Guanine-06 Methylation Reduces the Reactivity of d(GpG) Towards Platinum Complexes As&id F. Struik, Carla T. M. Zuiderwijk, Jacques H. van Boom, Lars I. Elding, and Jan Reedijk CTMZ, JHVB, JR. Department of Chemistry, Gorlaeus Laboratories, Leiden University, Leiden, The Netherlands.-LIE. Inorganic Chemistry I, Chemical Centre, University of Lund, Lund, Sweden
AFS,
ABSTRACT 06-methylated guanine dinucleotides were used to study the influence of hydrogen bonding on the specific binding of the antitumor drug cDDP, cis-PtCl,(NH,), , to DNA. In this interaction, the guanine-06 site appears to be important in explaining the preference for a pGpG-N7(1),N7(2) chelate, which results from H-bridge formation with the ammine ligand of cDDP. Guanine-06 methylated dinucleotides and the nonmodified dinucleotides were reacted with [Pt(dien)Cl]+, cis-PtCl,(NH,),, and cis-[Pt(NH,),(H,O),]*+ and the reaction products were characterized by ‘H NMR using pH titrations. Methylation at guanine-06 clearly reduces the preference for the guanine. In competition experiments monitored by NMR and experiments using UV spectrophotometry a decreasing reactivity towards [Pt(dien)(H,O)]*+ and c~~-[P~(NH,),(H,O),]~+ was found, in the order of d(GpG) > d(GomepG) > d(GpGome) > d(GomepGome). The difference in reactivity between 5’ guanine methylation and 3’ guanine methylation is ascribed to differences in the H-bond formation with the backbone phosphate. The resulting reduced stacking of the bases in both modified dinucleotides, compared to the bases in d(GpG), results in a preference for the 3’ guanine over 5’.
INTRODUCTION Despite
the fact that Rosenberg
dichloroplatinum
used clinically
(cDDP)
discovered
the antitumor
activity
of cis-diamtnine-
in 1965 [l], and that cDDP (also called cisplatin)
for chemotherapy
since 1972, the mechanism
has been for complex formation
Address reprint requests to: Professor Jan Reedijk, Department of Chemistry, Gorlaeus Laboratories, Leiden University, P.O. Box 9502, 2300 RA Leiden, Leiden, The Netherlands. Journal of Inorganic Biochemistry,
249 44,249-260 (1991) @ 1991 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NY, NY 10010 0162-0134/91/$3.50
250
A. F. Struik et al.
with DNA is not completely understood. Up to now it has been shown that for cDDP-type of compounds, binding to DNA is probably the most important mechanism for their antitumor activity. The active Pt antitumor complexes have several structural similarities, viz. a cis geometry with two moderately strongly bound leaving groups, like chlorides, and two strongly coordinating primary or secondary amine groups [2]. These amine ligands must have at least one NH group. Inside the cell. the chlorides of the complexes are exchanged for aqua ligands. The resulting reactive aqua complexes bind rapidly and covalently to DNA. As a result. cell division is inhibited with the concomitant reduction of the tumor (31. In binding to DNA, cDDP is known to form several different interstrand and intrastrand chelates. The most important adduct is probably the d(pGpG)-N7,Ni. 65% of which is formed after platinum binding; this is far wore than would be expected statistically (37%) 131. The preference for the guanine-N7 could be related to at least two factors: the influence of the 5’ phosphate [4] and guanine-06 [5], both of which can form hydrogen bonds to the amine groups It;]> and the nature of the neighboring bases [7] to this GG sequence. In this paper the influence of the hydrogen bonding on the preference of binding to the guanine-N7 is studied. using the 06-methylated guanine dinucleotides, d(G”n”pG) deoxyguanosine methylated at the 06). d(GpG”‘“‘” ). and (dG Onle denotes d(G”““pG”““). These modified DNA fragments will hamper hydrogen bonding between the amine ligand and the guanine-06. To eliminate possible complications when using bifunctional Pt-compounds, [Pt(dien)Cl]Cl (where dien stands for diethylenetriamine), a model compound for the first binding step of cDDP [S]. is used to analyze and calculate the product distribution. The reactivities of the N7 atoms of the modified and nonmodified guanines are compared by use of competition experiments analyzed by NMR and kinetic experiments monitored by I_V spectrophotometry. Several groups have investigated the kinetics between nucleosides or nucleobases and cDDP 19, lo-19J, but only a tew results have been published concerning dinucleotides 1211. In this paper the reactions between 06methylated guanine and nonmodified guanine dinucleotides and [Pt(dien)Clj t and /Pt(dienj(H,G)].‘ ’ are studied. Finally the interaction of cDDP and [cis-Pt(NN,),(H,0)21”+ with 06 methylated dinucleotides is examined and compared with d(GpG). Figure 1 presents a schematic picture of the modification d(G”“epG). The atomic numbering of the bases and sugar rings follow the recommendations bv the IUPACIUP Commission of Biochemical Nomenclature 1201.
FIGURE 1. d(GomepG).
