Bioconjugate Chem. 1901, 2, 403-406
403
A Novel Technique to Assay Adducts of DNA Induced by Anticancer Agent cis-Diamminedichloroplatinum(11) Minoti Sharma,' Rama Jain, and Thomas V. Isac Department of Biophysics, Roswell Park Cancer Institute, Buffalo, New York 14263. Received May 20, 1991
The dideoxynucleotides d(pGpG) and d(pApG) and the tetradeoxynucleotide d(CpTpApG) were synthesized in solution phase by a modified phosphotriester technique and reacted with the anticancer agent cis-diamminedichloroplatinum(I1)(cisplatin). The major products were isolated by HPLC and characterized by NMR and mass spectrometry as cross-link adducts of cisplatin with the neighboring purine bases. The cross-link adducts of d(pGpG) and d(pApG) were dansylated through a 5'-phosphoramidate linkage with ethylenediammine. The labeling efficiency of the adducts was quantitative as in the case of the normal dinucleotides. The modified tetramer was digested with nuclease PI. The excised adduct was enriched by HPLC and labeled with dansyl chloride. The analysis of the postlabeled adduct by HPCL, using a fluorescence detector, detected a peak with retention time corresponding to that of the dansylated cis-Pt(NH&d(pApG). Cochromatography with the authentic marker confirmed the identification. The same overall procedure was used to assay calf thymus DNA exposed to cisplatin. The major adducts were identified as cis-Pt(NH,),d(pGpG) and cis-Pt(NH&d(pApG). The quantitative labeling efficiency of platinum adducts combined with highly sensitive fluorescence detection technique (subfemtomol) suggests that fluorescence postlabeling assay could be a novel approach for real-time analysis of DNA modification induced by platinated drugs in biological system.
INTRODUCTION
cis-Dia"inedichloroplatinum( 11)(cisplatin) is a widely used chemotherapeutic agent that is effective against ovarian, testicular, and other tumors (1-3). It is now generally accepted that the antineoplastic activity of the drug is based on its interaction with genomic DNA ( 4 ) . Analysis of cisplatin-modified DNA shows two major adducts, cisPt(NH&d(pGpG) and cis-Pt(NH&d(pApG), both derived from intrastrand cross-link of cisplatin and neighboring nucleobases (5). I t is interesting that, in the latter adduct, deoxyadenosine is always the 5'-nucleotide. Investigations of these adducts by NMR spectroscopy reveal that cisplatin is bound to the purine base at the N-7position. Interstrand cross-links are relatively minor lesions (6, 7). The study of the molecular basis of the anticancer agent requires not only the detection but also identification and reliable quantitation of the various adducts derived from exposure of DNA to the agent. Correlation between the presence of specific cisplatin adducts in DNA from nucleated blood cells and clinical response could permit a predictive assay of the effectiveness of the agent for individual patient. Therefore, the need for a highly sensitive and straight-forward assay which would allow the detection, identification, and quantification of the level of adducts distribution in exposed DNA cannot be overstated. Large quantities of material (1108 cells) are needed for reliable determination of cisplatin adducts by atomic absorption spectroscopy (AAS) (8). This precludes many investigations on clinical material. The use of radiolabeled Pt compounds (9) is not suitable for human experimentation. A series of in vitro and in vivo studies were performed to characterize DNA damage recognized by an antiserum elicited against DNA modified with cisplatin. Adducts determined by cisplatin-DNA enzyme linked immunosorbent assay (ELISA) in blood cell DNA have been shown to correlate well with positive clinical outcome in testicular and ovarian cancer patients receiving platinum-drug-based chemotherapy (IO, 1I). However,
* To whom correspondence should be addressed.
