drug–dna interaction protocols

1 DNase I Footprinting Keith R Fox Introduction Footprmtmg provides a simple, quick, and reasonably mexpensive method for assessingthe sequence specific mteraction of ligands with DNA Although the techmque was developed in 1978 for studying the mteraction of DNAbinding proteins with then target sites (I), it has proved invaluable for determining the sequence specificity of many small hgands 1.1 Footprinting , Footprmting is essentially a protection assay, m which cleavage of DNA is inhibited at discrete locations by the sequence specific binding of a hgand or protein In this technique, a DNA fragment of known sequence and length (typically a restriction fragment of 100-200 bp), which has been selectively radiolabeled at one end of one strand, IS lightly dtgested by a suitable endonucleolytic probe m the presence and absence of the drug under investigation The cleavage agent is prevented from cutting around the drug-binding sites so that, when the products of reaction are separated on a denaturing polyacrylamide gel and exposed to autoradiography, the position of the ligand can be seen as a gap m the otherwise continuous ladder of bands (see Fig 1) In this figure, cleavage at position “a” will produce, after denaturing the DNA, one long fragment (9 bases) corresponding to the left hand strand, and two short fragments (7 bases and bases) from cleavage of the right hand strand Since the bands are located by autoradiography, only the shortest of these species bearing the radioactive label will be visualized The condittons of the cleavage reaction are adjusted so that, on average, each DNA fragment is cut no more than once As a result, each of the bands on the autoradiograph is produced by a single cleavage event, i.e., single-hit kmetics If an excessive amount of cleavage agent is used, then From Methods m Molecular Edited by Biology, K R Fox Vol 90 Drug-DNA Humana Press Interactron Inc , Totowa NJ Protocols Fox gel eleotrophoresis Fig Schemattc representation of the footprtntmg experiment The DNA is labeled (*) at the 3’ end of the right-hand strand labeled products can arose from more than one cleavage event, biasing the dlstribution of fragments toward short products In general, the extent of cleavage 1sadjusted so that between 60 and 90% of the radtolabeled DNA remains uncut, though longer fragments require greater amounts of digestion able band intensities to produce suit- DNase I footprmtmg has been successfully employed for mdentrfymg or conlirmmg the preferred DNA binding sites for several hgands mcludmg actinomycm (2-4), mtthramycin (5), quinoxalme antrbrotrcs (6,7), daunomycm (8,9), nogalamycin (1/J), vartous minor groove binding agents (2,3,12), and triplex binding ohgonucleottdes (12,13) Various other cleavage agents, both enzymrc and chemical, have also been used as footprinting probes for drugDNA interactions including micrococcal nuclease (24), DNase II (6,15), copper phenanthrolme (16,17), methtdiumpropyl-EDTA.Fe(II) (MPE) (18-21), uranyl photocleavage (22,23), and hydroxyl radicals (24-26) Each of these has a different cleavage mechanism, revealmg different aspects of drug-DNA interactions An ideal footprmtmg agent should be sequence neutral and generate an even ladder of DNA cleavage products in the absence of the hgand This property is almost achieved by certain chemical probes, such as MPE and hydroxyl radicals However, the most commonly used cleavage agent (because of its cost and ease of use) 1sthe enzyme DNase I, which produces an uneven cleavage pattern that varies according DNA sequence and local structure (see Subheading 1.2.) Cleavage at mdrvtdual phosphodiester bonds can vary by over an order DNase I Foo tprinting of magnitude m a manner determined by both local and global DNA structure (27,28) In addltlon, drugs that modrfy DNA structure can induce enhanced DNase I activity m regions surroundmg their binding sites if they alter the DNA structure so as to render it more suscepttble to cleavage (3,6,15,29,30) This ISmost frequently seen m regions that are particularly refractory to cleavage m the drug-free controls 1.2 DNase I DNase I 1sa monomeric glycoprotem of mol wt 30,400 It IS a double strandspecific endonuclease,which introduces single strand nicks m the phosphodiester backbone, cleaving the 03’-P bond Single stranded DNA is degraded at least four orders of magmtude more slowly (32,32) The enzyme requires divalent cations and shows opttmal actlvlty m the presence of calcmm and magnesium (33) Although it cuts all phosphodiester bonds, and it does not possess any simple sequence dependency, its cleavage pattern 1svery uneven and 1sthought to reflect variations m DNA structure (27,34) In particular, A, * T, tracts and GC-rich regions are poor substrates for the enzyme The most important factors affecting Its cleavage are thought to be mmor groove width (27,28) and DNA flexibility (35,36) Several crystal structures have been determined for both the enzyme and its complex with oligonucleotides (37-42) These show that DNase I bmds by inserting an exposed loop mto the DNA minor groove, Interacting with the phosphate backbone, as well as the walls of the groove This explains why cleavage is poor in regions, such as A,, * T, tracts on account of their narrow minor groove, to which the enzyme cannot bind An additional feature of these crystal structures 1sthat the DNA 1salways bent by about lo toward the major groove, away from the enzyme If this bendmg 1sa necessary feature of the catalytic reaction, then rigid regions, such as GC-rich sequences,may be refractory to cleavage However, these factors not explain the very different cutting rates that are often observed at adjacent dinucleotide steps.It 1spossible that this is determined by precise orientation of the sclssile phosphodlester bond, However, the crystal structures show that there may be other specific interactions between the exposed loop and DNA bases removed from the cutting site In particular, tyrosme-76 mteracts with the base posItIons to the 5’ side of the cutting site and arginme-4 binds to the base at position -3 This latter mteraction 1ssterically hindered by a GC base pair in thts position By examining the characteristics of several good DNase I cleavage sites, Herrera and Chaires (43) suggested that the best cleavage site was WYWIWVN (where W = A or T, Y = C or T, and V = any base except T) The DNA-binding surface of DNase I covers about 10 bp, i.e., one complete turn the DNA helix This has tmportant consequences for interpreting Fox A B Fig Schemattc representatron of the 3’staggered cleavage produced by DNase I The DNA helix has been opened out and IS viewed along the minor groove The hatched box represents DNase I the tilled box represents a DNA-binding ligand footprmtmg results and explams the