SvJ strain: identification of the promoter region and a comprehensive examination of tissue expression

Biochimica et Biophysica Acta 1678 (2004) 163 – 169 www.bba-direct.com

Short sequence-paper

Structural organization of the murine microsomal glutathione S-transferase gene (MGST1) from the 129/SvJ strain: identification of the promoter $ region and a comprehensive examination of tissue expression Michael J. Kelner a,*, Richard D. Bagnell a, Ralf Morgenstern b a

Department of Pathology, University of California, 8320, UCSD Medical Center, 200 West Arbor Drive, San Diego CA, USA b Institute of Environmental Medicine, Karolinska Institutet, S-17177, Stockholm, Sweden Received 2 July 2003; received in revised form 2 February 2004; accepted 1 March 2004 Available online 21 March 2004

Abstract The structure and regulation of the murine microsomal glutathione transferase gene (MGST1) from the 129/SvJ strain is described and demonstrates considerable difference in nucleotide sequence and consequently in restriction enzyme sites as compared to other mouse strains. A comparison of the amino acid sequence for MGST1 revealed one difference in exon 2 between the 129/SvJ strain (arginine at position 5) and the sequence previously reported for the Balb/c strain (lysine). The promoter region immediately upstream of the dominant first exon is functional, transcriptionally responds to oxidative stress, and is highly homologous to the human region. Oxidative stress also induced the production of endogenous MGST1 mRNA. The tissue-specific expression of MGST1 mRNA was studied, and as anticipated, was indeed highest in liver. There was, however, marked mRNA expression in several tissues not previously studied including smooth muscle, epidymus, ovaries, and endocrine glands in which the expression of various peroxidases is also very high (salivary and thyroid). Overall, there was a good agreement between the mRNA content detected and previous reports of MGST1 activity with the exception of brain tissue. D 2004 Elsevier B.V. All rights reserved. Keywords: MGST1; Promoter; Tissue

1. Introduction The glutathione transferases (GSTs) (E.C. 2.5.1.18) are a complex super gene family of enzymes catalyzing the reaction of the tripeptide glutathione with a wide range of endogenous and xenobiotic lipophilic electrophiles [1,2] The three membrane-bound or microsomal GSTs (MGST1, MGST2, MGST3) are also classified as members of the MAPEG (Membrane Associated Proteins involved in Eicosanoid and Glutathione metabolism). Other members of the MAPEG family include 5-lipoxygenase activating enzyme (FLAP), Leukotriene C4 synthase (LTC4S) and prostaglandin E synthase (PGES) [3].

$ The complete sequence of the MGST1 gene derived from the 129/ SvJ strain has been deposited in GenBank with accession number AY329626. * Corresponding author. Tel.: +1-619-543-5976; fax: +1-619-5433730. E-mail address: [email protected] (M.J. Kelner).

0167-4781/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbaexp.2004.03.001

MGST1 was first identified in 1982 [4], and displays non-selenium glutathione-dependent peroxidase activity [5]. Substrates for MGST1 include lipid and fatty acid hydroperoxides (linoleic acid/ester hydroperoxide), lipid peroxidation products (4-hydroxynon-2-enol), oxidized phospholipids (linoleoyl-palmitoyl phosphatidylcholine hydroperoxides and cholesterol linoleate hydroperoxides) [6 –9]. Thus, MGST1 demonstrates non-selenium dependent hydroperoxide activity similar to the selenium-dependent hydroperoxide activity of the phospholipid hydroperoxide glutathione peroxidase enzyme (GPX4), but differs in its subcellular location. MGST1 also conjugates a variety of potentially harmful xenobiotics including halogenated hydrocarbons, activated esters, and unsaturated carbonyls [10,11]. Although GST reactions are generally a detoxification mechanism, the nephrotoxicity of several haloalkenes is attributed to hepatic microsomal formation of GSH-conjugates, metabolism of these GSH-conjugates to corresponding cysteine-S-conjugates, subsequent translocation to the kidney, followed by bio-