with human samples, the characterization of the DNA damage recognized by an antiserum elicited against cisplatin-modified DNA measured a variable fraction of the total DNA-bound Pt determined by AAS study (12).The cisplatin-DNA antiserum has primary specificity for a three-dimensional lesion formed by intrastrand Pt-GG and Pt-AG adducts of DNA and oligomers but does not recognize adducts obtained from digested DNA. A variety of factors that affect a specificregion of DNA could possibly influence what is recognized by the antibody. Recently, Tilby et al. reported highly sensitive detection of DNA modification (1 Pt adduct/106 bases) induced by cisplatin and carboplatin [cis-diammine(1,l-cyclobutanedicarboxylato)platinum(II)] in both in vitro and in vivo using monoclonal antibody (13). But the immunoassay does not permit the resolution of the various DNA adducts formed by these agents. Fichtinger-Schepman et al. reported immunochemical quantitation of cis-Pt(NH&d(pGpG) but not cis-Pt(NH&d(pApG) after analytical separation of the modified DNA digest by FPLC on Mono Q anionexchange column ( 4 1 4 ) . The 32P-postlabelingtechnique introduced by Randerath et al. (15) has been a powerful tool for chromatographic resolution of low-level DNA adducts in experimental animal and human using only 1-5 pg of DNA digest. The successful application of this technique depends, however, on the efficiency of the radiolabeling of the modified substrates with [ Y - ~ ~ P ] A T P and polynucleotide kinase. The major cisplatin-DNA adduct has been reported recently to be phosphorylated only with an efficiency of 0.1% of the corresponding dinucleotide (16). This low level of phosphorylation does not appear encouraging for quantitative determination requiring a high sensitivity. We have developed a novel technique to assay DNA damage by combining enzymatic digestion of the modified DNA with fluorescence postlabeling (17). The fluorescent label is introduced chemically and the procedure, therefore, does not suffer from the vagaries of the enzymatic labeling. We observed that the two major intrastrand cis-Pt(NH& adducts of d(pGpG) and d(pApG) are labeled with dansyl chloride with the same efficiencyas their normal counterparts. The 0 1991 American Chemical Society
404
Sharma et al.
Bioconjugate Chem., Vol. 2, No. 6, 1991
labeling efficiencywas determined by previously published procedures (17,18). This report describes the fluorescence postlabeling assay of cisplatin-modified DNA model d(CTAG) and calf thymus DNA. EXPERIMENTAL PROCEDURES Protected mononucleotide phosphotriesters and protected nucleosides were obtained from Gallard Schesinger Biochemicals. Calf thymus DNA, dansyl chloride, bovine pancrease DNAse I, and nuclease PI from Penecilium citriums were purchased from Sigma. Cisplatin was obtained from Fulka. The syntheses of d(GpG), d(ApG), and d(CpTpApG) were carried out in solution phase by a modified phosphotriester technique previously reported (19). The dinucleoside monophosphates were 5'-phosphorylated by a facile chemical procedure previously described (17). The cisplatin adducts of the di- and tetradeoxynucleotides were prepared by reacting 2 OD units of the nucleotide in 50 pL of sodium phosphate buffer (4 mM, pH 6) with 50 pL of cisplatin in water (1mg/mL). The reaction was allowed to proceed at room temperature for 16 h. The products from the large-scale reactions were isolated by reversephase HPLC (C18,5 pm, 10 mm X 25 cm) using a linear gradient of 0-200/;, acetonitrile in 0.1 M ammonium acetate buffer. The isolated products were desalted in the same system using a linear gradient of 0-100% methanol in water. Calf thymus DNA was reacted with cisplatin and the modified DNA was degraded enzymatically following reported procedures (5). The modified tetramer (2 OD units) was degraded by nuclease PI (20 pg) at 37 "C in 100 pL of sodium acetate buffer (0.1 M, pH 5.5) containing 0.2 mM ZnClz. The FAB mass