dna protein cross linking

ANALYTICALBIOCHEMISTRY

Analytical Biochemistry 344 (2005) 204–215

www.elsevier.com/locate/yabio

A method for the isolation of covalent DNA–protein crosslinks suitable for proteomics analysis

Sharon Barker a, David Murray a, Jing Zheng b, Liang Li b, Michael Weinfeld a,¤

a Department of Experimental Oncology, Cross Cancer Institute, Edmonton, Alberta, Canada T6G 1Z2b Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2

Received 12 February 2005Available online 20 July 2005

Abstract

The covalent crosslinking of protein to DNA is a form of DNA damage induced by a number of commonly encountered agents,including metals, aldehydes, and radiation as well as chemotherapeutic drugs. DNA–protein crosslinks (DPCs) are potentially bulkyand helix distorting and have the potential to block the progression of translocating protein complexes. To fully understand theinduction and repair of these lesions, it will be important to identify the crosslinked proteins involved. To take advantage of dramaticimprovements in instrument sensitivity that have facilitated the identiWcation of proteins by proteomic approaches, improved meth-ods are required for isolation of DPCs. This article describes a novel method for the isolation of DPCs from mammalian cells thatuses chaotropic agents to isolate genomic DNA and stringently remove noncrosslinked proteins followed by DNase I digestion torelease covalently crosslinked proteins. This method generates high-quality protein samples in suYcient quantities for analysis bymass spectrometry. In addition, the article presents a modiWed form of this method that also makes use of chaotropic agents for pro-moting the adsorption of DNA (with crosslinked proteins) to silica Wnes, markedly reducing the DPC isolation time and cost. Theseapproaches were applied to radiation- and camptothecin-induced DPCs. 2005 Elsevier Inc. All rights reserved.

Keywords: Ionizing radiation; Covalent; DNA–protein crosslink; Proteomics; GRP78; DNA topoisomerase I

A DNA–protein crosslink (DPC)1 is created when aprotein becomes covalently bound to DNA. Theselesions are induced by UV and ionizing radiation, bymetals and metalloids (e.g., chromium, nickel, arsenic),and by various aldehydes and anticancer drugs [1]. It has

* Corresponding author. Fax: +1 780 432 8428.E-mail address: [email protected] (M. Weinfeld).

1 Abbreviations used: DPC, DNA–protein crosslink; SDS/K+, sodiumdodecyl sulfate/potassium; 2-D SDS–PAGE, two-dimensional poly-acrylamide gel electrophoresis; CHO, Chinese hamster ovary; PBS,phosphate-buVered saline; DMSO, dimethyl sulfoxide; SDS, sodiumdodecyl sulfate; EDTA, ethylenediamine tetraacetic acid; EGTA,ethyleneglycol-bis(�-aminoethylether)-N,N,N�,N�-tetraacetic acid; DTT,dithiothreitol; PMSF, phenylmethylsulfonyl Xuoride; 1-D, one-dimen-sional; TOF, time-of-Xight; MALDI, matrix-assisted laser desorptionionization; MS/MS, tandem mass spectrometry; GRP78, glucose-regu-lated protein 78.

0003-2697/$ - see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.ab.2005.06.039

been suggested that in mammalian cells, cellular stresses(e.g., illness, exposure to drugs, radiation, pollutants)result in the accumulation of diVerent types of DNAdamage, including DPCs, due to oxidative mechanisms[2]. There are numerous chemically distinct types ofDPCs; indeed, proteins can become crosslinked to DNAdirectly through oxidative free radical mechanisms orindirectly through aldehydes generated by oxidativestress, or they can be crosslinked through a chemical ordrug linker or through coordination with a metal atom[3]. These chemically distinct DPCs may also diVer intheir biological consequences, depending on their struc-ture and persistence in the genome. Gross DPC half-liveshave been measured in vitro and in vivo in mammaliancells and range from hours to days, depending on thesystem and agent being studied [4–8].

mailto: [email protected]

mailto: [email protected]

Method for the isolation of covalent DNA–protein crosslinks / S. Barker et al. / Anal. Biochem. 344 (2005) 204–215 205

Determining the biological relevance of DPCs is a com-plicated task. The covalent crosslinking of proteins toDNA is expected to physically block the access/assemblyor progression of replication, repair, recombination, ortranscription complexes. The induction of DPCs has beenshown to correlate with the incidence of genetic damagesuch as sister chromatid exchanges, transformation, andcytotoxicity [9–13], although the contribution of speciWcDPCs to these events remains to be determined. EVorts toelucidate the biological consequences of DPCs are con-founded by several factors, including the simultaneousinduction of other classes of lesions by DPC-inducingagents. DPCs, therefore, are inevitably induced in a back-ground of multiple types of damage, and ascribing partic-ular consequences to one type of damage is not yetpossible. A second complication is the background oftightly, but noncovalently, bound proteins. Methods thatwould permit the separation and study of genuinely cova-lently bound proteins would greatly facilitate this eVort.

Early studies of DPCs tended to focus on whether cel-lular proteins became associated with DNA followingexposure of a test system to a given genotoxic agent and,if so, to what extent. With the advent of high-throughputproteomics methodologies, the emphasis has shifted tothe possibility of recovering and identifying the proteinsthat become covalently linked to DNA. The latter studieswill, however, require methodologies that recover theDNA component and those (rare) covalently bound pro-teins that are extracted along with the DNA. The morecommonly used DPC investigation methods, such asnitrocellulose Wlter binding [14–16] and sodium dodecylsulfate/potassium (SDS/K+) precipitation [17,18], quanti-tate DPCs as the amount of DNA isolated when proteinsare trapped and, therefore, will not be informative for theisolation and study of speciWc crosslinked proteins with-out extensive modiWcation. A DPC isolation method thatisolates proteins by virtue of their association with DNAshould provide much cleaner DPC samples with respectto noncovalently associated proteins.

The stringency of isolating covalently bound proteinshas been part of the problem in assessing the biologicalrelevance of DPCs to date. For example, it is known thatnuclear matrix proteins are tightly associated with theDNA [19]; their complete dissociation, therefore, is cru-cial for the identiWcation of those proteins that are cova-lently crosslinked to DNA by a given agent. Previousstudies [20–22] have isolated cisplatin-crosslinked pro-teins and nuclear matrix fractions from mammalian cellsand have shown by two-dimensional polyacrylamide gelelectrophoresis (2-D SDS–PAGE) that the majority ofthe crosslinked proteins are present in the nuclear matrixfraction. However, this method involves binding ofDNA/DPCs to hydroxylapatite, which is also capable ofbinding noncrosslinked proteins.

