2'-O-methyl-modified phosphorothioate antisense oligonucleotides have
http://www.100md.com
《核酸研究医学期刊》
Department of Molecular and Cell Biology, Mail Stop FO31, The University of Texas at Dallas, PO Box 830688, Richardson, TX 75083-0688, USA, 1 Department of Biochemistry and Molecular Biology, The University of Oklahoma Health Sciences Center, Oklahoma City, OK 73190, USA and 2 Howard Hughes Medical Institute, The University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas, TX 75390, USA
*To whom correspondence should be addressed at Department of Molecular and Cell Biology, Mail Stop FO31, The University of Texas at Dallas, PO Box 830688, Richardson, TX 75083-0688, USA. Tel: +1 972 883 2513; Fax: +1 972-883-2409; Email: dongray@utdallas.edu
ABSTRACT
Antisense oligodeoxynucleotides (ODNs) have biological activity in treating various forms of cancer. The antisense effects of two types of 20mer ODNs, phosphorothioate-modified ODNs (S-ODNs) and S-ODNs with 12 2'-O-methyl groups (Me-S-ODNs), targeted to sites 109 and 277 of bcl-2 mRNA, were compared. Both types were at least as effective as G3139 (Genta, Inc.) in reducing the level of Bcl-2 protein in T24 cells following a 4 h transfection at a dose of 0.1 μM. Circular dichroism spectra showed that both types formed A-form duplexes with the complementary RNA, and the melting temperatures were in the order of Me-S-ODN·RNA > normal DNA·RNA > S-ODN·RNA. In comparison with the S-ODN, the Me-S-ODN had reduced toxic growth inhibitory effects, was less prone to bind the DNA-binding domain A of human replication protein A, and was as resistant to serum nucleases. Neither type of oligomer induced apoptosis, according to a PARP-cleavage assay. Hybrids formed with Me-S-ODN sequences were less sensitive to RNase H degradation than those formed with S-ODN sequences. Despite this latter disadvantage, the addition of 2'-O-methyl groups to a phosphorothioate-modified ODN is advantageous because of increased stability of binding and reduced non-specific effects.
INTRODUCTION
Phosphorothioate oligodeoxynucleotides (S-ODNs) were the most common type of the first generation antisense ODNs. These oligomers have resistance to nucleases and the ability to activate RNase H. The major shortcomings of S-ODNs are (i) the relatively low stability of duplexes with complementary RNA and (ii) the toxic effects, possibly related to non-specific protein–S-ODN interactions (1–8). Photocrosslinking experiments by others have shown that human replication protein A (RPA) is one of the major protein products crosslinked with single-stranded DNA oligomers (9).
A prevalent type of second-generation antisense ODN carries 2'-O-sugar modifications, which increases the stability of duplexes with complementary RNA (8). Duplexes formed with the 2'-O-sugar-modified S-ODNs themselves do not have the structure needed to activate RNase H cleavage of the target RNA, and such modified S-ODNs are routinely interrupted with a gap of phosphorothioate nucleotides that do not carry the sugar modification. Monia et al. (10) showed that 2'-O-modifications (such as pentyl, propyl, methyl and fluoro) in chimeric gap-mer S-ODNs had antisense effectiveness that correlated with the Tm values of hybrids formed with the target RNA, which was an Ha-ras oncogene in a reporter gene system. Stability of those gap-mers also correlated with the ability to inhibit Ha-ras mRNA expression in T24 cells, at a dose of 100 nM (11).
McKay et al. (12) found that a gap size of eight was required in 2'-O-methyl chimeras for optimal reduction of PKC mRNA in A549 cells. These authors advocated the use of 2'-O-methoxyethyl-modified sugars, because a 20 nt gap-mer with this modification had increased thermal stability when hybridized with the target mRNA and enhanced antisense effectiveness in reducing PKC mRNA (85% at 100 nM dose), when compared with an analogous 2'-O-methyl chimera (60% at 100 nM dose). In addition, maximum resistance to degradation by snake venom phosphodiesterase was shown to require that chimeras with either 2'-O-methyl or 2'-O-methoxyethyl sugars have a uniform phosphorothioate-modified backbone (12). The 2'-O-methyl modification remains one of the most valuable 2'-sugar modifications because it is cost-effective. In contrast, 2'-O-methoxyethyl-modified oligomers are not commercially available.
Various antisense modalities need to be carefully compared to determine their relative effectiveness (13). In this work, we compared 2'-O-methyl-modified and unmodified S-ODN sequences that are effective in reducing the level of Bcl-2 protein in T24 cells. The anti-apoptotic protein Bcl-2 is localized in the outer mitochondrial membrane, the endoplasmic reticulum and the nuclear envelope (14). Heterodimerization between pro- and anti-apoptotic members of the Bcl-2 family regulate the release of cytochrome c and subsequent apoptosis (15). Because of the Bcl-2 protein’s anti-apoptotic properties and abilities to inhibit tumor cell apoptosis induced by numerous drugs and radiation, the bcl-2 mRNA is an attractive antisense target for cancer treatment (7,15,16).
We show that non-specific effects of phosphorothioates on growth of T24 cells are reduced by the 2'-O-methyl modification. The 2'-O-methyl modification also reduces the binding of S-ODNs to domain A of RPA. Due to reduced non-specific effects, 2'-O-methyl-modified phosphorothioate ODN chimeras may be especially useful, in addition to S-ODNs, for comparative studies.
MATERIALS AND METHODS
Oligonucleotides and protein
Two oligomer sequences were chosen for the present study. Oligomers-109 (5'-CCGCATCCCACTCGTAGCCC-3') and oligomers-277 (5'-CCGGGCTGAGCGCAGGCCCC-3') were 20 nt long and were targeted at starting locations 109 and 277, respectively, of human bcl-2 mRNA (GenBank accession no. M14745 ). Negative control oligomers-C (5'-CTCA TTCCTACCGACACCCC-3') were also 20 nt long and consisted of a scrambled sequence that has no homology in the human genome and that was used as a control sequence by others (17). In preliminary experiments involving 29 S-ODNs targeted to various locations on bcl-2 mRNA, S-ODN-109 was the most effective in reducing Bcl-2 protein in A549 human lung cancer cells by the western blot protocol described below, but using an S-ODN concentration of 0.5 μM. S-ODN-277 was predicted to have the highest duplex stability among the tested oligomers and was as effective as S-ODN-109 in reducing the Bcl-2 level in T24 cells at a concentration of 0.1 μM. The positive control, compound G3139 of Genta, Inc. (Berkeley Heights, NJ), was 18 nt long and was targeted at starting location 32 of human bcl-2 mRNA .
The 11 DNA or chimeric oligomers used for this study and the modifications they contained are detailed in Table 1. For convenience, we denote the chimeric oligomers that contain 2'-O-methyl-modified nucleotides as ODNs, although they contain modified RNA nucleotides. Except for the unmodified ODN-109, all were 100% phosphorothioates. The unmodified ODN, S-ODNs, Me-S-ODN chimeras (with 12 2'-O-methyl sugars), CmeG-S-ODN chimera (with two 5-methyl cytosines in CpG sequences), and unmodified 20mer RNAs complementary to oligomers-109 and oligomers-277 were purchased from Midland Certified Reagent Co. (Midland, TX). The Me-S-ODN chimeras were HPLC-purified, and the unmodified RNA oligomers were purified using trityl selection. Other oligomers were purified by gel filtration. The unmodified RNAs were dissolved in water treated with diethyl pyrocarbonate as in previous work to reduce RNase degradation (18). Absorption spectra at 90°C (corrected for volume expansion due to heating) were used together with (260) values (M–1·cm–1, per mole of strand) to determine concentrations of the oligomers dissolved in water (19,20). Extinction coefficients at 260 nm were calculated as the sums of monomer extinction coefficients; see footnote to Table 1.
Table 1. Sequences, extinction coefficients, and melting temperatures of oligomers
Residues 181–291 of the 70 kDa RPA subunit were expressed in Escherichia coli and purified as described for other RPA domains (21).
Circular dichroism (CD) spectra and melting temperatures (Tm)
CD spectra and melting profiles were measured as in previous work (18–20). To form double-stranded hybrids for Tm measurements, the ODNs and complementary RNAs, at equimolar strand concentrations (2 μM) were mixed in 0.15 M K+ (phosphate buffer, pH 7.0) at room temperature for 2 h. Then, the mixtures were heated to 90°C for 90 s and allowed to cool at room temperature for 24 h. CD spectra were collected at 20°C using a Jasco Model J715 spectropolarimeter (Jasco, Easton, MD). To obtain melting profiles of absorbance as a function of temperature, the OD(260) (optical density at 260 nm) of each mixture was monitored in an Olis-modified Cary 14 (On-Line Instrument Co., Bogart, GA) as the temperature was increased from 20 to 95°C in increments of 1°C. Samples were incubated for 3 min at each temperature to allow for equilibration before taking OD readings. Reported Tm values were the maxima in first-derivative plots of the melting profiles.
Cell culture and transfections
T24 human bladder cancer cells were used because they have a relatively high concentration of Bcl-2 for detection by western blotting. The T24 cells were purchased from the American Type Culture Collection (Rockville, MD) and maintained in McCoy’s 5A medium (Mediatech, Inc., Herdon, VA), supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen, Carlsbad, CA), in 5% CO2 at 37°C. For transfections, T24 cells were plated 24 h prior to treatment in 25 cm2 flasks at a density of 8 x 105 cells per flask. A volume of 8 μl of the delivery reagent lipofectin (Invitrogen) was diluted in 1350 μl of McCoy’s 5A medium and incubated for 45 min at room temperature. ODNs were diluted into 650 μl of McCoy’s 5A media to give the required concentration and incubated for 15 min at room temperature. These two solutions were gently mixed and further incubated for 15 min. The concentrations of lipofectin and ODNs were 3 and 0.1 μM, respectively, in the final mixture. This gave a lipofectin/nucleotide (+/–) charge ratio of 1.5. (To determine whether the optimal lipofectin/charge ratio might differ for the methylated and unmethylated oligomers, electrophoretic mobility shift assay (EMSA) experiments were performed with S-ODN-109 and 5'-Me-S-ODN-109 mixed with increasing amounts of lipofectin to give (+/–) charge ratios of 0, 0.5, 1.0, 2.0 and 4.0. Both types of oligomer were partially neutralized at a charge ratio of 1.0 and were fully neutralized at a charge ratio of 2.0. Transfections in this work were therefore all performed with lipofectin/oligomer mixtures at the intermediate charge ratio of 1.5, which, in our experience with S-ODNs, was sufficient to maximize the antisense effect while minimizing non-specific effects of the lipofectin.) Cells were overlaid with this mixture after prewashing with McCoy’s 5A medium. Following a 4 h transfection of cells at 37°C, cells were washed and incubated in McCoy’s 5A medium supplemented with 10% FBS for an additional 24 h. Untreated (i.e. mock-treated) controls were identical except that during the treatment period they were overlaid with media without lipofectin or ODNs.
