1398
Chem. Res. Toxicol. 2002, 15, 1398-1406
Relative Toxicities of DNA Cross-Links and Monoadducts: New Insights from Studies of Decarbamoyl Mitomycin C and Mitomycin C Yolanda Palom,†,‡ Gopinatha Suresh Kumar,†,§ Li-Qian Tang,|,⊥ Manuel M. Paz,†,# Steven M. Musser,3 Sara Rockwell,| and Maria Tomasz*,† Department of Chemistry, Hunter College, City University of New York, New York, New York 10021, Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, Connecticut 06520, and U.S. Food and Drug Administration, 5100 Paint Branch Parkway, College Park, Maryland 20740 Received May 29, 2002
Mitomycin C (MC), a cytotoxic anticancer drug and bifunctional DNA DNA alkylating agent, induces cross-linking of the complementary strands of DNA. The DNA interstrand cross-links (ICLs) are thought to be the critical cytotoxic lesions produced by MC. Decarbamoyl mitomycin C (DMC) has been regarded as a monofunctional mitomycin, incapable of causing ICLs. Paradoxically, DMC is slightly more toxic than MC to hypoxic EMT6 mouse mammary tumor cells as well as to CHO cells. To resolve this paradox, EMT6 cells were treated with MC or DMC under hypoxia at equimolar concentrations and the resulting DNA adducts were analyzed using HPLC and UV detection. MC treatment generated both intrastrand and interstrand cross-link adducts and four monoadducts, as shown previously. DMC generated two stereoisomeric monoadducts and two stereoisomeric ICL adducts, all of which were structurally characterized; one was identical with that formed with MC, the other was new and unique to DMC. Overall, adduct frequencies were strikingly higher (20-30-fold) with DMC than with MC. Although DMC monoadducts greatly exceeded DMC cross-link adducts (∼10:1 ratio), the latter were equal or higher in number than the cross-link adducts from MC. DMC displayed a much higher monoadduct:cross-link ratio than MC. The similar cytotoxicities of the two drug show a correlation with their similar DNA cross-link adduct frequencies, but not with their total adduct or monoadduct frequencies. This provides specific experimental evidence that the ICLs rather than the monoadducts are critical factors in the cell death induced by MC. In vitro, overall alkylation of calf thymus DNA by DMC was much less efficient than by MC. Nevertheless, ICLs formed with DMC were clearly detectable. The chemical pathway of the cross-linking was shown to be analogous to that occurring with MC. These results also suggest that the differential sensitivity of Fanconi’s Anemia cells to MC and DMC is related to factors other than a selective defect in cross-link repair.
Introduction
Scheme 1
Mitomycin C (MC;1 Scheme 1) is an antitumor antibiotic and cytotoxic anticancer drug used in clinical chemotherapy regimens for the treatment of various carcinomas (1). On the molecular level MC acts as a bioreductive alkylating agent (2). As such, it is enzymatically reduced within the cell, yielding reactive species * To whom correspondence should be addressed. † Hunter College, City University of New York. ‡ Current address: 5653 Jacquiline Way, Livermore, CA 94550. § Current address: Indian Institute of Chemical Biology, Calcutta-700 032, India. | Yale University. ⊥ Current address: McGill University, Montreal PQ, H3G 1Y6 Canada. # Current address: Departmento de Qui´mica Orga ´ nica, Facultade de Cinecias, Universidade de Santiago de Compostela, Campus de Lugo, 27002 Lugo, Spain. 3 U.S. Food and Drug Administration. 1 Abbreviations: AUFS, absorbance unit full scale; ccr, cytochrome c reductase; CD, circular dichroism; CHO, Chinese hamster ovary; 2,7DAM, 2,7-diaminomitosene; DMC, 10-decarbamoyl mitomycin C; FA, Fanconi’s Anemia; GSH, reduced glutathione; ICL, interstrand crosslink; LC-ESIMS, liquid chromatography/electrospray ionization mass spectroscopy; MC, mitomycin C.
