160
Chem. Res. Toxicol. 1988, 1 , 160-168
Construction of an Escherichia coli Vector Containing the Major DNA Adduct of Activated Benzo[a]pyrene at a Defined Site Matt Benasutti, Z. Diala Ezzedine, and Edward L. Loechler* Department of Biology, Boston University, Boston, Massachusetts 02215 Received December 10, 1987
T h e mutagenic and carcinogenic substance benzo[a]pyrene reacts with DNA following activation t o its corresponding 7,8-diol 9,lO-epoxide (BPDE), and the major DNA adduct (BPN2-Gua) is formed when the C(10)-position of BPDE reacts with the N2-position of guanine. I t is unknown if this adduct is a premutagenic lesion in vivo. Herein, the construction and characterization of an M13mplS-based, E. coli vector that contains BP-N2-Gua located in the unique PstI restriction endonuclease recognition site at nucleotide position 6249 in the (-)-strand is described (designated, BP-N2-Gua-M13mpl9). First, the oligonucleotide 5’-TGCA-3’ was reacted with BPDE and a product (5’-T(BP-N2)GCA-3’)was isolated by HPLC that, when enzymatically digested to deoxynucleosides, yielded an adduct that comigrated on HPLC with an authentic BP-N2-Gua deoxynucleoside standard. Second, the 5’-hydroxyl group of 5’-T(BP-N2)GCA-3’was phosphorylated with ATP and T4 polynucleotide kinase, and the product (5’-pT(BP-N2)GCA-3’)was purified by HPLC. This product is stable when heated a t 80 “ C a t both neutral and alkaline pH. Third, M13mp19 was manipulated such that the sequence 5’-pTGCA-3’ was selectively removed from the (-)-strand in its unique PstI recognition site, and 5’-pT(BP-N2)GCA-3’was ligated into this gap with T 4 DNA ligase and ATP. The product of this reaction (BP-N2-Gua-M13mpl9) was shown to be insensitive t o clevage by PstI, which suggests that a modification is located in the PstI recognition site. The most likely modification is the adduct BP-N2-Gua.
Introduction Most mutagenic substances have been shown also to be carcinogenic, and the relationship between these two phenomena have become clearer in recent years; oncogenes are derived from their normal cellular counterparts, protooncogenes, via mutation (1-4). Mechanisms of mutagenesis have been studied in a variety of organisms and systems and have been shown both to occur spontaneously and to be induced by exogenous agents (5). Considerable progress has been made in the understanding of spontaneous mutagenesis where mutagenic mechanisms have been proposed for depurination (6-8) and deamination reactions of DNA (9),for polymerase errors (10-131, and for complex mutagenic events based upon the formation of unusual DNA structures (13-15). Simple alkylating agents induce mutations via oxygen adducts, principally 06-alkylguanine (16-20) and possibly 04-alkylthymine (21, 22). In contrast the fundamental basis of mutagenesis by UV light is controversial (23-26) and is totally unknown in the case of ionizing radiation (27, 28). Many of the more potently mutagenic and carcinogenic substances are bulky and three-dimensionally complex, such as benzo[a]pyrene (BP),’ aflatoxin B,, and 2aminofluorene ( 5 ) , and the mechanisms by which they induce mutations are not well understood. BP, a polycyclic aromatic hydrocarbon, reacts with DNA following cellular activation to its corresponding 7,8-diol 9,lO-epoxide (BPDE) (29-33), although in some cells other activation pathways may be dominant (34). The majority of BPDE adducts in vivo are derived from the (+)-anti derivative ((+)-anti-BPDE; Figure 1)(2S33). (+)-anti-BPDEforms primarily W-guanine adducts (BP-N2-Gua; Figure 1) (35-38), but BP-N(7)-Gua (39,40) and BP-N6-Ade (41) are * To whom correspondence should be addressed.
also formed at significant levels. In addition BPDE forms phosphotriesters (36, 42) and a cytosine adduct that is currently unidentified (38, 40). The adducts of (*)-anti-BPDE are known to be lethal, presumably by blocking DNA replication (43). In addition BP and (*I-anti-BPDE are known to be mutagenic and induce both base pairing and insertion/deletion mutations (44-47). The mutational specificity of (f)-anti-BPDE in bacteria has been determined both in the lacl system developed by Miller (48) and in a system developed by Ames (49),where GC to TA and AT to TA mutations are most prevalent. Recently, the mutagenic specificity of (*I-anti-BPDE has been determined in human cells by using a shuttle vector (50), where the specificity is reasonably similar to that determined in bacteria except that GC to CG and AT to TA mutations appear to be more and less prevalent, respectively. Finally, some information is available about the mutations induced in oncogenes (51-54).