Western blot analysis
Cells treated with ODNs were washed with TBS (0.05 M Tris–HCl, pH 7.4, 0.1 M NaCl), and extracted at 0°C for 25 min in lysis buffer with 5% (v/v) protease inhibitor cocktail (Sigma, St Louis, MO). Cell debris was removed by microcentrifugation at 4°C. Protein concentrations were determined using the Bio-Rad protein assay (Richmond, CA). Cell extracts (20 μg) were subjected to SDS–polyacrylamide gel electrophoresis and then transferred to PVDF membranes (Millipore, Bedford, MA). The portions of the membranes containing Bcl-2 (MW 26 kDa) and Raf-1 (MW 73 kDa) were blocked with 5% non-fat dry milk in TTBS , and probed with corresponding primary antibodies at room temperature overnight (for Bcl-2: Cat. sc-509, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; for C-raf-1: Cat. 610151, BD Transduction Laboratories, San Diego, CA). After washing in TTBS, the membranes were incubated for 1 h at room temperature with 5% non-fat dry milk in TTBS containing a 1:15 000 dilution of an alkaline phosphatase-conjugated secondary antibody (Jackson Immuno Research, West Grove, PA). The immunocomplexes were visualized by enhanced chemifluorescence reagent (ECF substrate; Amersham Pharmarcia Biotech, Piscataway, NJ), followed by scanning with a phosphoimager (Storm 860, Molecular Dynamics) and quantitation using ImageQuant version 5.0 (Molecular Dynamics). The portion of the gel containing actin (MW 42 kDa) was stained by a modification of the Fairbanks Coomassie staining method (22). To calculate the reduction of Bcl-2 protein levels, actin band densities of untreated control samples (at 50 and 100% concentrations) and the treated samples were used to calculate the loading ratios of the treated cell lysates. Bcl-2 band intensities of untreated control samples (at 50 and 100% concentrations), with background subtracted, were fitted with a second-order equation from which the apparent percentages of Bcl-2 protein in the treated samples were determined. These percentages were finally corrected for any unequal loading, using the loading ratios from the actin bands.
Growth inhibition
The effects on cell growth of the antisense ODNs were determined by thymidine incorporation (23,24). T24 cells were plated 24 h prior to treatment in 24 well plates at a density of 2 x 104 cells per well. Cells were treated with ODNs using the procedure described above, except that after the 4 h treatment, cells were washed and incubated in McCoy’s 5A medium supplemented with 10% FBS plus 1 μCi of thymidine for 24 h. After the 24 h incubation, media in the wells was removed by aspiration and cells were incubated in Hanks balanced salt solution with trypsin-EDTA (Invitrogen) for 5 min at 37°C. A cell harvester (Cambridge Technology, Watertown, MA) was used to harvest cells on glass fiber filters. The percentage of growth inhibition was calculated as 100 x thymidine incorporation by the treated cells)/(counts due to thymidine incorporation by untreated cells)].
Assay for apoptosis
As a marker of cells undergoing apoptosis, the percentage of PARP that was cleaved to give an 89 kDa cleavage product was determined in treated cell lysates by western blot analysis using a PARP antibody (BD Biosciences, San Diego, CA).
Stability of ODNs in serum
To test the stabilities of ODNs in serum, the ODNs (at 10 μM concentration) were incubated in McCoy’s 5A medium supplemented with 10% FBS at 37°C. At each time point (0, 24, 48 and 72 h), 12 μl of each mixture was taken and the reaction was terminated by addition of 12 μl of loading buffer (0.1 M Tris–HCl pH 8.4, 0.09 M boric acid, 7 M urea, 0.1% bromophenol blue, 0.01% xylene cyanol). Samples were subjected to electrophoresis in a denaturing 20% polyacrylamide gel with 7 M urea. After electrophoresis, the gels were stained with SYBR Green II (Molecular Probes, Eugene, OR) prepared in TBE (0.089 M Tris base, 0.089 M boric acid and 1 mM EDTA pH 8.0) and scanned with a phosphoimager.
EMSA of RPA70-A binding to ODNs
Two types of EMSA experiments were used to investigate the binding of RPA70-A to S-ODN-109 versus 5'-Me-S-ODN-109. ODNs were end-labeled with T4 polynucleotide kinase (Promega, Madison, WI). (i) Competition experiments were performed to compare the binding affinities of RPA70-A to S-ODN-109 and 5'-Me-S-ODN-109. RPA70-A was added to saturate the 32P end-labeled S-ODN-109 at 1 μM. After 15 min incubation at room temperature, unlabeled competitor, either S-ODN-109 or 5'-Me-S-ODN-109 was added, and the mixtures were incubated for an additional 15 min. After addition of competitor, the concentration of ODN that was initially saturated with RPA70-A was 0.5 μM in all mixtures. (ii) Further competition experiments were performed to compare binding affinities of RPA70-A to S-ODN-109 versus 5'-Me-S-ODN-109 in the presence of the complementary 20mer RNA. RPA70-A was added to complementary RNA, and, after 15 min incubation, either 32P-labeled S-ODN or 32P-labeled Me-S-ODN was added and incubated for an additional 5 h at room temperature. The ODN and complementary RNA concentrations were each 1 μM in the final mixtures. The final / ratios (molar protein monomer/molar nucleotide ratios) were 0, 0.2 and 0.5, where the nucleotide concentrations were the total of ODN plus RNA nucleotides.
In each of the above cases, the mixtures were subjected to 20% non-denaturing polyacrylamide gel electrophoresis in TBE buffer, the gels were scanned with a phosphoimager, and the data were analyzed using ImageQuant 5.0.
RNase H elicitation
To examine the ability of RNase H to degrade hybrids formed with modified ODNs, 5 pmol of each modified ODN and complementary 32P 5'-end-labeled RNA were hybridized in 60 mM Tris–HCl (pH 7.8) containing 2 mM dithiothreitol, 60 mM KCl and 2.5 mM MgCl2. The desired hybrid duplexes were formed by incubating at 90°C for 2 min and slow cooling to room temperature for at least 1 h. Then, solutions of hybrids were treated with 0.05 units of Escherichia coli RNase H (Invitrogen) at 37°C for 20 min. Samples were analyzed as in the EMSA experiments described above.
RESULTS
Reductions in Bcl-2 level
In order to determine whether antisense effectiveness correlates with hybridization stabilities of the different ODN modifications, T24 human bladder cancer cells were transfected with the modified ODNs targeted to sites 109 and 277 on bcl-2 mRNA. The ODNs were used at a concentration of 0.1 μM. (In preliminary experiments, we found that there were no substantial decreases of Bcl-2 level in T24 cells when the concentrations of S-ODNs or Me-S-ODNs targeted to these two sites were increased above 0.1 μM under our conditions.) As shown in Figure 1, the S-ODN-109 and 5'-Me-S-ODN-109 resulted in 57 ± 5 and 70 ± 2% (SD) reductions of Bcl-2 level, respectively, with the Me-S-ODN having the greater effect. The analogous oligomers targeted to site 277, S-ODN-277 and 5'-Me-S-ODN-277, resulted in 63 ± 4 and 45 ± 5% reductions of the Bcl-2 level, respectively. In that case, the Me-S-ODN had the lesser effect. The S-ODN-109 and S-ODN-277 had the same effect, within error, as an effective positive control, G3139, directed against the first six codons of the bcl-2 mRNA (7). The S-ODN and Me-S-ODN controls averaged 3 ± 10 and 12 ± 10%, respectively, and were not significantly different in their effects on the Bcl-2 level (Fig. 1).
Figure 1. Reduction of Bcl-2 level in T24 cells by ODNs targeted to different sites on bcl-2 mRNA. The percent reductions were relative to the amount of Bcl-2 in mock-treated cells from western blots. Oligomer sequences and S-ODN and 5'-Me-S-ODN modifications are shown in Table 1. G3139 is an 18mer S-ODN positive control. Error bars are ±1 SD of at least three determinations. The ODN doses were 0.1 μM for Figures 1 and 3.
The level of Raf-1 protein, a protein in a pathway separate from the Bcl-2 pathway, was monitored as an additional indication of non-specific cellular effects of the oligomer treatments. For the three Me-S-ODNs used to obtain data for Figure 1, the level of Raf-1 protein was reduced by an average of 21%, no different from the average of 24% reduction found after treatments by the three 20mer S-ODNs (data not shown).
The antisense effectiveness of three different Me-S-ODN chimeras, all targeted to site 109, were compared to study the effect of placing the 12 2'-O-methyl groups in one block at the 5' end, in one block at the 3' end or in two equal blocks at both ends to form a gap-mer. The Me-S-ODN chimeric sequences (5'-Me-S-ODN, 3'-Me-S ODN and gap-Me-S-ODN) are shown in Table 1. The hybrids of these chimeras paired with the target RNA had Tm values that differed by only 4°C (Table 1), and the reductions of Bcl-2 level caused by treatments with the three chimeras were all in the range 73 ± 4% (data not shown). Therefore, for this 20mer sequence, the location of the one or two blocks of 2'-O-methyl nucleotides was unimportant.
Because additional work (below) focused on the oligomers targeted to site 109, we examined whether the reduction of the Bcl-2 level by S-ODN-109 was influenced by the presence of CpG sequences. Synthetic DNAs containing unmethylated CpG dinucleotides can stimulate an immune response in vertebrates, and CpG islands in antisense ODNs can cause non-sequence-specific effects (7,25,26). Since sequence 109 contained two CpG islands, an ODN was prepared (CmeG-S-ODN; see Table 1) in which C-5-methyl groups were added to cytosines in these CpG sequences to suppress any non-specific effects. In simultaneous experiments, the reductions in the Bcl-2 level by S-ODN-109 and CmeG-S-ODN-109 were essentially identical (62 ± 4 and 60 ± 11%, respectively) demonstrating that the reduction of Bcl-2 level by the S-ODN-109 was not affected by CpG islands. This was not unexpected, since T24 cells are not immune system cells.
From all the data above, we take the reductions in Bcl-2 levels by the S-ODN and Me-S-ODN sequences targeted to sites 109 and 277 to denote primarily the antisense effects of these oligomers.
Although the relative effectiveness of the two types of ODN depended on the target site sequence, the oligomers targeted to site 109 show that an Me-S-ODN can have an apparent antisense effectiveness greater than that of a cognate S-ODN. The actual specific activity of the Me-S-ODN may be even greater due to decreased non-specific effects, such as a reduced toxic effect on cell growth (see below).
CD spectra of ODN·RNA hybrids
The ODN·RNA duplexes were characterized by CD spectral measurements. CD measurements were performed with the samples in 0.15 M K+ (phosphate buffer, pH 7.0), 20°C. The spectra of the ODN·RNA duplexes were significantly different from the averaged spectra of the individual strands (Fig. 2). Spectra of the hybrids showed an increase in the magnitude of the long wavelength positive band above 260 nm, an increase in the negative band at 210 nm, and a shift in the crossover near 250 nm to shorter wavelengths, relative to the averaged spectra. These CD characteristics are typical features of DNA·RNA duplexes (18,20) and confirm the formation of the desired duplexes.
Figure 2. Measured CD spectra of ODN·RNA hybrids compared with the averaged spectra of the individual strands. Hybrids were formed by pairing the following ODNs with the complementary RNA-109 strand: (A) Unmodified ODN-109, (B) S-ODN-109, (C) 5'-Me-S-ODN-109. (Solid line) Measured spectra; (dashed line) averaged spectra of the individual, unpaired strands. Spectra were taken at 20°C with samples in 0.15 M K+ (phosphate buffer, pH 7.0).
Tm values of ODN·RNA hybrids
The S-ODN and Me-S-ODN were further characterized by the stabilities (Tm values) of the hybrids they formed with complementary RNAs. The Me-S-ODN·RNA hybrids had relatively high Tm values (69°C to above 82°C), but the corresponding S-ODN·RNA hybrids had lower Tm values (55 and 66°C). The Tm difference between the three chimeric Me-S-ODN-109·RNA hybrids and the S-ODN-109·RNA hybrid was 14–18°C and that between one chimeric Me-S-ODN-277·RNA and the S-ODN-277·RNA was >16°C (Table 1). The Me-S-ODNs·RNA hybrids presumably have higher Tm values than the corresponding hybrids formed with the S-ODNs because the 2'-O-methyl group results in an RNA-like conformation, and RNA duplexes have higher Tm values than DNA·RNA hybrids (6,18,27,28).