Applying proteomic approaches to the study of DPCsrequires the development of novel methods that allow

the isolation of the proteins covalently crosslinked toDNA as a pure sample and in suYcient quantities forfurther analysis and detection. To this end, we havedeveloped two protocols to recover proteins covalentlybound to DNA. Both involve isolation of total genomicDNA using a commercial chaotrope/detergent mix(DNAzol) that lyses cells, hydrolyzes RNA, and dissoci-ates noncovalent protein–DNA complexes. In the Wrstmethod, the DNAzol–strip method (Fig. 1B), DNAzoltreatment is followed by salt washes to strip noncova-lently bound proteins from the DNA. In the DNAzol–silica method (Fig. 1C), the genomic DNA is adsorbedonto silica in the presence of a chaotrope (DNAzol, urea,and sodium chloride) under alkaline conditions toremove associated proteins. These DNA isolation meth-ods were followed by additional steps to allow the recov-ery of truly covalently crosslinked proteins.

Materials and methods

Cell culture

The Chinese hamster ovary (CHO) cell line, AA8, wasmaintained as a monolayer culture in �DMEM-F12medium (Invitrogen) with 10% fetal bovine serum (Invit-rogen) and 5% penicillin/streptomycin in a humidiWed5% CO2/95% air atmosphere at 37 °C.

Radiation and chemical treatments

Cells were grown to approximately 85% conXuency.For gamma radiation treatment, cells were irradiated ina 60Co irradiator (Gammacell 220, Atomic Energy ofCanada) with doses of 0–4 Gy. For formaldehyde treat-ment, 37% formaldehyde (Sigma) was added to themedium to a Wnal concentration of 1% and the samplewas incubated at 37 °C for 1 h. For topoisomerase Iinhibitor treatment, cells were washed with phosphate-buVered saline (PBS) and transferred to serum-freemedium (10 ml). The cultures received either 10�l ofdimethyl sulfoxide (DMSO) or 10�g/ml camptothecin(Sigma) in 10�l of DMSO and were incubated for 1.5 hat 37 °C. For proteasome inhibitor treatment, AA8 cellswere treated with 10�M MG132 (Cedarlane) in 10 mlmedium for 3 h at 37 °C. After 3 h, the medium wasreplaced with serum-free medium and both proteasomeinhibitor (to a Wnal concentration of 10 �M) and campto-thecin (to a Wnal concentration of 10�g/ml) were addedas above and the cells were incubated at 37 °C for 1.5 h.

DNAzol DPC isolation method

With this method (Fig. 1A), after treatment, the cul-ture medium was removed and the cells were washed onthe tissue culture dish with ice-cold PBS. Cells (or nuclei

206 Method for the isolation of covalent DNA–protein crosslinks / S. Barker et al. / Anal. Biochem. 344 (2005) 204–215

in later experiments) were lysed by the addition of 500�lDNAzol (Invitrogen) per 7 £ 107 cells. DNA was precip-itated from each sample using 0.5 volume of ice-cold99% ethanol. The pellets were resuspended in 8 mMNaOH (3 ml per 9 £ 106 cells) overnight at 37 °C with aprotease inhibitor mixture (Sigma). For DNA diges-tion, the pH of each DPC sample was adjusted to 5.5by the addition of 0.1 M sodium acetate, and MgCl2and ZnCl2 both were added to a Wnal concentration of10 mM. One ml of 5 £ digestion buVer (50 mM MgCl2,50 mM ZnCl2, 0.5 M sodium acetate, pH 5.0) wasadded to each sample, and the samples were digestedfor 1 h at 37 °C with 5 U of DNase I and 5 U of S1nuclease. After digestion, the DNA concentration wasdetermined by UV absorbance and the samples wereconcentrated to 1 ml using Centricon concentratorswith a molecular weight cutoV of 5000 Da (Millipore).Samples were then reduced to dryness by vacuumcentrifugation.

DNAzol–strip DPC isolation method

With this method (Fig. 1B), nuclei were isolated asdescribed below. Isolated nuclei were lysed by the addi-tion of 500 �l DNAzol per 7 £ 107 nuclei. DNA wasprecipitated from each sample using 0.5 volume of ice-cold 99% ethanol. The pellets were air-dried brieXy and

resuspended in 8 mM NaOH (3 ml per 9 £ 106 cells) at37 °C. An equal volume of 5 M urea was added, and thesamples were incubated at 37 °C for 30 min on a rotatingshaker. Sodium dodecyl sulfate (SDS, 10%) was addedto a Wnal concentration of 2%, and the samples wereincubated as above. The solute level was reduced usingCentricon concentrators with a cutoV of 3000 Da. Whenthe volume had been reduced to approximately 5 ml, anequal volume of 5 M NaCl was added. Samples weremixed at 37 °C for 30 min on a rotating shaker and thenwere Wltered and washed with distilled deionized waterthree times, using Centricon concentrators with a cutoV

of 3000 Da to reduce the volume and the salt concentra-tion. The DNA from each sample was then reprecipi-tated by the addition of 0.1 volume of 3 M sodiumacetate and 3 volumes of ice-cold 99% ethanol. Precipi-tated DNA was collected by centrifugation at 200g at4 °C for 30 min and dried. The DNA was dissolved in8 mM NaOH (3 ml per 9 £ 106 cells). For DNA diges-tion, the pH of each DPC sample was adjusted to 5.5 bythe addition of 0.1 M sodium acetate, and MgCl2 andZnCl2 both were added to a Wnal concentration of10 mM. Digestion buVer (5£, 1 ml of 50 mM MgCl2,50 mM ZnCl2, 0.5 M sodium acetate, pH 5.0) was addedto each sample, and the samples were digested for 1 h at37 °C with 5 U of DNase I and 5 U of S1 nuclease. Afterdigestion, DNA concentration was determined by UV

Fig. 1. DNAzol-based DPC isolation methods. The schematic representation shows the steps involved in the isolation and analysis of DPCs using theDNAzol method (A), the DNAzol–strip method (B), and the DNAzol–silica method (C). Proteins are represented by shaded circles/ovals.


absorbance and the samples were concentrated to 1 mlusing Centricon concentrators with a cutoV of 5000 Da.Samples were then reduced to dryness by vacuum centri-fugation.

