Nucleic Acids Research

INTRODUCTION

The use of chemical methods to probe protein-DNA interactions has greatly contributed to our understanding of the structure and DNA organization within macromolecular complexes (1,2). The hydroxyl radical has been a particularly useful reagent for investigating DNA structure and has been used extensively to map histone-DNA interactions precisely within nucleosomes (3-5). Typically, diffusible hydroxyl radicals are chemically generated via the reaction:

In this system, the radical-producing complex has an overall negative charge and is held out in solution, away from the DNA. Thus, radicals must diffuse some distance before encountering the DNA. The rate of DNA cleavage by [Fe(II)(EDTA)]^2--generated hydroxyl radicals is dependent upon accessibility of the minor groove of the double helix to bulk solvent (6). Thus, protein-DNA contacts or DNA conformations which reduce or block accessibility to the minor groove can be precisely mapped by this technique (6-8). However, protein-DNA contacts which occur primarily within the major groove may have only marginal affects on the rate of DNA cleavage by [Fe(II)(EDTA)]^2--produced hydroxyl radicals, and are consequently hard to identify by this technique (9,10). Convenient and general detection of such interactions may depend upon the development of new chemical probes for DNA.

Recently, Silver and Trogler (11) described a new oxidative DNA cleavage agent, (1,4,7-trimethyl-1,4,7-triazacyclononane)iron(III) chloride [also known as Fe(III)TACN, hereafter referred to as L'FeCl₃], which efficiently cleaves DNA at physiological pH and temperature with random, single-stranded nicks (Fig. 1). This compound is easily synthesized and stable in aqueous solution. Furthermore, the complex has an overall positive charge, is thought to bind directly to the phosphodiester backbone of DNA and is likely to effect DNA cleavage in a way very different to that of [Fe(II)(EDTA)]^2- (11,12; see below). Thus, L'FeCl₃ seemed to be an ideal candidate for probing protein-DNA interactions.

Evidence suggests that the active cleavage reagent is formed when chloride ions are replaced from L'FeCl₃ in aqueous solution by either aquo and/or hydroxo ligands. This produces a cationic complex with several labile and solvent-accessible coordination sites that could bind to the DNA backbone and mediate localized DNA cleavage. Spectral changes that occur when L'FeCl₃ is transferred from organic solvents to aqueous solutions support this conclusion (11). Another important feature of the complex is the three methyl groups which sterically prohibit more than one L' ligand from binding to the iron center. This maintains three open coordination sites to facilitate redox chemistry. Furthermore, the overall positive charge of the complex suggests favorable electrostatic binding to the DNA. Thus, L'FeCl₃ is expected to efficiently effect oxidative cleavage of DNA.

We were interested in determining the utility of L'FeCl₃ for studies of protein-DNA interactions. Since this compound binds directly to DNA, we reasoned that it would provide a complementary approach to studies with hydroxyl radicals produced from [Fe(II)EDTA]^-2. To test this, we used both [Fe(II)EDTA]^-2 and L'FeCl₃ to footprint two different types of protein-DNA interactions: (i) histone protein-DNA contacts within a nucleosome and (ii) two sequence-specific DNA-binding proteins, the bacteriophage Replication Initiatior A protein (RepA) and the bacterial repressor, DinR (13-15). Consistent with previous reports (11), we found evidence that L'FeCl₃ cuts DNA by a distinctly different mechanism to that of [Fe(II)EDTA]^-2. Moreover, we found that DNA cleavage by L'FeCl₃ is differentially sensitive to various types of protein-DNA interactions. Cleavage is relatively unaffected by the presence of histone-DNA interactions, but is clearly reduced by DNA contacts made by sequence-specific DNA binding proteins at selected positions.