Growth inhibition by thymidine incorporation
Thymidine incorporation in T24 cells was used to assay the inhibition of DNA synthesis and growth inhibitory effects of S-ODN-109 and 5'-Me-S-ODN-109. As shown in Figure 3, the same level of growth inhibition was measured for the controls as for the antisense ODNs. These effects were significant for the phosphorothioate-modified oligomer and were minimal for the Me-S-ODN oligomer. There was no effect on thymidine incorporation by treatment with lipofectin alone (data not shown). We conclude that the effects on growth inhibition by this assay were essentially caused by the cytotoxic effects of the ODN modifications and were not secondary to a reduction of the Bcl-2 level.
Figure 3. Growth inhibition of T24 cells following treatments with the S-ODN and the 5'-Me-S-ODN targeted to site 109 on bcl-2 mRNA. Growth inhibition was determined as the decrease in thymidine incorporation, relative to the incorporation in mock-treated cells. Error bars are ±1 SD of three determinations.
Apoptosis
We tested whether a reduction in the level of the Bcl-2 anti-apoptotic protein was correlated with apoptosis of T24 cells. PARP is one of the main cleavage targets of caspase-3, which is activated in the apoptosis cascade following release of cytochrome c, so that the cleavage of PARP serves as a marker of cells undergoing apoptosis (29). Cells were treated with ODNs as described in Materials and Methods for reducing the Bcl-2 protein level. The percentage of the 89 kDa PARP cleavage product after treatments with S-ODN-109, 5'-Me-S-ODN-109, and the two negative controls, ranged from 2 to 6%, not detectably different from the 2% cleavage product found in lysates of untreated controls (data not shown). Thus, according to this assay, there was no detectable apoptotic effect with either the S-ODN-109 of 5'-Me-S-ODN-109, although these oligomers reduce the Bcl-2 level by 57 and 70%, respectively, and they have differing effects on cell growth according to the thymidine incorporation assay.
Stability of ODNs in serum
An important consideration in the design of antisense ODN modifications is the degree to which the ODN is rendered nuclease resistant. The resistance to degradation in 10% FBS of the unmodified ODN-109 (phosphodiester) and three cognate modified ODNs (the S-ODN, 5'-Me-S-ODN and Gap-Me-S-ODN) was examined as described in Materials and Methods. The results are shown in Figure 4. While little of the unmodified ODN remained intact after 24 h, all three of the modified ODNs exhibited a half-life of >72 h in 10% FBS. This was in agreement with results by Grünweller et al. (30), who reported the half-life of an S-ODN as being >>48 h under similar conditions. Thus, the S-ODN and Me-S-ODN were completely stable over the time course of the transfections and incubations to reduce the level of Bcl-2, and the difference in antisense effect (Fig. 1) was not obviously due to a difference in nuclease stability.
Figure 4. Survival in FBS of oligomers 109. (A) Bands of intact ODNs in polyacrylamide gels after incubation in 10% FBS at 37°C for the indicated times. Bands were visualized by staining with SYBR Green II. (B) Percentages of uncleaved oligonucleotides as a function of incubation time.
RPA70-A binding affinities of S-ODN-109 and Me-S-ODN-109
One non-specific effect of ODNs is their binding to cellular proteins (1–9). The most abundant single-stranded DNA binding protein in human cells is RPA (31), and domain A of the 70 kDa subunit of RPA is the site of initial binding to single-stranded DNA (32–34). The binding site of RPA70-A accommodates 5 nt, similar to the binding site size of the Ff g5p, which has been used to assess the effects of different nucleotide modifications (6,32). Therefore, we determined the relative binding affinity of RPA for S-ODN-109 and 5'-Me-S-ODN-109 in both the absence and presence of the complementary RNA sequence.
To compare the RPA70-A binding affinities for the S-ODN and the Me-S-ODN, saturated S-ODN-109:RPA70-A complexes were prepared with 32P-labeled S-ODN and various amounts of unlabeled competing S-ODN-109 or Me-S-ODN-109 were added (in phosphate, pH 7.0, 37°C). The EMSA results after 15 min of incubation with the competitor are shown in Figure 5A and B. At an 8-fold excess of unlabeled competing S-ODN, 61% of the labeled S-ODN:RPA70-A complex was completely dissociated (Fig. 5A). The same results were obtained after up to 4 h of incubation with competing S-ODN-109, indicating that equilibrium had been reached (data not shown). The band of complex became more diffuse as the labeled S-ODN was dissociated because RPA70-A does not bind with high cooperativity, and up to five protein molecules can be bound per ODN 20mer. The same 8-fold excess of competing Me-S-ODN dissociated a maximum of 15% of the S-ODN:RPA70-A complex, 4-fold less than the same amount of competing S-ODN (Fig. 5B). Additional experiments in which the labeled S-ODN and unlabeled Me-S-ODN were premixed before adding the protein demonstrated that even at a 15-fold excess of Me-S-ODN the amount of completely dissociated labeled S-ODN was still only 15%. Finally, the ratio of 5'-Me-S-ODN-109 to S-ODN-109 needed for 50% dissociation of an unmodified ODN:RPA70-A complex was determined to be 9.9 ± 0.1 (range) in competition experiments (data not shown). In summary, the binding affinity of RPA70-A was estimated to be 10-fold greater for the S-ODN than for the 5'-Me-S-ODN, even though the latter has a stretch of eight phosphorothioate nucleotides.
Figure 5. Competition binding experiments. The relative binding affinity of RPA70-A for S-ODN and 5'-Me-S-ODN was determined by competition of (A) cold S-ODN-109 or (B) cold 5'-Me-S-ODN with the 32P end-labeled S-ODN-109 in a preformed complex with RPA70-A. The percent dissociation of S-ODN from the complex was calculated from the band intensity of free 32P-labeled S-ODN released from the 32P S-ODN:RPA70-A complex. (C) 32P end-labeled S-ODN-109 or 32P end-labeled 5'-Me-S-ODN-109 were titrated with RPA70-A in the presence of complementary RNA. In lanes 1 and 5, there were ODNs only, without RNA. In the other lanes, RNA and RPA70-A were added and, after 15 min, the complementary ODNs were added. The / ratios are the RPA70-A protein molar concentrations per total molar concentrations of nucleotides (DNA plus RNA). Multiple bands of the ODN·RNA duplexes were probably the result of some degradation of the RNA combined with a reduced stability of hybrids formed with the S-ODN. The complexes of RPA70-A with 5'-Me-S-ODN were more diffuse than those with the S-ODN, likely due to multiple arrangements of up to five protein molecules on the chimeric oligomer.
In a second type of experiment, we examined the ability of the S-ODN and the 5'-Me-S-ODN to bind to the complementary RNA in the presence of RPA70-A. At increasing RPA70-A concentrations, the S-ODN·RNA hybrid was less able to form than was the Me-S-ODN·RNA hybrid, as seen in Figure 5C. From quantitation of the results in this Figure, at a / ratio of 0.5, only 25% of the S-ODN was involved in the formation of a S-ODN·RNA hybrid while 80% of the Me-S-ODN was able to form an Me-S-ODN·RNA hybrid. These experiments were consistent with a higher affinity of RPA70-A for the S-ODN than for the 5'-Me-S-ODN shown in Figure 5A and B and with a higher Tm for the 5'-Me-S-ODN·RNA hybrid than for the S-ODN·RNA hybrid (Table 1).
RNase H elicitation
The ability of antisense ODNs to elicit RNase H degradation of the target mRNA is considered to be one of the most important criteria for antisense effectiveness. The S-ODNs and 5'-Me-S-ODNs targeted to RNA sites 109 and 277 were compared for their ability to recruit RNase H. RNase H degradation of the target RNAs (unmodified 20mer RNAs complementary to oligomers-109 and oligomers-277) was higher for hybrids formed with the S-ODNs than with the Me-S-ODNs (Fig. 6). For example, 80% of the hybrid formed with S-ODN-109 was degraded but only 30% of the hybrid formed with Me-S-ODN-109 was degraded under the same conditions. The lowered ability of RNase H to degrade hybrids with the Me-S-ODN was not unexpected. The Me-S-ODNs are chimeric, with 12 nt that are 2'-O-methyl and RNA-like. Therefore, a major portion of the hybrid formed with the Me-S-ODN sequesters the target RNA in an RNA·RNA duplex that is not recognized by RNase H. The most effective hybrid in recruiting RNase H was that formed by S-ODN-109, which interestingly had the lowest Tm.
Figure 6. RNase H elicitation of modified ODN·RNA duplexes. With the assumption of equal loading, the percent degradation was calculated from the reduction in the intensity of the DNA·RNA duplex band after RNase H treatment. The complementary 20mer RNAs were 32P end-labeled. Bands labeled as degraded products in the RNase H negative lanes were from labeled fragments of RNA and over-represent the amount of degradation in the RNA samples because at least two fragments result from a single cleavage. They did not affect the calculation of percent degradation of hybrids by RNase H.
DISCUSSION
Hybrids formed with 20mer S-ODNs and Me-S-ODNs and their cognate RNA sequences were shown to have the CD spectra of DNA·RNA and RNA·RNA A-form structures (Fig. 2). As expected, melting temperatures of the hybrids formed with the Me-S-ODNs were above those formed with the S-ODNs (Table 1). The two S-ODN sequences whose properties were compared in this work were as effective in reducing the level of Bcl-2 protein in T24 cells as the well studied Bcl-2 antisense sequence G3139 (7) (Fig. 1), independent of the presence of CpG islands. Treatments with S-ODN-109 and 5'-Me-S-ODN-109 did not lead to apoptosis according to a PARP-cleavage assay. G3139 also does not induce apoptosis in all cell lines (35), and Benimetskaya et al. (36) reported that G3139-induced down regulation of Bcl-2 in T24 cells is insufficient to decrease cellular viability. Neither S-ODN-109 nor Me-S-ODN-109 oligomers were significantly degraded by serum over the time course of the antisense treatments (Fig. 4). The latter is important in comparative studies, because without having the nuclease protection of a phosphorothioate backbone the 2'-O-methyl modification can be less effective in an antisense oligomer than a simple S-ODN (30).
Eight phosphorothioate nucleotides are needed in gap-mers with 2'-O-methyl-modified sugars and locked nucleic acids to fully activate RNase H (12,37). Oligomers containing 12 2'-O-methyl-modified nucleotides at the 5' ends were less able to induce RNase H degradation of hybridized RNA than were the analogous S-ODNs (Fig. 6). Nevertheless, the abilities of different chimeric arrangements of the 12 modified nucleotides, in one block at either end or as a gap-mer (5'-Me-S-ODN, 3'-Me-S ODN and gap-Me-S-ODN), were equally able to reduce the level of Bcl-2 protein. This implies that slight differences in the ability to activate RNase H are not as important as the ability to form stable hybrids with the target RNA and other factors in antisense effectiveness. Although we did not examine mRNA levels by northern blot analysis in this work, others have shown that antisense phosphorothioate oligomers carrying the 2'-O-methyl modification are unable to reduce endogenous PKC mRNA levels unless there is a gap without the sugar modification so that RNase H can be activated (12). Thus, we reasonably expect that the main mechanism of action of our oligomers in reducing the level of Bcl-2 protein was by RNase H inactivation of the encoding mRNA.
We have shown for the first time that the 2'-O-methyl modification results in a reduction of important non-specific effects. Thymidine incorporation into DNA has been used by others as a means of assaying cellular growth and proliferation, because dead and dying cells do not incorporate thymidine into their DNA (23,24). Under our transfection conditions, oligomers carrying just the phosphorothioate modification were about three times as inhibitory to the growth of T24 cells as oligomers that included 12 2'-O-methyl-modified nucleotides (Fig. 3). Moreover, we have determined that the reduced non-specific effects of Me-S-ODNs could be related to their reduced binding to cellular proteins, such as RPA (Fig. 5A and B). There was a decided decrease in binding affinity for the RPA70-A domain, by a factor of 10-fold, when 2'-O-methyl sugars were included in the phosphorothioate ODN. This decrease in binding affinity for RPA helps ensure that an ODN·RNA hybrid will remain stable in the presence of this cellular protein (Fig. 5C).