spectra and the 'H NMR measurements were done with Finnigan MAT 90 and Bruker WP200 spectrometers, respectively. The HPLC conditions used for the various analyses have been described in the figure legends. The authentic markers and the various fractions collected from the digests were labeled with dansyl chloride following reported procedures (17, 18). RESULTS AND DISCUSSION The major platinated adducts from the reactions of cisplatin with d(pGpG) and d(pApG) were isolated in 90% and 70% yields, respectively. The adducts were characterized by FAB mass spectrometry and lH NMR. Although a matrix of triethanolamine facilitated the formation of deprotonated nucleotides during negative-ion FAB mass spectrometry, platinated nucleotides showed poor solubility in this matrix. An equal volume mixture of triethanolamine and glycerol was superior to either pure triethanolamine or pure glycerol for the platinated nucleotide. The observed molecular ions (M-H)- at m / z 901 and 886 for the platinated dinucleotide adducts supported the expected molecular weights for cis-Pt(NH3)zd(pGpG) and cis-Pt(NHz)zd(pApG),respectively. The negative-ion mass spectrum of the monoplatinated tetradeoxynucleotide d(CpTpApG) contained fragment ions indicative of the oligonucleotidesequence and location of platination (Figure 1). The molecular ion (M - H)- at m l z 1400 agreed with the expected molecular weight for the tetramer containing the cross-link cis-Pt(NH&d(pApG) adduct. The molecular ion region of each platinated adduct was characterized by multiple isotope peaks due largely to isotopes of platinum (20). The 'H NMR spectra of both the adducts were in agreement with the reported results (5). Having been
1399.7
I
"1
MI2
Figure 1. Negative-ion FAB mass spectrum of monoplatinated adduct of d(CTAG) with cisplatin.
A.
cisplatin Modified d ( V A G1
I
9
.
.
.
.
I
8
.
.
,
.
l
7
.
.
~
.
~
.
6
8( ppm,TSP) Figure 2. Low-field resonances of 200-MHz lH NMR spectra of (A) cisplatin-modified d(CTAG)and (B)d(CTAG). Solutions were 10 mM base in DzO, pH 6.0 (uncorrected),and referenced
to TSP.
characterized, the dinucleotide adducts served to identify the same lesions from DNA exposed to cisplatin. The downfield resonances of the 'H NMR spectra of the tetradeoxynucleotide d(CTAG), before and after reaction with cisplatin, are shown in Figure 2. The integrated intensity of all the protons are in agreement with 1:l ratio of the four nucleobases in both the spectra. In general, the chemical shifts of the nonexchangeable base protons have been reported to move downfield upon binding of the platinumcompound to the nucleobases (21). As shown in Figure 2, the chemical shifts of both AH-8 and AH-2 protons showed a significant downfield shift from 8.28 and 8.17 ppm before modification to 9.20 and 9.14 ppm, respectively, after modification. The GH-8 proton and
.
.
.
Bioconjugate Chem., Vol. 2, No. 6, 1991 405
Assay of DNA Damage Induced by Cisplatin
?)
E
x
(u
0 0 C
0
e 2 a
n
33.3
26.7
40
Time (mln)
Figure 4. HPLC elution profiles of dansylated nucleotides generated with a reversed-phase C18 column (5 pm,4.6 m m X 25 cm) eluted with a 30-min linear gradient of 15-20% acetonitrile in 0.1 ammonium acetate buffer and detected with a fluorescent detector (excitation 340 nm, emission 520 nm): (A) cis-Pt(NH&d(pGpG) at 29 min and cis-Pt(NH&d(pApG) a t 36 min, (B) nuclease PI digested adduct from platinated d(CTAG), (C) nuclease P1 digested adduct from platinated calf thymus DNA, and (D) nuclease PI digested DNA control. 4
8
I2
16
MODIFIED DNA
20
Time (min ) Figure 3. HPLC elution profiles in a Radial-Pak LC cartridge 8MBC18 (10 pm, 8 mm X 10 cm) eluted with a linear gradient of 0-20% acetonitrile in 0.1 M ammonium acetate buffer. (A) A mixture of dCmp, Tmp, dGmp, dAmp, dC, cis-Pt(NHs)zd(pGpG), and cis-Pt(NH&d(pApG), (B) nuclease P1 digest of platinated d(CTAG),and (C)cochromatographyof nuclease P1digest of platinated d(CTAG) with cis-Pt(NH&d(pApG).