observatton that the enzyme overesttmates drug-binding site sizes Although DNA bases he perpendtcular to the hellcal axis, they are mclmed relative to the phosphodtester backbone As a result, closest phosphates, postttoned across the minor groove, are not attached to a single base pan, but are staggered by about 2-3 bases m the 3’ direction This is illustrated m Fig 2A, m which the DNA has been drawn lookmg along the minor groove, showmg the inclmatton of the DNA base pans Since DNase I (hatched box) binds across this groove, its bmdmg sate on the top strand 1s located bases to the 3’ side of that on the lower strand When a DNA-binding hgand is added (filled box in Fig 2B), it can be seen that the closest approach of the enzyme is not the same on each strand DNase I can approach closer to the enzyme on the lower strand; the region of the upper strand protected extends by about bases beyond the actual ligand-bmdmg sate As a result, DNase I footprmts are staggered by about 2-3 bases m the 3’ direction across the two strands Materials 2.1 DNase I For most footprintmg experiments the DNase I does not need to be especially pure There 1s ltttle advantage m purchasmg HPLC-pure, RNase-free enzymes Currently purchased 1s the type IV enzyme, from bovme pancreas, from Sigma (St Louis, MO) This should be dtssolved m 0.15 MNaCl contaming mMMgC1, at a concentratton of 7200 Kumtz U/mL Thts can be stored at -20°C, and is stable to frequent freezing and thawing The enzyme 1sdiluted to workmg concentrattons immedtately enzyme should be discarded before use; the remainder of the diluted DNase I Footprinting Table Sequence of the tyrT DNA Fragment AATTCCGGTTACCTTTAATCCGTTACGGATGAAAATTACGC~CCAGTTCATTTTTCTC~CGT~CAC 10 20 30 40 3'-AAGGCCAATGGAAATTAGGCAATGCCTACTACTTTT~TGCGTTGGTC~GT~GAGTTGCATTGTG 50 60 TTTACAGCGGCGCGTCATTTGATATGATGCGCCCCGCTTCCCGAT~GGGAGCAGGCCAGT~GCATT 70 80 90 100 110 AAATGTCGCCGCGCAGTAAACTATACTACGCGGGGCGAAG 120 130 ACCCCGTGGTGGGGGTTCCC 140 150 TGGGGCACCACCCCCAAGGGCT-5' The fragment ISobtainedby cutting with EcoRI andAvuI a-32P-dATP used label 3’endof IS to the the lower strand,whereas a-32P-dCTP used labelthe upperstrand IS to 2.2 Choice of DNA Fragment 2.2.1 Natural DNA Fragments For footprinting experiments, the length of fragment used depends on both convenience (how easily a specific fragment can be generated) and the resolution limit of the polyacrylamide gels The chosen fragment length is typically between 50 and 200 bp Although different laboratories have adopted different natural fragments as standard substratesfor footprmtmg experiments, a few have been used more widely Among these are the 160 bp tyrT fragment (sequence shown m Table 1) t&8)), the EcoRI-PvuII fragments from PBS (Stratagene) (4&M), and several fragments from pBR322 (HindIII-HueIII, HindIII-AM, or EcoRI-RsaI) The plasmids from which these can be prepared are available from commercial sources or from the author’s laboratory In many ways it would be convenient if a few fragments did become recognized standards, since this would facilitate direct comparison of the relattve specrfictttes of hgands prepared in different laboratories Since many sequence selective small molecules have recognition sites of between and bp, there is a reasonable probability that their preferred sites will be present in a lOO- to 200-bp restriction fragment However, it should be noted that there are different bp, 10 different dmucleotides, 32 trmucleotides, 136 tetranucleotides, 512 pentanucleotides, and 2080 hexanucleotides It can therefore be seen that the chance of finding a particular binding site within a given DNA fragment becomes more remote the greater the selectivity of the ligand A further complicatmg factor is that, although many ltgands spectfically recognize only a dmucleotlde step, their binding affinity is often influenced by the nature of the surrounding bases, Fox which alter the local DNA structure (47-49) It IS therefore possible that using a natural fragment may fail to detect the optimum bmdmg sites for the most selective hgands This becomes especially relevant since many novel synthetic ligands possessenhanced sequence recogmtton properties, with binding sites of eight or more base pairs 2.2.2 Synthetic Oligonucleotides As explamed, although footprmtmg experiments with natural DNA fragments provide a reasonable estimate of a ligand’s preferred bmdmg sites, these are complicated by the limited number of sequences studied, together with ambiguities over the exact bmdmg site within a larger footprmt The next step m confirmmg the sequence preference may be to prepare a synthetic DNA fragment containing the putative binding site and to use this as a substrate for footprmting experiments (50,51) In addition, for compounds that have been produced as the result of rational design, one may be able to predict their preferred bmdmg site Synthesis of suitable length ohgonucleotides (50 bases or longer) IS now routine However, the results obtained with short oligonucleotides need to be interpreted with caution and rigorously controlled for several reasons First, binding sites located close to the ends of short ohgonucleotides may not adopt the same configuration as when located within longer sequences because of “end effects.” Second, smce the synthetic fragments will contam only one or two binding sites, it is necessary to ensure that other sequences with equal or greater affinity have not been excluded This can be investigated by comparing the mteraction with other closely related sequences, m which one or two bases m or around the cognate sequence are altered m turn Analysis is simphfied further if the variant sites are contamed withm the same DNA fragment 2.2.3 Synthetic Fragments A frequent variant on the above is to clone the synthetic oligonucleottdes mto longer DNA fragments This removes the problems associated with end effects and provides other common flanking sequences to which ligand binding can be compared An added advantage is that, once it has been cloned, the sequence can be readily isolated from bacteria The authors usually clone synthetic ohgonucleotides mto the BamHI site of pUC plasmids They have prepared a wide range of such cloned inserts, containing central GC, CG, or (A/T),, sites (11,15,29,30), which are available from the authors’ laboratory on request DNA fragments contammg the synthetic inserts can be prepared and radiolabeled at either end (see Subheading 3.2.) by isolatmg the modified polylmker Once again a proper analysis will requtre fragments contammg both cognate and closely related noncognate sequences DNase I Footprinting 2.3 Buffers 2.3.1 Solutrons for Plasmid Preparation Resuspenston solution 50 mM Trts-HCl pH 5, contammg 10 mM EDTA Lysis solution 0.1% SDS, 0.1 MNaOH Neutralization solutton M potassium acetate, A4 acettc acid 2.3.2 