Others have found that S-ODNs and G-tetraplex-forming oligomers bind to nucleolin and that this binding can result in altered cell morphology and apoptosis (38–40). It remains to be determined whether Me-S-ODNs have reduced nucleolin-binding properties, consistent with the possibility that the 2'-O-methyl modification leads to reduced non-specific effects via reduced binding to cellular proteins.
Our previous work with the model Ff g5p single-stranded DNA-binding protein showed that a fully 2'-O-methyl-modified S-dA36 oligomer had about one-half of the binding affinity for g5p as did the simple S-dA36 oligomer (6). The use of a model protein such as the g5p, or the RPA70-A domain used in the present study, to assay relative binding affinities may help establish which nucleotide modifications are likely to have reduced non-specific effects in cells.
In summary, chimeric ODNs carrying 12 2'-O-methyl-modified nucleotides would be a good choice to include, in addition to simple phosphorothioates, in much needed comparative studies (4) of different antisense molecules. Our results also suggest that there could be a reduction in cost of synthesizing oligomers with 2'-O-modified sugars by including all 12 modified nucleotides at the 5' or the 3' end of the oligomers.
ACKNOWLEDGEMENTS
This work was performed by B.H.Y. in partial fulfillment of the requirements for the PhD degree in the Department of Molecular and Cell Biology, The University of Texas at Dallas. Support was provided in part by grants from the Robert A. Welch foundation (AT-503), eXegenics, Inc. (Dallas, TX) and the Texas Advanced Technology Program (009741-0021-1999) to D.M.G. and from the US National Institutes of Health to A.B. (RO1 GM061192).
REFERENCES
Stein,C.A. (1996) Phosphorothioate antisense oligodeoxynucleotides: questions of specificity. Trends Biotechnol., 14, 147–149.
Lebedeva,I. and Stein,C.A. (2001) Antisense oligonucleotides: promise and reality. Annu. Rev. Pharmacol. Toxicol., 41, 403–419.
Stein,C.A. (2001) The experimental use of antisense oligonucleotides: a guide for the perplexed. J. Clin. Invest., 108, 641–644.
Braasch,D.A. and Corey,D.R. (2002) Novel antisense and peptide nucleic acid strategies for controlling gene expression. Biochemistry, 41, 4503–4510.
Dove,A. (2002) Antisense and sensibility. Nat. Biotechnol., 20, 121–124.
Mou,T.C. and Gray,D.M. (2002) The high binding affinity of phosphorothioate-modified oligomers for Ff gene 5 protein is moderated by the addition of C-5 propyne or 2'-O-methyl modifications. Nucleic Acids Res., 30, 749–758.
Dias,N. and Stein,C.A. (2003) Potential roles of antisense oligonucleotides in cancer therapy. The example of Bcl-2 antisense oligonucleotides. Eur. J. Pharm. Biopharm., 54, 263–269.
Shen,L., Siwkowski,A., Wancewicz,E.V., Lesnik,E., Butler,M., Witchell,D., Vasquez,G., Ross,B., Acevedo,O., Inamati,G., Sasmor,H., Manoharan,M. and Monia,B.P. (2003) Evaluation of C-5 propyne-containing oligonucleotides in vitro and in vivo. Antisense Nucleic Acid Drug Dev., 13, 129–142.
Revers,F., Cario,M., Cao,T.L. and Cazenave,C. (1999) Detection of proteins binding to short RNA·DNA hybrids or short antisense oligonucleotides in Xenopus laevis oocytes and human macrophage cell extracts by photoaffinity radiolabeling. Antisense Nucleic Acid Drug Dev., 9, 317–331.
Monia,B.P., Lesnik,E.A., Gonzalez,C., Lima,W.F., McGee,D., Guinosso,C.J., Kawasake,A.M., Cook,P.D. and Freier,S.M. (1993) Evaluation of 2'-O-modified oligonucleotides containing 2'-O-deoxy gaps as antisense inhibitors of gene expression. J. Biol. Chem., 268, 14514–14522.
Monia,B.P., Johnston,J.F., Sasmor.H. and Cummins,L.L. (1996) Nuclease resistance and antisense activity of modified oligonucleotides targeted to Ha-ras. J. Biol. Chem., 271, 14533–14540.
McKay,R.A., Miraglia,L.J., Cummins,L.L., Owens,S.R., Sasmor,H. and Dean,N.M. (1999) Characterization of a potent and specific class of antisense oligonucleotide inhibitor of human protein kinase C- expression. J. Biol. Chem., 274, 1715–1722.
Braasch,D.A., Liu,Y. and Corey,D.R. (2002) Antisense inhibition of gene expression in cells by oligonucleotides incorporating locked nucleic acids: effect of mRNA target sequence and chimera design. Nucleic Acids Res., 30, 5160–5167.
Hockenbery,D., Nunez,G., Milliman,C., Schreiber,R.D. and Korsmeyer,S.J. (1990) Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature, 348, 334–336.
Hengartner,M.O. (2000) The biochemistry of apoptosis. Nature, 407, 770–776.
Chao,D.T. and Korsmeyer,S.J. (1998) BCL-2 family: regulators of cell death. Annu. Rev. Immunol., 16, 395–419.
Ziegler,A., Luedke,G.H., Fabbro,D., Altmann,K.H., Stahel,R.A. and Zangemeister-Wittke,U. (1997) Induction of apoptosis in small-cell lung cancer cells by an antisense oligodeoxynucleotide targeting the Bcl-2 coding sequence. J. Natl Cancer Inst., 89, 1027–1036.
Hashem,G.M., Pham,L., Vaughan,M.R. and Gray,D.M. (1998) Hybrid oligomer duplexes formed with phosphorothioate DNAs:CD spectra and melting temperatures of S-DNA·RNA hybirds are sequence-dependent but consistent with similar heteronomous conformations. Biochemistry, 37, 61–72.
Gray,D.M., Hung,S.H. and Johnson,K.H. (1995) Absorption and circular dichroism spectroscopy of nucleic acid duplexes and triplexes. Methods Enzymol., 246, 19–34.
Clark,C.L., Cecil,P.K., Singh,D. and Gray,D.M. (1997) CD, absorption and thermodynamic analysis of repeating dinucleotide DNA, RNA and hybrid duplexes 12·12 and the influence of phosphorothioate substitution. Nucleic Acids Res., 25, 4098–4105.
Bochkareva,E., Frappier,L., Edwards,A.M. and Bochkarev,A. (1998) The RPA32 subunit of human replication protein A contains a single-stranded DNA-binding domain. J. Biol. Chem., 273, 3932–3936.
Wong,C., Sridhara,S., Bardwell,J.C. and Jakob,U. (2000) Heating greatly speeds Coomassie blue staining and destaining. Biotechniques, 28, 426–428, 430, 432.
Robinson,L.A., Smith,L.J., Fontaine,M.P., Kay,H.D., Mountjoy,C.P. and Pirruccello,S.J. (1995) c-myc antisense oligodeoxyribonucleotides inhibit proliferation of non-small cell lung cancer. Ann. Thorac. Surg., 60, 1583–1591.
Yang,D.C., Jiang,X., Elliott,R.L. and Head,J.F. (2002) Antisense ferritin oligonucleotides inhibit growth and induce apoptosis in human breast carcinoma cells. Anticancer Res., 22, 1513–1524.
Krieg,A.M., Yi,A.K., Matson,S., Waldschmidt,T.J., Bishop,G.A., Teasdale,R., Koretzky,G.A. and Klinman,D.M. (1995) CpG motifs in bacterial DNA trigger direct B-cell activation. Nature, 374, 546–549.
Krieg,A.M. (1999) Mechanisms and applications of immune stimulatory CpG oligodeoxynucleotides. Biochim. Biophys. Acta, 1489, 107–116.
Hung,S.-H., Yu,Q., Gray,D.M. and Ratliff,R.L. (1994) Evidence from CD spectra that d(purine)·r(pyrimidine) and r(purine)·d(pyrimidine) hybrids are in different structural classes. Nucleic Acids Res., 22, 4326–4334.
Nishizaki,T., Iwai,S., Ohtsuka,E. and Nakamura,H. (1997) Solution structure of an RNA·2'-O-methylated RNA hybrid duplex containing an RNA·DNA hybrid segment at the center. Biochemistry, 36, 2577–2585.
Oliver,F.J., de la Rubia,G., Rolli,V., Ruiz-Ruiz,M.C., de Murcia,G. and Murcia,J.M. (1998) Importance of poly(ADP-ribose) polymerase and its cleavage in apoptosis. Lesson from an uncleavable mutant. J. Biol. Chem., 273, 33533–33539.
Grünweller,A., Wyszko,E., Bieber,B., Jahnel,R., Erdmann,V.A. and Kurreck,J. (2003) Comparison of different antisense strategies in mammalian cells using locked nucleic acids, 2'-O-methyl RNA, phosphorothioates and small interfering RNA. Nucleic Acids Res., 31, 3185–3193.
Wold,M.S. (1997) Replication protein A: a heterotrimeric, single-stranded DNA-binding protein required for eukaryotic DNA metabolism. Annu. Rev. Biochem., 66, 61–92.
Bochkarev,A., Pfuetzner,R.A., Edwards,A.M. and Frappier,L. (1997) Structure of the single-stranded-DNA-binding domain of replication protein A bound to DNA. Nature, 385, 176–181.
Arunkumar,A.I., Stauffer,M.E., Bochkareva,E., Bochkarev,A. and Chazin,W.J. (2003) Independent and coordinated functions of the replication protein A tandem high affinity ssDNA binding domains. J. Biol. Chem., 278, 41077–41082.
Bochkarev,A. and Bochkareva,E. (2004) From RPA to BRCA2: lessons from single-stranded DNA binding by the OB-fold. Curr. Opin. Struct. Biol., 14, 36–42.
Lai,J.C., Benimetskaya,L., Santella,R.M., Wang,Q., Miller,P.S. and Stein,C.A. (2003) G3139 (oblimersen) may inhibit prostate cancer cell growth in a partially bis-CpG-dependent non-antisense manner. Mol. Cancer Ther., 2, 1031–1043.
Benimetskaya,L., Miller,P., Benimetsky,S., Maciaszek,A., Guga,P., Beaucage,S.L., Wilk,A., Grajkowski,A., Halpern,A.L. and Stein,C.A. (2001) Inhibition of potentially anti-apoptotic proteins by antisense protein kinase C- (Isis 3521) and antisense bcl-2 (G3139) phosphorothioate oligonucleotides: relationship to the decreased viability of T24 bladder and PC3 prostate cancer cells. Mol. Pharm., 60, 1296–1307.
Kurreck,J., Wyszko,E., Gillen,C. and Erdmann,V.A. (2002) Design of antisense oligonucleotides stabilized by locked nucleic acids. Nucleic Acids Res., 30, 1911–1918.
Weidner,D.A., Valdez,B.C., Henning,D., Greenberg,S. and Busch,H. (1995) Phosphorothioate oligonucleotides bind in a non sequence-specific manner to the nuclear protein C23/nucleolin. FEBS Lett., 366, 146–150.
Bates,P.J., Kahlon,J.B., Thomas,S.D., Trent,J.O. and Miller,D.M. (1999) Antiproliferative activity of G-rich oligonucleotides correlates with protein binding. J. Biol. Chem., 274, 26369–26377.