the TH-6 proton also showed downfield shift of 0.24 and 0.10 ppm, respectively, after modification. The CH-5 and CH-6 protons of the self-complementary tetramer, on the other hand, showed upfield shifts suggesting increased shielding of these protons by the aromatic rings of the cross-linked purine bases after modification. Figure 3B shows HPLC analysis of the nuclease PI digest of the platinated d(CpTpApG). The profile contains peaks due to Tmp and dC along with an unknown at 17 min. Nuclease PI has been reported to excise bulky adducts from modified DNA as 5'-phosphorylated dinucleotide whereas the normal nucleotides are released as 5'-monophosphates (22). The peak at 17 min in profile B was identified as cis-Pt(NH&d(pApG) by cochromatography with the authentic marker (see profile 3C). Profile 3A shows the analysis of a mixture of four normal nucleotides, dC, and the two authentic markers cis-Pt(NH&d(pGpG) and cis-Pt(NHa)zd(pApG). The markers, although well resolved from the normal nucleotides, coelute under the HPLC condition used in the Figure 3. The DNA model d(CpTpApG) lacks two adjacent guanine residues. Therefore, the conclusion drawn from the cochromatographic analysis shown in profile C is valid for the identification of the peak at 17 min in profile B. Thus the nuclease PI digestion study supports the spectroscopic data in showing that the cisplatin reacted DNA model contains the crosslink adduct cis-Pt(NH&d(pApG). The peak at 17 min in profile B was collected, lyophilized, and dansylated. HPLC analysis of the postlabeled sample using fluorescence detection showed a peak at 36
[*
. .pNpNpXpNpN. ..] 1 Enzyme D i g e s t i o n pN
t
pX ( X
-
adduct)
I
HPLC Enrichment o f 1 Modified Nucleotide
'1
Fluorescence P o s t l a b e l i n g
a
HPLC Analysis w i t h 1 Fluorescence D e t e c t i o n
F-PX
Figure 5. Scheme of fluorescence postlabeling assay for DNA damage; F denotes fluorescence label.
min (Figure 4B) which coeluted with dansyl-labeled authentic marker cis-Pt(NH&d(pApG) (the results not shown). Profile A in the Figure 4 shows the HPLC analysis of the dansyl-labeled markers cis-Pt(NH&d(pGpG) and cis-Pt(NH&d(pApG). These two markers, which coeluted prior to the labeling (Figure 3A), resolve well after dansylation. The detection and identification of the adduct from cisplatin reacted DNA model study having been achieved, the fluorescence postlabeling technique was extended to assay damage induced by cisplatin in calf thymus DNA. Figure 5 outlines the steps involved in the fluorescence postlabeling assay. Calf thymus DNA was modified with cisplatin and resulted in a total input drug/nucleotide ratio in solution of 0.055. The modified DNA (20 pg) was digested with DNAse I and nuclease PI to excise the adducts. The normal nucleotides were released as 5'-monophosphates. A same-size sample of untreated DNA was also processed simultaneously as a control. The digests were filtered in a centrifugal ultrafree microunit with 10 000 NMWL polysulfone membrane. The filtrates were fractionated by HPLC using the same condition as described in the model study (see Figure 3). The HPLC profiles from the
406
Bloconjugate Chem., Vol. 2, No. 6, 1991