Schematic
representation
of
06-methylated
GUANINE-06
AND PLATINUM
COMPLEXES
251
Materials The 06-methylated guanine dinucleotides were synthesized via an improved phosphodiester method [21, 221, introducing the methyl group on the 06 according to Gaffney et al. [23]. Instead of diphenylacetyl the more alkaline labile phenylacetyl group was used to protect the amine group of guanine [24]. The dinucleotides were deprotected with a mixture of 1,8-diazabicyclo[5.4.O]undec-7-ene (DBU) and methanol in dioxane. Purification was performed by a DEAE sephadex A25 (Pharmacia) anionic exchange column (HCO,-form) chromatography. Elution was executed with a linear gradient of 0 to 0.7 M triethylammonium hydrogencarbonate (TEAB) and a flow rate of 12 ml per hr. The fractions were analyzed by FPLC, using a mono Q HR 5/5 LCCJOO column. The deprotected and purified dinucleotides were brought in the sodium form by passing them through a column containing Dowex WX4 (Na+ form). No further attempts to desalt the modified dinucleotides were undertaken because of considerable loss in yield. The remaining salt (mainly triethylammonium hydrogencarbonate) had no effect on the ‘H NMR spectra and the platinum-DNA reaction. For the competition reactions the modified dinucleotides were desalted using a HiLoad Sephadex SlOO HR26/60 column, eluted with 0.15 M TEAB (1.5 ml/min). The concentrations were determined with UV spectroscopy (Zeiss H30DS), using the molar absorption of nonmodified GpG. cDDP and [Pt(dien)Cl]Cl were prepared from K,PtCl, according to literature procedures [25, 261. Solutions of the corresponding aqua complexes were prepared by adding two equivalents of AgNO, to aqueous solutions of the chloro compounds in the dark. To a drop of the reaction mixtures, HCl or AgNO, was added to check whether Ag+ or Cl- ions were still present in the mixtures. The concentrations of the platinum aqua complexes were determined gravimetrically as dried AgCl precipitate [9].
NMR Experiments The modified and nonmodified dinucleotides d(G”““pG), d(GpGome), d(GomepGome), and d(GpG) (in 4.10w3 M and 6.10p5 M aqueous solutions) were reacted with one and two equivalents of [Pt(dien)Cl]Cl at pH between 5 and 6 (during the complete reaction). After one week in the dark at 37°C the 6.10e5 M solutions were concentrated by evaporation of the solvent and the samples of both concentrations were lyophilized three times from 99.7% D,O. The NMR samples were finally dissolved in 99.95% D,O. No ionic medium was used for the NMR experiments. In addition, competition experiments were carried out in a 1: 1:l ratio, for modified dinucleotides GpG and [Pt(dien)Cl]Cl, respectively (pH 6-7). These experiments were also performed at concentrations of 4.10e3 M and 6.10p5 M. To calculate the relative product distribution, the peak areas of the H8 resonances were measured with a Compensating Polar Planimeter (Kent). One equivalent of cDDP was reacted during one or two weeks with 6.1O-5 M modified dinucleotides at room temperature. The resulting reaction mixtures were evaporated and lyophilized three times from 99.7% D,O and recorded in 99.95% D,O. The reactions were also carried out with one equivalent of cisin an NMR tube (4.10p3 M) at 37°C (pH 6-7). WNH,MW3,2+
252
A. F. Struik et a/.
The ’ H NMR spectra (300 MHz) were recorded on a Bruker WM 300 spectrome,ter, connected to an Aspect 2000 computer. Resolution enhancement was performed by applying a Gaussian multiplication according to Ernst [27], except for spectra where the relative amounts of the different products were calculated. To assign the products, the pH* (pH’ denotes the uncorrected meter reading of solutions in D,G) dependence of the nonexchangeable base protons was followed by adding small amounts of 0.1 M DC1 and 0.1 M NaOD. Chemical shift values (set Table I) were recorded relative to TMA (3.18 ppm downfield from DSS) Kinetic
UV Measurements The DNA fragments (in 1.10 ’ M aqueous solutions) were reacted with [Pt(dien)(H20)]” (pH between 4.8 and 5.0 during the complete reaction time) between 36.6”C and 37°C. To ensure pseudo-first-order conditions an excess was used. The ionic strength was kept at 0.1 (20--500 times) of [Pt(dien)(H,O)]‘+ M with NaClO, (Merck). No buffers were used, to avoid possible complications due to coordination with platinum. Reactions were monitored at 280 and 254 nm by use of a Varian Cary 2200 spectrophotometer connected to a Victor PC; data were collected and analyzed using an OLIS Spectroscopy Operating System, version 9.12 (Ohs Inc.. Jefferson, U.S. ). Product distributions after the first and second reaction steps were checked with ‘14 NMR spectroscopy.