Genera/ Buffers 10 mA4Tris-HCl, pH 5, contannng mA4EDTA This is used for dtssolvmg DNA 10 mM Trts-HCl, pH 7.5, containing 10 mA4 NaCl This is used for preparing drug solutions DNase I buffer 20 mMNaC1,2 mM MgCl*, mM MnC& 2.3.3 Reagents for Electrophoresis TBE electrophorests buffer This should be made up as a 5X stock solutton containing 108 g Tns, 55 g Boric acid, and 9.4 g EDTA made up to L with water Acrylamide solutions Polyacrylamide sequencing gels are made from a mixture containing acrylamtde*btsacrylamtde in the ratio 19.1 Because of the toxic nature of these compounds acrylamide solution are best purchased from a commerctal supplier (National Diagnostics [Atlanta, GA], Anachem [Luton, Beds, UK]) and should be used according to the manufacturers mstructions DNase I stop solution Formamide containing 10 mM EDTA and 1% (w/v) bromophenol blue Methods 3.1 Plasmid Preparation Several methods are available for preparing plasmid DNA, which IS suitable for restriction digestion and radiolabeling, including several commerctal kits (including Qiagen or Wizard) and caesium chloride density gradient centrifugation It 1sbeyond the scope of this article to review the relative merits of each procedure, except to note that in many instances it is not necessary to generate high purity plasmid preparations Since the radtolabeled restrtction fragments are eventually isolated and purified by gel electrophoresis, prior purification of the plasmids may not be necessary, so long as the preparations not contain nucleases or any agents that inhibit restriction enzymes or polymerases As a result, plasmtds are usually prepared by standard alkaline lysts procedures, followed by extraction with phenol/chloroform A very brief protocol for extractmg pUC plasmids 1sdescribed as follows: Grow 50 mL bacteria overnight Spin culture at 3000g (I e., 5000 rpm m a Beckman JA20 rotor) for mm m Oakridge tube Fox Resuspend the bacterial pellet m mL cell resuspension solution (50 mM Tns-HCl, pH 7.5, containing 10 mM EDTA) Add mL cell lysis solution (0 1% SDS, MNaOH) and mix gently until the solution becomes clear Add mL neutralization solution (3 M potassium acetate, M acetic acid) Spin at 17,000g (12,000 rpm) for 15 mm Remove the supernatant and add vol of lsopropanol Spin at 17,OOOg(12,000 rpm) for 15 mm Remove the supernatant and wash the crude DNA pellet with 5-10 mL 70% ethanol Transfer the pellet to an Eppendorf tube and dry 10 Redissolve pellet m mL 10 mA4 Tns-HCl, pH 5, containing 0.1 mM EDTA and 100 pg/mL RNase Leave at 37°C to dissolve for at least 30 mm 11 Extract twice with mL phenol/chloroform (phenol forms the bottom layer and should be discarded) The interface will probably be very messy, leave the Junk behind 12 Remove any dissolved phenol by extracting twice with mL ether (which forms the top layer and should be discarded) Allow excess ether to evaporate by standing at 37°C for a few minutes 13 Precipitate with ethanol, dry and dissolve m 100-l 50 JJL Tns-HCI, pH 5, containing 0.1 mM EDTA 3.2 Radiolabeling the DNA DNA fragments can be efficiently labeled at either the 5’ end (using polynucleotlde kmase) or 3’ end using a DNA polymerase However, the results of DNase I digestion are easiest to interpret for 3’-end-labeled fragments Smce DNase I cuts the 03’-P bond, the products of dlgestlon possess a 3’-hydroxyl and 5’-phosphate group In contrast, Maxam-Gilbert sequencing reactions, which are used as markers in footprmtmg gels (see Subheading 3.3.), leave phosphate groups on both sides of the cleavage pomt (52) As a result, the radlolabeled products of DNase I cleavage and Maxam-Gilbert sequencmg reactions will be identical if the DNA 1s labeled at the 3’ end (i.e., both possess a phosphate at the 5’ end) However, if the DNA 1s labeled at the 5’ end then the labeled DNase I products will possess an extra phosphate group and so run slightly faster than the correspondmg Maxam-Gllbert products Although this difference 1s often overlooked in footprmtmg gels, it becomes significant for short fragments for which the difference m mobility may be as great as 2-3 bands For enzymes that cut the O-5’ bond, such as DNase II and mtcrococcal nuclease, 5’-end-labeled fragments comlgrate with the Maxam-Gilbert marker lanes 3.2.1 3’-End Labeling with Reverse Transcriptase The production of 3’-end-labeled DNA fragments can be achieved by cutting with a restrlction enzyme that generates sticky ends with 3’-overhanging DNase I Footprmting ends, followed by filling m with a polymerase using a suitable [a-32P]-dNTP The fragment of interest IS then released from the remamder of the plasmid by cleaving with a second enzyme that cuts the other side of the region of interest The two restriction enzymes usually cut at single locatlons in the plasmid, though this 1snot necessary so long as the various radiolabeled fragments can be separated from each other The most commonly used polymerase is the Klenow fragment However, it is found that the most efficient labeling is achieved using AMV reverse transcriptase, even though this 1s actually an RNA-dependent DNA polymerase However, not all commercially sources of this enzyme are equally reltable; consistent results are obtained with reverse transcrlptase from Promega or Pharmacia l RESTRICTION DIGESTION AND a’-END LABELING Using the aforementioned procedure for DNA isolation, the followmg 1s used for generating radlolabeled Hindlll-EcoRl polylmker fragments from pUC plasmids Mix 30 pL plasmld (about 50 pg DNA) with 10 pL of 10X restrlctlon enzyme buffer (as supplied by the manufacturer), 45 PL water Add pL HzndIII (A/AGCTT) and incubate at 37°C for h Add PL [a-32P]-dATP (3000 Wmmol, Amersham)together with PL reverse transcriptase and Incubate for a further h The reverse transcriptase IS then Inactivated (to prevent further mcorporatlon of radiolabel at the 3’ end of the second restrlctlon site) by heatmg at 65°C for mm After cooling to 37”C, pL EcoRI (G/AATTC) is added and the mixture mcubated for a further 1-2 h In this case, the DNA can be labeled on the opposite strand by reversing the order of addition of EcoRI and HzndIII If the second enzyme produces blunt ends or sticky ends with 5’ overhangs, or if the 3’ overhangs sites can not be filled m with dATP, then all the enzymes can be added simultaneously Examples of such combinations for pUC polylinker fragments are HzndlII-SacI, and EcoRI-&I The @rT fragment can be prepared by simultaneous digestion with EcoRl and Aval In this instance the EcoRl end is labeled with [a-32P]-dATP, whereas the Aval end can be labeled with [a-32P]dCTP Although various enzymes are supplied with dlfferent reaction buffers, it 1sfound that there IS usually no need to change buffers between the first and second enzymes The mixture of radlolabeled fragments is preclpltated by addmg 10 PL