164

M.J. Kelner et al. / Biochimica et Biophysica Acta 1678 (2004) 163–169

activation of the cysteine S-conjugates to a nephrotoxic intermediate [12]. Chlorotrifluoroethene is a potent nephrotoxin that undergoes such a hepatic microsomal bioactivation, and in rodent liver cells 85% of the GSH conjugation with chlorotrifluoroethene is catalyzed by MGST1 [13]. The regulation of MGST1 is complex and differs from other members of the MAPEG family [14]. The most closely related member to MGST1 is PGES, based on shared amino acid identity, but their organ distribution is quite distinct [15]. Both the MGST1 and PGES promoters are TATA-less and contain SP elements [14,16,17]. Whereas both Sp1 and Sp3 proteins appear important for basal expression of PGES, only Sp1 is critical for MGST1 basal expression [16,17]. The marked expression of MGST1 in the liver is due to the role of an upstream HNF element [16]. Recent studies suggest that MGST1 expression is increased in human prostate tumors, and that marked polymorphisms exist in the human gene in transcription-factor binding sites, including critical SP1 binding sites [18,19]. In order to further define the role of MGST1 in contributing to halogenated hydrocarbon nephrotoxicity, as well as provide insight into the unique tissue-specific expression of this enzyme and potential role in carcinogenesis, we recovered the murine MGST1 gene and investigated its tissue distribution in non-hepatic tissue. This information will allow generation of genetically modified mice with reduced expression of MGST1.

2. Materials and methods 2.1. Materials Human liver hepatoma HepG2 [20] and Chinese hamster ovarian derived AA8 cells [21] were obtained from the ATCC (Rockville, MD). The pGL3-enhancer, pSV-galactosidase, pGEM3, sequencing grade Taq polymerase, and fmol thermocycling DNA sequencing kit were obtained from Promega (Madison, WI). The A.L.L. lysis and luciferase reagents were obtained from Analytical Luminescence Laboratory (San Diego, CA). The lipofectin and galactosidase reagents were obtained from Gibco/BRL (Gaithersburg, MD). The P1 plasmid, with an insert of f110 kb containing the entire murine MGST1 gene from the 129/SvJ strain, was isolated and purified as previously described [22]. Murine multiple tissue arrays for examination of mRNA expression were obtained from BD Biosciences Clontech (Palo Alto, CA). 2.2. Cell culture HepG2 and AA8 cells were maintained in DMEM media, 10% bovine serum, and 2 mM glutamine. As variation in glutathione peroxidase activity influences cellular response to paraquat [23], the media was supplemented

with 50 ng/ml of selenium. This eliminates variation in endogenous cellular glutathione peroxidase due to posttranscriptional regulation by selenium content [24,25]. 2.3. Transfection HepG2 or AA8 cells were plated in 35-mm six-well plates at 100,000 cells per well and incubated overnight. Transfection mixture was freshly prepared by mixing equal amounts of solution A (2 Ag of the pGL3-derived plasmid, 3 Ag of SV-galactosidase vector, 100 Al of DMEM media) and solution B (10 Al of lipofectin in 100 Al of DMEM), incubating for 15 min at room temperature, then adding 800 Al of DMEM. The pGL3-derived plasmid was always co-transfected with the SV-galactosidase vector to correct for variations in transfection efficiency. Cells were rinsed twice with PBS, once with DMEM, then the lipofection transfection solution (1 ml) was overlaid onto the cells. Cells were incubated for 5 h at 37 jC in a humidified incubator. Transfection solution was removed by washing with PBS, cells removed by trypsinization, combined, then replated to allow for uniform distribution of transfected cells. Normal growth media (2 ml) containing various quantities of paraquat was added and cells were allowed to incubate for 18 h before luciferase and galactosidase assays. 2.4. Luciferase expression vector construction A 1.2-kb fragment contained the desired genomic region immediately upstream of the first exon was subcloned into pBluescript KS( ) and transferred into the pGL3-enhancer in the correct orientation to produce the desired luciferase expression vector pGLmMGST1(+)-luciferase. All inserted genomic regions in the pGL3 expression vectors were bidirectionally resequenced using Sanger’s dideoxynucleotide chain termination method to confirm identity. 2.5. Luciferase assay Media was removed, cells were washed three times with ice-cold PBS, 400 Al of 1:3 diluted A.L.L. lysis buffer added, followed by agitation for 15 min at 4 jC. Fluid was transferred to a microfuge tube, centrifuged at maximum speed for 5 min at 4 jC, and supernatant removed for luciferase analysis using a Monolight 2010 luminometer (Analytical Luminescence Laboratory). 2.6. Beta-galactosidase assay Media was removed, cells washed three times with icecold PBS, and 400 Al of diluted reporter lysis buffer was added. Cellular debris was removed by scraping into a microfuge tube and supernatant collected. To 150 Al of supernatant was added 150 Al of 2 galactosidase assay buffer, and mixture allowed to incubate for 2 h at 37 jC.