MATERIALS AND METHODS

5S DNA fragments

A 330 base pair (bp) BstXI-BstXI DNA fragment containing the Xenopus borealis somatic 5S rRNA gene was obtained from pJHX1.BstXI, constructed as follows. Asymmetric BstXI sites were introduced by PCR at each end of a 315 bp HindIII-XbaI fragment containing the X.borealis somatic 5S rRNA derived from pJHX1 (16). The BstXI site in pBSIIsk(+) (Stratagene) was altered to match the asymmetry in the insert, and a tandem dimer of the fragment was ligated into this plasmid to form pJHX1.BstXI. Digestion with BstXI (New England Biolabs) released the 330 bp fragment, which was isolated by preparative agarose gel electrophoresis and stored at -20°C until needed. The 5S DNA fragment was radioactively end-labeled by digesting the BstXI-BstXI fragments with HpaII, treating with calf intestinal alkaline phosphatase and subsequently incubating with T4 polynucleotide kinase (New England Biolabs) and [[gamma]-³²P]ATP (New England Nuclear). Digestion of this material with XbaI released a uniquely end-labeled 240 bp fragment.

Figure 1. Structure of (1,4,7-trimethyl-1,4,7-triazocyclononane)iron(III) chloride (L'FeCl₃).

RepA and DinR DNA fragments

A 109 bp fragment containing the 19 bp RepA consensus binding site prepared by the polymerase chain reaction (PCR) was provided by Dr Teresa E.Strzelecka (National Institute of Health, Bethesda, MD). The fragment was cloned into the XbaI-HindIII restriction sites within a pBSIIsk(+) Bluescript plasmid (Stratagene). We linearized the pBSII.RepA plasmid with HindIII, end-labeled the DNA as described above and released a 100 bp uniquely end-labeled DNA fragment by digestion with XbaI. A 130 bp DNA fragment containing the single DinR site found in the RecA promoter (15) was prepared as follows. Approximately 2 µg of PCR primer RECA01 (5[prime]-GCGAAGCTTACATGATTTTCTGATACATTA-3[prime]) was 5[prime] end-radiolabeled with T4 polynucleotide kinase by standard procedures. The labeled primer was used directly in standard PCR reactions with a complementary downstream primer RECA02 (5[prime]-CGCGAATTCCTTTTATGTTACACTACATA-3[prime]) and 300 ng BacillusYB886 chromosomal DNA. The 130 bp labeled double-stranded DNA product was purified by native 6% polyacrylamide gel electrophoresis.

Proteins

The RepA proteins (mol. wt 32 200 kDa) were provided by Dr Teresa E.Strzelecka (NIH, Bethesda, MD). The DinR protein was provided by Drs K.Winterling and R.Woodgate (NIH, Bethesda, MD). Histone proteins were purified from chicken erythrocyte nuclei as described (16).

Nucleosome reconstitution

Nucleosomes were reconstituted by dialysis from high salt (16) using radiolabeled 5S DNA fragments and chicken erythrocyte histones. The nucleosomes were ultimately dialyzed into TE. The efficiency of the reconstitution was monitored by nucleoprotein gel electrophoresis (16).

Formation of RepA-DNA and DinR-DNA complexes

The RepA protein-DNA binding reaction was performed in a volume of 35 µl containing [sim]200 fmol 5[prime]end-labeled DNA fragment in RepA binding buffer (20 mM Tris pH 7.5, 5 mM DTT, 1 mM EDTA and 150 mM NaCl) and 0.1 µg (3 pmol or [sim]90 nM final concentration) of RepA protein. The reaction was incubated at room temperature for 10 min before the footprinting. The DinR protein-DNA complex was formed by combining [sim]20 ng of the labeled recA DNA fragment in 35 µl of DinR binding buffer (15) with [sim]200 ng DinR protein. The binding reaction was incubated at room temperature for 25 min and then digested with Fe(II)EDTA or L'FeCl3 as described.