of M sodium acetate and 300 pL ethanol, followed by centrlfugatlon m a suitable microfuge, at top speed The pellet 1swashed with 70% ethanol, dried and dlssolved m 15-20 FL Tris-HCl containing mA4 EDTA Then PL of loading dye (20% F~oll, 10 mA4EDTA, 1% [w/v] bromophenol blue) is added before 10 Fox loading onto a polyacrylamide gel (typically 6-8%) The gel should be run cold, so as not to denature the DNA, it is usually run 3-mm-thick, 40-cm-long gels in 1X TBE at 800 V Samples are loaded into slots 10 mm wide by 15 mm deep After the bromophenol blue has reached the bottom of the gel (about h), the plates are separated and the gel covered with Saran wrap Scanning the gel with a hand-held Geiger counter should give a reading off scale (1 e , at least 3000 cps) over the radiolabeled bands The precise location of the radiolabeled bands is determined by short (2-10 min) autoradiography This autoradlograph IS placed under the glass plates and used to locate the band of Interest, which IS cut out using a sharp razor blade 3.2.1.2 EXTRACTION OF RADIOLABELED DNA FRAGMENTS The simplest, labeled DNA cheapest, and most efficient fragments from polyacrylamlde method for extracting gel slices IS by diffusion radioPlace a small glass wool plug m the bottom of a mL (PlOOO) pipet tip and seal the bottom end with parafilm Add the gel slice containing the radiolabeled DNA and cover this with 10 mA4 Tris-HCl, pH 5, containing 10 mM EDTA (about 300 pL is sufficient) Cover the top of the pipet tip with parafilm and incubate at 37°C with gentle agitation This is usually incubated overmght, though most of the DNA elutes after h Remove the parafilm from the top and bottom of the tip and expel the buffer mto an Eppendorf tube using a pipet and/or lowspeed centrifugation (15OOg m an Eppendorf centrifuge) The gel slice should be retamed in the pipet tip by the plug of glass wool, though a small amount of polyacrylamide does occasionally come through This can be removed by centrifugation For fragments shorter than 200 bp, this procedure recovers about 95% of the radiolabel m the gel slice, though the efficiency decreases for longer fragments The DNA should then be precipitated with ethanol and redissolved m Tris-HCI containing 0.1 mA4 EDTA so as to generate at least 10 cps per pL on a hand-held counter For most footprintmg experiments it is not necessary to know the absolute DNA concentration, since this is vamshmgly small The important factor is concentration of the radiolabel, which should be sufficient to produce an autoradiograph within l-2 d exposure 3.3 Maxam-Gilbert Marker Lanes Bands in the DNase I digestion patterns are identified by comparison with suitable marker lanes Since each DNA fragment produces a characteristic sequence dependent digestion pattern, it is sometimes possible to identify the bonds by comparison with a previous (published) pattern 3.3.1 G-Tracks The simplest and most commonly used marker lane is the dimethylsulfatepiperidme marker specific for guanine (52) Since the procedure is more time- Hopkms thermocouples or thermopiles are mcreased dramatically with high-sensmvity voltage amplifiers and monitored dtrectly (5) during the reaction An mstrument (5) usmg thermopiles in this manner and using stopped-flow mixing has been reported to have a resolution of 0.01 peal for heats as small as peal In a dtfferenttal calorimeter designed to keep a reference and sample cell at the same temperature, voltages produced by thermocouples or thermoptles connecting the reference and sample cells are applied to a feedback cncutt (4) providing power to the cells Power is added automattcally to the sample cell tf the process is endothern-nc or to the reference cell if the process 1sexothernnc With one of the commercial instruments that operates on the latter principle (4), heats as small as 10 peal can be measured to the nearest 0.5 peal In prmciple, differential scanning calortmeters (DSC) can be used to assess heats associated with the melting of the DNA structure m the presence and absence of the drug (7) Subtractton of the former quantity from the latter produces the desired heat associated with the drug-DNA mteraction Several comphcattons arise* endotherms for the bound and empty sttesmay be seen at low coverage, and the calculated AH is not for 25°C but for temperatures approachmg 90°C m many cases Only a few of the commercial DSC instruments (47) have the requisite sensitivity needed to the measurements 7.2 Concentration and Affinity Limitations When the drug aggregates at the concentrations of the measurements, the observed heats are a sum of at least two terms (5.8): AHots= &mdmg + (aggregation fractron)AHdeagg In many cases, AHblndlng1s exothermtc (negative thermodynammally) and AH deagg, enthalpy change for the productton of monomers from the aggrethe gate, 1susually endothermtc Problems associated with this phenomenon can be avoided by preforming the measurements at concentrattons at which the fraction of drug m the monomer state ts close to one A complete analysts of the observed heats when aggregation happens requnes the eqmhbrmm constant for the aggregation process ($8) Most AH values for drug-DNA mteractions are m the to -15 kcal/mol range Assummg a value of -2 kcal/mol for a hypothetical drug-DNA interaction, 0.005 pmol of the drug must bmd to the DNA lattice to produce 10 peal of heat Using a volume of 1.5 mL m the calorimeter cell, the minimum requtred concentration of sites would be 3.3 pM Intercalation requires at least bp per bmdmg site and groove bmdmg may require more, thus, the mmtmum concentration of basesm the calorimeter cell in this example would be 13.3 pil4 (6.6 PM m basepans) If posstble, the number of sites should be 50-200 times the number of drug molecules added at the beginning of the tttration When the drug is Calorimetric Techniques 261 added to the DNA solution m IO-yL volumes m a titration, then the concentration of the drug must be in the 1.O-mmol range In a the heat-flow instrument described (5) m the literature with stopped-flow inlection, the DNA and drug solutions are mixed in a 1:1 volume ratio, and one can use drug concentrations that are only twice as large as the concentration of occupied DNA sites after mixing However, the drug-to-basepair ratio is not easily changed m this instrument, and measuring the AH over a wide range of ratios requtres many individual experiments If the binding constant for the mteraction is lo6 or larger, then over 98% of the drug will be bound after the mitial addition With smaller binding constants or very low concentrations, the fraction bound at each addition of drug must be calculated from known binding constants (see Note 4) The apparent AH, calculated from the heat for bmdmg, must be divided by the calculated fraction to give the correct AH for attachment