RESULTS Reactions
AND DISCUSSION of the Modified Dinucleotides
with [Pt(dien)Cl]Cl
The ‘H NMR results of the reactions with the 06-methylated guanine modifications are summarized in Table 1. Identification of the products formed upon platinum binding was based on NMR and pH titrations. The pH dependence of the nonexchangeable base protons of the unreacted dinucleotides is depicted in Figures 2, 3, and 4 (parts A). The chemical shift changes at pH 10.0 and pH 2.5 are ascribed to the protonation of guanine-Nl and guanine,-N7, respectively, of nonmodified guanine HP protons [28]. Due to methylation of the guanine-06, the electron distribution in the purine ring changes. so that protonation of the Nl becomes less favorable. The Nl protonation of 06-methylated guanine bases is likely to occur below pH 2. The titration curves of these modified bases display only one chemical shift change at pH 3.4, most likely due to an N7 protonation. Assignments of the 5’ and 3’ 06-methylated guanines in d(GomepGome) were performed by comparing the chemical shifts of d(G”““pG) and d(GpG”“‘“) with the NMR spectrum of d(G”““pG”““). After one week of reaction of the modified dinucleotides with ]Pt(dien)Cl]+. under the same conditions as above, two products, 1 and 2, are formed as shown by the “H NMR spectra. The pH titrations of the nonexchangeable base protons, represented in Figures 2, 3, and 4, parts B (1) and C (2). were used to assign the reaction products. At low pH, no chemical shift changes, induced by N7 proton&on. ate observed for the 5’ guanine H8 resonances of products I a, b, and E (parts B). The Nl deprotonation effects for the nonmodified guanine bases have shifted to lower pH (around pH 8) as usually found [28]. For products 2 a, b, and c (parts C) a similar effect on the 3’ guanine H8 signals is found. Furthermore, all the other H8 signals
GUANINE-06
AND PLATINUM
TABLE 1. ‘H NMR Data at pH 6-7 (198 K) for the Modified Dinucleotides, and the Products Formed After Reaction with [Pt(dien)Cl]Cl and ck-Pt(NH,),(H,O),*+
6 O-W Products d(GompC) Pt(dien){d(GomepG)-N7(1)} la Pt(dien){d(G”‘pCi)-N7(2)} 2a [Pt(dien)]*{ F-d(G”“pG)-N7(l),N7(2)} cis-Pt(NH,),{d(GomepG)-N7(l),N7(2)}
38 4s
Pt(dien){d(GpGome)-N7(2)} 2b [Pt(dien)],{ cl-d(GpG”“)-N7(l),N7(2)} cis-Pt(NH,)2{d(GpG”m’)-N7(l),N7(2)} d(GomepGome) Pt(dien){d(GomCpG0me)-N7(l)} lc Pt(dien){d(GomepGome)-N7(2)} 2c [Pt(dien)]2{p-d(Go”‘epGome)-N7(l),N7(2)} cis-Pt(NH,),{d(GomepGome)-N7(l),N7(2)} The chemical
shifts (6) and the chemical
L.6
3b 4b
3c 4c shift differences
d(GpG),
6 0-w
5’
A6
3’
AS
4.46 5.29 4.82 5.36 5.28
+0.62 +0.15 +0.69 +0.61
4.78 4.69 5.21 5.32 5.41
-0.09 +0.43 +0.54 +0.63
4.56
d(GpG”“‘Y Pt(dien){d(GpGomC)-N7(1)} lb
253
COMPLEXES
4.91
5.15 4.60 5.26 5.05
+0.59 +0.04 +0.70 +0.49
4.88 5.32 5.40 5.25
-0.03 +0.41 +0.49 +0.34
4.62 5.23 4.72 5.32 5.01
+0.61 +0.10 +0.70 +0.39
4.82 4.81 5.31 5.40 5.32
-0.07 +0.43 +0.52 +0.50
(A 6) are given in ppm (relative to TMA)
L.6 !
-7-7-T
-
PH
-
PH
FIGURE 2. A-D. The pH dependence of the chemical shifts (6) of the nonexchangeable base protons in free d(G”“pG), Pt(dien){d(GomepG)-N7(1)} la, Pt(dien)(d(G”“‘pG)-N7(2)} 2a, Pt(dien)l,{ pd(GomepG)-N7(1),N7(2)} 3a. 5’ H8 protons (x ); 3’ H8 protons (0).
254
A. F. Struik et al.
FIGURE 3. A-D. The pH dependence of the chemical shifis (6) ot the nonexchangeable base protons in free d(GpG”““), Pt(dien)(d(GpG I’“‘“)-N7( 1)) lb. Pt(dien)id((;pG”““)-N7(1iJ 2b, Pt(dien)j,{ cc-d(GpG”‘“’ )-N7( 1),N7(2)} 3b. 5’ H8 protons ( x !; 3’ HX prcXons (:‘ ).