[Fe(II)EDTA]^-2 footprinting

A stock solution of [Fe(II)EDTA]^-2 was prepared by mixing equal volumes of 1 mM (NH₄)₂Fe(SO₄)₂[bull]H₂O and 2 mM EDTA and stored frozen. Reconstituted nucleosomes in TE or RepA-DNA complexes in binding buffer containing 50 000 c.p.m. of labeled DNA complexed with protein were placed in Eppendorf tubes in a total volume of 35 µl. To start the cleavage reaction, 5 µl each of the 0.5 mM [Fe(II)EDTA]^-2 solution, 10 mM sodium ascorbate and a 1:200 dilution of 30% hydrogen peroxide were mixed on the inner wall of the 1.5 ml Eppendorf tube then mixed immediately with the sample. The reaction was quenched after 2 min by addition of 5 µl 50% glycerol, 1 mM EDTA, and the protein-DNA complexes were separated from free DNA on nucleoprotein gels. The nucleosome samples were resolved on 0.7% agarose gels buffered with 0.5× TBE (1× = 90 mM Tris, 90 mM borate and 1 mM EDTA pH 8.3) run at 5 V/cm for 2 h. RepA-DNA complexes were separated on 5% acrylamide gels (acrylamide:bisacrylamide ratio 29:1) with the same buffer and electrophoresis conditions. Bound and free DNAs were detected by autoradiography of the wet gel, the complexed and free DNAs were isolated from the gel matrix, purified, and [sim]5000 c.p.m. of each was denatured at 90°C in formamide. The DinR-DNA complexes were precipitated with ethanol directly after footprinting. The samples were resolved on a 7% denaturing polyacrylamide sequencing gel and the autoradiographs of the dried gels analyzed by densitometry using a Bio-Rad Model GS-7000 Imaging Densitometer.

L'FeCl₃ complex preparation and DNA cleavage

L'FeCl₃ was prepared by mixing equal molar quantities (0.15 mM) of 1,4,7-trimethyl-1,4,7 triazacyclononane (Sigma Chemical Company) and ferric chloride (Sigma Chemical Company) in 1 ml methanol (17). The resulting brown crystals were incubated at 50°C for 1 h and then rinsed three times in methanol. The orange/marigold crystals obtained were dried and resuspended in 1 ml water. This solution was aliquoted and stored at -80°C until needed. UV/VIS absorbance spectra show that <5% of the original FeCl₃ remained after the synthesis (results not shown). Footprinting was conducted with 150 µM L'FeCl₃ using the same procedure described above for [Fe(II)EDTA]^-2.

RESULTS

To investigate the mechanism of DNA cleavage by L'FeCl₃ it was first necessary to determine the concentration of the complex required to cause, on average, approximately one DNA cleavage event per strand within the double-stranded DNA restriction fragment. This extent of cleavage insures that only native, undamaged complexes will be probed during footprinting experiments (6). Trial cleavage reactions performed with dilutions of the L'FeCl₃ stock and end-labeled naked DNA are shown in Figure 2A. A concentration of 150 µM consistently yielded an extent of cleavage within the desired range, since approximately two-thirds of the DNA remains uncut (8). This amount of reagent was used in all subsequent experiments and is comparable with the amount of [Fe(II)EDTA]^-2 [(100 µM)_f] used in the standard hydroxyl radical reactions.

Figure 2. Cleavage of naked DNA by L'FeCl₃. (A) L'FeCl₃ titration. Labeled DNA fragments were treated with L'FeCl₃ and then analyzed by denaturing polyacrylamide sequencing gel electrophoresis and autoradiography as described in the Materials and Methods. Lane 1, Maxam-Gilbert G-specific reaction marker; lanes 2-6, products of the cleavage reaction performed with 15 mM, 1.5 mM, 150 µM, 15 µM and 1.5 µM L'FeCl₃, respectively. (B) Comparison of cleavage products obtained with L'FeCl₃ and free iron(III). Freshly prepared aqueous 150 µM solutions of L'FeCl₃ and Fe(III)Cl₃(H₂O)₃ were used in the standard cleavage reaction and the products analyzed as in (A), as indicated above the lanes.

To confirm that the DNA cleavage resulted from action of the L'FeCl₃ reagent and not from trace amounts of unreacted iron in the stock solution, we compared the cleavage patterns of naked DNA produced by cleavage with L'FeCl₃ with that produced by free iron(III) (Fig. 2B). The specific pattern of cleavage was clearly different for each reagent, indicating that little if any of the L'FeCl₃ cleavage pattern was due to unliganded iron remaining in the preparation. Additionally, comparison of the UV/VIS absorbance spectra of a freshly prepared aqueous solution of FeCl₃[bull]123H₂O and L'FeCl₃ indicate that very little free Fe(III) remains in the L'FeCl₃ stock solution (data not shown).