that are capable of producing oxygen radicals through redox cycling (3) as well as DNA interstrand cross-links (4) and a variety of guanine monoadducts in the minor groove (5, 6). There is evidence that the cytotoxicity of MC is due primarily to the formation of the DNA adducts
10.1021/tx020044g CCC: $22.00 © 2002 American Chemical Society Published on Web 10/08/2002
Toxicities of DNA Cross-Links and Monoadducts
rather then oxygen radicals (7, 8). In particular, DNAinterstrand cross-links, resulting from bifunctional alkylation of DNA by MC have been proposed as the primary cause of cell death (4). Our past studies have established the structure of the DNA cross-link formed with MC (9), as the bis-guanine adduct 3a (Scheme 2), arising by consecutive alkylation Scheme 2
of two guanines at their N2-positions by a MC molecule, first by its C-1 aziridine, followed by the C-10 carbamate function of the activated form of the drug (Scheme 3). Scheme 3
We showed, furthermore, that treatment of living cells with MC produces six major DNA adducts (1-6; Schemes 2 and 4); of these, 3a represents the interstrand crossScheme 4
link. An identical bisadduct containing a phosphate link between two adjacent deoxyguanosines is formed in one strand, constituting an intrastrand cross-link (4). The monoadducts 5 and 6 are not formed directly with activated MC. Rather, they are products of alkylation by
Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1399
the metabolite of MC, 2,7-diaminomitosene (2,7-DAM) (Scheme 1), which is formed in situ and then itself acts as a monofunctional reductive alkylator of DNA (10). The elucidation of the array of DNA adducts of MC provides a system for investigating the relationship between the structure of MC adducts and MC cytotoxicity. Although the inherent cytotoxicity of DNA crosslinks is a widely held paradigm, the relative importance of MC cross-links and the monofunctional mitomycin adducts has not been experimentally addressed so far. There are important precedents that bulky minor groove adducts of natural products can be highly cytotoxic, e.g., the adducts of the anthramycin family (11), ecteinascidin743 (20), CC-1065 (11), and duocarmycin (12, 13). Since the MC cross-link adduct 3a represents only 10-20% of total adduct in EMT6 mouse mammary tumor cells (14), the significance of the more frequent MC monoadducts cannot be discounted a priori. A good correlation was observed between cross-linking activity and the cytotoxicity of MC analogues with different 7-substituents in EMT6 cells (8). However, all of these analogues were likely to give similar cross-link:monofunctional adduct ratios, and therefore the correlation is not necessarily specific to cross-links alone. Considering the class of synthetic chemotherapeutic cross-linking agents, monofunctional DNA lesions always predominate over bifunctional ones, and therefore, it has been inherently difficult to assess the cytotoxicities of the cross-links independently of the monoadducts (15). We investigated the structure-activity relationship of the various MC adducts by employing as tools two closely related monofunctional mitomycin analogues: 2,7-diaminomitosene (2,7-DAM), the major intracellular metabolite of MC, and 10-decarbamoyl mitomycin C (DMC), a chemically modified MC derivative. 2,7-DAM lacks the aziridine which is the primary alkylating function of MC while DMC lacks the C-10 carbamate, which forms the secondary “alkylating arm” (Schemes 1, 2, and 3). We found earlier that treatment of EMT6 cells with 2,7-DAM resulted in the formation of only two DNA adducts, 5 and 6, with high efficiency, and that these adducts were identical with two of the six major adducts formed in MCtreated cells (16). However, 2,7-DAM was far less toxic than MC to EMT6 cells (Figure 1). This indicated that the monofunctional DNA adducts 5 and 6 do not represent lethal DNA damage and, therefore, are not likely to contribute to the cytotoxicity of MC. Thus only four of the six MC adducts (1a, 2a, 3a, and 4) remained as the cytotoxic candidates (16). DMC generally has been regarded as another monofunctional derivative of mitomycin C, lacking the secondary alkylating function, i.e., the C-10 carbamate. Accordingly, on incubation in an in vitro system Micrococcus luteus DNA and chemically activated DMC yielded the monofunctional adduct 2a as the major product, with no evident DNA cross-link adduct detectable (6). Similarly, NaBH4-activated DMC failed to cross-link CHO DNA as shown using hydroxyapatite chromatography as the method of assay (17). Inconsistent with these in vitro results, however, our recent study of the cytotoxicity of DMC in mammalian cell lines revealed that DMC and MC had similar cytotoxicities to EMT6 cells under aerobic conditions. Under hypoxic conditions, DMC was more toxic than MC to EMT6 cells. In wild-type (AA8) Chinese hamster cells and two repair deficient mutants (UV4 and UV5), DMC
1400
Chem. Res. Toxicol., Vol. 15, No. 11, 2002
Figure 1. Survival of cells in exponentially growing EMT6 cultures treated with mitomycin C (MC, b), decarbamoyl mitomycin C (DMC, 9), or 2,7-diaminomitosene (2,7-DAM, 1). Points are means from two to eight different experiments. The mitomycin dose is plotted as concentration × time as described previously (46, 47), to allow the inclusion of data from experiments using 1 h treatments with MC or DMC and experiments using 2 h treatments with MC and DAM. Data are from experiments published in refs 16 and 18, plus unpublished data from Tang and Rockwell.