DNAzol–silica DPC isolation method

With this method (Fig. 1C), silica Wnes were activatedas detailed elsewhere [23]. BrieXy, silica Wnes (EM Sci-ence) were heated to near boiling in 5 M nitric acid,washed three times in distilled deionized water, andresuspended in an equal volume of distilled deionizedwater. The pH of the solution was adjusted to 7.0 using1 M Tris–HCl (pH 8.0), and the silica Wnes were sedi-mented, resuspended in an equal volume of distilleddeionized water, and autoclaved. After lysing the nucleiwith DNAzol as described above, 2 ml of prewarmed(65 °C) 10 mM Tris–HCl (pH 7.0) was added and eachsample was drawn through a 21-gauge needle threetimes, and then through a 25-gauge needle three times, toshear the DNA. NaCl (5 M) was added to a Wnal concen-tration of 4 M, and this mixture was incubated at 37 °Cwith shaking for 20 min. Urea (8 M) was added to a Wnalconcentration of 4 M, and the samples were incubated asabove. An equal volume of 99% ethanol was added toeach sample. The activated silica slurry was then added(1 ml per 7 £ 107 cells), and the samples were gentlyrocked for 20 min at room temperature to allow forbinding. The silica was collected by centrifugation for4 min at 35g, and the supernatant was discarded. The sil-ica was washed three times in 50% ethanol and collectedby gentle centrifugation each time. The DNA was elutedtwo times using 2 ml of 8 mM NaOH at 65 °C for 5 min,and eluates were combined. For DNA digestion, 1 ml of5£ digestion buVer (50 mM MgCl2, 50 mM ZnCl2, 0.5 Msodium acetate, pH 5.0) was added to each sample andthe samples were digested for 1 h at 37 °C with 5 U ofDNase I and 5 U of S1 nuclease. After digestion, DNAconcentrations were determined by UV absorbance andthe samples were concentrated to 1 ml using Centriconconcentrators with a cutoV of 5000 Da. Samples werethen reduced to dryness by vacuum centrifugation.

SDS/K+ DPC isolation method

We also employed the method of Zhitkovitch andCosta [18] to isolate DPCs. Nuclei were lysed by theaddition of 0.25 volume of 4% SDS in 20 mM Tris–HCl(pH 7.4), followed by heating at 65 °C for 10 min toallow complete binding of SDS to proteins. The SDSand protein-bound SDS were then precipitated by theaddition of an equal volume of 200 mM KCl in 20 mMTris–HCl (pH 7.4) and incubation on ice for 20 min. Pre-cipitated proteins and protein–DNA complexes werecollected by centrifugation at 12,000g at 4 °C for 10 min.The supernatant was discarded and the pellet was

resuspended in 8 mM NaOH (3 ml per 9 £ 106 cells)overnight at 37 °C.

Nuclei isolation

Cultures were trypsinized at room temperature for3 min and collected by centrifugation at 200g at 4 °C for5 min. Cells were washed in ice-cold PBS and collected asbefore. The cell pellet was gently resuspended in buVer 1(400 �l per 107 cells) using a wide-bore pipette tip (buVer1: 10 mM Hepes (pH 7.9), 10 mM KCl, 100 mM ethylene-diamine tetraacetic acid (EDTA), 100 mM ethylenegly-col-bis(�-aminoethylether)-N,N,N�,N�-tetraacetic acid(EGTA), 1 mM dithiothreitol (DTT), 0.5 mM phenyl-methylsulfonyl Xuoride (PMSF), 1% (v/v) aprotinin).Cells were chilled on ice for 15 min and then lysed by theaddition of 0.6% (v/v) Nonidet P-40 and mixing byinversion. Nuclei were pelleted at 200g for 5 min at 4 °C,and the supernatant was removed.

Nuclear extract preparation

Nuclear extracts of CHO AA8 cells were prepared forcontrol purposes. After nuclei isolation (above), the pel-let was resuspended gently in ice-cold buVer 2 (100 �l per107 cells) using a wide-bore pipette tip (buVer 2: 20 mMHepes (pH 7.9), 400 mM NaCl, 1 mM EDTA, 1 mMEGTA, 1 mM DTT, 1 mM PMSF, 1% (v/v) aprotinin,10% (v/v) glycerol) and incubated, with shaking, at 4 °Cfor 30 min. The lysate was centrifuged at 12,000g for10 min at 4 °C. The supernatant was aliquoted into ice-cold 1.5-ml microcentrifuge tubes supplemented with0.025 mg/ml leupeptin, and aliquots were Xash-frozen inliquid nitrogen and stored at ¡80 °C.

Quantitation of DNA

The UV absorbance at 260nm was measured for eachsample to determine the DNA concentration. A value of32�g (oligonucleotide) per 1 OD unit was used to calculatethe amount of DNA in each sample. The relative amountsof DNA were determined within each experiment and wereused to determine sample loads for SDS–PAGE analysis.The 260/280-nm absorbance ratios were also determined.Ratios of 1.5–1.7 were invariably obtained, indicating thatthe contribution of protein to the 260-nm reading was notsigniWcant. For optimization experiments, the amount ofDNA was also assessed by 1% agarose gel electrophoresisand ethidium bromide staining.

Quantitation of protein

Protein content of DPC isolates was determined usingthe Bradford reagent (Bio-Rad) and standard Bradfordassay procedure with bovine serum albumin as astandard.


SDS–PAGE analysis

Laemmli buVer (Bio-Rad) was added to each sample inamounts determined to equalize the DNA concentrationof each sample. In later experiments, dried protein sam-ples were dissolved in 20�l of Laemmli buVer and sampleloads were determined to equalize the DNA concentra-tion of each sample. Samples were analyzed by one-dimensional (1-D) SDS–PAGE using 12% separating gels(180£160£0.75 mm for mass spectrometry analysis or80£60£0.75mm for standard protein analysis) or 10–20% gradient gels (80£60£0.75 mm, Bio-Rad) in laterexperiments. In separate experiments, gels were stainedusing either the ammoniacal silver nitrate staining proce-dure, the standard Coomassie blue staining and destainingprocedures, or the SYPRO Tangerine (Invitrogen) stain-ing procedure according to the manufacturer’s protocol.

Mass spectrometry

Samples were analyzed at the Alberta Cancer BoardProteomics Facility (Department of Chemistry, Univer-sity of Alberta), where they were subjected to digestionwith trypsin. Peptide extracts were analyzed on a ReXexIII (serial no. FM 2413, Bruker) time-of-Xight (TOF)mass spectrometer using matrix-assisted laser desorptionionization (MALDI) in positive ion mode. The peptidemaps obtained were used for database searching to iden-tify proteins. Furthermore, selected peptides were frag-mented using MALDI tandem mass spectrometry (MS/MS) analysis with a PE Sciex API-QSTAR Pulsarinstrument (serial no. K0940105, MDS-Sciex). Theobtained partial sequence information for each peptidewas used to conWrm the previously obtained results fromthe peptide map search.

Results

DPC isolation by DNAzol

DPCs can be detected using the alkaline elution assay[24,25]. In the Wrst steps of this method, cells are lysed ona polycarbonate Wlter that traps DNA based on its highmolecular weight [26]. Repeated washing causes smallerfragments of DNA to be lost along with free proteins.Large fragments of DNA and any covalently bound pro-teins are trapped on the Wlter. We initially evaluated theutility of the polycarbonate Wlter trapping method forprotein recovery because these Wlters do not stronglybind DNA or protein and, therefore, might provide thestringency necessary for the isolation of pure DPCs.However, we found that this method resulted in poorprotein recovery and poor reproducibility (data notshown); therefore, we sought to develop an alternativeprocedure.