We next investigated the ability of L'FeCl₃ to detect protein-DNA contacts made by a sequence-specific DNA-binding protein (13,14). The hydroxyl radical footprint of bacteriophage repressor RepA-DNA complex has been characterized and thus provides a useful comparison (14,18). RepA-DNA complexes were formed on DNA fragments containing one RepA binding site and then footprinted with L'FeCl₃ or [Fe(II)EDTA]^-2 (Fig. 3). To control for the potential dissociation of complexes by the reagents, all results were obtained with stable complexes isolated from nucleoprotein gels. Comparisons of the cleavage patterns obtained with naked DNA and the RepA-DNA complex reveal footprints are obtained with both reagents (Fig. 3 compare lane 1 with 3, and 5 with 7).

Figure 3. Footprinting of the RepA-DNA complex with [Fe(II)EDTA]^-2 and L'FeCl₃. Naked DNA or the RepA-DNA complex were treated with the cleavage reagents and fragments analyzed as described in the Materials and Methods. Lanes 1-4 and 5-8 show products of [Fe(II)EDTA]^-2 and L'FeCl₃ cleavage reactions, respectively. Lanes 1, 2, 5 and 6 contain samples from reactions with naked DNA, while lanes 3, 4, 7 and 8 contain cleavage products obtained with the RepA-DNA complex, as indicated. Reaction products shown in even-numbered lanes were obtained in the presence of 5% glycerol, as indicated. The locations of the three main sites of protection of the DNA by RepA are indicated by the vertical bars, a, b and c.

The cleavage profile of these reagents is best visualized in the densitometer scans shown in Figure 4. Inspection of the [Fe(II)EDTA]^-2 cleavage patterns reveals three main regions where cleavage is reduced by the presence of RepA protein, marked as sites a, b and c (Fig. 4A). This pattern of protection is typical for dimeric bacterial repressors (19,20). Inspection of the L'FeCl₃ cleavage pattern also reveals a footprint pattern in which distinct backbone positions are protected from cleavage by the presence of RepA (Fig. 4B). Interestingly, the densitometric analysis reveals that the footprint pattern is somewhat different from that obtained with [Fe(II)EDTA]^-2. The positions protected in the L'FeCl₃ experiment correspond to sites identified previously as positions where the RepA protein contacts the major groove of the DNA (14,18,21).

Figure 4. Densitometric analysis of the RepA-DNA footprints. Scans of selected lanes from the autoradiograph shown in Figure 3 are shown. (A) [Fe(II)EDTA]^-2-cleavage of naked DNA (top), RepA-DNA complex (middle) and a merge of the two scans (bottom). The main sites of protection are indicated as in Figure 3. (B) L'FeCl₃ -cleavage of naked DNA (top), RepA-DNA complex (middle) and a merge of the two scans (bottom). Naked DNA and RepA-DNA scans are shown with light and bold lines, respectively.

DNA cleavage by [Fe(II)EDTA]^-2 mediates DNA cleavage by a diffusible hydroxyl radical mechanism, which can be efficiently quenched by the presence of radical scavengers such as glycerol (12). To investigate whether L'FeCl₃ cleaves DNA by a similar mechanism, we added glycerol to a 5% final concentration to each cleavage reaction. DNA cleavage by [Fe(II)EDTA]^-2 is virtually eliminated in the presence of glycerol (Fig. 3, compare lane 1 with 2, and 3 with 4). In marked contrast, this dramatic loss in the cutting rate is not observed when glycerol is added to the L'FeCl₃ reaction (Fig. 3, compare lane 5 with 6, and 7 with 8). Furthermore, the addition of glycerol does not change the features of the cleavage pattern of either the naked DNA or the RepA-DNA complex obtained with L'FeCl₃.