are hardly affected by platinum binding, proving that only one guanine is bound in each product. These observations indicate that reaction with [ Pt(dien)Clj’ results in both the platinated complexes 5’ guanine-N7 Cl a. b, and cl and platinatcd J guanine-N7 (2 a, b, and c). After assignment of H8 protons, the product distribution could be calculated from the peak areas of the base resonances. To ensure that no proton exchange of the H8 protons had occurred, the spectra were taken within 24 hr after dissolving the platinated dinucleotides in D,O. The relative amounts of adducts. formed after the 1: 1 reaction with [Pt(dien)Cl] +, are summarized in Scheme I, The recent results of Lempers et al. [28i shows that in reacrions between monofunctional platinum compounds and d(GpG), platination of the 5’ guanine is favored. Comparing these results with those above. it is evident that methylation of the the guanine-06 decreases the reactivity of the guanine-N7 atom. The d(GpG”““) modification results in a decrease in the preference for the 3 base of 15% in comparison with d(GpG); for d(Gilm“ pG) a decrease of as much as 35% for the 5’ guanine is observed. B> methylating both guanines-04 in d(G”““pCY’““) a sumilar reactivity ratio for the 5’ and 3’ guanine was found compared to nonmodified d(GpG), assuming that the experimental error in the measurements of the peak dress is approximately 5 %). When a second equivalent of [Pt(dien)CIJi was added to the 1--to-l reaction products, one end product (3) for each modified dinucleotide was found. The large downfield shifts of the HX protons after reaction with platinum and the pH titrations (see Figs. 2, 3, and 4; parts 11) prove the presence of dinucieotides containing two
GUANINE-06
i,_,
, L
, -
, 8
,
AND PLATINUM
COMPLEXES
255
4
,
I
+I
12
8 -
PH
I
T
12
PH
FIGURE 4. A-D. The pH dependence of the chemical shifts (6) of the nonexchangelc, F’t(dien) able base protons in free d(GomepG Ome), Pt(dien){d(GomepGome)-N7(1)} {d(GomepG”m”)-N7(2)} 2c, F’t(dien)],{ ~-d(GomepG“me)-N7(1),N7(2)} 3c. 5’ H8 protons (x); 3’ H8 protons (0).
platinum fragments. None of the H8 base protons showed a chemical shift change at low pH due to N7 protonation, giving further evidence for the formation of the dinuclear [Pt(dien)],{ p-d(modified-dimer)-N7(1),N7(2)} products. No evidence for bonding to platinum via Nl is found after addition of a large excess of [Pt(dien)Cl]+. The chemical shift changes are not in agreement with Nl platination in e.g. d(ApG) [29]. This indicates that the chemical shift changes for the 06-methylated guanines at low pH (3.4) is most likely caused by N7 protonation.
Pt
Pt
1”
I
GPG
GPG
G”p G
GDG
Pt
Pt I G-p G”
I GpG’ 80
60 /
// G*pG*
G pG* 7’ \
20
GpG’
SCHEME 1. Schematic representation of d(GpG), d(GomepG), d(GpGom’), and d(GomepGome) in a l-to-l ratio with [F’t(dien)Cl]Cl. G* denotes 06-methylated guanine. Values are given in percentages.
256
A. F. Struik et al.
Competition
Experiments
To obtain a better insight into the reactivity of the methylated bases with respect to [Pt(dien)Cl] +, competition experiments were carried out in which modified dinucleotides and nonmodified d(GpG) reacted with the complex. The modified dinucleotides were first desalted by use of a sephadex column resulting in products still containing about 60% triethylammonium salts; this did not disturb the measurements. In large volumes (6.10 ’ M) and also in NMR tubes (4.10 ’ M) d(GpG), 06-methylated guanine dinucleotides and [Pt(dien)Cl]Cl were allowed to react in a 1: i : 1 ratio during a period of one week. The nonexchangeable H8 protons in the reaction mixtures could be assigned on the basis of the results described above. Immediately after dissolving the mixtures in D,G, the peak areas of the H8 protons were plotted and calculated, see Scheme 2. For each competition experiment the results of Lempers et al. [38] on nonmodified d(GpG) could he reproduced within experimental error. These competition data again proved a significant preference of [Pt(dien)ClJ+ for binding to N7 in nonmodified guanines. Doubly methylated d(G”““pG”““‘i. in competition with d(GpG), gave hardly any platinated 06methylated guanine, whereas the modification of d(G”me pG) showed a reactivity close to that of d(GpGj. Summarizing these experiments, a decreasing reactivity was observed with respect to [Pt(dien)ClJ ’ following the order: d(GpG) > d(G”“‘epG) > d(GpGOi”e) > d(G 01”epG O”X )
The difference in reactivity for d(G”““pG) and d(GpG”“‘“) may appear unexpected when the results with the difference in reactivity between 5’ guanine and 3’ guanine in nonmodified d(GpG) are considered. One would expect a much higher reactivity for d(GpG”““) with a nonmodified 5 guanine than for d(G”“‘“pG) with no methylation on the 3’ guanine. This discrepancy can be explained by assuming a difference in conformation between nonmodified d(GpG) and 06methylated guanine dinucleotides. In d(GpG) the purine bases are efficiently stacked so that the phosphate does not have any significant influence on the binding of platinum to the dinucleotide. In the modified dinucleotides, however. the bases are less stacked [30. 3 I, 321, so that the phosphate can play a role in the binding of platinum by forming a hydrogen bond to the amine group on platinum, Kinetic Experiments Analysis of the absorbance vs time data for the reactions between the (non-)modihed DNA fragments and [Pt(dien)(H20)]*‘-, revealed two consecutive reactions for each dinucleotide, interpreted by the two exponentials in Eq. (3) of Scheme 3 133. 341. Plots of the observed pseudo first-order rate constants for the slow and fast
P!
+ G*p G -
PI
I
+ GpG‘-
0 p G + G p G .