As outlined above, considerable information is known about both the adduct profile and the mutational specificity of (A)-anti-BPDE;however, little is known about the relationship between these two. This relationship is further confused by the fact that some (or all) of these adducts may form breakdown products, most significantly Abbreviations: BP, benzo[a]pyrene; (+)-anti-BPDE, (+)-r-7,t-8-di-
hydroxy-t-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (anti) (Figure 1); BP-N2-Gua,the trans addition product between the C(lO)-position of
(+)-anti-BPDE and the N2-position of guanine (Figure 1); AP-sites, apurinic/apyrimidinic sites; 5’-pT(BP-N2)GCA-3’,an oligonucleotide containing BP-N2-Gua;BP-N2-Gua-M13mp19,the vector M13mp19 with the oligonucleotide 5’-pT(BP-N2)GCA-3’ligated into the PstI recognition site in the (-)-strand; Gua-M13mpl9, the vector M13mp19 with the oligonucleotide 5’-pTGCA-3’ligated into the PstI recognition site in the (-)-strand; DNase-I, deoxyribonuclease I; CIP, calf intestinal alkaline phosphatase; SVPD, snake venom phosphodiesterase; HAP, hydroxylapatite.
0893-228~/88/2701-0160$01.50/0 0 1988 American Chemical Society
Chem. Res. Toxicol., Vol. 1, No. 3, 1988 161
Construction of an E. coli Vector Pstl
J.
CTGCAG
d s M13mp19
ai
(+)-anti-BPDE
(+)-anti -BP - N 2 Gua (Uans. addition)
a-l
s s M13mp19
(-)-anti-BPDE
(+)-anti-BP-N2Gua (& addition)
LpTGCA-3’
F i g u r e 1. Structures of (+)-anti-BPDE and (-)-anti-BPDE and the N2-guanine adducts of (+)-anti-BPDE derived from trans and cis addition (structure I and 11, respectively).
apurinic/apyrimidinic sites (AP-sites) (39,401, which are known to be premutagenic lesions (6-8,55). One approach that is being employed to probe the relationship between DNA adducts and mutation is to build individual adducts of known structure into DNA vectors at defined genome locations in vitro, then to place these vectors into cells where cellular processing can occur in vivo, and finally to isolate progeny DNA in order to determine if mutations resulted at or near the original genome location of the adduct. This approach has been used successfully in the study of the mutagenic consequences of 06-methylguanine (16-20), 06-butylguanine (18), 04-methylthymine (21), inosine (20), thymine glycol2 and 2-(acetylamino)fluorene3 in vivo, and AP-sites (56,57)and 8-hydroxyguanine (58) in vitro (reviewed in ref 59). Other vectors have been constructed that contain individual adducts and work on their mutagenic consequences is in progress (59-63). Herein we describe building the major DNA adduct of (+)-anti-BPDE (i.e., BP-N2-Gua)into the double stranded form of the E. coli phage, M13mp19 (64),by using a combination of chemical synthetic and recombinant DNA techniques.