To confirm the general utility of the L'FeCl₃ reagent for detection of sequence-specific protein-DNA interactions, we next footprinted the bacterial repressor DinR bound to its cognate DNA element (15). This protein is known to bind with high affinity to several elements upstream of DNA damage inducible genes (15). Consistent with previous work, hydroxyl-radical footprinting experiments with this complex revealed tight interactions with a 12 bp element (Fig. 5). Moreover, as found for the RepA complex, several clear but slightly less obvious protected regions were observed when L'FeCl₃ was used in the footprinting experiments. Likewise, the presence of glycerol completely abolished the Fe(II)EDTA cleavage pattern while having little affect on the L'FeCl₃ pattern (results not shown).

Figure 5. Footprinting of the DinR-DNA complex with [Fe(II)EDTA]^-2 and L'FeCl₃. Naked DNA or the DinR-DNA complex were treated with the cleavage reagents and fragments analyzed as described in the Materials and Methods. Lane 1, Maxam-Gilbert G-specific reaction; lanes 2 and 3, Fe(II)EDTA cleavge pattern of the RecA DNA fragment in the absence and presence of DinR protein, respectively; lanes 4 and 5, L'FeCl₃-cleavage pattern of the RecA DNA fragment in the absence and presence of DinR protein, respectively. Cleavage with L'FeCl₃ was performed in the presence of 5% glycerol. Bars on the side of the gel indicate the regions protected from L'FeCl₃ by the DinR protein. Hatched bar indicates the 12 bp sequence recognized by the DinR protein.

We next wished to determine whether L'FeCl₃ was useful for detection of the non-sequence-specific protein-DNA contacts within a nucleosome. To this end, the cleavage patterns of naked 5S DNA and 5S DNA reconstituted into nucleosomes were obtained with both reagents (see Materials and Methods). As expected, cleavage of nucleosomal DNA with [Fe(II)EDTA]^-2 yields a cleavage pattern exhibiting a striking sinusoidal modulation (Fig. 6, lane 5) (3). This modulation indicates the positions on the DNA backbone that are alternatively exposed to and shielded from bulk solvent due to histone-DNA contacts. Much to our surprise, the nucleosomal footprints produced by L'FeCl₃ were markedly different. The pattern produced by L'FeCl₃ cleavage lacks the pronounced periodicity seen in the [Fe(II)EDTA]^-2-mediated hydroxyl radical cleavage pattern, and indicates a much more uniform cleavage of the DNA backbone (Fig. 6, lane 6). Inspection of the cleavage pattern of naked 5S DNA produced by L'FeCl₃ shows the small amount of modulation observed in the nucleosome pattern is evident even in the absence of protein (Fig. 6, lane 4). This modulation is probably due to the effect of inherent sequence-dependent structural variations within 5S DNA on L'FeCl₃ cleavage efficiency (3,19). It is important to note that these cleavage patterns were obtained from gel-isolated, stable protein-DNA complexes, ruling out the possibility that L'FeCl₃ causes dissociation of histone-DNA interactions. In addition, experiments in which hydroxyl radical and L'FeCl₃ cleavage was carried out simultaneously on the same complex suggest that L'FeCl₃ does not cause local disruptions of histone-DNA interactions (results not shown; see Discussion). The lack of an obvious nucleosomal footprint, even in densitometer scans of this gel (Fig. 7), is quite remarkable and suggests that histone-DNA contacts and the compression of the minor groove due to the drastic bending of the DNA within the nucleosome apparently have little effect on the ability of L'FeCl₃ to cleave DNA. Thus, the presence of histone-DNA contacts provides little or no shielding of the DNA backbone from cleavage by L'FeCl₃.

