New methods for detection of low levels of DNA damage in human populations.

The use of a postlabeling method to characterize and to detect infrequent base modifications in DNA is outlined. This method has the advantage that low levels of DNA modifications, approximately 1 modified base per 10(5) nucleotides, can be detected. Moreover, a broad spectrum of modification can be identified by using this methodology. The basis for the method involves transfer of a radioactive phosphate from the gamma position of ATP to the 5'-hydroxyl terminus of 3'-phosphoryl nucleotides that are derived from modified DNA by appropriate nuclease digestion. The second method involves use of a defined DNA sequence within human cells. The alpha sequence is used as a probe for DNA damage to specific nucleotides. The alpha DNA sequence is reiterated approximately 300,000 times in the human genome and exists in tandem arrays. It comprises approximately 1% of the entire genome. The reiterated sequence is sufficiently homogeneous to permit its use as a probe for a site specific in DNA damage. Examples of the application of both of these methodologies to DNA damage inflicted in human cells by chemicals and ultraviolet light are provided. ImagesFIGURE 9.FIGURE 11.FIGURE 12.


Introduction
Exposure of humans to toxic substances poses an increasing health risk throughout the industrialized world. Measurement of the health risk associated with such exposure is bedeviled by a myriad of problems. Problems begin with the accurate assessment of exposure. Exposures often occur over long periods of time to low levels of complex mixtures of chemical substances. Compounding the complexity of the problem is the likelihood that individuals respond differently to different sets of substances. Differences in individual health and metabolism may dramatically alter the affect of a given exposure.
What tools must be developed to grapple with this intractable problem? Accurate means for individual dosimetry are needed. Methods for assessment of biological damage to individuals are necessary. Means for determination of which chemical, among a variety of chemical exposures, leads to observable biological changes must be developed. Over the past few years, we have developed several new techniques that meet some of these needs. The methods were designed to provide insights into mechanisms of DNA damage and repair. Applied to the area of human toxicology, they should provide sensitive, analytical tools for determination of individual loads of DNA damage.
Postlabeling Methods to Detect and Characterize Infrequent Base Modifications in DNA We sought to develop a method for the detection of DNA lesions that could measure low levels of damage in DNA of exposed individuals. The postlabeling method described below has the following attractive features for such work.
Low levels of DNA modifications can be detected. The limit of sensitivity is probably about one modified base per 105.
A broad spectrum of modifications can be detected. Both small and large base adducts can be detected. Alterations in sugars and in the bases can also be resolved. Intrastrand and interstrand crosslinks can be detected as well.
The specific type of damage can be determined. Each DNA modification can be analyzed independently. The contribution to DNA damage of each component of a complex mixture can be identified.
Consequently, the damage created by specific environments should yield a characteristic "fingerprint" of lesions. Such fingerprint analysis could permit tracing of DNA damage to particular types of toxic waste, and permit cause-effect relationships to be established.
The persistence of each type of lesion can be measured. Measurements taken over time would permit analysis ofthe persistence of specific lesions.
The method would provide an individual dosimeter of effective exposure. Measurement of exposure based on ambient levels of a substance in the air or water do not measure accurately individual exposure. Measurement of serum and urine levels of substances do measure individual burdens. However, the biological affects of a given level of exposure may vary considerably from person to person. Measurement of the level of DNA damage is likely to provide a better indicator of the biological effects of exposure than other methods.
The measurements can be performed on very small tissue samples. A 1or 2-,ug quantity of DNA is all that is required for complete analysis. Such amounts of DNA can be collected from routine blood samples or from microbiopsy samples.

The Method
A new method is presented that allows the detection and characterization of infrequent base modifications in DNA. The method combines the technique of in vitro labeling of nucleotides derived from cellular DNA with the methods of high pressure liquid chromatography (HPLC). Previous methods for the detection of modified bases in DNA have included the use of radioactive drugs or modifying agents (1) and the uniform labeling of DNA by feeding cells radioactive DNA precursors. The method presented here differs from these previous methods as DNA is postlabeled after being treated with a DNA base-modifying agent.
An example of the DNA postlabeling method is demonstrated. DNA was treated in vitro with dimethyl sulfate (DMS), a DNA alkylating agent. DMS reacts with purines in DNA (2), the principal adducts formed being 7-methylguanine, 3-methyladenine and 1-methyladenine (3). Following DMS treatment, the DNA was isolated, and digested to nucleotides. The nucleotides were then postlabeled and were resolved by HPLC. DMS-methylated bases, in nucleotide form, were readily resolved from the four normal DNA bases and were detected along with two other minor products, 5-methylcytosine and uracil.