Genomic DNA isolation kits currently available arenot useful for DPC isolation because most of theminclude a proteinase K digestion step (which will destroythe crosslinked proteins) and because they are not ame-nable to scaling up to the level necessary to isolate suY-cient quantities of DPCs for protein identiWcationpurposes. One currently available genomic DNA isola-tion reagent is DNAzol, a proprietary reagent (US pat-ent No. 5,945,515) that contains a guanidine salt anddetergent in alkali conditions. This reagent lyses cells,dissociates proteins, and hydrolyzes RNA. The DNA isprecipitated by the addition of ethanol (DNAzolmethod, Fig. 1A). The DNAzol reagent does not containproteinases and is not overtly damaging to proteins(Fig. 2). No obvious degradation of proteins was seenafter incubation of AA8 nuclear extract with DNAzol(5 min at room temperature) and analysis of the proteinsby SDS–PAGE and Coomassie blue staining (Fig. 2,lane 2). In contrast, the complete degradation of proteinswas apparent following incubation of AA8 nuclearextract with proteinase K (Fig. 2, lane 3), as expected.

Although the DNAzol reagent contains detergent andthe chaotropic agent guanidine hydrochloride, the isola-tion of DNA from untreated AA8 cells using the stan-dard DNAzol method does not fully dissociate proteins

Fig. 2. EVect of DNAzol on protein integrity. AA8 nuclear extract wasincubated with DNAzol for 5 min at room temperature and analyzedby SDS–PAGE. For comparison, an equal amount of AA8 nuclearextract was incubated with proteinase K for 5 min at room tempera-ture to fully digest the proteins. For reference, both the proteinase Kreagent and the AA8 nuclear extract were also run on their own. Pro-tein molecular weights are indicated in kilodaltons.


from the DNA as detected by both SDS–PAGE andBradford analyses (Fig. 3, lane 1). We attempted to opti-mize the stripping of proteins from the DNA by varyingboth the amounts of the detergent/chaotrope (i.e., geno-mic DNA isolation using 2 or 4 times the volume ofDNAzol used in the standard DNAzol method) and eth-anol (i.e., DNA precipitation using 2 or 4 times the vol-ume of ethanol used in the standard DNAzol isolationmethod) (Fig. 3, lanes 2–5). Although these modiWca-tions did reduce the background level of associated pro-teins, the purity of the samples was not adequate becausethere was still a signiWcant level of protein isolated fromthe untreated sample. To further modify this method toobtain the level of stringency that would be necessary forDPC isolation, we Wrst carried the AA8 cells through anuclei isolation procedure and then isolated DNA usingthe DNAzol method (Fig. 3, lane 6), and this greatlyreduced the background level of protein isolated. How-ever, these modiWcations were not suYcient to removeall noncovalently associated proteins from the DNA, asthere was still some staining observed on the SDS–PAGE gel as well as protein detected in these samples byBradford analysis (Fig. 3, lane 6). Nonetheless, these

experiments demonstrated that the isolation of nucleiand the use of an increased volume of chaotropic agentgreatly reduced the level of background proteins iso-lated, and these modiWcations formed the basis for fur-ther method development.

DPC isolation by DNAzol–strip method

We developed a method from this point (DNAzol–strip method, Fig. 1B) exclusively using isolated nuclei.We combined the DNAzol reagent to lyse nuclei, hydro-lyze RNA, and dissociate bulk proteins from DNA withadditional chaotropic agents to strip noncovalentlybound proteins from the DNA. Isolated nuclei werelysed by the addition of DNAzol, and the DNA (withattached proteins) was precipitated with ethanol. TheDNA was then resuspended and washed in an SDS/urea/sodium chloride mixture to optimize removal of nonco-valently bound proteins. This was followed by extensivedesalting and volume reduction. The DNA was then iso-lated by ethanol precipitation and resuspended. TheDNA was digested with DNase I and S1 nucleases, andthe proteins were collected and reduced to dryness.

Fig. 3. ModiWcation of the DNAzol protocol to reduce the level of background protein. To minimize the isolation of noncovalently bound protein,several modiWcations to the standard DNAzol method were tested. DNA and bound protein was isolated from untreated AA8 cells using diVerentvolumes of DNAzol or diVerent volumes of ethanol to precipitate the DNA and associated protein. We also tested the DNAzol protocol using iso-lated nuclei. Sample loads were normalized based on the amount of digested DNA present in the sample (»35 �g of DNA loaded for each sample),and proteins were analyzed by 12% SDS–PAGE and silver staining (A) and Bradford protein quantitation (B) to assess the level of recovered pro-tein. Protein molecular weights are indicated in kilodaltons.


Because this DPC isolation method isolates crosslinkedproteins as a function of their attachment to DNA, sam-ple loads were always normalized for DNA contentwithin each experiment as determined after DNA diges-tion. Proteins were separated by 1-D SDS–PAGE, andthe gels were stained for visualization.

Using nuclei from untreated AA8 cells, various formsof the DNAzol isolation method were compared with theSDS/K+ precipitation method to assess the backgroundlevel of proteins isolated (Fig. 4). (It should be noted thatothers have combined additional isolation and washsteps with the SDS/K+ protocol [3,27] to reduce the back-ground level of noncovalently bound proteins.) Proteinsample loads were normalized based on the cell numberdetermined at plating (24 million cells plated per sample)(Fig. 4A, lane 2) or DNA content determined after DNAdigestion (Fig. 4B, lane 2). As expected, the unmodiWedSDS/K+ method resulted in the recovery of a high level ofnoncovalently associated protein. In contrast, the DNA-zol–strip method (Fig. 4A, lane 4, and Fig. 4B, lane 4) iso-lated relatively little noncovalently associated protein.Thus, combining the use of DNAzol with additionalwash steps reduces the background level of proteins tonearly zero and represents an improvement over the

SDS/K+ and standard DNAzol (Fig. 4A, lanes 2 and 3,and Fig. 4B, lanes 2 and 3) isolation methods.

DPC isolation by DNAzol–silica method

We have modiWed the DNAzol-based DPC isolationprocedure to make it considerably faster and more eco-nomical (Fig. 1C). The modiWcation relies on the abilityof DNA (but not proteins) to bind to silica in the pres-ence of chaotropic/dissociative agents such as guanidi-nium hydrochloride, sodium chloride, and urea, whichstrip the DNA of associated proteins [23]. This protocolsubstitutes an adsorption step for the desalting/concen-tration step, resulting in the DNA (and covalentlyattached proteins) being bound to the silica and the non-covalently associated proteins being removed in thesupernatant and subsequent wash steps. The DNA (withDPCs) is then eluted from the silica and digested, releas-ing the proteins, which are collected and analyzed by 1-D SDS–PAGE.