M.J. Kelner et al. / Biochimica et Biophysica Acta 1678 (2004) 163–169

The reaction was stopped with 50 Al of sodium carbonate and the absorbance determined at 420 nm. All values were corrected for endogenous galactosidase cellular activity. 2.7. Northern dot blot analysis Northern dot blot analysis of mouse multiple tissue expression array (BD Biosciences) was used to allow a direct comparison to reports previously describing the expression of other MAPEG enzymes such as PGES-1, which used this technique [15]. The mRNA amount for each tissue of the array was previously normalized by the manufacturer to yield similar hybridization signals for eight housekeeping genes. Three different arrays were used to determine relative tissue expression of MGST1. The membranes were hybridized with a random primed cDNA probe of mouse MGST1 recovered from a Balb/c cDNA hepatic lambda phage library. This cDNA probe was labeled using the Stratagene Prime-It RmT Random Primer Labeling Kit and a-32dCTP (3000 Ci/mmol) per manufacturer’s instructions. The probe was purified using a Qiaquick Nucleotide Removal Kit. The array was pre-hybridized in Express Hyb solution with sheared salmon testes DNA for 1 h. The radiolabeled MGST1 cDNA probe added, followed by hybridization overnight at 65 jC with continuous rolling. The array was washed five times with 2 SSC plus 1% SDS at 65 jC for 20 min, followed by two washed with 0.1 SSC plus 0.5% SDS for 20 min in a hybridization oven with rotation at 55 jC. The array blot was mounted in a phosphoimaging cassette with a prepped phosphoimaging screen for 24 h, and images then quantified on a Storm 80 phosphoimager. The results for MGST1 mRNA expression were standardized to GAPDH expression. 2.8. RT-PCR analysis RT-PCR was performed using the Amplifluork Detection System (Serologicals, Inc.). All primers for RT-PCR were designed using Primer Expressk 1.5 from Applied Biosystems. The assay was performed in a 25-Al reaction made up of 12.5 Al of 2 Platinum QPCR Supermix-UDG (Invitrogen), 7.5 Al of primers and probe (Integrated DNA Technologies), and 5 Al of template. The master mix includes PlatinumR Taq DNA Polymerase, ROX internal control, UDG (to prevent carryover contamination), dNTPs, and buffers. The primer concentrations were 10 nM for the 5V primer, 100 nM for the 3V primer, and 100 nM for the 6FAM Amplifluor Uniprimer. The primers for MGST1 were TTGGCCTCCTGTATTCCTTGA and TGCTCCGACAAATAGTCTGAAGTG. To determine the relative increase of MGST1 upon oxidative stress, the HepG2 cells were exposed to paraquat at 7 Ag/ml for up to 24 h, and RNA was then harvested for RT-PCR analysis. Results were again standardized to GAPDH mRNA expression.