Materials and Methods All manipulations involving (+)-anti-BPDE and involving DNA in the presence of ethidium bromide were conducted under yellow lights in order to minimize photodegradation reactions. T4 DNA polymerase, T4 polynucleotide kinase, T4 DNA ligase, PstI, and other restriction endonucleases were obtained from New England Biolabs. Calf intestinal alkaline phosphatase (CIP) was obtained from Pharmacia. Snake venom phosphodiesterase (phosphodiesterase I from Crotalus atrox; EC 3.1.4.1.) was obtained from Sigma (type VII). DNase-I from bovine pancreas (EC 3.1.21.1.) was obtained from Sigma (type 11-s). (f)-r-7,t-8Dihydroxy-t-9,10-epoxy-7,,8,9,10-tetrahydrobenzo[a]pyrene (anti) (Le., (+)-anti-BPDE) was obtained from Chemsyn Science Laboratories through the National Cancer Institutes Carcinogen Reference Standard Repository. Unlabeled (catalogue number, P702) and 3H-labeled (catalogue number, R702H) (+)-anti-BPDE were from lots CSL-83-344-49-5and CSL-83-344-84,respectively. All other chemicals and materials were manufacturers highest grade purity. Gapped Duplex Genome Construction. The construction of the vector containing BP-N2-Gua involved the ligation of a (+)-anti-BPDE-modified oligonucleotide into a four-base, sinBasu, A. K., Loechler, E. L., Leadon, S. A,, and Essigmann, J. M., personal communication. Romano, L. J., and King, C. M., personal communication.
Ligase + ATP
-
CTGCAG G ACGT C
7
BP-N 2 - G u a - M 1 3 m ~ 1 9 Figure 2. Strategy for situating BP-N2-Guain the (-)-strand of the PstI recognition site of M13mp19 (see text); the product is BP-N2-M13mp19. gle-stranded gap in the (-)-strand of the PstI recognition site of M13mp19 (Figure 2) (16, 17); this product is referred to as a gapped duplex, and its construction occurs as follows. M13mp19 (60 pg) was treated with 180 units of PstI to give linear DNA with a four-base, 3’-overhanging end of sequence, 5’-TGCA-3’ (Figure 2). These four bases were removed in a blunt-ending reaction with T4 DNA polymerase; material cleaved with PstI (60 wg/mL) in 50 mM Tris-HC1, pH 8.0/50 mM NaCl/10 mM MgC1, was incubated a t 37 “C with 3 units of T 4 DNA polymerase for 30 min to remove approximately 22 bases from both 3’-ends. After 30 min an additional 3 units of T 4 DNA polymerase were added with dATP, dCTP, dGTP (120 pM each) and d T T P (60 wM). DNA synthesis proceeded for 30 min at 37 “C after which the solution was adjusted to 12 mM disodium EDTA and cooled on ice and the DNA precipitated with ethanol. This material was denatured/renatured with 300 pg of singlestranded circular M13mp19 DNA (a tenfold molar excess of (+)-stranded material) by a procedure described previously (16), giving a duplex molecule missing the bases 5’-pTGCA-3’ in the (-)-strand at the PstI site (Figure 2), as well as starting materials (i.e., circular (+)-strands and PstI, blunt-ended linears) and a small amount of displaced linear (+)-strands. Double-stranded material was purified from contaminating single-stranded DNA by hydroxylapatite (HAP) chromatography and contains a mixture of gapped duplex (Le., M13mp19 missing the sequence 5’-pTGCA-3’ in the (-)-strand of the PstI reconition site; Figure 2) and linear, PstI blund-ended material in the ratio of approximately 1O:l. A 3.6-mL HAP column was used. Single-stranded material was eluted in six 1.5-mL fractions with 0.17 M phosphate; single strands were predominantly found in fractions 2 and 3. Double-stranded material was eluted in 15 0.5-mL fractions with 0.40 M phosphate; double strands were predominantly found in fractions 4-6. The phosphate solutions were equimolar mixtures of mono- and disodium phosphate. The gapped duplexes were dialyzed, ethanol precipitated, resuspended (10 mM Tris-HC1 (pH 8.0), 1 mM EDTA and 20% glycerol), and stored at -20 “C. Synthesis of an Oligonucleotide Containing BP-N2-Gua. The tetranucleotide ~’-HoT,G,C,A~H-~’ (5’-TGCA-3’) was synthesized on a Milligen 6500 DNA synthesis device using phosphoramidite chemistry. Following deprotection, 5’-TGCA-3’ was purified twice by HPLC; a linear 1-10.570aqueous ethanol gradient was employed over 30 min in 0.1 M NH,OAc, pH 5.8, a t room temperature with a Waters C-18 p-Bondapak column (flow rate, 1 mL/min). 5’-TGCA-3’ eluted at 32 min (data not shown).