DNA Postlabeling Procedure
The basic procedure for postlabeling DNA is presented in Figure 1. DNA is modified with a DNA damaging agent that produces stable base modifications. The modified DNA is isolated and is digested by two enzymes, micrococcal nuclease and spleen phosphodiesterase, to fonn 3'-deoxynucleoside monophosphates (4). The 3'-nucleotides are then labeled with 32P at the 5' position by the addition of high specific activity y32P ATP and T4 polynucleotide kinase (5). A mutant enzyme that lacks the endogenous 3'-phosphatase activity (6)  nucleotides, the products produced being 3'5'-deoxynucleoside diphosphates. The 3'-phosphate group is then removed by the addition of normal T4 polynucleotide kinase which contains 3'-phosphatase activity (7). The normal four DNA nucleotides are labeled, as are any nucleotides containing modified bases.
Reaction of DMS with DNA DNA was made up at a concentration of 400 ,ug/mL in lOOmM Tris-Cl, pH 7.5. DMS (20 FL) was reacted with 180 ,uL of DNA for 1.5 hr at 37°C. As a control, 20 puL of lOOmM Tris-Cl, pH 7.5, was added to 180 pL of DNA and was also incubated at 37°C for 1.5 hr. Following the incubation, 22 ,uL of 3M Na acetate, pH 5.5, and 450 ,uL of ethanol were added to the incubation mixture to precipitate the DNA. The tube was well mixed and placed in a dry ice/ethanol bath at -70°C for 10 min and was then centrifuged at 12000g for 10
In order to remove any unreacted _y32P ATP, the labeled nucleotides were separated from ATP by the use of a Bond-Elut mini-extraction column containing 100 mg of an aminopropyl sorbent (Analytichem. International, Harbor City, CA). The column packing material consists of aminopropyl moieties bonded to silica. The reaction mixture of 5'-labeled nucleotides (90 p,L) was added to the top of the column, along with 500 puL of water. The column, which contains the sorbent in a 1-cm3 syringe, was placed in a 15 mL centrifuge tube and was centrifuged for 3 min at 5000 rpm in a clinical centrifuge. The elutant was discarded, and 500 ,L of 0.1M KH2PO4, pH 4.5, was added to the top of the column to elute the 5'-labeled nucleotides. The column was again placed in a 15 mL centrifuge tube and was centrifuged at 5000 rpm for 3 min. The elutant was collected and lyophilized. The radioactive 5'-nucleotides were resuspended in 200 ,uL of water.

HPLC Instrumentation and Separation Procedures
HPLC instrumentation included a Varian 5020 liquid chromatograph equipped with a Valco loop injector fitted with a 50 ,uL sample loop. Additional equipment included a Varian UV-50 variable wavelength detector and an ISCO model 328 fraction collector equipped with a stop-flow valve.
Separation of 5'-deoxynucleotides was performed on a Waters Bondapak C18 reverse phase column (300 x 4mm). A gradient elution was utilized in which an initial elution buffer concentration of 100% 75mM KH2PO4, pH 4.5, was maintained for 15 min, followed by a linear gradient for 15 min to a final elution mixture of40% methanol, 60% 75mM KH2PO4, pH 4.5. The flow rate was maintained at 1 mL/min. UV detection of 5'-deoxynucleotides was at a wavelength of 260 nm. For radioactivity measurements, fractions were obtained of liquid eluting from the column at intervals of 0.3 min each. Fractions were collected in 1.5 mL micro test tubes (Eppendorff). The 32P radioactivity in each fraction was determined by Cerenekov counting using a Becknan LS8000 scintillation counter.