was more cytotoxic than MC under both aerobic and hypoxic conditions (18). Analogous findings were reported earlier by others (19). In CHO cells, mutagenicity and the induction of sister chromatid exchange by MC and DMC were also indistinguishable (17), although the relationship of these properties to cytotoxicity is not known. The apparent contradiction between similarities of the biological activities of MC and DMC and their different cross-linking abilities in chemical systems prompted us to investigate the DNA adducts formed in vivo after treatment of EMT6 cells with DMC. We report here that, in EMT6 cells treated with DMC, DNA alkylation occurs at a much higher frequency than in cells treated with MC under the same conditions. Although the products of the alkylation were predominantly DMC monoadducts, DNA cross-link adducts were also generated by DMC at relatively low frequencies. However, their frequencies were similar to or greater than those of the DNA cross-link adducts observed in MC-treated cells. The DNA mono- and cross-link adducts of DMC were structurally characterized at atomic resolution. These findings constitute new evidence that DNA interstrand cross-links are primarily responsible for the cytotoxicity of the mitomycins.
Experimental Section Materials. Calf thymus DNA was purchased from Sigma Chemical Co. (St. Louis, MO) and was sonicated before use. NADH-cytochrome c reductase (NADH-FMN oxidoreductase, EC 1.6.99.3) and NADH were obtained from Sigma. Nuclease P1 (Penicillium citrinum, EC 3.1.30.1) and poly(dG-dC)‚poly(dG-dC) were purchased from Amersham Pharmacia Biotech (Piscataway, NJ). Phosphodiesterase I (snake venon diesterase, Crotalus adamenteus venom, EC 3.1.4.1) and alkaline phosphatase (Escherichia coli, EC 3.1.3.1) were obtained from Worthington Biochemical Corp. (Freehold, NJ). Nonradioactive MC was supplied by D. M. Vyas (Bristol-Myers Squibb Co., Wallingford, CT). 2,7-Diaminomitosene was synthesized from MC as previously described (21). 10-Decarbamoyl mitomycin C was synthesized from MC by a published procedure (22). Coliphage T-7 DNA and Hoechst 33258 dye were purchased from Sigma.
