation Software. Suitable blank control injections with running and/or injection buffers were performed, and the resulting sensorgrams subtracted from the experimental sensorgrams. The apparent stoichiometry of the surface complex (i.e. the number of analyte molecules that can bind to one ligand molecule) was calculated from the saturating binding capacity of the surface by the equation (Jonsson et al., 1991; Vadgama and Cump, 1992; Malmqvist, 1993):

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Detection of Binding of Mithramycin to the SA-Sp1 Sensor Chip. To determine whether binding of mithramycin to DNA is detectable by SPR technology, the experiments shown in Fig. 2 were performed. We first produced a SA5 sensor chip containing Sp1 target double-stranded DNA. Second, we allowed mithramycin binding to (Sp1)DNA. Third, we studied stability of mithramycin-(Sp1)DNA complexes by injecting binding buffer.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   A, binding of mithramycin to double-stranded HIV-1 Sp1 DNA. This experiment was conducted with CH/MT buffer as running buffer. Sp1c mer was injected in HBS buffer (A,a). The sensorgram shows that injection of 20 µl of HIV-1 LTR Sp1b mer (0.25 µg/50 µl of CH buffer) generate an increase (A,a) in RU. The binding of mithramycin to the SA5 sensor chip containing double-stranded HIV-1 LTR Sp1 DNA is shown in A,b. An increase of 833 RU was observed following a 30-µl injection of 5 µM mithramycin. The decrease of RU during the washing step (injection of CH/MT binding buffer) is shown in A,c. B, after injection of Sp1c mer, a 20-µl injection of increasing concentrations (a = 0.5 µM; b = 1 µM; c = 2 µM; d = 4 µM; e = 5 µM; f = 8 µM) of mithramycin was performed. The sensorgrams have been elaborated using the BIA-evaluation Software (Pharmacia Biosensors). C, relationship between bound mithramycin (RUmax  RUi) and amount of ligand Sp1 mer (RUi  RUo) present on the flow cell surface. In this experiment, injections of 20 µl of 5 µM mithramycin were performed to different flow cells containing increasing amounts of target Sp1 mer.

Detection of Binding of Distamycin to the SA-NFkB Sensor Chip. To determine whether binding of distamycin to DNA is detectable by SPR technology, the experiment shown in Fig. 2 was repeated using a SA sensor chip carrying a double-stranded oligonucleotide containing the HIV-1 LTR NF-kB sequence. We previously demonstrated by DNase I footprinting experiments that distamycin binds to the NF-kB binding sites of the HIV-1 LTR (Feriotto et al., 1994b). The SA5 sensor chip containing NF-kB target double-stranded DNA was prepared as described for Sp1. Distamycin was injected in HBS buffer. The stability of distamycin/DNA complexes was studied with a further injection of HBS buffer.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   A, binding of distamycin to double-stranded HIV-1 NF-kB DNA. This experiment was conducted with HBS buffer as running buffer. NF-kBc mer was injected in HBS buffer (A,a). The sensorgram shows that injection of 20 µl of HIV-1 LTR NF-kBc mer (0.25 µg/50 µl of HBS buffer) generate an increase (A,a) in RU. The binding of distamycin to the SA5 sensor chip containing double-stranded HIV-1 LTR NF-kB DNA is shown in A,b. An increase of 147 RU was observed following a 30-µl injection of 5 µM distamycin. The decrease of RU during the washing step (injection of HBS binding buffer) is shown in A,c. B and C, after injection of NF-kBc mer, a 20-µl injection of increasing concentrations (a = 0.125 µM; b = 0.25 µM; c = 0.5 µM; d = 1 µM; e = 2 µM; f = 5 µM;) of distamycin was performed. The sensorgrams have been elaborated using the BIA-evaluation Software (Pharmacia Biosensors). C, relationship between bound distamycin (RUmax  RUi) and amount of ligand NF-kB mer (RUi RUo) present on the flow cell surface. In this experiment, injections of 20 µl of 5 µM distamycin were performed to different flow cells containing increasing amounts of target NF-kB mer.