Example of Postlabeling Method DMS Modification of DNA
To demonstrate the DNA postlabeling method, human DNA was treated in vitro with DMS to allow sufficient alkylation ofpurine bases. The DNA was then digested to both 3'-and 5'-deoxynucleoside monophosphates. The 3'-deoxynucleotides were then postlabeled with 32P and were subjected to the 3'-phosphatase activity of T4 polynucleotide kinase, forming 5'-labeled deoxynucleotides. This step is included in the labeling procedure as there is much more selectivity in the separation of nucleoside monophosphates than nucleoside diphosphates by HPLC. The 5'-deoxynucleotides formed from the digestion of DNA by DNAse I and snake venom phosphodiesterase are used as markers for the radioactive nucleotides.
The use of reverse-phase chromatography as a highly selective method for the separation of nucleotides is demonstrated in Figure 2. In this separation, the four common deoxynucleotides 5'-dCMP, 5'-dTMP, 5'-dGMP and 5'-dAMP are readily separated and resolved, as are two nucleotides containing bases less frequently found in DNA-5'-dUMP and 5'-methyl-5'-dCMP. The occurrence of 5'-dUMP is found in DNA as a result of the deamination of cytosine. The occurrence of 5-methylcytosine results from methylation of cytosine and represents about 0.4% of the total bases in DNA of HeLa cells (10). The separation of5'-deoxynucleotides from digested human DNA is shown in Figure 3. The separation ofthe postlabeled nucleotides from untreated human DNA is shown in Figure 4. The composition of each nucleotide is shown in Table 1. 10 15 20 EL£/J/ON T/ME, Minutes ELUT/ON TIME, Minutfes  It is noted that the content of 5-methylcytosine is 0.6%, which is comparable to a published value of 0.4% for human DNA (10). The treatment of human DNA with DMS resulted in the formation of modified bases, as is evident in Figure 5. Three new eluting products are seen in the HPLC separation of the labeled 5'-deoxynucleotides. The three new peaks most likely result from purine methylation, as the percentage of normal purine bases is less than in the untreated DNA (see Table 1). The identification of each DMS product was not possible, as markers were not available for 5'-deoxynucleotides containing methylated purines. However, preliminary evidence from the treatment of 5'-dGMP and 5'-dAMP with DMS indicates that the product which elutes prior to 5'-dTMP results from the methylation of adenine, as does the product which elutes immediately after 5'-dAMP (data not shown). The product eluting most distant from the solvent front may be formed from the methylation of guanine.

Discussion
A new method has been presented allowing the detection of infrequent base modifications in DNA without the use of either radioactive DNA precursors or radiolabeled drugs or mutagens. Given the high specific activity of y2P-ATP available and the high resolution of HPLC reverse-phase chromatography, infrequent base modifications in the range of  II one in 105 should be detectable. The method is also useful for the detection of modified nucleotides which can be incorporated into DNA. Once such nucleotide is cytosine arabinoside monophosphate (ara-CMP). The nucleoside cytosine arabinoside (araC) is used as an antitumor agent in the treatment of several kinds of leukemia (11). Once incorporated into DNA, ara-CMP can be isolated and has been resolved by HPLC (12). The 3'ara-CMP can be labeled by the DNA postlabeling method, and the resulting formation of 5'-ara-CMP is shown in Figure 6. It should also be possible to label other nucleotide analogs which can be incorporated into DNA.
Use of the Alphoid Sequence as an Indicator of DNA Damage We have also developed an analytical method for detection of DNA damage within defined sequences of human DNA. This method has the advantage that it permits determination of distribution of DNA tL damage within defined DNA sequences. It also has the advantage that such damage can be determined within intact human cells. Thus, cells from individuals that have been exposed to high levels of toxic substances can be examined at the level of individual nucleotides for DNA damage.
It is unlikely that the use of the alphoid sequence will be suitable for routine analysis of field samples. Limit of detection of DNA lesions using this method is about one alteration per thousand nucleotides. However, the method will serve as an important tool to determine the types of DNA damage that are created by noxious chemicals. For this analysis, the chemical composition -of the damaging mix need not be known.