Palom et al. Cells, Treatment of Cells with MC or DMC and Isolation of Nuclear DNA from Drug-Treated Cells. EMT6 mouse mammary tumor cells were cultured and maintained as exponentially growing monolayers in Waymouth’s medium supplemented with 15% fetal bovine serum and antibiotics, as described previously (23). Treatment of cells with MC or DMC and isolation of nuclear DNA from drug-treated cells were carried out by our published procedures (23). Cell survival was assayed using a clonogenic assay, as described previously (18). Treatment of Calf Thymus DNA with MC or DMC Using NADH-Cytochrome c Reductase as the Activator. A solution of calf thymus DNA (12 mM), MC (1 mM), or DMC (1 mM), and NADH (2 mM) in 15 mM Tris-HCl buffer (pH 7.4) was deaerated. Cytochrome c reductase in 20 units/mL stock solution was added to give a concentration of 0.58 unit/mL in the reaction mixture. Incubation under a positive pressure of argon was continued for 50 min at 37 °C. The reaction mixture was extracted with phenol/CHCl3 (1:1. v/v), and the drug-DNA complex was isolated by ethanol precipitation. Treatment of Calf Thymus DNA Using Na2S2O4 as Activator. A solution of calf thymus DNA (12 mM), MC (1 mM) or DMC (1 mM) was deaerated in 15 mM Tris-HCl buffer (pH 7.4). A deaerated solution of Na2S2O4 (1.5 mM) was added and the mixture was incubated under a positive pressure of argon for 50 min at room temperature. The drug-DNA complexes were isolated as above. Treatment of Synthetic Poly(dG-dC)•poly(dG-dC) with DMC Using NADH-Cytochrome c Reductase as the Activator. A solution of poly(dG-dC)‚poly(dG-dC) (12 mM), 1 mM DMC, 2 mM NADH, and NADH-cytochrome c reductase (0.6 units/mL) in 15 mM Tris-HCl (pH 7.4) buffer was deaerated and incubated for 50 min at 37 °C under a positive pressure of argon. The drug-polynucleotide complex was isolated as above. Enzymatic Digestion of the Calf Thymus DNA-Drug Complex to Nucleosides and Drug-Nucleoside Adducts. The lyophilized MC or DMC-DNA complex was digested to the nucleoside level by the following protocol. Nuclease P1 (1.0 unit/ A260 unit of complex) was added to the complex in dilute aqueous acetic acid at pH 5.0 (2.5 A260 units/mL), followed by incubation for 4 h at 37 °C. The pH was adjusted to 8.2 by addition of 0.5 M Tris, and MgCl2 was added to a concentration of 0.9 mM. Addition of snake venom diesterase (2.25 units/A260 unit of complex) and a 2 h incubation at 37 °C were followed by the addition of alkaline phosphatase (1.6 units/A260 unit of complex) and incubation overnight at 37 °C. The digest was then heated at 90 °C for 1 h to hydrolyze guanine N7 adducts. Samples were lyophilized and redissolved in varying volumes of water for HPLC analysis. HPLC Analysis of Adducts in Digests. A Beckman System Gold 125 instrument, equipped with a Beckman System Gold 168 diode array detector, set at 254 and/or 320 nm and controlled by System Gold Chromatography Software, was used. A Rainin Microsorb MV C18 column (4.6 mm × 250 mm) was employed, the column being eluted with a 6 to 18% acetonitrile gradient in 30 mM KH2PO4 or 20 mM NH4OAc (pH 5.5) for 60 min at a flow rate of 1 mL/min. Adduct frequency (moles of adduct per mole of DNA nucleotide) was calculated from areas of appropriate peaks in the HPLC chromatogram of the digest using the formula
[mol of adduct/mol of DNA nucleotide] ) [Area of adduct peak × (E254 of dT/E254 of adduct) × 0.28]/ [Area of dT peak] where E254 is the molar extinction coefficient at 254 nm and 0.28 is the molar ratio of dT to total deoxynucleoside in calf thymus DNA (23). The dG peak instead of the dT peak was used analogously to calculate adduct frequencies in the synthetic G-C polynucleotide. Enzymatic Digestion of DNA and Analysis of DNA Adducts from Drug-Treated Cells. The relatively low adduct frequency of EMT6 DNA isolated from drug-treated cells neces-
Toxicities of DNA Cross-Links and Monoadducts
Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1401
sitated using a larger amount of DNA digest for the analysis compared to calf thymus DNA-drug complexes. A modified twostage digestion and adduct separation was used for this purpose, as follows. The EMT6 DNA-drug complex (20-30 A260 units of DNA) was digested first to nucleotides by nuclease P1 (1 unit/ A260 unit) overnight at 37 °C, as above. The digest was fractionated on a 1 cm × 1 cm Sep-Pak cartridge into the unmodified nucleotide fraction, which was eluted with water, and the adducted nucleotides were eluted with 60% methanol. Some unmodified adenine nucleotide was also present in the latter fraction. Both fractions were concentrated, lyophilized and further digested by snake venom diesterase (2.25 units/A260 unit) and alkaline phosphatase (1.6 units/A260 unit) in 30 mM TrisHCl-7 mM MgCl2 (pH 8.2) buffer overnight. Each was then analyzed by HPLC as described in the preceding section. Spectroscopic Techniques and LC/ESIMS Analysis. These were described in a previous publication (10). Synthesis and Characterization of Authentic Adduct 2b. In analogy to the known conversion of adduct 1a to adduct 2a by hydrolysis of the 10-carbamate group an aqueous solution of authentic adduct 1b (24) was heated at 90 °C for 1 h. HPLC of the reaction mixture indicated 70-90% conversion of 1b (elution time 28.8 min) to 2b (elution time 28.0 min) with no other products present. The structure of 2b was confirmed based on (i) the identity of its UV spectrum with that of 2a, (ii) its CD spectrum showing a Cotton effect at 570 nm with the opposite sign of that of 2a, indicating 1′′-β chirality of the mitosene link to N2 of the deoxyguanosine (25), and (iii) its ESIMS spectrum. In Vitro Assay of DNA Cross-Linking by DMC. The method was based on fluorescence enhancement of DNA-bound Hoechst 33258 dye as described (26).