Using nuclei from untreated AA8 cells, the DNAzol–silica method (Fig. 4C, lane 4) was compared with theDNAzol (Fig. 4C, lane 2) and DNAzol–strip (Fig. 4C,lane 3) methods to assess the background level of

Fig. 4. Comparison of background protein levels using diVerent DPC isolation methods. Untreated AA8 cell nuclei were subjected to DPC isolationby the SDS/K+ method, the DNAzol method, or the DNAzol–strip method. Sample volumes were adjusted for either cell number (A, equal numberof cells determined at plating, 24 million cells per plate) or DNA content (B, determined after DNA digestion, »70 �g of DNA loaded for each sam-ple) and were analyzed by 12% SDS–PAGE and silver staining. (C) Using untreated AA8 cell nuclei, the DNAzol (Fig. 1A), DNAzol–strip (Fig. 1B),and DNAzol–silica (Fig. 1C) methods were compared directly. Sample volumes were adjusted for DNA content (measured after DNA digestion),and proteins were analyzed by 10–20% SDS–PAGE gradient gel and SYPRO Tangerine staining. The silica-based isolation method was also per-formed using a noncommercial genomic DNA isolation reagent (G-HCl solution) instead of DNAzol. The “M” lanes are the molecular weightmarkers with the molecular weights indicated in kilodaltons. Lanes within each panel are from the same gel with intervening lanes removed.


proteins isolated by each protocol. Protein sample loadswere normalized based on DNA content determinedafter DNA digestion. As demonstrated in Fig. 4, theDNAzol–strip and DNAzol–silica methods both iso-lated relatively little noncovalently associated protein.

We also examined the impact of substitution ofDNAzol by a noncommercial DNA extraction solution(Fig. 4C, lane 5, “G-HCl solution”) composed of 6 Mguanidinium hydrochloride, 0.5% SDS, and 8 mMsodium hydroxide in this silica-based isolation method.As seen in Fig. 4C, this solution proved to be less eVec-tive than DNAzol in reducing background proteinrecovery.

Isolation of ionizing radiation-induced DPCs by DNAzol–strip and DNAzol–silica methods

Both the DNAzol–strip and DNAzol–silica methodssuccessfully dissociated noncovalently bound proteinsfrom genomic DNA. The utility of each of these meth-ods for the isolation of covalently crosslinked proteinsfrom biological samples was investigated. The DNAzol–strip method was used in preliminary experiments to iso-late and analyze DPCs induced in AA8 cells exposed toformaldehyde or ionizing radiation (Fig. 5). AA8 cellswere exposed to 0 or 1 Gy of gamma radiation, or to 1%formaldehyde at 37 °C for 1 h, and nuclei were isolated.DPCs were isolated using the DNAzol–strip method asoutlined above. Dried protein samples were resuspendedin Laemmli loading buVer, and volumes were adjustedbased on DNA content. Proteins were analyzed by 12%SDS–PAGE and silver staining. Only a few faint distinctprotein bands were visible in the unirradiated sample

Fig. 5. Isolation of formaldehyde- and gamma ray-induced DPCs fromCHO cells using the DNAzol–strip method. AA8 cells received 0 or 1Gy of gamma radiation or were treated with 1% formaldehyde(HCHO) at 37 °C for 1 h. DPCs were isolated using the DNAzol–stripmethod. Sample volumes were adjusted for DNA content (determinedafter DNA digestion), and proteins were analyzed by 12% SDS–PAGE and silver staining. The “M” lane is the molecular weightmarkers with the molecular weights indicated in kilodaltons. Lanes inthe Wgure are from the same gel with intervening lanes removed.

(Fig. 5, lane 2), whereas a greater number and intensityof distinct protein bands were observed routinely in bothirradiated and formaldehyde-treated samples (Fig. 5,lanes 3 and 4), demonstrating that the DNAzol–stripmethod isolated reasonably pure, presumably covalentlycrosslinked proteins and little background protein. Theprotein concentration measurements routinely demon-strated that the level of protein in the irradiated sample(1 Gy) was approximately threefold higher than that inthe unirradiated sample. The suitability of the SDS–PAGE bands for further analysis by mass spectrometrywas then addressed.

We also assessed the potential utility of the more con-venient DNAzol–silica method in isolating DPCs frombiological samples. Additional modiWcations in the anal-ysis involved the use of a quantitative reversible proteinstain, SYPRO Tangerine. AA8 cells were exposed to 0 or1 Gy of gamma radiation or to 1% formaldehyde at37 °C for 1 h. DPCs were isolated using the DNAzol–sil-ica method as outlined above. Dried protein sampleswere resuspended in Laemmli loading buVer, and vol-umes were adjusted based on DNA content. Proteinswere analyzed by 10–20% gradient SDS–PAGE andSYPRO Tangerine staining (Fig. 6). This method gener-ated results similar to those of the DNAzol–stripmethod. There was a low level of background proteinisolated, as evidenced by the few distinct protein bandsobserved in the untreated sample (Fig. 6, lane 2). TheDNAzol–silica method allowed the isolation of rela-tively pure, presumably covalently crosslinked proteinsfrom both the 1 Gy-irradiated and formaldehyde-treated

Fig. 6. Isolation of formaldehyde- and gamma ray-induced DPCs fromCHO cells using the DNAzol–silica method. AA8 cells received 0 or 1Gy of gamma radiation or were treated with 1% formaldehyde(HCHO) at 37 °C for 1 h. DPCs were isolated using the DNAzol–silicamethod. Sample volumes in the “0 Gy,” “1 Gy,” and “HCHO” laneswere adjusted to equalize the DNA concentrations determined by UVabsorbance after DNA digestion. Proteins were analyzed by 10– 20%gradient SDS–PAGE and SYPRO Tangerine staining. “M” representsmolecular weight markers with the molecular weights indicated inkilodaltons, and NE represents AA8 nuclear extract from untreatedcells. Lanes in the Wgure are from the same gel with intervening lanesremoved.


samples (Fig. 6, lanes 3 and 4), as evidenced by theappearance of distinct protein bands. Some smearing ofthe protein bands is expected on 1-D SDS–PAGEbecause there may be multiple protein species of similarsize. This behavior was more marked in the case of form-aldehyde crosslinking, probably due to the potency ofthis crosslinking agent and the extended treatment inter-val used (1 h) compared with irradiation (10 s) as well asthe diVerent lifetimes of the various intermediatesinvolved in these two chemically distinct crosslinkingmechanisms. Nonetheless, individual bands were readilyvisible.

Preliminary identiWcation of a crosslinked protein by mass spectrometry

The DNAzol–strip method yielded excellent qualityprotein samples of suYcient quantity to allow identiWca-tion of a number of radiation-crosslinked proteins bymass spectrometry. Fig. 7 shows an example of a massspectrum of peptides isolated from pooled SDS–PAGEgel bands excised from identical samples of irradiatedCHO cells (inset). Several of the peptides isolated fromthe excised bands (Table 1) led to the identiWcation ofthe hamster heat shock protein, glucose-regulated pro-tein 78 (GRP78), which has previously been shown to becrosslinked to DNA by the antitumor antibiotic gilvo-carcin [27]. A more extensive analysis of radiation-induced DPCs will be the subject of another study (inpreparation).