DISCUSSION

We tested the ability of (1,4,7-trimethyl-1,4,7-triazacyclononane)-Fe(III)Cl₃, which cuts DNA by an unknown cleavage mechanism, and [Fe(II)EDTA]^-2, which generates diffusible hydroxyl radicals, to cleave the DNA and to detect protein-DNA interactions in two different protein-DNA complexes. An important factor is that each of these complexes is known to have distinctly different types of protein-DNA interactions. Raman spectroscopy, dimethyl sulfate reactivity and nuclease digestion studies have shown that the DNA in the nucleosome is remarkably accessible to solvent, and both the major and minor grooves are available for interactions with small molecules (22,23). Access to the DNA backbone is not occluded by the bound histone proteins, because the DNA is thought to be suspended on charged amino acid residue side chains emanating from the histone octamer, much like a train is suspended on a trestle (24,25). Conversely, the putative helix-turn-helix protein, RepA, makes strong and intimate contacts in two consecutive major grooves of the DNA, where access to the backbone deoxyriboses is blocked by the bound protein (18,20). Likewise, the DinR protein is thought to bind to DNA in a similar fashion (15). These differences between the histone-DNA and repressor-DNA complexes correlate with the different way the L'FeCl₃ cuts DNA within each complex. The binding of histone proteins to nucleosomal DNA does not appear to provide any protection to the DNA from L'FeCl₃ cleavage, while the sequence-specific DNA binding proteins protect DNA from L'FeCl₃ cleavage to a measurable degree at specific locations within their binding sites.

It is unlikely that the L'FeCl₃ cleavage pattern of nucleosomal DNA is due to simple displacement of histones from DNA by the cationic iron complex. First, we have isolated complexes from gels after cleavage, indicating that no global dissociation of the complex is caused by the presence of the reagent. Secondly, local histone-DNA interactions are known to be stable to relatively high concentrations of multivalent cations, such as Mg²⁺, which have much higher charge/mass ratios than the L'FeCl₃ reagent (23 and refs therein). Thirdly, simultaneous Fe(II)EDTA/L'FeCl₃ cleavage experiments indicate no detectable change in the stability of local histone-DNA interactions, as evidenced by the hydroxyl radical cleavage pattern (results not shown).

Although the present results are not conclusive with regard to exactly how L'FeCl₃ cleaves DNA, two key results give clues as to the mechanism of DNA cleavage by this complex. First, as reported previously (11), the addition of glycerol to the L'FeCl₃ reactions had little effect on the efficiency of DNA cleavage. Secondly, only very intimate protein-DNA contacts inhibit cleavage of the DNA by L'FeCl₃. Since the L'FeCl₃ carries an overall positive charge and is likely to directly bind the DNA, our results are consistent with the view that L'FeCl₃ cuts the DNA through a mechanism of DNA cleavage in which a reduced oxygen species such as a hydroxyl radical or a multi-valent iron-oxo complex is generated immediately adjacent to the DNA (11). Thus, these species do not diffuse sufficiently through bulk solvent for reaction with scavengers such as glycerol (12).

Figure 6. Footprinting of histone-DNA interactions within the nucleosome. Nucleosomes were prepared containing radioactively end-labeled DNA, subjected to cleavage with [Fe(II)EDTA]^-2 and L'FeCl₃, and products analyzed as in Figure 3. Lane 1, Maxam-Gilbert G-specific reaction marker; lane 2, unreacted DNA control; lanes 3 and 4, naked DNA cleaved with [Fe(II)EDTA]^-2 and L'FeCl_3, respectively; lanes 5 and 6, nucleosomal DNA cleaved with [Fe(II)EDTA]^-2 and L'FeCl_3, respectively.

Figure 7. Densitometric analysis of the nucleosome DNA footprints. Scans of selected lanes from the autoradiograph shown in Figure 5 are shown. Top, [Fe(II)EDTA]^-2-cleavage of naked DNA (light line) and nucleosome DNA (bold line). Bottom, L'FeCl₃ cleavage of naked DNA (light line) and nucleosome DNA (bold line).