Results Comparison of the Survival of EMT6 Cells Treated with DMC or MC. Figure 1 shows data from series of experiments comparing the effects of MC, DMC, and 2,7DAM on the viability (colony forming ability) of EMT6 cells. In the studies with hypoxic cells shown in Figure 1, 2,7-DAM was much less toxic than MC, with IC90s of 55 µM × h and 0.8 µM × h, respectively. In contrast, DMC was slightly more toxic than MC under these conditions, with an IC90 of 0.4 µM × h. MC, DMC, and 2,7-DAM are all more toxic to EMT6 cells treated under hypoxia than to cells treated in air (16, 18). Under aerobic conditions, 2,7-DAM is again much less toxic than MC (16), while the toxicities of DMC and MC are similar (18). The surprising toxicity of DMC, relative to MC and 2,7DAM, led to the studies of the adduct patterns described below. Adducts of DMC and MC Formed in EMT6 Cells. HPLC of the digested DNA indicated that DMC treatment at 20 µM drug concentration generated two predominant adducts, 2a and 2b. A third, minor adduct was also detectable (Figure 2a). The same HPLC run is depicted in Figure 3a at a 10-fold lower AUFS scale. At this greater magnification, several additional minor adducts were discernible, most notably adduct 3a. The structures of the adducts 2a, 2b, 3a, and 3b were identified by several experimental criteria as follows. LCESIMS established their appropriate molecular weights (Figure S1, Supporting Information). Authentic standards of 2a (6), 2b (Experimental Section), and 3a (9) had identical UV spectra and HPLC elution times with corresponding adducts from the cells. The predominant monoadduct 2b was independently characterized as the 1′′-β isomer of 2a by CD spectra (Figure 4b). The use of a method for establishing the stereochemistry of mitomycin adducts has been described (25).
Figure 2. HPLC of DNA adducts formed in EMT6 cells treated with (a) DMC, (b) MC, and (c) 2,7-DAM. DNA was isolated from cells after drug treatment and digested to nucleosides. Drug treatment conditions: (a) 20 µM DMC, 4 h, under hypoxia. (b) 20 µM MC, 1 h, under hypoxia. Reprinted with permission from ref 23. Copyright 1998 Cognizant Communications Corp. (c) 40 µM 2,7-DAM, 4 h, aerobic conditions. Reprinted with permission from ref 16. Copyright 2001 Elsevier. The numbers by the peaks refer to the adduct structures in Schemes 3 and 4.