Differential Binding of Distamycin and Mithramycin to SA5-Sp1 and SA5-NF-kB Sensor Chips. In the experiment shown in Fig. 4 (A and B), 5 µM mithramycin was injected onto SA5-Sp1 (Fig. 4A) and SA5-NF-kB (Fig. 4B) sensor chips. After mithramycin injection (A,a and B,a), we performed an injection of binding buffer (A,b and B,b). The results obtained demonstrate that binding of mithramycin to the SA5-Sp1 sensor chip is more efficient, although binding occurs also to the NF-kB sensor chip. In both cases, the drug/DNA complexes rapidly dissociate. The sequence-selective binding of distamycin is shown in Fig. 4 (C and D). In this experiment 5 µM distamycin was injected onto SA5-NF-kB (C) and SA5-Sp1 (D) sensor chips. After distamycin injection (C,a and D,a), we injected HBS binding buffer (C,b and D,b). The results obtained demonstrate that binding of distamycin to the SA5-Sp1 sensor chip does not occur, whereas binding to the NF-kB sensor chip is readily detectable.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Differential binding of mithramycin and distamycin to SA5-Sp1 and SA5-NF-kB sensor chips. 5 µM mithramycin (A, B) and 5 µM distamycin (C, D) were injected onto SA5-Sp1 (A, D) and SA5-NF-kB (B, C) sensor chips. After injection of DNA-binding drugs (Aa, Ba, Ca, and Da), a 15 µl injection of binding buffer (Ab, Bb, Cb, and Db) was performed. Binding buffers were CH/MT and HBS buffers in the case of mithramycin and distamycin, respectively.

Differential Binding of Distamycin to SA5-NF-kB and SA5-TATA Box Sensor Chips. In the experiment shown in Fig. 5, 5 µM distamycin was injected onto SA5-NF-kB (Fig. 5A) and SA5-TATA box (Fig. 5B,b) sensor chips. After distamycin injection (A,b and B,b), we injected HBS running buffer (A,c and B,c). The results obtained demonstrate that binding of distamycin to the SA5-TATA box sensor chip is more efficient than binding to the NF-kB sensor chip. In addition, the data obtained clearly show that the distamycin/TATA box complexes are more stable than distamycin-NF-kB mer complexes. The values of RUres  RUi were found to be 4 (2.45% of RUmax  RUi) in the case of binding of distamycin to SA5-NF-kB flow cell; the values of RUres  RUi were found to be 117 (59.6% of RUmax  RUi) in the case of binding of distamycin to the TATA box flow cell. These results suggest that SPR technology and BIA are useful for the study of the activities of DNA-binding drugs toward different promoter sequences.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Differential binding of distamycin to SA5-NF-kB and SA5-TATA box sensor chips. Distamycin (5 µM) was injected onto SA5-NF-kB (A) and SA5-TATA box (B) sensor chips. After distamycin injection (Ab and Bb), a 15-µl injection of HBS running buffer (Ac and Bc) was performed. The sensorgrams have been elaborated using the BIA-evaluation Software (Pharmacia Biosensors).

Differential Stability of Drug/DNA Complexes Generated After Injection of Mithramycin and Chromomycin onto SA5-Sp1 Sensor Chip. In the experiment shown in Fig. 6, 5 µM mithramycin (Fig. 6A,b) and chromomycin (Fig. 6B,b) were injected onto SA5-Sp1 sensor chips. After injection of DNA-binding drugs, we performed a 30-µl injection of CH/MT running buffer (Fig.6 A,c and B,c). The results obtained demonstrate that chromomycin/Sp1 complexes are more stable than mithramycin/Sp1 mer complexes. The values of RUres  RUi were 33 (6.18% RUmax  RUi) for mithramycin/Sp1 mer interactions and 531 (73.6% RUmax  RUi) for chromomycin/Sp1 mer interactions.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Differential stability of drug/DNA complexes generated after injection of mithramycin and chromomycin onto SA5-Sp1 sensor chip. Mithramycin (5 µM) (A) and chromomycin (5 µM) (B) were injected onto SA5-Sp1 sensor chips. After injection of DNA-binding drugs (Ab and Bb), a 30-µl injection of CH/MT running buffer (Ac and Bc) was performed. The sensorgrams have been elaborated using the BIA-evaluation Software (Pharmacia Biosensors).