of the drug to the site on the DNA lattice 1.3 Heat-Pulse Analysis One commercial (4) instrument determmes the differential power (peal/s) needed to keep a reference and sample cell at the same temperature while very slowly increasing the temperature of both cells The observed differential power values measured with such an mstrument are plotted vs time m Fig Data are shown for two injections of propidmm iodide mto a poly(dA) * poly(dT) solution at the beginning of a titration When the solutton containing proprdmm iodide was injected mto the sample cell, the production of heat caused by the association of propidium iodide with the DNA duplex caused a series of deflections on the power axis (heat-pulse) When the rate of attachment of the drug to DNA is fast and efficient stirrmg is present, the shape of the pulse is determined by the electrical response of the calorimeter (approx s in this case) The areasunder the pulses are the total heatsassociatedwith all processesin the calorimeter cell These areas are determined by connecting the regions before and after the injection by a straight line, and performing an integration of the curves by a numerical procedure, which for the data shown produces heats of 38.1 peal In the commercial instrument (see Subheading 2.1 for company address) used to produce the data shown m Fig 1, programs automatically perform the tasksfor the investigator Similar shapedheat pulses are observed in heat-flow calorimeters (5) without differential feedback Both type of calorimeters must be calibrated routmely with known heats produced by electrical currents m resistors on the calorimeter cells and periodically with a chemical or physical process(seeNote 7) 1.4 Calorimetric Data and Analysis for AHbinding In a separate titration, the heats associatedwith the dilution of the drug in the sample cell with only the buffer present is determmed for a series of injections Hopkins 262 02 &!I 80- 00 0 00 78- 0 O 760 0 0 0 74- 72- 0 0 0 0 0 0 66 Time (min) Fig Power plotted vs time for two qections (2 pL each) of propldlum iodide poly(dT)(O.OOl M m dA) m PIPES buffer at (0.0024 M) mto 43 mL of poly(dA) pH = 7.0 and 300 IS with [NaCI] = 015 A4 The areas under the two pulses are 38 peal and are for endothermic heats If these heats are small relative to the heats observed when DNA is present, one assumesthat the average of these values 1sthe best estimate for the heat of dilution correction When better estimates of the heats of drug-DNA mteraction are required, the aggregation constants must be determined and used in the analysis for the heat of dllutton correction ($8) The observed heats (I-5,8) minus the dilution heats are converted to the apparent AH of binding values by dividing these heat values by the number of moles of drug delivered in the injections When the binding constant IS available or can be estimated from literature values, the apparent AH values can be converted to the thermodynamic parameter for attaching a drug molecule to the DNA latttce at this point m the titration The apparent AH values can be plotted vs the mole ratio of drug to DNA base pair, as IS shown in Fig for the titration (9) of propidmm iodide mto poly(dAdT) at low [NaCI] at 308 K From a mole ratio of 0.002 all the way to 0.15, the AH IS nearly constant at -7.5 kcal/mol of propldium added Between 0.15 and 0.25, the AH rapidly approaches -1 kcal/mol and remams near this value out to a ratio of 0.4 This plot IS in accord with there being one highaffinity site on the DNA lattice, which is saturated after a mole ratio of 0.3 Calorimetric Techniques 263 m 003 004 -2 -4 rn 00 -6 5- m n I 000 1(1m 001 002 005 I I I 01 02 03 04 Ratio (mol Propidium lodidelmol dA) Fig Plot of the apparent AH for binding of propidium dication to poly(dAdT) in PIPES buffer at pH = 7.0 and 308 K with [NaCl] = 015 M The Insert IS a plot for ratios below 0.05 and 1s used to determine the intersection of the curve at zero ratio, I.e., the AH for attachment of the drug to the unperturbed DNA lattice Assuming this site to be intercalatton, the apparent AH value for the mtercalation of a propidium dication into the unperturbed lattice IS easily evaluated from the mtersectton of the curve with the y-axis near zero ratio (see the insert) At 16°C lower (292 K) proprdium iodide titrated (9) into poly(dAdT) at low [NaCl] (Fig 3) produces a AH vs mole ratio curve that is similar in shape at the beginning and end to the curve shown in Fig However, between 0.04 and 0.20 mol ratio, the AH becomes dramatically more negative before sharply rising to near zero at 0.25 mol ratro This unexpected shape observed in similar experiments (3,9) may be because of interactions between occupied sites or a distribution between different type sttes Nevertheless, rt is seen that all the high-affinity sites are nearly saturated, and that the AH can be calculated for attachment of the drug to a single site on the unperturbed DNA lattice (see insert at left of Fig 3) The two plots shown in Figs and for the tttratton of proptdium iodide into poly(dAdT) also illustrate the dramatic variations that can be observed by varying the temperature by only 15OC.Apparently, the AH for attachment of propidium iodide to this particular DNA lattice IS very dependent on temperature and the AC,, for the process is large 264 n 8 % -6- I = d - 8 8 -6 - 01 00 I 02 Ratio (mol Propidium 03 lodidelmol t t dA) Fig Plot of the apparent AH for bmdmg of proptdtum dicatton to poly(dAdT) in PIPES buffer at pH = and 292 K with [NaCl] = 0 15 M The Insert left 1sa plot for ratios below 05 and is used to determme the mtersectton of the curve at zero ratio, I e , the AH for attachment of the drug to the unperturbed DNA latttce Materials 2.1 Calorimetric Apparatus Only the recently developed calorimeters (4,s) with microcalorie or better resolution and small total volumes can be used effectively m drug-DNA studies Several manufacturers have offered calorimeters that can perform heat measure- ments at the low levels needed for evaluating AH for drug-DNA interactions A titration Instrument, such as the one produced by Mtcrocal (Northampton, MA), can perform (see Fig 1) many injections mto a DNA solution to quickly produce the entire tttratton curve Breslauer and coworkers (10) used a Mtcrocal, mstrument m their recently reported studies on bereml binding to DNA and RNA du- plexes, instead of the stopped-flow mstrument used by Breslauer and coworkers (5) m earlier studies The stopped-flow instrument (5) can perform multiple mjections at a fixed drug-to-DNA ratio with a total volume after mixing nearly 200 pL The volume of drug and DNA solutions needed in this instrument to fill the syringe and lines connecting the mixing chamber is much larger 2.2 Buffers and Salts All buffers and salts should be analytical grade, and any of the