165

2.9. Sequence and statistical analysis Computer analysis of upstream MGST1 regions for potential transcription factor and regulatory binding sites was by several approaches. The regions were analyzed by a nonmatrix method including the WDNASISR DNA motif version 2.5 database (latest version), and by matrix methods. Matrix methods included MatInspectorR 6.1 software using the Matrix Library version 3.1.1 (latest version) utilizing both normal homology scanning (core similarity at 0.80 and matrix at 0.85), and high homology scanning (core similarity 1.00 and matrix similarity at 0.95). The matrix approach was also accomplished by accessing the Signal Scan IMD Search Service of BIMAS using the MAMMALIAN matrix data base. For direct comparison of mouse and human sequences, both the WDNASIS multi-alignment and the MultiAlign software were used. The response of endogenous MGST1 mRNA to paraquat, and the ability of the MGST1 promoter to transcriptionally direct heterologous gene expression in response to paraquat, was analyzed by the unpaired t-test using Prism (version 3.0) software (Graph Pad, La Jolla, CA).

3. Results 3.1. Recovery of the murine MGST1 gene and cDNA clone Overlapping fragments were recovered from the murine genomic P1 plasmid as previously described [22]. Analysis revealed that the murine MGST1 gene contained four exons, and that the first exon contained only 5V-untranslated nucleotides (no coding region), in agreement with a report on the human MGST1 gene [14]. There were marked differences in the nucleotide sequence of the MGST1 gene from the 129/SvJ strain (GenBank accession AY329626 for the entire murine MGST1 gene), as compared to the gene reconstructed from sequences for the ‘‘Black 6’’ strain, in The International Mouse Genome Sequencing Consortium Project http://genome.wustl.edu/ pub/seqmgr/mouse/), and these differences substantially altered the restriction enzyme pattern (data not shown). In addition to single nucleotide changes, there were other major changes such as an additional 215-bp sequence immediately after exon 3 in the 129/SvJ strain, as well as numerous changes in repetitive elements (repetitive CT and AG sequences). The promoter region immediately upstream of the transcription initiation site, however, was highly conserved as there was significant homology between the murine and human sequences, including the regions containing HNF1 and GC-boxes previously demonstrated to be critical for expression of human MGST1 [16,17] (Fig. 1). Comparison of the amino acid sequence for MGST1 revealed one difference in exon 2 between the 129/SvJ strain (arginine at position 5) and the sequence previously reported for the Balb/c strain (lysine) [26].

166

M.J. Kelner et al. / Biochimica et Biophysica Acta 1678 (2004) 163–169

Fig. 1. Comparison of the human and murine MGST1 promoter region including G-C boxes critical for basal expression.

Fig. 2. The ability of the murine MGST1 promoter region to direct transcription of the luciferase gene product in response to oxidative stress induced by paraquat exposure, and comparison to endogenous MGST1 mRNA expression. Luciferase expression directed by the pGLmMGST1(+) vector in HepG2 cells (.) and in AA8 cells in the presence of paraquat (o). Luciferase expression directed by the control vector SV-luc in HepG2 cells (E) and in AA8 cells (D) in the presence of paraquat. Variation in transfection efficency was controlled by co-transfection with the SVgalactosidase vector. The response of endogenous MGST1 mRNA (- - -*- - -) upon exposure to 7 Ag/ml paraquat as determined by RT-PCR with GAPDH expression as the control. The symbol ‘‘*’’ indicates a P<0.05 for the pGLMGST1(+) as compared to the control SV-luciferase vector results, or a P<0.01 for endogenous MGST1 expression after 24 h of exposure to paraquat as compared to no exposure. Values represent the meanF1 S.D.of results from three different experiments.

Fig. 3. Northern dot blot analysis of MGST1 in murine tissues on a multiple tissue array hybridized with a cDNA probe specific for murine MGST1. The various dots represent the following: A1, whole brain; A2, eye; A3, liver; A4, lung; A5, kidney; B1 heart; B2 skeletal muscle; B3, smooth muscle; B4 and B5, empty; C1, pancreas; C2, thyroid; C3, thymus; C4, submaxillary gland; C5, spleen; D1, testis, D2, ovary; D3, prostate; D4, epidymus; D5, uterus; E1, whole embryo at 7 days, E2, whole embryo at 11 days; E3, whole embryo at 15 days; E4, whole embryo at 17 days; E5, blank; F1, yeast total RNA; F2, yeast tRNA; F3, E. coli rRNA; F4, E. coli DNA; F5, blank.