Xu,X., Hamhouyia,F., Thomas,S.D., Burkel,T.J., Girvant,A.C., McGregor,W.G., Trent,J.O., Miller,D.M. and Bates,P.J. (2001) Inhibition of DNA replication and induction of S phase cell cycle arrest by G-rich oligonucleotides. J. Biol. Chem., 276, 43221–43230.(Byong Hoon Yoo, Elena Bochkareva1, Alexe)
*To whom correspondence should be addressed at Department of Molecular and Cell Biology, Mail Stop FO31, The University of Texas at Dallas, PO Box 830688, Richardson, TX 75083-0688, USA. Tel: +1 972 883 2513; Fax: +1 972-883-2409; Email: dongray@utdallas.edu
ABSTRACT
Antisense oligodeoxynucleotides (ODNs) have biological activity in treating various forms of cancer. The antisense effects of two types of 20mer ODNs, phosphorothioate-modified ODNs (S-ODNs) and S-ODNs with 12 2'-O-methyl groups (Me-S-ODNs), targeted to sites 109 and 277 of bcl-2 mRNA, were compared. Both types were at least as effective as G3139 (Genta, Inc.) in reducing the level of Bcl-2 protein in T24 cells following a 4 h transfection at a dose of 0.1 μM. Circular dichroism spectra showed that both types formed A-form duplexes with the complementary RNA, and the melting temperatures were in the order of Me-S-ODN·RNA > normal DNA·RNA > S-ODN·RNA. In comparison with the S-ODN, the Me-S-ODN had reduced toxic growth inhibitory effects, was less prone to bind the DNA-binding domain A of human replication protein A, and was as resistant to serum nucleases. Neither type of oligomer induced apoptosis, according to a PARP-cleavage assay. Hybrids formed with Me-S-ODN sequences were less sensitive to RNase H degradation than those formed with S-ODN sequences. Despite this latter disadvantage, the addition of 2'-O-methyl groups to a phosphorothioate-modified ODN is advantageous because of increased stability of binding and reduced non-specific effects.
INTRODUCTION
Phosphorothioate oligodeoxynucleotides (S-ODNs) were the most common type of the first generation antisense ODNs. These oligomers have resistance to nucleases and the ability to activate RNase H. The major shortcomings of S-ODNs are (i) the relatively low stability of duplexes with complementary RNA and (ii) the toxic effects, possibly related to non-specific protein–S-ODN interactions (1–8). Photocrosslinking experiments by others have shown that human replication protein A (RPA) is one of the major protein products crosslinked with single-stranded DNA oligomers (9).
A prevalent type of second-generation antisense ODN carries 2'-O-sugar modifications, which increases the stability of duplexes with complementary RNA (8). Duplexes formed with the 2'-O-sugar-modified S-ODNs themselves do not have the structure needed to activate RNase H cleavage of the target RNA, and such modified S-ODNs are routinely interrupted with a gap of phosphorothioate nucleotides that do not carry the sugar modification. Monia et al. (10) showed that 2'-O-modifications (such as pentyl, propyl, methyl and fluoro) in chimeric gap-mer S-ODNs had antisense effectiveness that correlated with the Tm values of hybrids formed with the target RNA, which was an Ha-ras oncogene in a reporter gene system. Stability of those gap-mers also correlated with the ability to inhibit Ha-ras mRNA expression in T24 cells, at a dose of 100 nM (11).
McKay et al. (12) found that a gap size of eight was required in 2'-O-methyl chimeras for optimal reduction of PKC mRNA in A549 cells. These authors advocated the use of 2'-O-methoxyethyl-modified sugars, because a 20 nt gap-mer with this modification had increased thermal stability when hybridized with the target mRNA and enhanced antisense effectiveness in reducing PKC mRNA (85% at 100 nM dose), when compared with an analogous 2'-O-methyl chimera (60% at 100 nM dose). In addition, maximum resistance to degradation by snake venom phosphodiesterase was shown to require that chimeras with either 2'-O-methyl or 2'-O-methoxyethyl sugars have a uniform phosphorothioate-modified backbone (12). The 2'-O-methyl modification remains one of the most valuable 2'-sugar modifications because it is cost-effective. In contrast, 2'-O-methoxyethyl-modified oligomers are not commercially available.
Various antisense modalities need to be carefully compared to determine their relative effectiveness (13). In this work, we compared 2'-O-methyl-modified and unmodified S-ODN sequences that are effective in reducing the level of Bcl-2 protein in T24 cells. The anti-apoptotic protein Bcl-2 is localized in the outer mitochondrial membrane, the endoplasmic reticulum and the nuclear envelope (14). Heterodimerization between pro- and anti-apoptotic members of the Bcl-2 family regulate the release of cytochrome c and subsequent apoptosis (15). Because of the Bcl-2 protein’s anti-apoptotic properties and abilities to inhibit tumor cell apoptosis induced by numerous drugs and radiation, the bcl-2 mRNA is an attractive antisense target for cancer treatment (7,15,16).
We show that non-specific effects of phosphorothioates on growth of T24 cells are reduced by the 2'-O-methyl modification. The 2'-O-methyl modification also reduces the binding of S-ODNs to domain A of RPA. Due to reduced non-specific effects, 2'-O-methyl-modified phosphorothioate ODN chimeras may be especially useful, in addition to S-ODNs, for comparative studies.
MATERIALS AND METHODS
Oligonucleotides and protein
Two oligomer sequences were chosen for the present study. Oligomers-109 (5'-CCGCATCCCACTCGTAGCCC-3') and oligomers-277 (5'-CCGGGCTGAGCGCAGGCCCC-3') were 20 nt long and were targeted at starting locations 109 and 277, respectively, of human bcl-2 mRNA (GenBank accession no. M14745 ). Negative control oligomers-C (5'-CTCA TTCCTACCGACACCCC-3') were also 20 nt long and consisted of a scrambled sequence that has no homology in the human genome and that was used as a control sequence by others (17). In preliminary experiments involving 29 S-ODNs targeted to various locations on bcl-2 mRNA, S-ODN-109 was the most effective in reducing Bcl-2 protein in A549 human lung cancer cells by the western blot protocol described below, but using an S-ODN concentration of 0.5 μM. S-ODN-277 was predicted to have the highest duplex stability among the tested oligomers and was as effective as S-ODN-109 in reducing the Bcl-2 level in T24 cells at a concentration of 0.1 μM. The positive control, compound G3139 of Genta, Inc. (Berkeley Heights, NJ), was 18 nt long and was targeted at starting location 32 of human bcl-2 mRNA .
The 11 DNA or chimeric oligomers used for this study and the modifications they contained are detailed in Table 1. For convenience, we denote the chimeric oligomers that contain 2'-O-methyl-modified nucleotides as ODNs, although they contain modified RNA nucleotides. Except for the unmodified ODN-109, all were 100% phosphorothioates. The unmodified ODN, S-ODNs, Me-S-ODN chimeras (with 12 2'-O-methyl sugars), CmeG-S-ODN chimera (with two 5-methyl cytosines in CpG sequences), and unmodified 20mer RNAs complementary to oligomers-109 and oligomers-277 were purchased from Midland Certified Reagent Co. (Midland, TX). The Me-S-ODN chimeras were HPLC-purified, and the unmodified RNA oligomers were purified using trityl selection. Other oligomers were purified by gel filtration. The unmodified RNAs were dissolved in water treated with diethyl pyrocarbonate as in previous work to reduce RNase degradation (18). Absorption spectra at 90°C (corrected for volume expansion due to heating) were used together with (260) values (M–1·cm–1, per mole of strand) to determine concentrations of the oligomers dissolved in water (19,20). Extinction coefficients at 260 nm were calculated as the sums of monomer extinction coefficients; see footnote to Table 1.
Table 1. Sequences, extinction coefficients, and melting temperatures of oligomers
Residues 181–291 of the 70 kDa RPA subunit were expressed in Escherichia coli and purified as described for other RPA domains (21).
Circular dichroism (CD) spectra and melting temperatures (Tm)
CD spectra and melting profiles were measured as in previous work (18–20). To form double-stranded hybrids for Tm measurements, the ODNs and complementary RNAs, at equimolar strand concentrations (2 μM) were mixed in 0.15 M K+ (phosphate buffer, pH 7.0) at room temperature for 2 h. Then, the mixtures were heated to 90°C for 90 s and allowed to cool at room temperature for 24 h. CD spectra were collected at 20°C using a Jasco Model J715 spectropolarimeter (Jasco, Easton, MD). To obtain melting profiles of absorbance as a function of temperature, the OD(260) (optical density at 260 nm) of each mixture was monitored in an Olis-modified Cary 14 (On-Line Instrument Co., Bogart, GA) as the temperature was increased from 20 to 95°C in increments of 1°C. Samples were incubated for 3 min at each temperature to allow for equilibration before taking OD readings. Reported Tm values were the maxima in first-derivative plots of the melting profiles.
Cell culture and transfections
T24 human bladder cancer cells were used because they have a relatively high concentration of Bcl-2 for detection by western blotting. The T24 cells were purchased from the American Type Culture Collection (Rockville, MD) and maintained in McCoy’s 5A medium (Mediatech, Inc., Herdon, VA), supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen, Carlsbad, CA), in 5% CO2 at 37°C. For transfections, T24 cells were plated 24 h prior to treatment in 25 cm2 flasks at a density of 8 x 105 cells per flask. A volume of 8 μl of the delivery reagent lipofectin (Invitrogen) was diluted in 1350 μl of McCoy’s 5A medium and incubated for 45 min at room temperature. ODNs were diluted into 650 μl of McCoy’s 5A media to give the required concentration and incubated for 15 min at room temperature. These two solutions were gently mixed and further incubated for 15 min. The concentrations of lipofectin and ODNs were 3 and 0.1 μM, respectively, in the final mixture. This gave a lipofectin/nucleotide (+/–) charge ratio of 1.5. (To determine whether the optimal lipofectin/charge ratio might differ for the methylated and unmethylated oligomers, electrophoretic mobility shift assay (EMSA) experiments were performed with S-ODN-109 and 5'-Me-S-ODN-109 mixed with increasing amounts of lipofectin to give (+/–) charge ratios of 0, 0.5, 1.0, 2.0 and 4.0. Both types of oligomer were partially neutralized at a charge ratio of 1.0 and were fully neutralized at a charge ratio of 2.0. Transfections in this work were therefore all performed with lipofectin/oligomer mixtures at the intermediate charge ratio of 1.5, which, in our experience with S-ODNs, was sufficient to maximize the antisense effect while minimizing non-specific effects of the lipofectin.) Cells were overlaid with this mixture after prewashing with McCoy’s 5A medium. Following a 4 h transfection of cells at 37°C, cells were washed and incubated in McCoy’s 5A medium supplemented with 10% FBS for an additional 24 h. Untreated (i.e. mock-treated) controls were identical except that during the treatment period they were overlaid with media without lipofectin or ODNs.
Western blot analysis
Cells treated with ODNs were washed with TBS (0.05 M Tris–HCl, pH 7.4, 0.1 M NaCl), and extracted at 0°C for 25 min in lysis buffer with 5% (v/v) protease inhibitor cocktail (Sigma, St Louis, MO). Cell debris was removed by microcentrifugation at 4°C. Protein concentrations were determined using the Bio-Rad protein assay (Richmond, CA). Cell extracts (20 μg) were subjected to SDS–polyacrylamide gel electrophoresis and then transferred to PVDF membranes (Millipore, Bedford, MA). The portions of the membranes containing Bcl-2 (MW 26 kDa) and Raf-1 (MW 73 kDa) were blocked with 5% non-fat dry milk in TTBS , and probed with corresponding primary antibodies at room temperature overnight (for Bcl-2: Cat. sc-509, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; for C-raf-1: Cat. 610151, BD Transduction Laboratories, San Diego, CA). After washing in TTBS, the membranes were incubated for 1 h at room temperature with 5% non-fat dry milk in TTBS containing a 1:15 000 dilution of an alkaline phosphatase-conjugated secondary antibody (Jackson Immuno Research, West Grove, PA). The immunocomplexes were visualized by enhanced chemifluorescence reagent (ECF substrate; Amersham Pharmarcia Biotech, Piscataway, NJ), followed by scanning with a phosphoimager (Storm 860, Molecular Dynamics) and quantitation using ImageQuant version 5.0 (Molecular Dynamics). The portion of the gel containing actin (MW 42 kDa) was stained by a modification of the Fairbanks Coomassie staining method (22). To calculate the reduction of Bcl-2 protein levels, actin band densities of untreated control samples (at 50 and 100% concentrations) and the treated samples were used to calculate the loading ratios of the treated cell lysates. Bcl-2 band intensities of untreated control samples (at 50 and 100% concentrations), with background subtracted, were fitted with a second-order equation from which the apparent percentages of Bcl-2 protein in the treated samples were determined. These percentages were finally corrected for any unequal loading, using the loading ratios from the actin bands.