buffers (e.g., PIPES) that not interact appreciably with metal ions can be used (see Note 1) Calorimetric Techniques 265 The preferred buffer for these thermodynamtc studies is cacodylate (10 mM, pH 6.8) which may mhtbtt microorganisms from growing m the calorimeter Heat effects from mtcroorgamsm in the soluttons or on the walls of the calortmeter cells can cause very erratic baselmes (see Note 2) All soluttons should be prepared with water that has been treated to remove organic matter, and which has been degassed (see Note 3) 2.3 DNA and Drug Characterization As with all thermodynamtc studtes, it is extremely important to work with well-characterized systemsbecause all heat effects will be observed m the calorimeter For some studies, a random sequence DNA such as calf thymus DNA can be used, but the AH calculated from the heat effects is the average for all possible basepair sequences for a binding site Most studies today are being performed with specific polymers, e.g., poly(dAdT), and some studies are possible with ohgomers Complimentary strands that form duplexes have the advantage that dissolving these m a buffer with the appropriate salt concentration produces primartly duplexes Any change m DNA structures durmg the calorimetrtc titration may produce heat effects that cannot be easily separated from the drug-DNA interaction heats Methods 3.1 Preparation of Samples In order for one to calorimetric studies successfully on drug-DNA mteracttons, the composition of the buffer-salt solution containing the DNA must be exactly the same as the solutton contammg the drug One way to this is the classical method of dtalysts of the DNA solution against the buffer-salt solution used m the experiments There are several disadvantages to this procedure: Small volumes are involved, making it difficult to perform the dialysis without losing some solutton; the DNA samples can stick to the dialyses materials and oligomers will pass through most dialysis membranes A much easier procedure involves a lyophihzation of both the DNA and drug sample before dtssolvmg these samplesm the buffer-salt solutton employed in the calorimeter If the DNA sample has been purchased as a solid salt material, thts can be dissolved m water to form a stock solutton A stock solutton for the drug can also be prepared m pure water (see Subheading 1.2 for estimating the required drug and DNA concentrattons) Determine the concentratton of each of the stock solutions and calculate how many microliters of each is neededto prepare the calortmetric solutions (see Note 6) Transfer each solutton mto a vial of suffictent volume to hold the final solution Place these vials mto a tube that can be attached to a vacuum system, attach the tube to the vacuum system, freeze the soluttons wtth hqutd nitrogen, 266 Hopkins and evacuate After all water has been removed, add the required amount of buffer/salt solution to each vial Wtth this procedure, the two solutions that are to be mixed m the calortmeter have the same buffer and salt composition, and any heat effects associated wtth changes m salt concentrattons will be mmtmum This 1sextremely tmportant because the heats accompanying changes m salt, DNA, buffer, and drug concentrations can be large relative to the heats for the drug-DNA mteractions (see Note 7) If the concentrations of all these components are low, the effects are mmimal When the [NaCl] or concentration of other salts are above 0.1 A4, small changes m the concentration can produce peal levels of heat (see Note 4) 3.2 Calorimeter Operation All calorimeters are dehcate mstruments, so be prepared to allocate some space for it where there is little passage of people, minimal airflow, and small temperature variations Allow the calorimeter to be operational for several days before starting a study with a valuable sample Measure tts heat capacity many times durmg this period; this will probably be called a calibration procedure by the manufacturer Also remember that a stable baselme IS required and, if the baseline is not stable for a long period before startmg, the titration will probably fail to produce acceptable data (see Note 3) 3.3 Heats of Dilution for the Drug Before attempting a titration of the drug mto the DNA sample, determine the heat of dilution associated with the change in concentration accompanying the injection of the drug into the DNA solution Place the drug solutton mto the syringe used to inject rt mto the sample solution, load the sample compartment with a solutton contammg only the buffer-salt solution, watt unttl the baseline is level, and inject a volume of the drug solution mto the sample cell In some casesheat will be below the detection hmits of the mstrument, and one will not observe a heat pulse for the mjection If a flow system is bemg used, also determme the heat of dtlutton for the DNA solution because the DNA system is also being diluted by a factor of one-half 3.4 Calorimetric Reaction Place the DNA and drug solutions prepared as described m Subheading 3.1 m the sample and reference compartments (see Note 5) At least two of the commercial calorimeters have stirrers m the sample cells and these must be activated properly Normally one should wait approx h for the calorimeter to come to a steady state and the baselme to stabilize If a computer is mcorporated mto the measurement circuit, a numerical value can be used to decide when the instrument is ready for the titration to begin Make a small injection Calorimetric Techniques 267 to ensure that the calorimeter is functioning and that mixing at the end of the syringe has not caused changes in concentrations Now start the injection of the desired volume, or for the stopped-flow instrument a series of mixing events If sufficiently large heats are observed, continue the procedure In order to mcrease the relative preciston when the AH values are low in magnitude, the size of the injection volume must be increased or the concentrations must be increased accordingly Notes The chelation agent EDTA is added to the buffer solutions at mA4 m order to bmd divalent cations that might affect the DNA lattice structures If any microorgamsms enter the calorimeter cells, these can produce heat effects that vary with time When this happens the baseline on the instrument will not be stable A thorough cleanmg of the cells with a commercial detergent must be performed to remove all living matter m the solutions in the cell and possibly on the walls Using a buffer made from cacodyhc acid can mmimize this problem, but it is toxic and must be used with caution Tmy an bubbles m the calorimeter cell ~111cause the baseline to fluctuate and all solutions should be routinely degassed with a small vacuum desiccator before bemg placed m the syringes and cells of the calorimeter If an an bubble is present in a