M.J. Kelner et al. / Biochimica et Biophysica Acta 1678 (2004) 163–169

Both 129/SvJ and Balc/c genomic DNA were obtained from a second independent source (UCSD Cancer Center Core Facility), and analysis confirmed that there was indeed a difference in the fifth amino acid between the 129/SvJ strain (arginine) and the Balb/c strain (lysine). 3.2. Identification of the promoter region and response to oxidative stress The pGLmMGST1(+)-luciferase expression vector, or the control SV-luciferase expression vector, was transfected individually into either human liver HepG2 or Chinese hamster ovarian AA8 cells. The HepG2 line was chosen to represent high expressing tissue (liver) and the AA8 line to represent low expression tissue (ovarian). The pGLMGST(+)-luciferase expression vector is identical to the SV-luciferase vector except that the SV promoter was replaced with a 1.2-kb fragment from the genomic region immediately upstream of the first exon, and containing the mRNA initiation site (see Fig. 1). Cells were also co-transfected with the SV-h-gal vector to allow for correction in transfection efficiency. After 18 h, the cells were exposed to varying concentrations of paraquat for 48 h. Analysis of luciferase expression revealed that in both cell lines the murine MGST1 promoter region responded to the oxidative stress induced by paraquat by directing transcriptional production of the luciferase reporter gene (Fig. 2) in both liver HepG2 and ovarian AA8 cells ( P<0.03 and <0.05, respectively). To determine the physiological relevance of the promoter response, the

167

increase in endogenous MGST1 mRNA upon exposure of cells to paraquat was determined by RT-PCR. Cells were exposed to 7 Ag/ml of paraquat for up to 24 h. There was a detectable increase in the quantity of endogenous MGST1 mRNA after 8 h of exposure to paraquat, and the magnitude of increase was similar to the response noted for luciferase expression (Fig. 2). After 24 h of exposure to paraquat, there was significant increase in endogenous MGST1 mRNA response ( P<0.01). 3.3. Comprehensive analysis of mRNA tissue expression A partial tissue distribution of MGST1 in rodents has been reported, based upon Northern blot analysis [27], actual enzymatic activity utilizing 1-chloro-2,4-dinitrobenzene as substrate, or by immunolocalization [28,29]. The microsomal GST activity detected in non-hepatic tissue, in contrast to hepatic tissue, was not increased upon activation with N-ethylmaleimide, although it is clearly observed by immunochemistry [28]. These observations could be partly explained by the fact that cytosolic GSTs adhere to membranes [30,31] and that endogenous inhibitors could be present [32]. Furthermore, there was a discrepancy between the activity and the protein content (determined by antibody studies) for several tissues. Thus, we desired to independently determine the mRNA expression of MGST1 in a variety of tissues to allow correlation to previous studies which determined rat and human mRNA levels, protein content and enzymatic activity [27 –29].

Fig. 4. Relative expression of MGST1 mRNA in various murine tissues as compared to expression in the liver (liver MGST1 mRNA defined as 100%). A fulllength cDNA probe (containing all four exons) was used to assess mRNA expression (see text for details). There was no detectable expression of MGST1 mRNA in brain tissue. Values represent the meanF1 S.D. of results from three different blots.