Growth inhibition
The effects on cell growth of the antisense ODNs were determined by thymidine incorporation (23,24). T24 cells were plated 24 h prior to treatment in 24 well plates at a density of 2 x 104 cells per well. Cells were treated with ODNs using the procedure described above, except that after the 4 h treatment, cells were washed and incubated in McCoy’s 5A medium supplemented with 10% FBS plus 1 μCi of thymidine for 24 h. After the 24 h incubation, media in the wells was removed by aspiration and cells were incubated in Hanks balanced salt solution with trypsin-EDTA (Invitrogen) for 5 min at 37°C. A cell harvester (Cambridge Technology, Watertown, MA) was used to harvest cells on glass fiber filters. The percentage of growth inhibition was calculated as 100 x thymidine incorporation by the treated cells)/(counts due to thymidine incorporation by untreated cells)].
Assay for apoptosis
As a marker of cells undergoing apoptosis, the percentage of PARP that was cleaved to give an 89 kDa cleavage product was determined in treated cell lysates by western blot analysis using a PARP antibody (BD Biosciences, San Diego, CA).
Stability of ODNs in serum
To test the stabilities of ODNs in serum, the ODNs (at 10 μM concentration) were incubated in McCoy’s 5A medium supplemented with 10% FBS at 37°C. At each time point (0, 24, 48 and 72 h), 12 μl of each mixture was taken and the reaction was terminated by addition of 12 μl of loading buffer (0.1 M Tris–HCl pH 8.4, 0.09 M boric acid, 7 M urea, 0.1% bromophenol blue, 0.01% xylene cyanol). Samples were subjected to electrophoresis in a denaturing 20% polyacrylamide gel with 7 M urea. After electrophoresis, the gels were stained with SYBR Green II (Molecular Probes, Eugene, OR) prepared in TBE (0.089 M Tris base, 0.089 M boric acid and 1 mM EDTA pH 8.0) and scanned with a phosphoimager.
EMSA of RPA70-A binding to ODNs
Two types of EMSA experiments were used to investigate the binding of RPA70-A to S-ODN-109 versus 5'-Me-S-ODN-109. ODNs were end-labeled with T4 polynucleotide kinase (Promega, Madison, WI). (i) Competition experiments were performed to compare the binding affinities of RPA70-A to S-ODN-109 and 5'-Me-S-ODN-109. RPA70-A was added to saturate the 32P end-labeled S-ODN-109 at 1 μM. After 15 min incubation at room temperature, unlabeled competitor, either S-ODN-109 or 5'-Me-S-ODN-109 was added, and the mixtures were incubated for an additional 15 min. After addition of competitor, the concentration of ODN that was initially saturated with RPA70-A was 0.5 μM in all mixtures. (ii) Further competition experiments were performed to compare binding affinities of RPA70-A to S-ODN-109 versus 5'-Me-S-ODN-109 in the presence of the complementary 20mer RNA. RPA70-A was added to complementary RNA, and, after 15 min incubation, either 32P-labeled S-ODN or 32P-labeled Me-S-ODN was added and incubated for an additional 5 h at room temperature. The ODN and complementary RNA concentrations were each 1 μM in the final mixtures. The final / ratios (molar protein monomer/molar nucleotide ratios) were 0, 0.2 and 0.5, where the nucleotide concentrations were the total of ODN plus RNA nucleotides.
In each of the above cases, the mixtures were subjected to 20% non-denaturing polyacrylamide gel electrophoresis in TBE buffer, the gels were scanned with a phosphoimager, and the data were analyzed using ImageQuant 5.0.
RNase H elicitation
To examine the ability of RNase H to degrade hybrids formed with modified ODNs, 5 pmol of each modified ODN and complementary 32P 5'-end-labeled RNA were hybridized in 60 mM Tris–HCl (pH 7.8) containing 2 mM dithiothreitol, 60 mM KCl and 2.5 mM MgCl2. The desired hybrid duplexes were formed by incubating at 90°C for 2 min and slow cooling to room temperature for at least 1 h. Then, solutions of hybrids were treated with 0.05 units of Escherichia coli RNase H (Invitrogen) at 37°C for 20 min. Samples were analyzed as in the EMSA experiments described above.
RESULTS
Reductions in Bcl-2 level
In order to determine whether antisense effectiveness correlates with hybridization stabilities of the different ODN modifications, T24 human bladder cancer cells were transfected with the modified ODNs targeted to sites 109 and 277 on bcl-2 mRNA. The ODNs were used at a concentration of 0.1 μM. (In preliminary experiments, we found that there were no substantial decreases of Bcl-2 level in T24 cells when the concentrations of S-ODNs or Me-S-ODNs targeted to these two sites were increased above 0.1 μM under our conditions.) As shown in Figure 1, the S-ODN-109 and 5'-Me-S-ODN-109 resulted in 57 ± 5 and 70 ± 2% (SD) reductions of Bcl-2 level, respectively, with the Me-S-ODN having the greater effect. The analogous oligomers targeted to site 277, S-ODN-277 and 5'-Me-S-ODN-277, resulted in 63 ± 4 and 45 ± 5% reductions of the Bcl-2 level, respectively. In that case, the Me-S-ODN had the lesser effect. The S-ODN-109 and S-ODN-277 had the same effect, within error, as an effective positive control, G3139, directed against the first six codons of the bcl-2 mRNA (7). The S-ODN and Me-S-ODN controls averaged 3 ± 10 and 12 ± 10%, respectively, and were not significantly different in their effects on the Bcl-2 level (Fig. 1).
Figure 1. Reduction of Bcl-2 level in T24 cells by ODNs targeted to different sites on bcl-2 mRNA. The percent reductions were relative to the amount of Bcl-2 in mock-treated cells from western blots. Oligomer sequences and S-ODN and 5'-Me-S-ODN modifications are shown in Table 1. G3139 is an 18mer S-ODN positive control. Error bars are ±1 SD of at least three determinations. The ODN doses were 0.1 μM for Figures 1 and 3.
The level of Raf-1 protein, a protein in a pathway separate from the Bcl-2 pathway, was monitored as an additional indication of non-specific cellular effects of the oligomer treatments. For the three Me-S-ODNs used to obtain data for Figure 1, the level of Raf-1 protein was reduced by an average of 21%, no different from the average of 24% reduction found after treatments by the three 20mer S-ODNs (data not shown).
The antisense effectiveness of three different Me-S-ODN chimeras, all targeted to site 109, were compared to study the effect of placing the 12 2'-O-methyl groups in one block at the 5' end, in one block at the 3' end or in two equal blocks at both ends to form a gap-mer. The Me-S-ODN chimeric sequences (5'-Me-S-ODN, 3'-Me-S ODN and gap-Me-S-ODN) are shown in Table 1. The hybrids of these chimeras paired with the target RNA had Tm values that differed by only 4°C (Table 1), and the reductions of Bcl-2 level caused by treatments with the three chimeras were all in the range 73 ± 4% (data not shown). Therefore, for this 20mer sequence, the location of the one or two blocks of 2'-O-methyl nucleotides was unimportant.
Because additional work (below) focused on the oligomers targeted to site 109, we examined whether the reduction of the Bcl-2 level by S-ODN-109 was influenced by the presence of CpG sequences. Synthetic DNAs containing unmethylated CpG dinucleotides can stimulate an immune response in vertebrates, and CpG islands in antisense ODNs can cause non-sequence-specific effects (7,25,26). Since sequence 109 contained two CpG islands, an ODN was prepared (CmeG-S-ODN; see Table 1) in which C-5-methyl groups were added to cytosines in these CpG sequences to suppress any non-specific effects. In simultaneous experiments, the reductions in the Bcl-2 level by S-ODN-109 and CmeG-S-ODN-109 were essentially identical (62 ± 4 and 60 ± 11%, respectively) demonstrating that the reduction of Bcl-2 level by the S-ODN-109 was not affected by CpG islands. This was not unexpected, since T24 cells are not immune system cells.
From all the data above, we take the reductions in Bcl-2 levels by the S-ODN and Me-S-ODN sequences targeted to sites 109 and 277 to denote primarily the antisense effects of these oligomers.
Although the relative effectiveness of the two types of ODN depended on the target site sequence, the oligomers targeted to site 109 show that an Me-S-ODN can have an apparent antisense effectiveness greater than that of a cognate S-ODN. The actual specific activity of the Me-S-ODN may be even greater due to decreased non-specific effects, such as a reduced toxic effect on cell growth (see below).
CD spectra of ODN·RNA hybrids
The ODN·RNA duplexes were characterized by CD spectral measurements. CD measurements were performed with the samples in 0.15 M K+ (phosphate buffer, pH 7.0), 20°C. The spectra of the ODN·RNA duplexes were significantly different from the averaged spectra of the individual strands (Fig. 2). Spectra of the hybrids showed an increase in the magnitude of the long wavelength positive band above 260 nm, an increase in the negative band at 210 nm, and a shift in the crossover near 250 nm to shorter wavelengths, relative to the averaged spectra. These CD characteristics are typical features of DNA·RNA duplexes (18,20) and confirm the formation of the desired duplexes.
Figure 2. Measured CD spectra of ODN·RNA hybrids compared with the averaged spectra of the individual strands. Hybrids were formed by pairing the following ODNs with the complementary RNA-109 strand: (A) Unmodified ODN-109, (B) S-ODN-109, (C) 5'-Me-S-ODN-109. (Solid line) Measured spectra; (dashed line) averaged spectra of the individual, unpaired strands. Spectra were taken at 20°C with samples in 0.15 M K+ (phosphate buffer, pH 7.0).
Tm values of ODN·RNA hybrids
The S-ODN and Me-S-ODN were further characterized by the stabilities (Tm values) of the hybrids they formed with complementary RNAs. The Me-S-ODN·RNA hybrids had relatively high Tm values (69°C to above 82°C), but the corresponding S-ODN·RNA hybrids had lower Tm values (55 and 66°C). The Tm difference between the three chimeric Me-S-ODN-109·RNA hybrids and the S-ODN-109·RNA hybrid was 14–18°C and that between one chimeric Me-S-ODN-277·RNA and the S-ODN-277·RNA was >16°C (Table 1). The Me-S-ODNs·RNA hybrids presumably have higher Tm values than the corresponding hybrids formed with the S-ODNs because the 2'-O-methyl group results in an RNA-like conformation, and RNA duplexes have higher Tm values than DNA·RNA hybrids (6,18,27,28).
Growth inhibition by thymidine incorporation
Thymidine incorporation in T24 cells was used to assay the inhibition of DNA synthesis and growth inhibitory effects of S-ODN-109 and 5'-Me-S-ODN-109. As shown in Figure 3, the same level of growth inhibition was measured for the controls as for the antisense ODNs. These effects were significant for the phosphorothioate-modified oligomer and were minimal for the Me-S-ODN oligomer. There was no effect on thymidine incorporation by treatment with lipofectin alone (data not shown). We conclude that the effects on growth inhibition by this assay were essentially caused by the cytotoxic effects of the ODN modifications and were not secondary to a reduction of the Bcl-2 level.