syringe, then an erroneous heat will be observed sometime during the experiment At high [NaCl] the bmdmg constants for the drug-DNA mteractions decrease dramatically (3) compared to the [NaCl] = 015 used m experiments used to construct Figs and 3; thus, a substantial fraction of the drug molecules added may not bmd to a DNA site Recent experience (1,3,9) with ethidmm and propidmm binding studies m the calorimeter have demonstrated that the observed AH for drug-DNA interactions can depend on the salt concentration, DNA concentration, and temperature Making measurements at only one set of conditions may provide misleadmg information If the drug IS not sufficiently soluble in water to perform the studies, then ethanol can be added to increase the solubility Up to 20 mass percent ethanol m an aqueous solution did not effect the AH for dissociation of poly(dA) * poly(dT) and poly(dAdT), but did lower the meltmg temperatures (11) for these duplexes Protonation of the hydroxide ion (6) (addition of a small quantity of HCl into excess NaOH) or the standard base THAM (1) (addition of a small quantity of HCl into excess THAM that is 50% protonated) has been used because both reactions have large eqmhbrmm constants and substantially negative AH values Careful preparation of the solutions are required when using either of the aforementioned processes for producing a known heat m the calorimeter cell A simpler procedure involves dilution (5,lU) of a solution of NaCl, sucrose, or HCl m the calorimeter 268 Hopkins References Hopkins, H P , Jr, Fumero, J , and Wilson, W D (1990) Temperature dependence of enthalpy changes for ethldtum and propldmm bmdmg to DNA: Effect of alkylamme chains Btopolymers 29,449-459 Hopkms, H P , Jr., Mmg, Y , Wilson, W D , and Boykm, D W (1991) Intercalation binding of 6-substituted naphthothlopheneamldes to DNA* enthalpy and entropy components Bzopolymers 31, 115&l 114 Marky, L A and Macgregor, R B (1990) Hydration of dA*dT polymers role of water m the thermodynamics @f ethldmm and propldmm mtercalatlon Bzochemlstry 29,4805-48 1 Wiseman, T , Brandts, J F , dllhston, S , and Lm, L N (1989) Rapid measurement of binding constants and heats of bindmg using a new tltratlon calorimeter Anal Biochem 179, 13 1-137 Remeta, D P., Mudd, C P , Berger, R L , and Breslauer, K J (199 1) Thermodynamic characterization of daunomycin-DNA interactions mlcrocalorlmetrlc measurements of daunomycm-DNA bmdmg enthalples Bzochemwtry 30,9799-9807 Vlckers, L P , Hopkins, H P., Jr., Ah, S Z , and Carey, V (1984) Error analysis m titration mlcrocalonmetry of biochemical systems Anal Blochem 145,257-265 Marky, L A , Blumenfeld, K S , and Breslauer K J (1983) Calorlmetrlc and spectroscoptc mvestlgatlon of drug-DNA mteractlons I Bmdmg of netropsm to poly d(AT) Nucleic Acrd,v Res 11,2857-2870 Hopkins, H P , Jr , Stevenson, K A., and Wilson, W D (1986) Enthalpy and entropy changes for the intercalation of small molecules to DNA I Substituted naphthalene monolmldes and naphthalene dlimldes J Sol Chem 15, 563-579 Morgan, W B and Hopkins, H P , Jr (1995) Calorimetric studies on the mteractlon of ethldmm and propldmm with duplex and triple helix structures formed with poly(dA), poly(dT) and poly(dAdT) Unpublished data from M S Thesis at Georgia State University 10 Pllch, D S , Klrolos, M A Llu, X Plum, G E , and Breslauer, K J (1995) Bereml [ 3-Bis(4’-amldmophenyll)tnazene] binding to DNA duplexes and to a RNA duplex evidence for both mtercalatlve and minor groove bmdmg properties Blochemlstry 34,9962-9976 11 Hopkins, H P , Jr, Hamilton, D D , Wilson, W D , Campbell, J , and Fumero, J (1993) Effect of C2HSOH, Na+(aq), N(CH$H#(aq) and Mg2+(aq) on the thermodynamics of double-helix-to-random-coil transitions ofpoly(dA)*poly(dT) and poly(dAdT) J Chem Therm 25, 111-126 Methods for the Studies of Drug Dissociation from DNA Fu-Ming Chen Introduction In addition to bmdmg affinity, the on and off rates of drug-DNA mteractions are important in determining the biological activities of a drug For example, the rate of dissociation of a drug from DNA has been shown to be related to its pharmacological activities (1) Various techniques have been employed to study the dissociation kinetics of drugs from DNA These include detergent sequestration technique pioneered by Muller and Crothers (I), a modification of the footprinting technique for examming the dissoclatlon of ligands from mdlvldual binding sites (2), relaxation methods such as T-Jump for measuring fast kinetics (3), and a procedure that can yield drug-DNA dissociation kinetics under conditions of active transcrlptlon of the DNA (4) The simplest and most widely used method IS detergent-induced dissociation rate measurement incorporating the detergent sodium dodecyl sulfate (SDS) This chapter will thus focus on the SDS-sequestration technique and include only conventional spectrophotometric and stopped-flow methods An example of an actinomycin D dissociation measurement from the author’s laboratory will be used to illustrate the methodology Materials Appropriate buffer for the system of interest For example, a buffer of pH 2X that contains (l-10 mM) Mg*+ wdl be reqmred for the studies of chromomycm A, and mithramycm A 20% SDS solution can be prepared by dlssolvmg 20 g SDS m 80 mL of buffer solutlon (see Note 7) From Methods m Molecular Edlted by Bfology, K R Fox Vol 90 Drug-DNA Humana 269 Press Interact/on Inc , Totowa NJ Protocols Chen 270 Method 3.1 Non-Stopped-Flow Technique A drug-DNA solutton mtxture IS either obtained as an end product of the association kmettc measurement or prepared by mrxmg together appropriate amounts of drug and DNA solutions The solutton mrxture IS usually allowed to reach equtlibrmm The waiting time depends on the rate of assoctatron A time period of 25 x z, (the characterlstrc assocratton time) should be sufficrent (see Note 4) Record the mmal value A,, where A represents any measurable physical properties such as absorbance, fluorescence, or elhptrcrty at the wavelength of Interest (see Note 3) The dtssoctatton of the drug IS imtiated by the addition of an appropriate volume of 20% SDS to the DNA&ug mrxture to result in a 1% final SDS concentratron (see Note 5) The solutron 1sthen thoroughly mrxed by either rtgorous manual shaking or mechamcal stnrmg (see Notes and 2) Data collectron should commence as soon as rt IS feasible, via computer or chart recorder The run IS terminated when reasonable A, can be estimated, usually with t > 2-3 rd (charactertsttc dtssociatton time) 3.2 Stopped-Flow Technique Prepare a 2% SDS solution vta a lo-fold dilution from the 20% stock F111the two reservoir syringes with the drug-DNA and 2% SDS solutions, respectively Carefully fill the driving syringes (see Note 8) and wait until the temperature reaches eqmlibrmm (see Note 6) Actuate the