Pt I GpG
L5 Pi .
G*p G‘-
Pt
I
36
GPG
PI
I G p G
56
11
2s
I + GpG”.
GpG-
?6
Pt I + GpG
Pt
23
Pt
?2
G
Pt I
1
AFS,
ABSTRACT 06-methylated guanine dinucleotides were used to study the influence of hydrogen bonding on the specific binding of the antitumor drug cDDP, cis-PtCl,(NH,), , to DNA. In this interaction, the guanine-06 site appears to be important in explaining the preference for a pGpG-N7(1),N7(2) chelate, which results from H-bridge formation with the ammine ligand of cDDP. Guanine-06 methylated dinucleotides and the nonmodified dinucleotides were reacted with [Pt(dien)Cl]+, cis-PtCl,(NH,),, and cis-[Pt(NH,),(H,O),]*+ and the reaction products were characterized by ‘H NMR using pH titrations. Methylation at guanine-06 clearly reduces the preference for the guanine. In competition experiments monitored by NMR and experiments using UV spectrophotometry a decreasing reactivity towards [Pt(dien)(H,O)]*+ and c~~-[P~(NH,),(H,O),]~+ was found, in the order of d(GpG) > d(GomepG) > d(GpGome) > d(GomepGome). The difference in reactivity between 5’ guanine methylation and 3’ guanine methylation is ascribed to differences in the H-bond formation with the backbone phosphate. The resulting reduced stacking of the bases in both modified dinucleotides, compared to the bases in d(GpG), results in a preference for the 3’ guanine over 5’.
INTRODUCTION Despite
the fact that Rosenberg
dichloroplatinum
used clinically
(cDDP)
discovered
the antitumor
activity
of cis-diamtnine-
in 1965 [l], and that cDDP (also called cisplatin)
for chemotherapy
since 1972, the mechanism
has been for complex formation
Address reprint requests to: Professor Jan Reedijk, Department of Chemistry, Gorlaeus Laboratories, Leiden University, P.O. Box 9502, 2300 RA Leiden, Leiden, The Netherlands. Journal of Inorganic Biochemistry,
249 44,249-260 (1991) @ 1991 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NY, NY 10010 0162-0134/91/$3.50
250
A. F. Struik et al.
with DNA is not completely understood. Up to now it has been shown that for cDDP-type of compounds, binding to DNA is probably the most important mechanism for their antitumor activity. The active Pt antitumor complexes have several structural similarities, viz. a cis geometry with two moderately strongly bound leaving groups, like chlorides, and two strongly coordinating primary or secondary amine groups [2]. These amine ligands must have at least one NH group. Inside the cell. the chlorides of the complexes are exchanged for aqua ligands. The resulting reactive aqua complexes bind rapidly and covalently to DNA. As a result. cell division is inhibited with the concomitant reduction of the tumor (31. In binding to DNA, cDDP is known to form several different interstrand and intrastrand chelates. The most important adduct is probably the d(pGpG)-N7,Ni. 65% of which is formed after platinum binding; this is far wore than would be expected statistically (37%) 131. The preference for the guanine-N7 could be related to at least two factors: the influence of the 5’ phosphate [4] and guanine-06 [5], both of which can form hydrogen bonds to the amine groups It;]> and the nature of the neighboring bases [7] to this GG sequence. In this paper the influence of the hydrogen bonding on the preference of binding to the guanine-N7 is studied. using the 06-methylated guanine dinucleotides, d(G”n”pG) deoxyguanosine methylated at the 06). d(GpG”‘“‘” ). and (dG Onle denotes d(G”““pG”““). These modified DNA fragments will hamper hydrogen bonding between the amine ligand and the guanine-06. To eliminate possible complications when using bifunctional Pt-compounds, [Pt(dien)Cl]Cl (where dien stands for diethylenetriamine), a model compound for the first binding step of cDDP [S]. is used to analyze and calculate the product distribution. The reactivities of the N7 atoms of the modified and nonmodified guanines are compared by use of competition experiments analyzed by NMR and kinetic experiments monitored by I_V spectrophotometry. Several groups have investigated the kinetics between nucleosides or nucleobases and cDDP 19, lo-19J, but only a tew results have been published concerning dinucleotides 1211. In this paper the reactions between 06methylated guanine and nonmodified guanine dinucleotides and [Pt(dien)Clj t and /Pt(dienj(H,G)].‘ ’ are studied. Finally the interaction of cDDP and [cis-Pt(NN,),(H,0)21”+ with 06 methylated dinucleotides is examined and compared with d(GpG). Figure 1 presents a schematic picture of the modification d(G”“epG). The atomic numbering of the bases and sugar rings follow the recommendations bv the IUPACIUP Commission of Biochemical Nomenclature 1201.
FIGURE 1. d(GomepG).