Benasutti et al.
162 Chem. Res. Toxicol., Vol. 1, No. 3, 1988
1
I
L 0
20
40 Minutes
60
Figure 3. HPLC chromatogram (254 nm) showing the purification of the oligonucleotidecontaining BP-N2-Gua (designated, 5’-T(BP-N2)GCA-3’),which eluted a t 45 min (see Materials and Methods for details). All of the peaks eluting between 37 and 58 min showed absorbance a t 254 and 313 nm as well as 3Hcounts; the latter two are indicative of material containing the BP moiety. This chromatogram shows the purification of ‘/2 of the total sample, and the peak height was 0.066 OD on a Waters Model 440 detector equipped with a 1-cm pathlength flow cell. The peak eluting a t 45 min (5’-T(BP-N2)GCA-3’) from both purifications had approximately 200 000 cpm and following SEP-PAK purification the yield was estimated to be approximately 0.9% from starting material 5’-TGCA-3’. Approximately 250 pg (210 nmol) of 5’-TGCA-3’ was reacted with 195 pg (588 nmol) of (+)-anti-BPDE for 2 h a t room temperature in 50 mM potassium acetate, pH 5.2. The total volume of the reaction was 480 pL and the (+)-anti-BPDE was added in 110 pL and was 0.047 Ci/mmol. Fourteen ether extractions were performed to remove byproducts, and the aqueous phase was purified by HPLC (Figure 3). The HPLC conditions employed were a linear 10.5-20% aqueous ethanol gradient over 30 min in 0.1 M NH,OAc, pH 5.8, at room temperature with a Waters C-18 p-Bondapak column (flow rate, 1 mL/min). Peaks eluting at 37 and 45 min were collected and desalted with SEP-PAK, C18 cartridges by the following procedure. The SEP-PAK cartridge was washed in 5 mL of 80% acetonitrile and 10 mL of 1% acetonitrile; the sample containing the peak of interest was rotary evaporated to 1mL, then concentrated to 500 pL with a Speed Vac concentrator, loaded onto the cartridge, washed with 7 mL of water, and eluted in 5 mL of 50% acetonitrile. The 5-mL sample was concentrated to 100 pL with a Speed Vac concentrator. The material from the major peak eluting a t 45 min contained approximately 1.8 nmol (approximate yield, 0.9%) and proved to be the adducted oligonucleotide of interest (see below). A portion of the purified material that eluted a t 45 min in Figure 3 (designated, 5’-T(BP-N2)GCA-3’)was digested overnight with SVPD (0.011 units) and calf intestinal alkaline phophatase (CIP; 17 units) to mononucleosides. (SEP-PAK purification was performed in some cases but not in the one reported in this paper. The SEP-PAK purification step eliminated the enzymes but decreased the yield of adduct and did not affect the clarity of the HPLC chromatogram in the vicinity of the adduct. Thus,we have now eliminated it from our protocol.) The deoxynucleoside products were separated by HPLC using three different chromatographic conditions. One procedure (A) was described by Marnett and co-workers as being optimal for the seperation of (f)-anti-BPDE-deoxynucleosideadducts (65). The second (B) was a variant of this procedure that was designed to further improve resolution: 45-50% aqueous methanol in 40 min (0.6 mL/min), isocratic for 30 min (0.6 mL/min), and then 50-60% methanol in 20 min (1.0 mL/min) a t room temprature with a Beckman Ultrasphere ODS column (Figure 4). The third procedure (C) was 45% aqueous methanol (isocratic) for 45 min (1.0 mL/min), followed by a linear gradient to 60% methanol over 60 min (1mL/min) a t room temperrature with a Beckman U1trasphere ODS column. The deoxynucleoside adduct from 5’T(BP-N2)GCA-3’eluted a t approximately 35, 58, and 125 min using conditions A, B, and C, respectively, and comigrated in each case with authentic (+)-anti-BP-N2-Guastandard. (+)-anti-BP-N2-Gua standard was prepared by reacting (+)-anti-BPDE with calf thymus DNA, followed by digestion with DNase-I, SVPD, and CIP and purification of deoxynucleoside adducts by HPLC. (+)-anti-BP-N2-Guastandard was identified as the major adduct and was confirmed by sQlvent partitioning pK, determination (36),which gave pK, 9.7. We found ethyl acetate to be a poor choice for the organic phase of this procedure and used 25% diethyl ether, 25% n-butanol, and 50% methyl ethyl ketone instead.