Isolation of topoisomerase I inhibitor-induced DPCs by DNAzol–strip and DNAzol–silica methods

The utility of each of these methods for the isolationof genuinely covalently crosslinked proteins from bio-logical samples was also investigated using a known tar-get DPC. Mammalian DNA topoisomerase I (91 kDa)becomes transiently covalently crosslinked to DNA dur-ing DNA processing [28], and these DPCs can betrapped using inhibitors such as camptothecin. AA8 cellswere treated with 10 �g/ml camptothecin for 1.5 h at37 °C, and nuclei were isolated. DPCs were isolatedusing the DNAzol–strip method (Fig. 8, lanes 1–3) or theDNAzol–silica method (Fig. 8, lanes 5–7). Dried proteinsamples were resuspended in Laemmli loading buVer,and volumes were adjusted based on DNA content pre-viously determined after DNA digestion. Proteins wereanalyzed by 10–20% gradient SDS–PAGE and SYPROTangerine staining. As shown in lane 3 of Fig. 8, theDNAzol–strip method primarily isolated smaller bandsfrom the camptothecin-treated cells, many of which wereprobably degradation products. The DNAzol–silicamethod (Fig. 8, lane 6), on the other hand, isolated aband of approximately 100 kDa as well as severalsmaller and larger sized protein bands. The highermolecular weight species probably represent ubiquitina-ted topoisomerase I [29], which are seen to an evengreater extent in cells treated simultaneously with cam-ptothecin and a proteasome inhibitor, MG132 (Fig. 8,lane 7, and [29]). The presence of DNA topoisomerase I

Fig. 7. Mass spectrometric identiWcation of GRP78 as an ionizing radiation-crosslinked protein in CHO AA8 cells. The mass/charge (m/z) ratios forpeptides isolated from the indicated protein band are shown. Database searching identiWed several of these peptides as part of the amino acidsequence of the hamster heat shock protein, GRP78 (Table 1). The intensely stained band at approximately 35 kDa is due to DNase I. In the inset,peptides were obtained from the 12% SDS–PAGE silver-stained gel band indicated by the arrow in the irradiated “IR” lane and pooled with thesame band from multiple irradiated samples. The “C” lane is the unirradiated control. Protein molecular weights are indicated in kilodaltons.


in DPCs isolated from these experiments (Fig. 8, lane 3,sample A; lane 6, sample E; and lane 7, samples B and C)was conWrmed by mass spectrometry (Table 1). SampleD in Fig. 8 (lane 7) was not found to contain topoiso-merase I. (Pooling of excised bands within sample areasA, B, and E was performed to ensure suYcient materialfor mass spectrometry identiWcation because we wereconcerned that the recovery of topoisomerase I might belimited by its rapid degradation. Furthermore, the detec-tion limit of SYPRO Tangerine is lower than the detec-tion limit of the mass spectrometry technology.)

Discussion

The study of covalent protein–DNA complexes hasbeen limited by the lack of availability of techniques thatovercome a number of challenges [1]. DPCs will involvea small fraction of the proteome and may involve low-abundance proteins and proteins of diVering solubilitiesand stability. DPC isolation must be rigorous because

Table 1Peptides used for mass spectral identiWcation of proteins

Note. The table lists the peptides and peptide masses that were used toidentify GRP78 from the samples in Fig. 7 (inset) and to conWrm thepresence of DNA topoisomerase I in the samples indicated in Fig. 8.An asterisk(s) indicates the presence of oxidized methionine(s) in thepeptide.

Protein Peptide Mass observed

GRP78 SDIDEIVLVGGSTR 1460.10ITPSYVAFTPEGER 1566.10KSDIDEIVLVGGSTR 1588.20

Topoisomerase I, sample A

EMTNDEK* 882.15ITVAWCKK 1004.69QRAVALYFIDK 1323.16YIMLNPSSRIK* 1338.04AVQRLEEQLMK* 1360.14CDFTQMSQYFKDQSEAR 2139.48MSGDHLHNDSQIEADFRLNDSHK* 2680.77

Topoisomerase I, sample B

IEPPGLFR 927.27DQLADARR 943.93ILSYNRANR 1105.34TYNASITLQQQLK 1507.71IMPEDIIINCSKDAK* 1761.16

Topoisomerase I,sample C

EENKQIALGTSK 1316.53RIMPEDIIINCSK* 1602.52QIALGTSKLNYLDPR 1687.44LNYLDPRITVAWCK 1747.47LLKEYGFCVMDNHR 1782.03

Topoisomerase I, sample E

EDIKPLK 841.67GNHPKMGMLK* 1128.21QRAVALYFIDK 1322.48WGVPIEKIYNK 1345.41SMMNLQSKIDAK* 1380.74TFEKSMMNLQSK** 1473.44IMPEDIIINCSKDAK 1745.77CDFTQMSQYFKDQSEAR 2140.21

DPCs must be distinguished from various DNA–proteinassociations that are noncovalent but may nonethelessbe relatively abundant and strong enough to resist disso-ciation by commonly used isolation methods. Detectionand chemical analysis of DPCs will require suYcientlylarge and pure samples and sensitive protein analyticaltechniques. Current methods used for the isolation ofDPCs from cells fail to provide adequate stringency,speciWcity, and scalability of isolation [1]. We havedescribed the development of novel methods for the iso-lation of pure, enriched, and intact DPCs that is applica-ble to large numbers of cells and is economical, rapid,and amenable to high-throughput. The methods arebased on the use of the DNAzol reagent and high con-centrations of additional chaotropes to dissociate non-covalent DNA–protein associations. The two variationsof the DNAzol-based DPC isolation procedure allowthe isolation of highly pure, covalently crosslinked pro-teins from cells.

The SDS–PAGE analyses (Figs. 5 and 6) indicate thatthe background level of protein isolated from untreatedAA8 cells, although extremely low, is not zero. However, itshould be noted that endogenous DPC-inducing agents(e.g., free radicals, aldehydes, lipid peroxidation products)will be present in a cell at any given time. Indeed, it hasbeen proposed that DNA crosslink repair mechanismsactually evolved in response to the damage induced bysuch intracellular crosslinking agents [30]. We have nowused both the DNAzol–strip and DNAzol–silica methodsextensively for DPC isolations and have routinely observedvery little signal in the unirradiated samples on SDS–PAGE analysis in a larger study of radiation-crosslinked

Fig. 8. Isolation of camptothecin-induced DPCs from CHO cells. AA8cells received no treatment (NO), 1 �l/ml DMSO (DM), 10 �g/ml cam-ptothecin for 1.5 h at 37 °C (CP), or 10 �M MG132 for 3 h at 37 °C fol-lowed by 10 �g/ml camptothecin and 10 �M MG132 for 1.5 h at 37 °C(CP-MG). DPCs were isolated using the DNAzol–strip method (lanes1–3) or the DNAzol–silica method (lanes 5–7). Dried protein sampleswere resuspended in Laemmli loading buVer, and sample volumeswere adjusted for DNA content (measured after DNA digestion). Pro-teins were analyzed by 10–20% gradient SDS–PAGE and SYPROTangerine staining. Bands from within the indicated sample regions(A–E) were excised, and bands within sample regions A, B, and E werepooled separately. The “M” lane is the molecular weight markers withmolecular weights indicated in kilodaltons.


proteins. Considering the evidence from the current studyshowing that measured background DPC levels are verysensitive to small methodological alterations, it is not sur-prising that the level of background endogenously inducedDPCs reported in diVerent studies varies greatly with themethod used for DPC detection [31–33].