Schematic
representation
of
06-methylated
GUANINE-06
AND PLATINUM
COMPLEXES
251
Materials The 06-methylated guanine dinucleotides were synthesized via an improved phosphodiester method [21, 221, introducing the methyl group on the 06 according to Gaffney et al. [23]. Instead of diphenylacetyl the more alkaline labile phenylacetyl group was used to protect the amine group of guanine [24]. The dinucleotides were deprotected with a mixture of 1,8-diazabicyclo[5.4.O]undec-7-ene (DBU) and methanol in dioxane. Purification was performed by a DEAE sephadex A25 (Pharmacia) anionic exchange column (HCO,-form) chromatography. Elution was executed with a linear gradient of 0 to 0.7 M triethylammonium hydrogencarbonate (TEAB) and a flow rate of 12 ml per hr. The fractions were analyzed by FPLC, using a mono Q HR 5/5 LCCJOO column. The deprotected and purified dinucleotides were brought in the sodium form by passing them through a column containing Dowex WX4 (Na+ form). No further attempts to desalt the modified dinucleotides were undertaken because of considerable loss in yield. The remaining salt (mainly triethylammonium hydrogencarbonate) had no effect on the ‘H NMR spectra and the platinum-DNA reaction. For the competition reactions the modified dinucleotides were desalted using a HiLoad Sephadex SlOO HR26/60 column, eluted with 0.15 M TEAB (1.5 ml/min). The concentrations were determined with UV spectroscopy (Zeiss H30DS), using the molar absorption of nonmodified GpG. cDDP and [Pt(dien)Cl]Cl were prepared from K,PtCl, according to literature procedures [25, 261. Solutions of the corresponding aqua complexes were prepared by adding two equivalents of AgNO, to aqueous solutions of the chloro compounds in the dark. To a drop of the reaction mixtures, HCl or AgNO, was added to check whether Ag+ or Cl- ions were still present in the mixtures. The concentrations of the platinum aqua complexes were determined gravimetrically as dried AgCl precipitate [9].
NMR Experiments The modified and nonmodified dinucleotides d(G”““pG), d(GpGome), d(GomepGome), and d(GpG) (in 4.10w3 M and 6.10p5 M aqueous solutions) were reacted with one and two equivalents of [Pt(dien)Cl]Cl at pH between 5 and 6 (during the complete reaction). After one week in the dark at 37°C the 6.10e5 M solutions were concentrated by evaporation of the solvent and the samples of both concentrations were lyophilized three times from 99.7% D,O. The NMR samples were finally dissolved in 99.95% D,O. No ionic medium was used for the NMR experiments. In addition, competition experiments were carried out in a 1: 1:l ratio, for modified dinucleotides GpG and [Pt(dien)Cl]Cl, respectively (pH 6-7). These experiments were also performed at concentrations of 4.10e3 M and 6.10p5 M. To calculate the relative product distribution, the peak areas of the H8 resonances were measured with a Compensating Polar Planimeter (Kent). One equivalent of cDDP was reacted during one or two weeks with 6.1O-5 M modified dinucleotides at room temperature. The resulting reaction mixtures were evaporated and lyophilized three times from 99.7% D,O and recorded in 99.95% D,O. The reactions were also carried out with one equivalent of cisin an NMR tube (4.10p3 M) at 37°C (pH 6-7). WNH,MW3,2+
252
A. F. Struik et a/.
The ’ H NMR spectra (300 MHz) were recorded on a Bruker WM 300 spectrome,ter, connected to an Aspect 2000 computer. Resolution enhancement was performed by applying a Gaussian multiplication according to Ernst [27], except for spectra where the relative amounts of the different products were calculated. To assign the products, the pH* (pH’ denotes the uncorrected meter reading of solutions in D,G) dependence of the nonexchangeable base protons was followed by adding small amounts of 0.1 M DC1 and 0.1 M NaOD. Chemical shift values (set Table I) were recorded relative to TMA (3.18 ppm downfield from DSS) Kinetic
UV Measurements The DNA fragments (in 1.10 ’ M aqueous solutions) were reacted with [Pt(dien)(H20)]” (pH between 4.8 and 5.0 during the complete reaction time) between 36.6”C and 37°C. To ensure pseudo-first-order conditions an excess was used. The ionic strength was kept at 0.1 (20--500 times) of [Pt(dien)(H,O)]‘+ M with NaClO, (Merck). No buffers were used, to avoid possible complications due to coordination with platinum. Reactions were monitored at 280 and 254 nm by use of a Varian Cary 2200 spectrophotometer connected to a Victor PC; data were collected and analyzed using an OLIS Spectroscopy Operating System, version 9.12 (Ohs Inc.. Jefferson, U.S. ). Product distributions after the first and second reaction steps were checked with ‘14 NMR spectroscopy.