50 OO
20
40
60
80
Minutes
Figure 4. HPLC chromatogram showing the comi ation of the
T
deoxynucleoside adduct(s) isolated from 5’-T(BP-N )GCA-3’ and authentic (+)-anti-BP-N2-Gua(see Materials and Methods for details). Authentic (+)-anti-BP-N*-Guaadduct standard (not containing 3H-label) was coinjected with the 3H-labeled, deoxynucleoside adduct(s) derived from 5’-T(BP-N?)GCA-3’by digestion with SVPD and CIP (see Materials and Methods). The tracing shows absorbance a t 313 nm, while the column gra h shows 3H-counts from the adduct derived from 5’-T(BP-N )GCA-3’. Absorbance at 313 nm was 0.002 OD in height (see legend to Figure 3) of which approximately 70% was contributed by the adduct standard; the adduct derived from 5’-T(BP-N2)GCA-3’yielded a total of 1333 cpm. One-minute fractions were collected from 0 to 50 min and from 67 to 90 min, while half-minute fractions were collected from 50 to 67 min. The OD curve is a tracing and does not give a good indication of the level of noise, which was approximately OD.
B
The material eluting a t 45 min in Figure 4 (i.e., 5’-T(BPN2)GCA-3’)was phosphorylated in the presence of 69 pM ATP ( I O nmol; [y-32P]-containingof specific activity 39.5 Ci/mmol) by using T4 polynucleotide kinase (50 units) for 45 min a t 37 “C (140 mM Tris, 20 mM MgC12, 10 mM DTT, pH 8.0) to give 5’-pT(BP-N2)GCA-3’. Purification (utilizing two separate HPLC separations) was performed as described for 5’-T(BP-N2)GCA-3’, except that a 15-min isocratic step was included before the gradient was started, which permitted 5’-pT(BP-N2)GCA-3’(elution time, 46 min) to be somewhat better resolved from [32P]-ATP. A small amount (4%) of unphosphorylated 5’-T(BP-N2)GCA-3’ remained and eluted at 51 min (Figure 3). (Following purification, the HPLC column had to be purged for approximately 10 h to lower 32P-countsto 1000 cpm per mL.) The peak eluting a t 46 min in Figure 4 was desalted by using SEP-PAK cartridges as follows. The SEP-PAK cartridge was washed in 5 mL of 80% acetonitrile and 10 mL 1%acetonitrile; the sample containing the peak of interest (in 1 mL) was loaded in 25 mM triethylammonium carbonate (pH 7.0), washed with 11mL of water, and eluted in 5 mL of 30% acetonitrile. (Loading in the presence of triethylammonium carbonate helps phosphorylated oligonucleotides elute in a smaller volume. A larger wash is included to remove all salt. A second elution with 5 mL of 30% acetonitrile usually gives an additional 10% material.) The 5-mL sample was rotary evaporated to 1mL and then concentrated to 100 pL on a Speed Vac concentrator. The product (0.34 nmol) was isolated (overall yield, 0.16%). pT(BP-N2)GCA-3’(40 ng, 0.03 nmol) was analyzed by HPLC in order to determine the amount of contaminating 5’-pTGCA-3’ that remained in the sample. To ensure the recovery of low levels of material, 1.0 pg of cold 5’-pTGCA-3’ was added. A linear 1-20% aqueous ethanol gradient was employed over 60 min in 0.1 M NH,OAc, pH 5.8, a t room temperature with a Waters C-18 pBondapak column (flow rate, 1 mL/min). 5’-pTGCA-3’ eluted at 30 min. The level of contamination was estimated to be