Differential Binding of Distamycin Analogues to SA5-NF-kB Sensor Chip. In the experiment shown in Fig. 7, 2 µM distamycin (Fig. 7A,b) and MEN 10567 (Fig. 7B,b) were injected onto the SA5-NF-kB sensor chip. In this experiment the washing step was performed with HBS running buffer (Fig. 7, A,c and B,c). The results obtained demonstrate that the dissociation of distamycin from NF-kB mer is fast, whereas MEN 10567 binds to the SA5-NF-kB sensor chip generating stable complexes. The values RUres  RUi were 10 (6.5% RUmax  RUi) for distamycin/NF-kB mer interactions and 158 (70.2% RUmax  RUi) for MEN 10567/NF-kB mer interactions.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Differential binding of distamycin analogues to SA5-NF-kB sensor chip. Distamycin (2 µM) (A) and MEN 10567 (2 µM) (B) were injected onto the SA5-NF-kB sensor chip. Binding and washing steps were performed with HBS running buffer. After injection of DNA-binding drugs (Ab and Bb), a 30-µl injection of HBS buffer (Ac and Bc) was performed. The sensorgrams have been elaborated using the BIA-evaluation Software (Pharmacia Biosensors).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

DNA-binding drugs have recently been the object of many investigations aimed at developing sequence-selective inhibitors of transcription (Chen et al., 1997). This issue is of great interest in order to design antitumor compounds and antiviral agents, as well as compounds able to induce cellular differentiation (Gambari and Nastruzzi, 1994; Nilsson et al., 1995). For instance, chromomycin binds to GC-rich DNA sequences (van Dyke and Dervan, 1983), inhibits the interaction between the transcription factor Sp1 and target DNA (Snyder et al., 1991; Bianchi et al., 1996), and blocks the transcription directed by promoters that are regulated by Sp1, such as SV40 (Ray et al., 1989), c-myc (Snyder et al., 1991), and HIV-1 (Bianchi et al., 1996). On the other hand, distamycin was found to bind A+T-rich gene sequences and to inhibit the interactions between Oct-1, GATA-1, TBP/TFIID, and Eryf-1 DNA-binding proteins to the relative DNA target sites (Broggini et al., 1989; Feriotto et al., 1994a; Welch et al., 1994; Bianchi et al., 1996).

The major conclusion gathered from the experiments reported in this paper is that interactions of DNA-binding drugs to target DNA sequences can be monitored in real-time using SPR technology and a commercially available biosensor. This is supported by the finding that the increase of RUmax  RUi values is proportional to the amounts of specifically recognized target DNA immobilized on the sensor chip after injection of a same concentration of DNA-binding drugs. This experimental approach has been shown to be useful for detection of low-molecular weight analytes as well as for characterization of weak interactions by BIA (Karlsson and Stahlberg, 1995). Please note that the data obtained during the association phases of the BIA experiments described in the present study are in agreement with previously published results. In particular, distamycin was shown to be able to interact with the NF-kB binding sites and the TATA-box element of the HIV-1 LTR, but unable to bind to the G+C-rich Sp1 binding sites. Furthermore, as found in the present study, mithramycin and chromomycin do not bind to single-stranded DNA and do require Mg²⁺ for efficient interaction with target Sp1 binding sites.

To our knowledge, this is the first report concerning the use of SPR and biosensor technology to compare real-time biospecific interactions between different DNA-binding drugs and specifically recognized target DNA elements. From the practical point of view, this experimental approach appears to be of great interest for the following reasons: (a) results are obtained within 1 h; (b) unlike footprinting and gel retardation, this technology does not have need of ³²P-labeled probes; and (c) BIA allows kinetics studies of both association and dissociation.

We think that this approach could be applied to develop drugs exhibiting differential affinity to target DNA sequences and differential ability to generate stable drugs/DNA complexes.

The results shown in the present study demonstrate indeed that BIA allows identification of different stabilities of molecular complexes generated by the interactions between target DNA sequences and different DNA-binding drugs exhibiting similar chemical structure [i.e., chromomycin and mithramycin (Fig.6); distamycin and MEN 10567 (Fig.7)]. In particular, it is interesting to point out that the different number of pyrroles (three in the case of distamycin, four in the case of MEN 10567) lead to a differential ability to generate stable complexes with the NF-kB binding sites of the HIV-1 LTR. In addition, BIA analysis demonstrates that a same DNA-binding drug (distamycin in our study) interacts with different target DNA sequences (the NF-kB and TATA box binding sites of the HIV-1 LTR) generating complexes that exhibit different stabilities (Fig. 5).

In conclusion, this report is one of the few reports demonstrating the possible use of biosensor technology and SPR to investigate molecular interactions involving low-molecular weight ligands (DNA-binding drugs) and target molecules (DNA). The determination of the stability of molecular complexes generated from the interactions between DNA-binding drugs and target DNA sequences is important in the design of antitumor and antiviral drugs exhibiting a well defined mechanism of action.