RESULTS Reactions
AND DISCUSSION of the Modified Dinucleotides
with [Pt(dien)Cl]Cl
The ‘H NMR results of the reactions with the 06-methylated guanine modifications are summarized in Table 1. Identification of the products formed upon platinum binding was based on NMR and pH titrations. The pH dependence of the nonexchangeable base protons of the unreacted dinucleotides is depicted in Figures 2, 3, and 4 (parts A). The chemical shift changes at pH 10.0 and pH 2.5 are ascribed to the protonation of guanine-Nl and guanine,-N7, respectively, of nonmodified guanine HP protons [28]. Due to methylation of the guanine-06, the electron distribution in the purine ring changes. so that protonation of the Nl becomes less favorable. The Nl protonation of 06-methylated guanine bases is likely to occur below pH 2. The titration curves of these modified bases display only one chemical shift change at pH 3.4, most likely due to an N7 protonation. Assignments of the 5’ and 3’ 06-methylated guanines in d(GomepGome) were performed by comparing the chemical shifts of d(G”““pG) and d(GpG”“‘“) with the NMR spectrum of d(G”““pG”““). After one week of reaction of the modified dinucleotides with ]Pt(dien)Cl]+. under the same conditions as above, two products, 1 and 2, are formed as shown by the “H NMR spectra. The pH titrations of the nonexchangeable base protons, represented in Figures 2, 3, and 4, parts B (1) and C (2). were used to assign the reaction products. At low pH, no chemical shift changes, induced by N7 proton&on. ate observed for the 5’ guanine H8 resonances of products I a, b, and E (parts B). The Nl deprotonation effects for the nonmodified guanine bases have shifted to lower pH (around pH 8) as usually found [28]. For products 2 a, b, and c (parts C) a similar effect on the 3’ guanine H8 signals is found. Furthermore, all the other H8 signals
GUANINE-06
AND PLATINUM
TABLE 1. ‘H NMR Data at pH 6-7 (198 K) for the Modified Dinucleotides, and the Products Formed After Reaction with [Pt(dien)Cl]Cl and ck-Pt(NH,),(H,O),*+
6 O-W Products d(GompC) Pt(dien){d(GomepG)-N7(1)} la Pt(dien){d(G”‘pCi)-N7(2)} 2a [Pt(dien)]*{ F-d(G”“pG)-N7(l),N7(2)} cis-Pt(NH,),{d(GomepG)-N7(l),N7(2)}
38 4s
Pt(dien){d(GpGome)-N7(2)} 2b [Pt(dien)],{ cl-d(GpG”“)-N7(l),N7(2)} cis-Pt(NH,)2{d(GpG”m’)-N7(l),N7(2)} d(GomepGome) Pt(dien){d(GomCpG0me)-N7(l)} lc Pt(dien){d(GomepGome)-N7(2)} 2c [Pt(dien)]2{p-d(Go”‘epGome)-N7(l),N7(2)} cis-Pt(NH,),{d(GomepGome)-N7(l),N7(2)} The chemical
shifts (6) and the chemical
L.6
3b 4b
3c 4c shift differences
d(GpG),
6 0-w
5’
A6
3’
AS
4.46 5.29 4.82 5.36 5.28
+0.62 +0.15 +0.69 +0.61
4.78 4.69 5.21 5.32 5.41
-0.09 +0.43 +0.54 +0.63
4.56
d(GpG”“‘Y Pt(dien){d(GpGomC)-N7(1)} lb
253
COMPLEXES
4.91
5.15 4.60 5.26 5.05
+0.59 +0.04 +0.70 +0.49
4.88 5.32 5.40 5.25
-0.03 +0.41 +0.49 +0.34
4.62 5.23 4.72 5.32 5.01
+0.61 +0.10 +0.70 +0.39
4.82 4.81 5.31 5.40 5.32
-0.07 +0.43 +0.52 +0.50
(A 6) are given in ppm (relative to TMA)
L.6 !
-7-7-T
-
PH
-
PH
FIGURE 2. A-D. The pH dependence of the chemical shifts (6) of the nonexchangeable base protons in free d(G”“pG), Pt(dien){d(GomepG)-N7(1)} la, Pt(dien)(d(G”“‘pG)-N7(2)} 2a, Pt(dien)l,{ pd(GomepG)-N7(1),N7(2)} 3a. 5’ H8 protons (x ); 3’ H8 protons (0).
254
A. F. Struik et al.
FIGURE 3. A-D. The pH dependence of the chemical shifis (6) ot the nonexchangeable base protons in free d(GpG”““), Pt(dien)(d(GpG I’“‘“)-N7( 1)) lb. Pt(dien)id((;pG”““)-N7(1iJ 2b, Pt(dien)j,{ cc-d(GpG”‘“’ )-N7( 1),N7(2)} 3b. 5’ H8 protons ( x !; 3’ HX prcXons (:‘ ).