digests showed only the normal nucleotides. Calculation of the peak area of each nucleotide multiplied by the response factor (area/nmol) of the corresponding standard nucleotide showed that the digestion efficiency of the modified DNA was 75 9% of the control value. A fraction was collected from each treated and control DNA digest corresponding to the retention time of the authentic markers as shown in the profile A in Figure 3. Collection of such a fraction from the treated DNA samples, prior to labeling, enriches both the adducts from the normal nucleotides. The collected fractions were lyophilized and dansylated. HPLC analysis (Figure 4) showed that the profile C from the modified postlabeled sample had two peaks in addition to those present in the control (profile D in Figure 4). The peaks at 29 and 36 min in profile C were identified as cis-Pt(NH&d(pGpG) and cis-Pt(NH&d(pApG) respectively by cochromatography with the dansylated authentic markers (results not shown). We observed that in cisplatin-modified calf thymus DNA, the area of the d(pGpG) adduct is twice the area of the d(pApG) adduct. The cross-link adducts constituted nearly 90% of total platinum bound to DNA. A similar observation has been reported for cisplatin-modified salmon sperm DNA using FPLC of the DNA digest followed by measurement of platinum content by AAS (5). The artifactual peaks at 34 and 38 min shown in the control profile D are also present in profiles A-C. The large peak at 25 min is due to excess labeling reagent. The profiles shown in Figure 4C,D correspond to the analysis from 5-pg DNA digests. Recently we have developed an analytical system for fluorescence postlabeling assay by combining microbore HPLC with laserinduced fluorescence detection (18). The preliminary investigation shows that the new system has a subfemtomole detection limit and is 4 orders of magnitude more sensitive than the conventional system used in the current study (23). The present report, though preliminary, suggests that fluorescence postlabeling assay could be a novel approach in correlating clinical outcome with the efficacy of those drugs that bind covalently with DNA resulting in molecular lesions. ACKNOWLEDGMENT
Supported in part by National Cancer Institute Grant CA46896. LITERATURE CITED (1) Harder, H. C., and Rosenberg, B. (1970) Inhibitory effects of anti-tumor platinum compounds on DNA, RNA and protein synthesis in mammalian cells in vitro. Int. J. Cancer 6,207216. (2) Rosecweig, M., Von Hoff, D. D., Slavik, M., and Muggia, F. M. (1977)Cis-diamminedichloroplatinum(I1):A new anticancer drug. Ann. Int. Med. 86, 803-812. (3) Gottlieb, J. A., and Drewinko, B. (1975)Review of the current clinical status of platinum coordination compounds in cancer chemotherapy. Cancer Chemother. Rep. 59,621-628. (4) Roberts, J. J. (1981) Mechanism of action of anti-tumor platinum coordination compounds. Molecular Actions and Targets for Cancer Chemotherapeutic Agents (A. C. Sartorelli, J. S. Lazo, and J. R. Bertino, Eds.) pp 17-43, Academic Press, New York. (5) Fichtinger-Schepman, A. M. J., Van der Veer, J. L., den Hartog, J. H. J., Lohman, P. H. M., and Reedijk, J. (1985) Adducts of the antitumor drug cis-diamminedichloroplatinum(I1)with DNA: Formation, identification, and quantitation. Biochemistry 24, 707-713. (6) Lemaire, M-A., Schwartz, A. M., Rahmouni, A. R., and Leng, M. (1991) Interstrand cross-links are preferentially formed at
Sharma et al.