The difference seen in the DNA cleavage signatures produced by [Fe(II)EDTA]^-2 and L'FeCl₃ is probably due to the DNA binding activity of the latter reagent. When DNA is bent, as is the case in a nucleosome, both the major and minor grooves become tightly compressed on the concave side of the DNA (25). Likewise, the presence of protein restricts access to the minor groove of DNA. This prevents hydroxyl radicals produced in bulk solution from easily diffusing into and abstracting hydrogen atoms from positions on the minor groove face of the deoxyribose ring (7,12). Conversely, the positively charged L'FeCl₃ molecule binds tightly to the negatively charged DNA. When reactive species are produced, they have to travel only a short distance and the DNA is cleaved at the point of attachment. Moreover, it seems that only strong protein-DNA interactions which penetrate the counter-ion condensation layer and result in a substantial release of counter ions from the DNA can lead to protection (26). In the RepA-DNA or DinR-DNA complexes, the protein forms much tighter contacts with the DNA and it is likely that DNA binding results in significant displacement of bound counter-ions from the DNA (26). Interestingly, it is known that histones displace surprisingly few bound counter-ions from DNA (27). Since L'FeCl₃ binds directly to DNA, we propose that the different extents of protection afforded by these two types of protein-DNA interactions are due to different extents of displacement of the cation [L'Fe(III)(OH₂)_n(OH)_m]ⁿ⁺ (where n + m = 3) from the DNA.

Other evidence also corroborates this proposal. The cleavage efficiency of the L'FeCl₃ complex is unaffected by the coupling of psoralen, a DNA intercalent that covalently binds DNA upon UV radiation, to the compound (11). This suggests that the L'FeCl₃ by itself has a strong affinity for the negatively charged DNA, probably due to the discharge of its chloride ions, which results in the positively charged complex. Cleavage by triazacyclononane complexes was not affected by the known free radical scavenger, DMSO, even at a concentration of 75 mM, while similar concentrations of DMSO completely quench the [Fe(II)EDTA]^-2 DNA cleavage reaction, consistent with the results presented here (11). Additionally, desferrioxamine, an extremely efficient chelator of Fe³⁺, and sodium formate, a hydroxyl radical scavenger and potential ligand for Fe³⁺, were only partially effective in quenching the L'FeCl₃ DNA cleavage, even at a concentration of 100 mM. Reductants, such as sodium ascorbate or dithiothreitol, significantly increased the amount of cleavage, while removing oxygen from the solution had the opposite effect. However, no change in cleavage was seen with or without the oxidant (NH₄)₂Ce^IV(NO₃)₆ (11). This implies that the active species in the cleavage reaction results from hydrolysis of the original complex and that DNA cleavage is mediated by a reaction between iron (II) center and dissolved oxygen.

CONCLUSIONS

We conclude that L'FeCl₃ may be a powerful tool for probing protein-DNA interactions. The mechanism of cleavage remains unknown, but it probably involves the formation of iron-oxo species immediately adjacent to the DNA backbone. Cleavage of DNA by this complex is: (i) sensitive to DNA sequence or some of the sequence-dependent structural variations found in naked B-DNA in solution, (ii) not affected by histone-DNA interactions or severe DNA bending found in the nucleosome and (iii) can be diminished by intimate protein-DNA contacts such as are found when sequence-specific DNA binding proteins are bound to their cognate elements. Experiments to determine if this reagent can detect the presence of a sequence-specific DNA binding protein bound to DNA assembled into nucleosome complexes are presently underway.

ACKNOWLEDGEMENTS

We thank Mr Woong Kim for preparation of histone proteins. We are grateful to Dr Teresa E.Strzelecka, National Institute of Health, Bethesda, MD for her generous gift of RepA protein and to Drs K.Winterling and R.Woodgate for their gift of DinR protein. This work was supported by NIH grant GM52426. D.C. was supported by University of Rochester Cancer Center training grant #CA09363D-16A1.