168

M.J. Kelner et al. / Biochimica et Biophysica Acta 1678 (2004) 163–169

A full-length MGST1 cDNA clone was independently recovered from a Balb/c cDNA hepatic lambda phage library and used to determine relative expression of MGST1 mRNA in various murine tissues. The mRNA expression, as anticipated, was indeed the highest in liver (Figs. 3 and 4). There was, however, marked mRNA expression in several tissues not previously studied including smooth muscle, epidymus, and ovaries (Fig. 4). Significant mRNA expression was also noted in several endocrine glands (salivary and thyroid) in which the expression of various peroxidases is also very high. Overall there was good agreement between the mRNA content detected and previous reports of MGST1 activity with the exception of brain. Brain tissue was previously reported to express MGST1 activity, but the protein content was almost nondetectable. In addition, in contrast to other organs, treatment of brain microsomes with NEM inhibited activity (not enhancing by three- to sixfold). We were unable to detect mRNA expression in the murine brain, in agreement with the previous protein studies. This suggests that enzymatic activity previously noted in brain may derive from other unidentified gene products.

4. Discussion In summary, our findings define the murine MGST1 gene from the 129/SvJ strain, and allows for production of genetically modified mice that lack expression of microsomal glutathione transferase-1. There were significant differences between the MGST1 gene for the 129/ SvJ strain as compared to the gene reconstructed from sequence data obtained from the International Mouse Genome Sequencing Consortium Project for the ‘‘Black 6’’ strain, such that it would be difficult to use the consortium data to generate a null or MGST1 ‘‘knockout’’ mouse. There was high homology between the human and murine promoter immediately upstream of the transcription initiation site, including the regions containing the GC boxes and HNF1 sites, responsible for controlling expression of the human MGST1 gene [16,17]. The promoter region of the murine MGST1 gene, as anticipated, does respond to oxidative stress induced by paraquat. This transcriptional response to oxidative stress is physiologically relevant as exposure to paraquat also increased the endogenous MGST1 expression. The expression of MGST1 mRNA activity was highest in murine hepatic tissue. Substantial expression, however, is also noted in other organs including those that express a variety of peroxidases (thyroid and salivary glands). As MGST1 represents >3% of total microsomal protein in rodent liver [28], the quantity of MGST1 mRNA in tissues such as thyroid or smooth muscle corresponds to a significant expression in these tissues. Thus, MGST1 may play a substantial role in protecting those tissues from endogenously generated oxidative stress. The inabil-

ity to detect mRNA expression of MGST1 in the brain is consistent with studies by others that failed to detect protein. Enzymatic activity, however, is present in brain microsomes, indicating that this enzymatic activity derives from other members of the MAPEG family, cytosolic GST, or perhaps as yet unidentified gene(s).

Acknowledgements This study was supported by NIEHS Award 1 P42 ES10337 and the Swedish Cancer Society.