The two method variations presented here involve rel-atively mild conditions for the elution and resuspensionof DNA and DPCs, and they release the crosslinked pro-teins by nuclease digestion, making these methods suit-able for analysis of DPCs induced by various agents. Inaddition, the crosslinked proteins isolated in the currentstudy are in a suYciently pure and enriched form to beuseful for proteomics analysis because we were able touse MALDI–TOF mass spectrometry and MS/MS toidentify GRP78 as a protein crosslinked to DNA bygamma radiation in two independent experiments.GRP78 was described here only to illustrate the useful-ness of this protein isolation method for interfacing withhigh-throughput proteomics technologies; to date, infact, we have isolated and identiWed 29 cellular proteinsthat appear to participate in such lesions (in preparation).

The comparison of the DNAzol–strip and DNAzol–silica methods for isolation of the DNA topoisomerase Icomplex revealed that the DNAzol–strip method proba-bly isolated degraded complex (Fig. 8 and Table 1, sam-ple A). This was most likely due to the lengthyprocessing time in high salt conditions involved in theDNAzol–strip method. Previous studies have shownthat (i) camptothecin treatment induces a time- anddose-dependent degradation of topoisomerase I [34] and(ii) camptothecin-induced DNA topoisomerase I com-plexes are rapidly lost once the drug is removed in vivo[34] and are reversed in vitro with the addition of 0.5 Msodium chloride [28]. The DNAzol–silica method iso-lated these degradation products as well as larger prod-ucts that may represent ubiquitinated forms oftopoisomerase I (Fig. 8 and Table 1, samples B and C),which are seen as higher molecular weight bands onSDS–PAGE analysis [29]. The comparison of the DNA-zol–strip and DNAzol–silica methods indicates that thelatter is faster, thereby permitting the isolation of theshorter lived population of DPCs.

With respect to the resolution of exogenously inducedDPCs, the DNAzol–strip and DNAzol–silica methodscan detect DPCs at gamma ray doses as low as 1 Gy(Figs. 5 and 6). Bradford analyses performed on DNA-zol–strip experiments demonstrated an average of three-fold more protein isolated from 1 Gy-irradiated samplesover the unirradiated samples. This can be comparedwith the alkaline elution/polycarbonate Wlter method, inwhich DPC detection was possible only at much higherradiation doses (50 Gy) [35], and with the nitrocelluloseWlter binding technique, which can detect DPCs in irra-diated cells at doses as low as 30 Gy but which does notallow speciWc protein recovery [35].

During preparation of this article, we became awareof a method devised for the isolation of cisplatin-induced crosslinks [20–22]. The method used was devel-oped for analyzing a speciWc type of DPC rather than forisolating any/all crosslinked proteins, but it underscoresthe utility of chaotropic agents for dissociating noncova-lently crosslinked proteins. However, the backgroundlevel of proteins isolated from untreated cells was notreported in those studies [20–22]; therefore, the contribu-tion of noncovalently bound, but tightly associated,nuclear matrix proteins cannot be evaluated. Thatmethod involved binding DNA with attached proteinsto a hydroxylapatite matrix that can also bind proteins.The DNAzol–strip method described here does notinvolve adsorption to a solid phase, and the DNAzol–sil-ica method uses a solid phase that does not bind proteinsigniWcantly.

In summary, we have developed DPC isolation meth-ods that optimize the isolation of proteins covalentlycrosslinked to DNA in a pure and concentrated sample.The isolation procedures are readily scalable and eco-nomical. DPCs were isolated in suYcient quantities foruse with proteomics technology for protein separationand identiWcation.

Acknowledgments

This work was supported by grants from the NationalCancer Institute of Canada with funds from the Cana-dian Cancer Society to Michael Weinfeld (Grant013104) and to David Murray (Terry Fox research grant8067). This work was also supported by the AlbertaCancer Board (ACB) through Pilot Project Grants R-294 and R-465 to Murray, a New Initiatives award toLiang Li to support the ACB Proteomics Facility, andstudentship support for Sharon Barker.

References

[1] S.L. Barker, M. Weinfeld, D. Murray, DNA–protein crosslinks:Their induction, repair and biological consequences, Mutat. Res.589 (2005) 111–135.

[2] S. Gebicki, J.M. Gebicki, Crosslinking of DNA and proteinsinduced by protein hydroperoxides, Biochem. J. 338 (1999) 629–636.

[3] C.A.I. Miller, M.D. Cohen, M. Costa, Complexing of actin andother nuclear proteins to DNA by cis-diamminedichloroplati-num(II) and chromium compounds, Carcinogenesis 12 (1991)269–276.

[4] J.R. Kuykendall, M.S. BogdanVy, Reaction kinetics of DNA-his-tone crosslinking by vinyl acetate and acetaldehyde, Carcinogene-sis 13 (1992) 2095–2100.

[5] J.R. Kuykendall, M.S. BogdanVy, Formation and stability of acet-aldehyde-induced crosslinks between poly-lysine and poly-deoxy-guanosine, Mutat. Res. 311 (1994) 49–56.

[6] M. Costa, A. Zhitkovich, M. Harris, D. Paustenbach, M. Gargas,DNA–protein cross-links produced by various chemicals in


cultured human lymphoma cells, J. Toxicol. Environ. Health 50(1997) 433–449.

[7] V. Voitkun, A. Zhitkovich, Analysis of DNA–protein crosslinkingactivity of malondialdehyde in vitro, Mutat. Res. 424 (1999) 97–106.

[8] S. Toyokuni, T. Mori, H. Hiai, M. Dizdaroglu, Treatment ofWistar rats with a renal carcinogen, ferric nitrilotriacetate, causesDNA–protein cross-linking between thymine and tyrosine in theirrenal chromatin, Int. J. Cancer 62 (1995) 309–313.

[9] A.J. Fornace Jr., J.B. Little, Malignant transformation by theDNA–protein crosslinking agent trans-Pt(II) diamminedichloride,Carcinogenesis 1 (1980) 989–994.