are hardly affected by platinum binding, proving that only one guanine is bound in each product. These observations indicate that reaction with [ Pt(dien)Clj’ results in both the platinated complexes 5’ guanine-N7 Cl a. b, and cl and platinatcd J guanine-N7 (2 a, b, and c). After assignment of H8 protons, the product distribution could be calculated from the peak areas of the base resonances. To ensure that no proton exchange of the H8 protons had occurred, the spectra were taken within 24 hr after dissolving the platinated dinucleotides in D,O. The relative amounts of adducts. formed after the 1: 1 reaction with [Pt(dien)Cl] +, are summarized in Scheme I, The recent results of Lempers et al. [28i shows that in reacrions between monofunctional platinum compounds and d(GpG), platination of the 5’ guanine is favored. Comparing these results with those above. it is evident that methylation of the the guanine-06 decreases the reactivity of the guanine-N7 atom. The d(GpG”““) modification results in a decrease in the preference for the 3 base of 15% in comparison with d(GpG); for d(Gilm“ pG) a decrease of as much as 35% for the 5’ guanine is observed. B> methylating both guanines-04 in d(G”““pCY’““) a sumilar reactivity ratio for the 5’ and 3’ guanine was found compared to nonmodified d(GpG), assuming that the experimental error in the measurements of the peak dress is approximately 5 %). When a second equivalent of [Pt(dien)CIJi was added to the 1--to-l reaction products, one end product (3) for each modified dinucleotide was found. The large downfield shifts of the HX protons after reaction with platinum and the pH titrations (see Figs. 2, 3, and 4; parts 11) prove the presence of dinucieotides containing two
GUANINE-06
i,_,
, L
, -
, 8
,
AND PLATINUM
COMPLEXES
255
4
,
I
+I
12
8 -
PH
I
T
12
PH
FIGURE 4. A-D. The pH dependence of the chemical shifts (6) of the nonexchangelc, F’t(dien) able base protons in free d(GomepG Ome), Pt(dien){d(GomepGome)-N7(1)} {d(GomepG”m”)-N7(2)} 2c, F’t(dien)],{ ~-d(GomepG“me)-N7(1),N7(2)} 3c. 5’ H8 protons (x); 3’ H8 protons (0).
platinum fragments. None of the H8 base protons showed a chemical shift change at low pH due to N7 protonation, giving further evidence for the formation of the dinuclear [Pt(dien)],{ p-d(modified-dimer)-N7(1),N7(2)} products. No evidence for bonding to platinum via Nl is found after addition of a large excess of [Pt(dien)Cl]+. The chemical shift changes are not in agreement with Nl platination in e.g. d(ApG) [29]. This indicates that the chemical shift changes for the 06-methylated guanines at low pH (3.4) is most likely caused by N7 protonation.
Pt
Pt
1”
I
GPG
GPG
G”p G
GDG
Pt
Pt I G-p G”
I GpG’ 80
60 /
// G*pG*
G pG* 7’ \
20
GpG’
SCHEME 1. Schematic representation of d(GpG), d(GomepG), d(GpGom’), and d(GomepGome) in a l-to-l ratio with [F’t(dien)Cl]Cl. G* denotes 06-methylated guanine. Values are given in percentages.
256
A. F. Struik et al.
Competition
Experiments
To obtain a better insight into the reactivity of the methylated bases with respect to [Pt(dien)Cl] +, competition experiments were carried out in which modified dinucleotides and nonmodified d(GpG) reacted with the complex. The modified dinucleotides were first desalted by use of a sephadex column resulting in products still containing about 60% triethylammonium salts; this did not disturb the measurements. In large volumes (6.10 ’ M) and also in NMR tubes (4.10 ’ M) d(GpG), 06-methylated guanine dinucleotides and [Pt(dien)Cl]Cl were allowed to react in a 1: i : 1 ratio during a period of one week. The nonexchangeable H8 protons in the reaction mixtures could be assigned on the basis of the results described above. Immediately after dissolving the mixtures in D,G, the peak areas of the H8 protons were plotted and calculated, see Scheme 2. For each competition experiment the results of Lempers et al. [38] on nonmodified d(GpG) could he reproduced within experimental error. These competition data again proved a significant preference of [Pt(dien)ClJ+ for binding to N7 in nonmodified guanines. Doubly methylated d(G”““pG”““‘i. in competition with d(GpG), gave hardly any platinated 06methylated guanine, whereas the modification of d(G”me pG) showed a reactivity close to that of d(GpGj. Summarizing these experiments, a decreasing reactivity was observed with respect to [Pt(dien)ClJ ’ following the order: d(GpG) > d(G”“‘epG) > d(GpGOi”e) > d(G 01”epG O”X )
The difference in reactivity for d(G”““pG) and d(GpG”“‘“) may appear unexpected when the results with the difference in reactivity between 5’ guanine and 3’ guanine in nonmodified d(GpG) are considered. One would expect a much higher reactivity for d(GpG”““) with a nonmodified 5 guanine than for d(G”“‘“pG) with no methylation on the 3’ guanine. This discrepancy can be explained by assuming a difference in conformation between nonmodified d(GpG) and 06methylated guanine dinucleotides. In d(GpG) the purine bases are efficiently stacked so that the phosphate does not have any significant influence on the binding of platinum to the dinucleotide. In the modified dinucleotides, however. the bases are less stacked [30. 3 I, 321, so that the phosphate can play a role in the binding of platinum by forming a hydrogen bond to the amine group on platinum, Kinetic Experiments Analysis of the absorbance vs time data for the reactions between the (non-)modihed DNA fragments and [Pt(dien)(H20)]*‘-, revealed two consecutive reactions for each dinucleotide, interpreted by the two exponentials in Eq. (3) of Scheme 3 133. 341. Plots of the observed pseudo first-order rate constants for the slow and fast
P!
+ G*p G -
PI
I
+ GpG‘-
0 p G + G p G .
Pt I GpG
L5 Pi .
G*p G‘-
Pt
I
36
GPG
PI
I G p G
56
11
2s
I + GpG”.
GpG-
?6
Pt I + GpG
Pt
23
Pt
?2
G
Pt I
1