References [1] C.P. Tu, A. Matsushima, N.Q. Li, D.M. Rhoads, K. Srikumar, A.P. Reddy, C.C. Reddy, J. Biol. Chem. 261 (1986) 9540 – 9545. [2] C.B. Pickett, A.Y. Lu, Ann. Rev. Biochem. 58 (1989) 743 – 764. [3] P.J. Jakobsson, R. Morgenstern, J. Mancini, A. Ford-Hutchinson, B. Persson, Protein Sci. 8 (1999) 689 – 692. [4] R. Morgenstern, C. Guthenberg, J.W. DePierre, Eur. J. Biochem. 128 (1982) 243 – 248. [5] C.C. Reddy, C.P. Tu, J.R. Burgess, C.Y. Ho, R.W. Scholz, E.J. Massaro, Biochem. Biophys. Res. Commun. 101 (1981) 970 – 978. [6] E. Mosialou, F. Piemonte, C. Andersson, R.M. Vos, P.J. van Bladeren, R. Morgenstern, Arch. Biochem. Biophys. 320 (1995) 210 – 216. [7] E. Mosialou, G. Ekstrom, A.E. Adang, R. Morgenstern, Biochem. Pharmacol. 45 (1993) 1645 – 1651. [8] E. Mosialou, C. Andersson, G. Lundqvist, G. Andersson, T. Bergman, H. Jornvall, R. Morgenstern, FEBS Lett. 315 (1993) 77 – 80. [9] E. Mosialou, R. Morgenstern, Arch. Biochem. Biophys. 275 (1989) 289 – 294. [10] C. Andersson, E. Mosialou, R. Weinander, R. Morgenstern, Adv. Pharmacol. 27 (1994) 19 – 35. [11] R. Morgenstern, J.W. DePierre, Microsomal glutathione transferase, in: E. Hodgson, J.R. Bend, R.M. Philpot (Eds.), Reviews in Biochemical Toxicology, 7th ed., Elsevier, New York, 1985, pp. 67 – 104. [12] M.W. Anders, L. Lash, W. Dekant, A.A. Elfarra, D.R. Dohn, Crit. Rev. Toxicol. 18 (1988) 311 – 341. [13] S.J. Hargus, M.E. Fitzsimmons, M.W. Anders, Adv. Exp. Med. Biol. 283 (1991) 289 – 293. [14] M.J. Kelner, R.D. Bagnell, M.A. Montoya, L.A. Estes, L. Forsberg, R. Morgenstern, J. Biol. Chem. 275 (2000) 13000 – 13006. [15] P.J. Jakobsson, S. Thoren, R. Morgenstern, B. Samuelsson, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 7220 – 7225. [16] L. Ekstrom, L. Lyrena¨sa, P.-J. Jakobsson, R. Morgenstern, M.J. Kelner, Biochem. Biophys. Acta 1627 (2003) 79 – 84. [17] L. Forsberg, L. Persson, M.J. Kelner, P.-J. Jakobsson, R. Morgenstern, Chem.-Biol. Interact. 133 (2001) 301 – 303. [18] A. Iida, S. Saito, A. Sekine, S. Harigae, S. Osawa, C. Mishima, K. Kondo, Y. Kitamura, Y. Nakamura, J. Hum. Genet. 46 (2001) 590 – 594. [19] H. Chaib, E.K. Cockrell, M.A. Rubin, J.A. Macoska, Neoplasia 3 (2001) 43 – 52. [20] B.B. Knowles, C.C. Howe, D.P. Aden, Science 209 (1980) 497 – 499. [21] L.H. Thompson, S. Fong, K. Brookman, Mutat. Res. 74 (1980) 21 – 36. [22] M.J. Kelner, M.N. Stokely, N.E. Stovall, M.A. Montoya, Genomics 36 (1996) 100 – 103. [23] M.J. Kelner, R.D. Bagnell, S.F. Uglik, M.A. Montoya, G.T. Mullenbach, Arch. Biochem. Biophys. 323 (1995) 40 – 46.

M.J. Kelner et al. / Biochimica et Biophysica Acta 1678 (2004) 163–169 [24] F.F. Chu, R.S. Esworthy, S. Akman, J.H. Doroshow, Nucleic Acids Res. 18 (1990) 1531 – 1539. [25] P.J. Hornsby, D.W. Pearson, A.P. Autor, K.A. Aldern, S.E. Harris, J. Cell. Physiol. 123 (1985) 33 – 38. [26] H. Raza, J. Mullick, A. John, S.V. Bhagwat, N.G. Avadhani, Oncol. Rep. 7 (2000) 645 – 649. [27] M. Estonius, L. Forsberg, O. Danielsson, R. Weinander, M.J. Kelner, R. Morgenstern, Eur. J. Biochem. 260 (1999) 409 – 413. [28] R. Morgenstern, G. Lundqvist, G. Andersson, L. Balk, J.W. DePierre, Biochem. Pharmacol. 33 (1984) 3609 – 3614. [29] M.A. Otieno, R.B. Baggs, J.D. Hayes, M.W. Anders, Drug Metab. Dispos. 25 (1997) 12 – 20.

169

[30] R. Morgenstern, C. Guthenberg, B. Mannervik, J.W. DePierre, FEBS Lett. 160 (1983) 264 – 268. [31] S.P. Singh, A.J. Janecki, S.K. Srivastava, S. Awasthi, Y.C. Awasthi, S.J. Xia, P. Zimniak, J. Biol. Chem. 277 (2002) 4232 – 4239. [32] T.D. Boyer, D. Zakim, D.A. Vessey, Biochem. J. 207 (1982) 57 – 64.