The Method
For this work we used the alphoid sequence of human DNA (13)(14)(15)(16). The alphoid sequence is ideal for this purpose. It is a highly reiterated sequence present in about 300,000 copies per haploid genome. The alphoid sequence is present in tandem arrays up to 40 repeat units in length (15). The basic repeat length of the human alphoid sequence is 342 nucleotides (13). This sequence is itself derived from an imperfect repeat unit of 171 nucleotides.
The alphoid sequence comprises about 1% of the total mass of human DNA. For a thorough treatment of the structure of the human alphoid sequences, see the article by Wu and Manuelidis (15). The alphoid DNA can be readily obtained from total genomic extracts. The detailed methods of isolation are presented below. Figures 7 and 8 are schematic illustrations of the isolation and labeling procedure. Total DNA is prepared from intact cells. Any human tissue or cultured cell line may be used, as the alphoid sequence is present in all normal and tumor cells. After purification, total cellular DNA is treated with a restriction enzyme, Eco RI. The alphoid sequence contains an Eco RI cleavage site. Treatment with Eco RI releases a major fraction of the alphoid DNA as a fragment of 342 nucleotide length as a direct consequence of the tandem array organization of the alphoid DNA. This 342 base pair fragment is easily separated from the bulk of the cellular DNA that is cleaved into fragments of between 3000 and 6000 base pairs (Fig. 9). For the analysis, the Eco RI alphoid fragment can be labeled at either the 3' or 5' terminus. DNA fragments that contain a single labeled terminus are obtained by treatment of the Eco RI fragment with restriction enzymes that cleave the DNA. A restriction enzyme cleavage map of the Eco RI alphoid DNA is pictured in Figure 10. We have used both the Eco RI* of 92 base pair length and the Mbo I fragments of 52 base pair length for our work.
If the human alphoid sequence is to be useful for these studies, it must have a defined sequence. Fortunately this is the case for both the Eco RI/Eco RI* and the Eco RI/Mbo I alphoid fragments. Figure   in part by our work is shown in Figures 10 and 11. The alphoid sequence of human DNA can be used as an indicator to compare DNA damage inflicted in vitro to damage created upon treatment of the same agent in vivo. We have used this procedure to study DNA damage in both situations created either by ultraviolet light (18) or by nitrogen mustard (9). The example of ultraviolet light damage taken from our previous work (18) is given below. Any human tissue or human cell in culture can be used for preparation of the alphoid sequence. We have used FIGURE 9. Human DNA digested with the restriction endonuclease Eco RI run on a 5% acrylamide gel stained with ethidium bromide. This gel illustrates the highly reiterated 342 and 680 base pair tandem repeats. a number of cell lines including the HeLa cells, the TK6 line of human B lymphoblasts (8) and the CEM human lymphoblastic leukemia cell line (19). We have also used normal and transformed human fibroblasts grown in cell culture.