the d(GC) sites in the reaction between cis-diamminedichloroplatinum(I1) and DNA. Proc. Natl. Acad. Sci. U.S.A. 88, 1982-1985. (7) Hopkins, P. B., Millard, T. J., Woo, J.,Weidner, M. F., Kirchner, J. J., Sigurdsson, S. T., and Raucher, S. (1991) Sequence preferences of DNA interstrand cross-linking agents: Importance of minimal DNA structural reorganization in the crosslinking reaction of merchlorethamine, cisplatin, and mitomycin C. Tetrahedron 47, 2475-2489. (8) Fichtinger-Schepman, A. M. J.,Lohman, P. H. M.,andReedijk, J. (1982) Detection and quantitation of adducts formed upon interaction of cis-diamminedichloroplatinum(I1)with DNA by ion-exchange chromatography after enzymatic degradation. Nucleic Acids Res. 10, 5345-5356. (9) Eastman, A. (1983)Characterization of the adducts produced in DNA by cis-diamminedichloroplatinum(I1). Biochemistry 22,3927-3933. (10) Reed, E., Ozols, F., Tarone, R., Stuart, H. Y., and Poirier, M. C. (1988)The measurement of cisplatin-DNAadduct levels in testilcular cancer patients. Carcinogenesis 9, 1909-1911. (11) Reed, E., Litterst, C. L., Thill, C. C., Yuspa, S. H., and Poirier, M. C. (1987) Platinum-DNA adducts in leukocyte correlate with disease response in ovarian cancer patients receiving platinum-based chemotherapy. Proc. Natl. Acad. Sci. U.S.A. 84, 5024-5028. (12) Reed, E., Gupta-Burt, S., Litterst, C. L., and Poirier, M. C. (1990)Characterization of the DNA damage recognized by an antiserum elicited against cis-diamminedichloroplatinum(I1)modified DNA. Carcinogenesis 11, 2117-2121. (13) Tilby, M. J., Johnson, C., Knox, R. J., Cordell, J., Roberts, J. J., and Dean, C. J. (1991) Sensitive detection of DNA modifications induced by cisplatin and carboplatin in vitro and in vivo using a monoclonal antibody. Cancer Res. 51, 123-129. (14) Fichtinger-Schepman, A. M. J., Baan, R. A., Luiten-Schuite, A., Van Dijk, M., and Lohman, P. H. M. (1985) Immunochemical quantitation of adducts induced in DNA by cis-diamminedichloroplatinum(I1)and analysis of adduct-related DNAunwinding. Chem.-Biol. Interact. 55, 275-288. (15) Randerath, K., Reddy, M. Y., and Gupta, R. C. (1981) 32Ppostlabeling test for DNA damage. Proc. Natl. Acad. Sci. U.S.A. 78, 6126-6129. (16) Hemminki, K., Peltonin, K., and Mustonin, R. (1990) 32Ppostlabeling of 7-methyl-dGmp, ring-opened 7-methyl-dGmp and platinated d(GpGp). Chem.-Biol. Interact. 74, 45-54. (17) Kelman, D. J., Lilga, K. T.,and Sharma, M. (1988)Synthesis and application of fluorescent labeled nucleotides to assay DNA damage. Chem.-Biol. Interact. 66, 85-100. (18) Sharma,M., and Freund, H. G. (1991)Development of laserinduced fluorescence detection to assay DNA damage. In Optical Methods for Ultrasensitive Detection and Analysis: Techniques and Applications (B. L. Fearey, Ed.) Proc. SPIE 1435, pp 280-291, SPIE, Bellingham, WA. (19) Sharma, M., and Box, H. C. (1985)Synthesis, modification with N-acetoxy-2-acetylaminofluorene and physicochemical studies of DNA model compound d(TACGTA). Chem.-Biol. Interact. 56, 73-88. (20) Martin, L. B., 111, Schverner, A. F., and van Breeman, R. B. (1991) Characterization of cisplatin adducts of oligonucleotides by fast atom bombardment mass spectrometry. Anal. Biochem. 193,6-15. (21) Caradonna, J. P., and Lippard, S. J. (1982) The antitumor drug cis-[Pt(NH&C12] forms an intrastrand d(GpG) crosslink upon reaction with [d(ApGpGpCpCpT)]z. J.Am. Chem. SOC. 104,5793-5795. (22) Randerath, K., Randerath, E., Danna, T. F., Van Golen, K. L., and Putnam, K. L. (1989) A new sensitive 32P-postlabeling assay based on the specificenzymatic conversionof bulky DNA lesions to radiolabeled dinucleotides and nucleoside monophosphates. Carcinogenesis 10, 1231-1239. (23) Sharma, M. (1991)A new technique to assay DNA damage using microbore HPLC with laser-induced fluorescence detector. The Pittsburgh Conference, Chicago, IL, March 4-8, Abstracts 072.