[10] A.J. Fornace Jr., Detection of DNA single-strand breaks producedduring the repair of damage by DNA–protein cross-linkingagents, Cancer Res. 42 (1982) 145–149.

[11] M.O. Bradley, I.C. Hsu, C.C. Harris, Relationship between sisterchromatid exchange and mutagenicity, toxicity, and DNA dam-age, Nature 282 (1979) 318–320.

[12] M.O. Bradley, K.W. Kohn, X-ray induced DNA double strandbreak production and repair in mammalian cells as measured byneutral Wlter elution, Nucleic Acids Res. 7 (1979) 793–804.

[13] O. Merk, G. Speit, SigniWcance of formaldehyde-induced DNA–protein crosslinks for mutagenesis, Environ. Mol. Mutagen. 32(1998) 260–268.

[14] L.Y. Xue, L.R. Friedman, N.L. Oleinick, Repair of chromatindamage in glutathione-depleted V-79 cells: Comparison of oxicand hypoxic conditions, Radiat. Res. 116 (1988) 89–99.

[15] S.M. Chiu, N.M. Sokany, L.R. Friedman, N.L. Oleinick, DiVeren-tial processing of ultraviolet or ionizing radiation-induced DNA–protein cross-links in Chinese hamster cells, Int. J. Radiat. Biol.Relat. Stud. Phys. Chem. Med. 46 (1984) 681–690.

[16] A.E. Cress, K.M. Kurath, B. Stea, G.T. Bowden, The crosslinkingof nuclear protein to DNA using ionizing radiation, J. Cancer Res.Clin. Oncol. 116 (1990) 324–330.

[17] M. Costa, A. Zhitkovich, M. Gargas, D. Paustenbach, B. Finley, J.Kuykendall, R. Billings, T.J. Carlson, K. Wetterhahn, J. Xu, S.Patierno, M. BogdanVy, Interlaboratory validation of a new assayfor DNA–protein crosslinks, Mutat. Res. 369 (1996) 13–21.

[18] A. Zhitkovich, M. Costa, A simple, sensitive assay to detect DNA–protein crosslinks in intact cells and in vivo, Carcinogenesis 13(1992) 1485–1489.

[19] J.W. Bodnar, C.J. Jones, D.H. Coombs, G.D. Pearson, D.C. Ward,Proteins tightly bound to HeLa cell DNA at nuclear matrixattachment sites, Mol. Cell. Biol. 3 (1983) 1567–1579.

[20] A. Ferraro, P. Grandi, M. Eufemi, F. Altieri, C. Turano, Crosslink-ing of nuclear proteins to DNA by cis-diamminedichloroplatinumin intact cells: Involvement of nuclear matrix proteins, FEBS Lett.307 (1992) 383–385.

[21] S.K. Samuel, V.A. Spencer, L. Bajno, J.M. Sun, L.T. Holth, S. Oest-erreich, J.R. Davie, In situ cross-linking by cisplatin of nuclearmatrix-bound transcription factors to nuclear DNA of humanbreast cancer cells, Cancer Res. 58 (1998) 3004–3008.

[22] V.A. Spencer, S.K. Samuel, J.R. Davie, Altered proWles in nuclearmatrix proteins associated with DNA in situ during progressionof breast cancer cells, Cancer Res. 61 (2001) 1362–1366.

[23] M.S. Elphinstone, G.N. Hinten, M.J. Anderson, C.J. Nock, Aninexpensive and high-throughput procedure to extract and purifytotal genomic DNA for population studies, Mol. Ecol. 3 (2003)317–320.

[24] K.W. Kohn, DNA as a target in cancer chemotherapy: measure-ment of macromolecular DNA damage produced in mammaliancells by anticancer agents and carcinogens, in: H. Busch, V.T.DeVita (Eds.), Methods in Cancer Research, vol. 16, AcademicPress, New York, 1979, pp. 291–345.

[25] K.W. Kohn, R.A. Ewig, L.C. Erickson, L.A. Zwelling, Measure-ment of strand breaks and crosslinks by alkaline elution, in: E.C.Friedberg, P.C. Hanawalt (Eds.), DNA Repair: A LaboratoryManual of Research Procedures, Marcel Dekker, New York,1981, pp. 379–401.

[26] D. Murray, S.C. van Ankeren, R.E. Meyn, Applicability of thealkaline elution procedure as modiWed for the measurement ofDNA damage and its repair in nonradioactively labeled cells,Anal. Biochem. 160 (1987) 149–159.

[27] A. Matsumoto, P.C. Hanawalt, Histone H3 and heat shock pro-tein GRP78 are selectively cross-linked to DNA by photoacti-vated gilvocarcin V in human Wbroblasts, Cancer Res. 60 (2000)3921–3926.

[28] Y.H. Hsiang, R. Hertzberg, S. Hecht, L.F. Liu, Camptothecininduces protein-linked DNA breaks via mammalian DNA topoi-somerase I, J. Biol. Chem. 260 (1985) 14873–14878.

[29] Q. Fu, S.W. Kim, H.X. Chen, S. Grill, Y.C. Cheng, Degradation oftopoisomerase I induced by topoisomerase I inhibitors is depen-dent on inhibitor structure but independent of cell death, Mol.Pharmacol. 55 (1999) 677–683.

[30] G. Weeda, I. Donker, J. de Wit, H. Morreau, R. Janssens, C.J. Vis-sers, A. Nigg, H. van Steeg, D. Bootsma, J.H. Hoeijmakers, Dis-ruption of mouse ERCC1 results in a novel repair syndrome withgrowth failure, nuclear abnormalities, and senescence, Curr. Biol.7 (1997) 427–439.

[31] R.K. Zahn, G. Zahn-Daimler, S. Ax, M. Hosokawa, T. Takeda,Assessment of DNA–protein crosslinks in the course of aging intwo mouse strains by use of a modiWed alkaline Wlter elutionapplied to whole tissue samples, Mech. Ageing Dev. 108 (1999)99–112.

[32] A. Izzotti, C. Cartiglia, M. Taningher, S. De Flora, R. Balansky,Age-related increases of 8-hydroxy-2�-deoxyguanosine andDNA–protein crosslinks in mouse organs, Mutat. Res. 446 (1999)215–223.

[33] R. Olinski, Z. Nackerdien, M. Dizdaroglu, DNA–protein cross-linking between thymine and tyrosine in chromatin of �-irradiatedor H2O2-treated cultured human cells, Arch. Biochem. Biophys.297 (1992) 139–143.

[34] D.R. Beidler, Y.C. Cheng, Camptothecin induction of a time- andconcentration-dependent decrease of topoisomerase I and itsimplication in camptothecin activity, Mol. Pharmacol. 47 (1995)907–914.

[35] I. Al-Nabulsi, K.T. Wheeler, Temperature dependence of radia-tion-induced DNA–protein crosslinks formed under hypoxic con-ditions, Radiat. Res. 148 (1997) 568–574.

dna protein cross linking

Documents