Figure 3. Growth inhibition of T24 cells following treatments with the S-ODN and the 5'-Me-S-ODN targeted to site 109 on bcl-2 mRNA. Growth inhibition was determined as the decrease in thymidine incorporation, relative to the incorporation in mock-treated cells. Error bars are ±1 SD of three determinations.
Apoptosis
We tested whether a reduction in the level of the Bcl-2 anti-apoptotic protein was correlated with apoptosis of T24 cells. PARP is one of the main cleavage targets of caspase-3, which is activated in the apoptosis cascade following release of cytochrome c, so that the cleavage of PARP serves as a marker of cells undergoing apoptosis (29). Cells were treated with ODNs as described in Materials and Methods for reducing the Bcl-2 protein level. The percentage of the 89 kDa PARP cleavage product after treatments with S-ODN-109, 5'-Me-S-ODN-109, and the two negative controls, ranged from 2 to 6%, not detectably different from the 2% cleavage product found in lysates of untreated controls (data not shown). Thus, according to this assay, there was no detectable apoptotic effect with either the S-ODN-109 of 5'-Me-S-ODN-109, although these oligomers reduce the Bcl-2 level by 57 and 70%, respectively, and they have differing effects on cell growth according to the thymidine incorporation assay.
Stability of ODNs in serum
An important consideration in the design of antisense ODN modifications is the degree to which the ODN is rendered nuclease resistant. The resistance to degradation in 10% FBS of the unmodified ODN-109 (phosphodiester) and three cognate modified ODNs (the S-ODN, 5'-Me-S-ODN and Gap-Me-S-ODN) was examined as described in Materials and Methods. The results are shown in Figure 4. While little of the unmodified ODN remained intact after 24 h, all three of the modified ODNs exhibited a half-life of >72 h in 10% FBS. This was in agreement with results by Grünweller et al. (30), who reported the half-life of an S-ODN as being >>48 h under similar conditions. Thus, the S-ODN and Me-S-ODN were completely stable over the time course of the transfections and incubations to reduce the level of Bcl-2, and the difference in antisense effect (Fig. 1) was not obviously due to a difference in nuclease stability.
Figure 4. Survival in FBS of oligomers 109. (A) Bands of intact ODNs in polyacrylamide gels after incubation in 10% FBS at 37°C for the indicated times. Bands were visualized by staining with SYBR Green II. (B) Percentages of uncleaved oligonucleotides as a function of incubation time.
RPA70-A binding affinities of S-ODN-109 and Me-S-ODN-109
One non-specific effect of ODNs is their binding to cellular proteins (1–9). The most abundant single-stranded DNA binding protein in human cells is RPA (31), and domain A of the 70 kDa subunit of RPA is the site of initial binding to single-stranded DNA (32–34). The binding site of RPA70-A accommodates 5 nt, similar to the binding site size of the Ff g5p, which has been used to assess the effects of different nucleotide modifications (6,32). Therefore, we determined the relative binding affinity of RPA for S-ODN-109 and 5'-Me-S-ODN-109 in both the absence and presence of the complementary RNA sequence.
To compare the RPA70-A binding affinities for the S-ODN and the Me-S-ODN, saturated S-ODN-109:RPA70-A complexes were prepared with 32P-labeled S-ODN and various amounts of unlabeled competing S-ODN-109 or Me-S-ODN-109 were added (in phosphate, pH 7.0, 37°C). The EMSA results after 15 min of incubation with the competitor are shown in Figure 5A and B. At an 8-fold excess of unlabeled competing S-ODN, 61% of the labeled S-ODN:RPA70-A complex was completely dissociated (Fig. 5A). The same results were obtained after up to 4 h of incubation with competing S-ODN-109, indicating that equilibrium had been reached (data not shown). The band of complex became more diffuse as the labeled S-ODN was dissociated because RPA70-A does not bind with high cooperativity, and up to five protein molecules can be bound per ODN 20mer. The same 8-fold excess of competing Me-S-ODN dissociated a maximum of 15% of the S-ODN:RPA70-A complex, 4-fold less than the same amount of competing S-ODN (Fig. 5B). Additional experiments in which the labeled S-ODN and unlabeled Me-S-ODN were premixed before adding the protein demonstrated that even at a 15-fold excess of Me-S-ODN the amount of completely dissociated labeled S-ODN was still only 15%. Finally, the ratio of 5'-Me-S-ODN-109 to S-ODN-109 needed for 50% dissociation of an unmodified ODN:RPA70-A complex was determined to be 9.9 ± 0.1 (range) in competition experiments (data not shown). In summary, the binding affinity of RPA70-A was estimated to be 10-fold greater for the S-ODN than for the 5'-Me-S-ODN, even though the latter has a stretch of eight phosphorothioate nucleotides.
Figure 5. Competition binding experiments. The relative binding affinity of RPA70-A for S-ODN and 5'-Me-S-ODN was determined by competition of (A) cold S-ODN-109 or (B) cold 5'-Me-S-ODN with the 32P end-labeled S-ODN-109 in a preformed complex with RPA70-A. The percent dissociation of S-ODN from the complex was calculated from the band intensity of free 32P-labeled S-ODN released from the 32P S-ODN:RPA70-A complex. (C) 32P end-labeled S-ODN-109 or 32P end-labeled 5'-Me-S-ODN-109 were titrated with RPA70-A in the presence of complementary RNA. In lanes 1 and 5, there were ODNs only, without RNA. In the other lanes, RNA and RPA70-A were added and, after 15 min, the complementary ODNs were added. The / ratios are the RPA70-A protein molar concentrations per total molar concentrations of nucleotides (DNA plus RNA). Multiple bands of the ODN·RNA duplexes were probably the result of some degradation of the RNA combined with a reduced stability of hybrids formed with the S-ODN. The complexes of RPA70-A with 5'-Me-S-ODN were more diffuse than those with the S-ODN, likely due to multiple arrangements of up to five protein molecules on the chimeric oligomer.
In a second type of experiment, we examined the ability of the S-ODN and the 5'-Me-S-ODN to bind to the complementary RNA in the presence of RPA70-A. At increasing RPA70-A concentrations, the S-ODN·RNA hybrid was less able to form than was the Me-S-ODN·RNA hybrid, as seen in Figure 5C. From quantitation of the results in this Figure, at a / ratio of 0.5, only 25% of the S-ODN was involved in the formation of a S-ODN·RNA hybrid while 80% of the Me-S-ODN was able to form an Me-S-ODN·RNA hybrid. These experiments were consistent with a higher affinity of RPA70-A for the S-ODN than for the 5'-Me-S-ODN shown in Figure 5A and B and with a higher Tm for the 5'-Me-S-ODN·RNA hybrid than for the S-ODN·RNA hybrid (Table 1).
RNase H elicitation
The ability of antisense ODNs to elicit RNase H degradation of the target mRNA is considered to be one of the most important criteria for antisense effectiveness. The S-ODNs and 5'-Me-S-ODNs targeted to RNA sites 109 and 277 were compared for their ability to recruit RNase H. RNase H degradation of the target RNAs (unmodified 20mer RNAs complementary to oligomers-109 and oligomers-277) was higher for hybrids formed with the S-ODNs than with the Me-S-ODNs (Fig. 6). For example, 80% of the hybrid formed with S-ODN-109 was degraded but only 30% of the hybrid formed with Me-S-ODN-109 was degraded under the same conditions. The lowered ability of RNase H to degrade hybrids with the Me-S-ODN was not unexpected. The Me-S-ODNs are chimeric, with 12 nt that are 2'-O-methyl and RNA-like. Therefore, a major portion of the hybrid formed with the Me-S-ODN sequesters the target RNA in an RNA·RNA duplex that is not recognized by RNase H. The most effective hybrid in recruiting RNase H was that formed by S-ODN-109, which interestingly had the lowest Tm.
Figure 6. RNase H elicitation of modified ODN·RNA duplexes. With the assumption of equal loading, the percent degradation was calculated from the reduction in the intensity of the DNA·RNA duplex band after RNase H treatment. The complementary 20mer RNAs were 32P end-labeled. Bands labeled as degraded products in the RNase H negative lanes were from labeled fragments of RNA and over-represent the amount of degradation in the RNA samples because at least two fragments result from a single cleavage. They did not affect the calculation of percent degradation of hybrids by RNase H.
DISCUSSION
Hybrids formed with 20mer S-ODNs and Me-S-ODNs and their cognate RNA sequences were shown to have the CD spectra of DNA·RNA and RNA·RNA A-form structures (Fig. 2). As expected, melting temperatures of the hybrids formed with the Me-S-ODNs were above those formed with the S-ODNs (Table 1). The two S-ODN sequences whose properties were compared in this work were as effective in reducing the level of Bcl-2 protein in T24 cells as the well studied Bcl-2 antisense sequence G3139 (7) (Fig. 1), independent of the presence of CpG islands. Treatments with S-ODN-109 and 5'-Me-S-ODN-109 did not lead to apoptosis according to a PARP-cleavage assay. G3139 also does not induce apoptosis in all cell lines (35), and Benimetskaya et al. (36) reported that G3139-induced down regulation of Bcl-2 in T24 cells is insufficient to decrease cellular viability. Neither S-ODN-109 nor Me-S-ODN-109 oligomers were significantly degraded by serum over the time course of the antisense treatments (Fig. 4). The latter is important in comparative studies, because without having the nuclease protection of a phosphorothioate backbone the 2'-O-methyl modification can be less effective in an antisense oligomer than a simple S-ODN (30).
Eight phosphorothioate nucleotides are needed in gap-mers with 2'-O-methyl-modified sugars and locked nucleic acids to fully activate RNase H (12,37). Oligomers containing 12 2'-O-methyl-modified nucleotides at the 5' ends were less able to induce RNase H degradation of hybridized RNA than were the analogous S-ODNs (Fig. 6). Nevertheless, the abilities of different chimeric arrangements of the 12 modified nucleotides, in one block at either end or as a gap-mer (5'-Me-S-ODN, 3'-Me-S ODN and gap-Me-S-ODN), were equally able to reduce the level of Bcl-2 protein. This implies that slight differences in the ability to activate RNase H are not as important as the ability to form stable hybrids with the target RNA and other factors in antisense effectiveness. Although we did not examine mRNA levels by northern blot analysis in this work, others have shown that antisense phosphorothioate oligomers carrying the 2'-O-methyl modification are unable to reduce endogenous PKC mRNA levels unless there is a gap without the sugar modification so that RNase H can be activated (12). Thus, we reasonably expect that the main mechanism of action of our oligomers in reducing the level of Bcl-2 protein was by RNase H inactivation of the encoding mRNA.
We have shown for the first time that the 2'-O-methyl modification results in a reduction of important non-specific effects. Thymidine incorporation into DNA has been used by others as a means of assaying cellular growth and proliferation, because dead and dying cells do not incorporate thymidine into their DNA (23,24). Under our transfection conditions, oligomers carrying just the phosphorothioate modification were about three times as inhibitory to the growth of T24 cells as oligomers that included 12 2'-O-methyl-modified nucleotides (Fig. 3). Moreover, we have determined that the reduced non-specific effects of Me-S-ODNs could be related to their reduced binding to cellular proteins, such as RPA (Fig. 5A and B). There was a decided decrease in binding affinity for the RPA70-A domain, by a factor of 10-fold, when 2'-O-methyl sugars were included in the phosphorothioate ODN. This decrease in binding affinity for RPA helps ensure that an ODN·RNA hybrid will remain stable in the presence of this cellular protein (Fig. 5C).