plungers via pressured gas and commence the data collection The run 1s terminated when the decay curve shows sign of leveling 3.3 Data Analysis 3.1 Manual Graphical Method (see Fig 1) A, at the monitoring wavelength 1sfirst estimated or obtained experimentally by wattmg until there 1sno longer any change m the value of A Values of L4 E IA, - AmI are then calculated and plotted vs time on a semtlogrithmic graph paper, where A, 1sthe value of A at time t A reasonable value of A, will yield a straight lme plot for a single-exponenttal rate process and the rate constant 1sthen obtained by -2 303 x (slope) Extrapolation to t = will yield A,4,, the total measurable change of A as a result of this process The percentage contrtbutton of this process can then be obtained by the followmg formula 100 x A& / 1A, - A,(dtlution corrected)1 For a multiexponential process, a curved plot with a straight line portion at the longtime data region will result The slowest rate constant k( 1) and the measur- 271 Drug Dissociation from DNA a -24 d, $ A -2 0.3 10 15 20 25 TIME (set) Fig An example tllustratmg how kinetic parameters can be extracted via graphtcal method for a multtexponenttal kinetic profile Estimate a reasonable A, Calculate AA = IA &I Plot A.4 vs t on a semtlogrtthmtc graph paper (Note: The vertical scale 1s a lmear scale of log A.4 and not the logrtthmtc scale of a semtlog graph paper ) Repeat steps l-3 untd a linear plot IS obtained for the long-ttme region (s) A best stratght lme 1s drawn through these long-time data points and extended to t = (connected line) to obtam k and aA, for this process AA value correspondmg to each experimental time point IS read directly from the stratght lme and subtracted from the experimental value to obtain new AA These new AA values are replotted on the semilog paper (open squares) Repeat steps 5-7 with this new data set The process 1s contmued until the new plot IS a straight lme without the presence of a curvature at the short-time region able total change associated with this process AI!,, are then obtamed from the slope of the straight lme and its intercept at zero time, respectrvely The values of the straight lme at each time-point are read dtrectly from the graph and then subtracted point-by-point from the ortgmal A,4 These new values are then replotted to obtain k(2) and AA,, and the process continues unttl the last straight line plot IS obtained 3.3.2 Nonlinear Least-Squares Curve Fit (see Figs and 3) The kmettc data can be fitted dnectly with any commerctally available nonlmear least-squares program The equations to be used are AA = aA,e-kt + B for a smgleexponenttal and AA = AA,,e- k(l)t + M02e-k(2)t + B for a double-exponenttal process, and so forth 272 Chen 00425 , I d TIME (WC) Fig (A) Compartson of experimental data and a smgle-exponential fit (B) The corresponding residual plot It is apparent that the kmettcs cannot be adequately described by a single exponential process The kmettc parameters k’s and A&‘s are obtamed directly from the fit The goodness of the fit can be appraised by visual comparison of the experimental data with the fitted curve, the value of sum of square deviation, or the restdual plot Since a higher order exponential mode1 will m general result m a better fit because of the larger number of parameters, use of a higher order mode1 may not be warranted unless significant improvement, such as several-fold reduction m the sum of square deviation, IS obtained Drug Dlssoclatlon from DNA 002 d YE % 273 L B 000 K I -0 yj 50 TIME (set) Ftg (A) Comparison of expertmental data and a two-exponenttal nonlinear leastsquares fit The extracted parameters are k( 1) = 0.038 + 0.00 16 ss’, A,4,, = 0064 f 0.00005; k(2) = 0.444 iT 015 s-1, Aid02 = 0.00548 * 0.00009 (B) The correspondmg residual plot Judgmg the goodness of fit IS best done by restdual plot, since the vtsual comparison of the experrmental data and the fitted curve can sometrmes gave an erroneous Impressron of a good fit 3.3.3 Global Analysis If a serves of spectra can be measured during a kinetic run, kinetic profiles at different wavelengths can be obtained and analyzed after the experiment 274 Chen Furthermore, global analysis using the data at every wavelength can be performed which can sometimes help eliminate or suggest certain mechanistic models Notes A gentle mversion of the cuvet after rigorous manual shaking may help mmtmize the numbers of tmy au bubbles sticking on the cell walls The use of a mechanical stirrer is preferable as it shortens the dead time, mmimizes the bubble generation, and assures contmuous uniform mixing during the course of a kinetic run Absorbance momtormg should preferably be at the wavelength that corresponds to the isosbestic point of free and SDS-sequestered drug spectra so that the measured intensity changes reflect the drug dtssocration from DNA more accurately This can be determined via spectral titrations of drng-vs-stock SDS solutions Smce reaction kinetics are temperature sensitive, maintaining a constant temperature during the run is essential Although the 1% SDS strength is usually sufficient for most purposes, it may be a good idea to experimentally confirm it for a particular system of interest Measurements should not be made below 15°C as SDS forms precipitates near or below this temperature SDS powders are extremely fine and can easily get mto the nasal passages to cause irritation Thus, SDS should be handled very gently during weighing and the use of a nose-mask is strongly recommended Careful and slow tillmg of the driving syringes can help minimize the bubble formation in the stopped-flow experiment References Muller, W and Crothers, D M (1968) Studies of the Binding of Actmomycm and Related Compounds to DNA J MOE Bzol 35,25 l-290 Fletcher, M C and Fox, K R (1993) Visuahsmg the Kinetics of Dissociation of Actmomycm from Individual Sites m Mixed Sequence DNA by Dnase I Footprmtmg Nuclezc Aczds Res 21, 1339-1344 Chaires, J B , Dattagupta, N , and Crothers, D M (1985) Kinetics of the Daunomycm-DNA Interaction Bzochemzstry 24,26&267 Phillips, D R and Crothers, D M (1986) Kmetics and Sequence Specificity of Drug-DNA Interactions An in Vitro Transcription Assay Bzochemzstry 25,73557362 ... cleavage agents, both enzymrc and chemical, have also been used as footprinting probes for drugDNA interactions including micrococcal nuclease (24), DNase II (6,15), copper phenanthrolme (16,17),... (24-26) Each of these has a different cleavage mechanism, revealmg different aspects of drug-DNA interactions An ideal footprmtmg agent should be sequence neutral and generate an even ladder of... sclssile phosphodlester bond, However, the crystal structures show that there may be other specific interactions between the exposed loop and DNA bases removed from the cutting site In particular,