DNA Purification
The first step ofthe procedure involves purification of the cellular DNA. For this work the starting cells should be washed free of culture medium.
Then 1 x 106 cells/mL in 100 mL grown in suspension are pelleted and washed in phosphate buffered saline. The pellet is resuspended in SSC (0.15M NaCl/0.015 sodium citrate). Sodium dodecyl sulfate (SDS) is added to a final concentration of 1%. Proteinase K (2.5 mg/mL, predigested for 30 min at 37°C) is added to a concentration of 0.01 mg/mL and is incubated for 30 min at 50°C. Proteinase K is   ture is then extracted two times with phenol (redistilled) that has been saturated with SSC and extracted two additional times with a mixture of chlorophorm and isoamyl alcohol (19:1, vol/vol). The nucleic acids are precipitated by addition of two volumes of cold 95% ethanol. The DNA is collected by spooling on a glass rod. The precipitate is resuspended in 2 mL of SSC, and ribonuclease A (2 mg/mL) is added to a final concentration of 0.04 mg/mL. The mixture is incubated at 37°C for 30 min. Proteinase K (2.5 mg/mL) is added to a concentration of 0.025 mg/mL and incubated at 37°C for 30 min. The mixture is then brought to a final volume of between 10 and 15 mL with SSC and extracted two times with phenol and two additional times with the chlorophorm isoamyl alcohol mix indicated above. The DNA is precipitated with 2 volumes of 95% ethanol and spooled. The DNA is then lyophilized.
Eco RI Restriction Enzyme Cleavage of Human DNA.
The lyophilized DNA is resuspended in the Eco RI cleavage buffer (100mM Tris-HCI, pH 7.5; 50mM NaCl; 5mM MgC92; 100 ,ug/mL bovine serum albumin) and is digested with the restriction enzyme, Eco RI, at a concentration of 3 units of enzyme to 1 ,ug of DNA at 37°C for 3 hr. Sodium ethylenediaminetetraacetic acid (EDTA) was added to a final concentration of 25mM to terminate the reaction.
To isolate the alphoid DNA from the bulk of the cellular DNA, the Eco RI digest is centrifuged through a 4.6 mL 15-30% sucrose gradient at 50,000 rpm in a Beckman VTi-65 rotor for 90 min at 20°C.
The buffer for the sucrose gradient is 1X Agarose buffer (see below). The gradient is fractionated, and 20 ,uL of each fraction is analyzed on a 2% agarose gel containing ethidium bromide (0.003 mg/mL). The agarose gel contains 3 g of agarose in 150 mL. The gel buffer is 0.04M Tris-HCl, pH 8.0; 0.018 M NaCI, 0.02M EDTA. The fractions containing the 342 base pair alphoid sequence are located by visualization of the ethidium bromide stained bands using shortwave ultraviolet light. The gel is photographed with a Polaroid MP-3 industrial camera and type 57 (4 x 5) Land film. There should be a discrete band visible in the fractions containing the 342 base pair fragment. These fractions are then pooled and precipitated with ethanol. The precipitate is lyophilized and stored in lOmM Tris-HCl, pH 8.0, lmM EDTA at -20°C. An alternative to visualization of the alphoid sequence using ethidium bromide stained gels is the inclusion of a labeled restriction fragment of suitable length in the sucrose gradient.

Terminal Labeling of the Alphoid 342 Base Pair Fragment
The 3'-termini can be labeled in reactions that contain the Klenow fragment of DNA polymerase (11 units), and 30 puCi each of (ao-32P)dATP and (a-32P)dTTP, 7mM Tris-HCl, pH 7.5, 7mM MgCl2, 50mM NaCl, O.lmM DTT, 0.01mM dGTP, 0.01mM dCTP and that are incubated for 50 min at 10°C. Following this incubation, 0.01mM dATP and O.OmM dTTP (unlabeled) are added, and the reaction mixture is incubated for an additional 10 min at 10°C (20). The DNA is then precipitated by addition of sodium acetate to a final concentration of 0.3M, 10 ,ug of the tRNA and two volumes of 95% ethanol. The mixture is chilled to -70°C in a dry ice-ethanol bath for 15 min. The mixture is then centrifuged at 15,000g in an Eppendorf centrifuge for 15 min. The supernatant is discarded and the pellet is washed with 95% ethanol and then lyophilized.
For labeling of the 5' terminus the DNA is incubated in a reaction containing lOOmM imidizole, pH 6.6, 20mM MgC92, 8mM DTT, 0.2mM spermine, 0.2mM EDTA, 0.6mM ADP, 16mM MgC92, and 1.5 units of polynucleotide kinase. The reaction also includes 250 ,uCi of (y-32P)dATP that has been lyophilized in the tube prior to addition of the other reaction components. The reaction mixture is incubated at 37°C for 60 min. The labeled DNA is precipitated by the addition of ammonium acetate to a final concentration of 0.3M and 2 volumes of 95% ethanol.