REFERENCES

1. Hayes, J. J. (1995) Chem. Biol. 2, 127-135. MEDLINE Abstract

2. Tullius, T. D. (1989) Annu. Rev. Biophys. Biophys. Chem. 18, 213-237. MEDLINE Abstract

3. Hayes, J. J., Tullius, T. D. and Wolffe, A. P. (1990) Proc. Natl. Acad. Sci. USA 87, 7405-7409. MEDLINE Abstract

4. Roberts, M. S., Gilberto, F. and Hager, G. (1995) Biochemistry 34, 12470-12480. MEDLINE Abstract

5. Nightingale, K., Dimitrov, S., Reeves, R. and Wolffe, A. P. (1996) EMBO J. 15, 548-561. MEDLINE Abstract

6. Tullius, T. D. (1987) Trends Biochem. Sci. 12, 297-300.

7. Tullius, T. D. (1988) Nature 332, 663-664. MEDLINE Abstract

8. Dixon, W. J., Hayes, J. J., Levin, J. R., Weidner, M. F., Dombroski, B. A. and Tullius, T. D. (1991) Methods Enzymol. 208, 380-413. MEDLINE Abstract

9. Vinson, C. R., Sigler, P. B. and McKnight, S. L. (1989) Science 246, 911-916. MEDLINE Abstract

10. Hayes, J. J. and Tullius, T. D. (1992) J. Mol. Biol. 227, 407-417. MEDLINE Abstract

11. Silver, G. C. and Trogler, W. C. (1995) J. Am. Chem. Soc. 117, 3983-3993.

12. Pogozelski, W. K., McNeese, T. J. and Tullius, T. D. (1995) J. Am. Chem. Soc. 117, 6428-6433.

13. Wickner, S., Hoskins, J. and McKenney, K. (1991) Proc. Natl. Acad. Sci. USA 88, 7903-7907. MEDLINE Abstract

14. Papp, P. P. and Chattoraj, D. K. (1993) J. Mol. Biol. 233, 219-230. MEDLINE Abstract

15. Winterling, K. W., Chafin, D. C., Hayes, J. J., Ji, S., Levine, A. S., Yasbin, R. E. and Woodgate, R. (1998) J. Bacteriol. 180, in press.

16. Hayes, J. J. and Lee, K. (1997) Methods Comp. Meth. Enzymol. 12, 2-9.

17. Chaudhuri, P., Winter, M., Wieghardt, K., Gehring, S., Haase, W., Huber, B. and Weiss, J. (1988) Inorg. Chem. 27, 1564-1569.

18. Ehmann, A. (1997) Thesis, University of Rochester, Rochester, NY.

19. Tullius, T. D. and Dombroski, B. A. (1986) Proc. Natl. Acad. Sci. USA 83, 5469-5473. MEDLINE Abstract

20. Hayes, J. J. and Tullius, T. D. (1989) Biochemistry 28, 9521-9527. MEDLINE Abstract

21. Papp, P. P. and Chattoraj, D. K. (1994) Nucleic Acids Res. 22, 152-157. MEDLINE Abstract

22. McGhee, J. D. and Felsenfeld, G. (1979) Proc. Natl. Acad. Sci. USA 76, 2133-2137. MEDLINE Abstract

23. vanHolde, K. E. (1989) Chromatin. Springer-Verlag, New York.

24. Arents, G. and Moudrianakis, E. N. (1993) Proc. Natl. Acad. Sci. USA 90, 10489-10493. MEDLINE Abstract

25. Luger, K., Mader, A. W., Richmond, R. K., Sargent, D. F. and Richmond, T. J. (1997) Nature, 389, 251-260. MEDLINE Abstract

26. Record, M. T., Jr, Lohman, T. M. and De Haseth, P. (1976) J. Mol. Biol. 107, 145-158. MEDLINE Abstract

27. McGhee, J. D. and Felsenfeld, G. (1980) Nucleic Acids Res. 8, 2751-2769. MEDLINE Abstract

---

*To whom correspondence should be addressed. Tel: +1 716 273 4887; Fax: +1 716 271 2683; Email: jjhs@uhura.cc.rochester.edu

---

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 17 Apr 1998
Copyright©Oxford University Press, 1998.
CiteULike    Connotea    Del.icio.us    What's this?

This Article Abstract Print PDF (197K) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (2) Request Permissions Commercial Re-use Guidelines for Open Access NAR Content Google Scholar Articles by Ehmann, A. Articles by Hayes, J. J. Search for Related Content PubMed PubMed Citation Articles by Ehmann, A. Articles by Hayes, J. J. Social Bookmarking          What's this?