Others have found that S-ODNs and G-tetraplex-forming oligomers bind to nucleolin and that this binding can result in altered cell morphology and apoptosis (38–40). It remains to be determined whether Me-S-ODNs have reduced nucleolin-binding properties, consistent with the possibility that the 2'-O-methyl modification leads to reduced non-specific effects via reduced binding to cellular proteins.
Our previous work with the model Ff g5p single-stranded DNA-binding protein showed that a fully 2'-O-methyl-modified S-dA36 oligomer had about one-half of the binding affinity for g5p as did the simple S-dA36 oligomer (6). The use of a model protein such as the g5p, or the RPA70-A domain used in the present study, to assay relative binding affinities may help establish which nucleotide modifications are likely to have reduced non-specific effects in cells.
In summary, chimeric ODNs carrying 12 2'-O-methyl-modified nucleotides would be a good choice to include, in addition to simple phosphorothioates, in much needed comparative studies (4) of different antisense molecules. Our results also suggest that there could be a reduction in cost of synthesizing oligomers with 2'-O-modified sugars by including all 12 modified nucleotides at the 5' or the 3' end of the oligomers.
ACKNOWLEDGEMENTS
This work was performed by B.H.Y. in partial fulfillment of the requirements for the PhD degree in the Department of Molecular and Cell Biology, The University of Texas at Dallas. Support was provided in part by grants from the Robert A. Welch foundation (AT-503), eXegenics, Inc. (Dallas, TX) and the Texas Advanced Technology Program (009741-0021-1999) to D.M.G. and from the US National Institutes of Health to A.B. (RO1 GM061192).
REFERENCES
Stein,C.A. (1996) Phosphorothioate antisense oligodeoxynucleotides: questions of specificity. Trends Biotechnol., 14, 147–149.
Lebedeva,I. and Stein,C.A. (2001) Antisense oligonucleotides: promise and reality. Annu. Rev. Pharmacol. Toxicol., 41, 403–419.
Stein,C.A. (2001) The experimental use of antisense oligonucleotides: a guide for the perplexed. J. Clin. Invest., 108, 641–644.
Braasch,D.A. and Corey,D.R. (2002) Novel antisense and peptide nucleic acid strategies for controlling gene expression. Biochemistry, 41, 4503–4510.
Dove,A. (2002) Antisense and sensibility. Nat. Biotechnol., 20, 121–124.
Mou,T.C. and Gray,D.M. (2002) The high binding affinity of phosphorothioate-modified oligomers for Ff gene 5 protein is moderated by the addition of C-5 propyne or 2'-O-methyl modifications. Nucleic Acids Res., 30, 749–758.
Dias,N. and Stein,C.A. (2003) Potential roles of antisense oligonucleotides in cancer therapy. The example of Bcl-2 antisense oligonucleotides. Eur. J. Pharm. Biopharm., 54, 263–269.
Shen,L., Siwkowski,A., Wancewicz,E.V., Lesnik,E., Butler,M., Witchell,D., Vasquez,G., Ross,B., Acevedo,O., Inamati,G., Sasmor,H., Manoharan,M. and Monia,B.P. (2003) Evaluation of C-5 propyne-containing oligonucleotides in vitro and in vivo. Antisense Nucleic Acid Drug Dev., 13, 129–142.
Revers,F., Cario,M., Cao,T.L. and Cazenave,C. (1999) Detection of proteins binding to short RNA·DNA hybrids or short antisense oligonucleotides in Xenopus laevis oocytes and human macrophage cell extracts by photoaffinity radiolabeling. Antisense Nucleic Acid Drug Dev., 9, 317–331.
Monia,B.P., Lesnik,E.A., Gonzalez,C., Lima,W.F., McGee,D., Guinosso,C.J., Kawasake,A.M., Cook,P.D. and Freier,S.M. (1993) Evaluation of 2'-O-modified oligonucleotides containing 2'-O-deoxy gaps as antisense inhibitors of gene expression. J. Biol. Chem., 268, 14514–14522.
Monia,B.P., Johnston,J.F., Sasmor.H. and Cummins,L.L. (1996) Nuclease resistance and antisense activity of modified oligonucleotides targeted to Ha-ras. J. Biol. Chem., 271, 14533–14540.
McKay,R.A., Miraglia,L.J., Cummins,L.L., Owens,S.R., Sasmor,H. and Dean,N.M. (1999) Characterization of a potent and specific class of antisense oligonucleotide inhibitor of human protein kinase C- expression. J. Biol. Chem., 274, 1715–1722.
Braasch,D.A., Liu,Y. and Corey,D.R. (2002) Antisense inhibition of gene expression in cells by oligonucleotides incorporating locked nucleic acids: effect of mRNA target sequence and chimera design. Nucleic Acids Res., 30, 5160–5167.
Hockenbery,D., Nunez,G., Milliman,C., Schreiber,R.D. and Korsmeyer,S.J. (1990) Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature, 348, 334–336.
Hengartner,M.O. (2000) The biochemistry of apoptosis. Nature, 407, 770–776.
Chao,D.T. and Korsmeyer,S.J. (1998) BCL-2 family: regulators of cell death. Annu. Rev. Immunol., 16, 395–419.
Ziegler,A., Luedke,G.H., Fabbro,D., Altmann,K.H., Stahel,R.A. and Zangemeister-Wittke,U. (1997) Induction of apoptosis in small-cell lung cancer cells by an antisense oligodeoxynucleotide targeting the Bcl-2 coding sequence. J. Natl Cancer Inst., 89, 1027–1036.
Hashem,G.M., Pham,L., Vaughan,M.R. and Gray,D.M. (1998) Hybrid oligomer duplexes formed with phosphorothioate DNAs:CD spectra and melting temperatures of S-DNA·RNA hybirds are sequence-dependent but consistent with similar heteronomous conformations. Biochemistry, 37, 61–72.
Gray,D.M., Hung,S.H. and Johnson,K.H. (1995) Absorption and circular dichroism spectroscopy of nucleic acid duplexes and triplexes. Methods Enzymol., 246, 19–34.
Clark,C.L., Cecil,P.K., Singh,D. and Gray,D.M. (1997) CD, absorption and thermodynamic analysis of repeating dinucleotide DNA, RNA and hybrid duplexes 12·12 and the influence of phosphorothioate substitution. Nucleic Acids Res., 25, 4098–4105.
Bochkareva,E., Frappier,L., Edwards,A.M. and Bochkarev,A. (1998) The RPA32 subunit of human replication protein A contains a single-stranded DNA-binding domain. J. Biol. Chem., 273, 3932–3936.
Wong,C., Sridhara,S., Bardwell,J.C. and Jakob,U. (2000) Heating greatly speeds Coomassie blue staining and destaining. Biotechniques, 28, 426–428, 430, 432.
Robinson,L.A., Smith,L.J., Fontaine,M.P., Kay,H.D., Mountjoy,C.P. and Pirruccello,S.J. (1995) c-myc antisense oligodeoxyribonucleotides inhibit proliferation of non-small cell lung cancer. Ann. Thorac. Surg., 60, 1583–1591.
Yang,D.C., Jiang,X., Elliott,R.L. and Head,J.F. (2002) Antisense ferritin oligonucleotides inhibit growth and induce apoptosis in human breast carcinoma cells. Anticancer Res., 22, 1513–1524.
Krieg,A.M., Yi,A.K., Matson,S., Waldschmidt,T.J., Bishop,G.A., Teasdale,R., Koretzky,G.A. and Klinman,D.M. (1995) CpG motifs in bacterial DNA trigger direct B-cell activation. Nature, 374, 546–549.
Krieg,A.M. (1999) Mechanisms and applications of immune stimulatory CpG oligodeoxynucleotides. Biochim. Biophys. Acta, 1489, 107–116.
Hung,S.-H., Yu,Q., Gray,D.M. and Ratliff,R.L. (1994) Evidence from CD spectra that d(purine)·r(pyrimidine) and r(purine)·d(pyrimidine) hybrids are in different structural classes. Nucleic Acids Res., 22, 4326–4334.
Nishizaki,T., Iwai,S., Ohtsuka,E. and Nakamura,H. (1997) Solution structure of an RNA·2'-O-methylated RNA hybrid duplex containing an RNA·DNA hybrid segment at the center. Biochemistry, 36, 2577–2585.
Oliver,F.J., de la Rubia,G., Rolli,V., Ruiz-Ruiz,M.C., de Murcia,G. and Murcia,J.M. (1998) Importance of poly(ADP-ribose) polymerase and its cleavage in apoptosis. Lesson from an uncleavable mutant. J. Biol. Chem., 273, 33533–33539.
Grünweller,A., Wyszko,E., Bieber,B., Jahnel,R., Erdmann,V.A. and Kurreck,J. (2003) Comparison of different antisense strategies in mammalian cells using locked nucleic acids, 2'-O-methyl RNA, phosphorothioates and small interfering RNA. Nucleic Acids Res., 31, 3185–3193.
Wold,M.S. (1997) Replication protein A: a heterotrimeric, single-stranded DNA-binding protein required for eukaryotic DNA metabolism. Annu. Rev. Biochem., 66, 61–92.
Bochkarev,A., Pfuetzner,R.A., Edwards,A.M. and Frappier,L. (1997) Structure of the single-stranded-DNA-binding domain of replication protein A bound to DNA. Nature, 385, 176–181.
Arunkumar,A.I., Stauffer,M.E., Bochkareva,E., Bochkarev,A. and Chazin,W.J. (2003) Independent and coordinated functions of the replication protein A tandem high affinity ssDNA binding domains. J. Biol. Chem., 278, 41077–41082.
Bochkarev,A. and Bochkareva,E. (2004) From RPA to BRCA2: lessons from single-stranded DNA binding by the OB-fold. Curr. Opin. Struct. Biol., 14, 36–42.
Lai,J.C., Benimetskaya,L., Santella,R.M., Wang,Q., Miller,P.S. and Stein,C.A. (2003) G3139 (oblimersen) may inhibit prostate cancer cell growth in a partially bis-CpG-dependent non-antisense manner. Mol. Cancer Ther., 2, 1031–1043.
Benimetskaya,L., Miller,P., Benimetsky,S., Maciaszek,A., Guga,P., Beaucage,S.L., Wilk,A., Grajkowski,A., Halpern,A.L. and Stein,C.A. (2001) Inhibition of potentially anti-apoptotic proteins by antisense protein kinase C- (Isis 3521) and antisense bcl-2 (G3139) phosphorothioate oligonucleotides: relationship to the decreased viability of T24 bladder and PC3 prostate cancer cells. Mol. Pharm., 60, 1296–1307.
Kurreck,J., Wyszko,E., Gillen,C. and Erdmann,V.A. (2002) Design of antisense oligonucleotides stabilized by locked nucleic acids. Nucleic Acids Res., 30, 1911–1918.
Weidner,D.A., Valdez,B.C., Henning,D., Greenberg,S. and Busch,H. (1995) Phosphorothioate oligonucleotides bind in a non sequence-specific manner to the nuclear protein C23/nucleolin. FEBS Lett., 366, 146–150.
Bates,P.J., Kahlon,J.B., Thomas,S.D., Trent,J.O. and Miller,D.M. (1999) Antiproliferative activity of G-rich oligonucleotides correlates with protein binding. J. Biol. Chem., 274, 26369–26377.
Xu,X., Hamhouyia,F., Thomas,S.D., Burkel,T.J., Girvant,A.C., McGregor,W.G., Trent,J.O., Miller,D.M. and Bates,P.J. (2001) Inhibition of DNA replication and induction of S phase cell cycle arrest by G-rich oligonucleotides. J. Biol. Chem., 276, 43221–43230.(Byong Hoon Yoo, Elena Bochkareva1, Alexe)