Isolation of the 92 Base Pair End-Labeled Eco RI/Eco RI* Fragment of Alphoid DNA
The alphoid DNA which is labeled at both termini is then digested with another restriction enzyme to obtain fragments that are labeled at a single terminus. To obtain a 92 base pair alphoid fragment derived by cleavage of the 342 base pair fragment with Eco RI*, the labeled DNA is resuspended in the Eco RI* reaction buffer which contains 25mM Tris-HCI, pH 8.6, 2mM MgCl2, and 20% glycerol (21). The reaction is incubated for 1 hr at 37°C with 70 units of Eco RI. A mixture of two dyes, bromophenol blue and xylene cyanol, both at a concentration of 0.05%, is added to the reaction mix. This preparation is layered directly onto a polyacrylamide gel for separation of the cleaved fragments.
For isolation of the 92 base pair fragment, the cleaved DNA is layered onto a 5% polyacrylamide gel made with TEB buffer (50mM Tris-HCl, 50mM boric acid, lmM EDTA). The gel is polymerized with 0.07% ammonium persulfate and TEMED (N,N,N'N'-tetramethylethylenediamine). The gel is run at 400 V until the slow-moving dye has migrated halfway down a 40 cm long gel. In order to isolate the gel fragment, the 92 base pair sequence is located by autoradiography and the fragment excised from the gel. DNA is removed from the gel fragment by crushing the gel piece with a siliconized glass rod in a 1.5 mL Eppendorf tube. A solution of 0.3M sodium acetate is added to the crushed gel and the mixture is rotated overnight at room temperature. The gel mixture is spun for 3 min in an Eppendorf centrifuge, and the supernatant is removed and saved. The crushed gel is washed with 0.3M sodium acetate twice until most of the radioactivity has been eluted from the crushed gel. The supernatants are pooled and filtered through a 0.45 pum cellulose acetate filter. The DNA is precipitated by an addition of two volumes of 95% ethanol. The 92 base pair fragment is then ready for use.

Results
The alphoid sequence can be used to compare DNA damage created by treatment of DNA in vitro to damage deposited in the same DNA sequence when that sequence is part of the intact cells (Fig.  12). An experiment of this type is described using ultraviolet light as the insulting agent. The 3' end-labeled 92 base pair Eco RI/Eco RI* alphoid sequence was prepared from cells that were either untreated or treated with 5000 J/m2 of 254 nanometer light from a germicidal lamp. A fraction of the DNA extracted from the untreated cells was also treated with 5000 J/m2 from the same source. These three samples were analyzed on a polyacrylamide gel after (1) no treatment, (2) treatment with a preparation of the ultraviolet-specific endonuclease purified from M. luteus that cleaves the DNA at sites of cyclobutane pyrimidine dimers or (3) treatment with 1M piperidine for 20 min at 90°C [a treatment that cleaves the DNA at PyC lesions (18)] and treatment with the M. luteus enzyme followed by treatment with 1M piperidine as before.
The results of such an experiment are illustrated in Figure 12. The alphoid DNA extracted from untreated cells is not broken by any of the treatments. Treatment of ultraviolet-irradiated alphoid DNA with piperidine or ultraviolet-specific endonuclease creates strand breaks at specific sites which are evident as DNA fragments that migrate more rapidly than the intact 92-mer. The sites of damage in DNA irradiated in vitro and within intact cells can be compared by analysis of the cleavage patterns. The relative amount of damage at each site can be determined by measurement of the amount of radioactivity in gel slices that correspond to discrete bands in the autoradiograms. The amount of damage at six specific sites in the alphoid DNA as a function of increasing ultraviolet dose is plotted in Figures 13 and 14. These data demonstrate that the sites at which DNA damage occurs are similar in ---. the two situations. However, the effective dose seen by the DNA in intact cells is about 50% of the effective dose for the same fragment when it is irradiated as naked DNA. This is the case over the entire dose range of 50-5000 J/m2. This method can also be used to judge the physiological stability of individual DNA lesions. Thus, the PyC and cyclobutane dimer lesions are stable in cellular DNA when the cells are exposed to very high doses, i.e., 5000 J/m2 of ultraviolet light ( Table  2). It is hoped that this method will be useful for detection of intermediates in the DNA repair pathways when lower doses of ultraviolet light are used. This method is capable of detecting damage induced by low ultraviolet fluences. We have detected lesions in alphoid DNA exposed to as little as 15 J/m2 of ultraviolet light.

Summary
The alphoid DNA provides a convenient tool for the analysis of DNA damage within intact cells at the level ofindividual nucleotide sequences. It should provide a sensitive tool for the study of DNA.