Although no authentic standard existed for isomeric cross-link adduct 3b, its structure is assigned with confidence, based on its mass (Figure S1, Supporting Information) and UV spectra (Figure 4a), both of which are characteristic of two dG residues per mitosene chromophore (5). Furthermore, the UV spectrum of 3b was essentially identical with the UV spectrum of the known (1′′-R isomer) cross-link 3a (Figure 4), indicating that the same positions of the guanines are linked to the mitomycin as in 3a. MC treatment of the cells at 20 µM drug concentration resulted in the HPLC adduct pattern seen in Figure 3b, at the same magnification as that for HPLC of the DMC adducts in Figure 3a. All six major MC adducts known to be formed in EMT6 cells (10, 14) were present; the HPLC peaks are marked accordingly (see Schemes 3 and 4 for the structures). In our previous work, the analyses
1402
Chem. Res. Toxicol., Vol. 15, No. 11, 2002
Figure 3. HPLC of DNA adducts formed in EMT6 cells treated with (a) 20 µM DMC and (b) 20 µM MC, for 4 h under hypoxic conditions. Comparison of DMC and MC adducts on the same absorbance scale. DNA was isolated from the cells after drug treatment and digested to nucleosides.
Palom et al.
Figure 4. (a) UV spectrum of adduct 3b isolated from EMT6 cells (;) superimposed with UV spectrum of authentic adduct 3a (‚‚‚). (b) CD spectra of the isomeric pair of adducts 2a (authentic standard) and 2b (isolated from EMT6 cells). Buffer: 0.1 M Tris‚HCl; pH 7.4 Table 1. DNA Adducts Formed in EMT6 Cells Treated with DMC or MC at 10 µM Concentrationa
3
of MC adducts formed in vivo employed H-labeled MC (e.g., Figure 2b) which resulted in better adduct signalto-noise ratios in the HPLC patterns than the 254 nm absorbance used in the present work. However, parallel monitoring of the absorbance at 254 and 320 nm distinguished the drug-nucleoside adducts from nonadduct material, since the latter absorbed light only at 254 nm but not at 320 nm. The HPLC patterns of the drugnucleoside adducts show exellent agreement with the patterns observed earlier with radiolabeled MC (e.g., Figure 3b compared with Figure 2b). In summary (Table 1), DMC formed two kinds of bifunctional cross-link adducts with DNA in EMT6 cells. One was identical with the bis-guanine cross-link adduct formed also with MC (3a), and the other was its 1′′-β stereoisomer 3b, unique to DMC. The intrastrand crosslink adduct (4) observed with MC was not generated by DMC. DMC also formed monofunctional DNA adducts. Of these, 2a, 5, and 6 were common to DMC and MC, while 2b, the 1′′-β isomer of 2a was again unique to DMC. Quantitative Analysis of the Adducts. Figures 2a, 3a, and 3b depict results of single experiments of adduct formation in cells which illustrate typical HPLC patterns. A number of additional experiments were carried out with cells treated with DMC, at concentrations of 2 µM (one experiment), 10 µM (total of 3 experiments), and 20 µM (two experiments), and with MC, at a concentration of 10 µM (total of three experiments). The frequencies (mol of adduct/mol of DNA nucleotide) of individual adducts were determined in each experiment. From these
drug
monoadduct
monoadduct of 2,7-DAM (in situ metabolite)
DMC MC
2a, 2b 1a, 2a
5, 6 5, 6
a
bisadduct (interstrand cross-link)
bisadduct (intrastrand cross-link)
3a, 3b 3a
4
Hypoxic conditions, 4 h, 37 °C.
data the frequencies of total adducts, total monoadducts, and total cross-link adducts were calculated and are shown in Table 2. DMC acts as a strikingly efficient DNA alkylating agent in EMT6 cells compared to MC: at 2, 10, and 20 µM cell treatment concentrations, DMC total adducts frequencies are 28-, 20-, and 18.5-fold higher, respectively, than the corresponding MC total adduct frequencies. When monoadduct frequencies alone are compared, DMC is a 63-, 38-, and 28-fold stronger DNA monoalkylator, respectively, than MC. DMC is also a more intensive cross-linker, but only by a factor of 7.5, as calculated from the 10 µM drug treatment (triplicate) data sets. Within this average value, the cross-link adduct 3a alone, common to DMC and MC, is formed in approximately equal amounts upon cell treatments with DMC or MC (0.7:1 ratio), while the DMC-specific crosslink isomer 3b was 10 times more abundant then the common cross-link 3a. Adducts of DMC with Calf Thymus DNA Formed Under Enzymatic or Chemical Reductive Activation. Enzymatic activation of DMC using NADH-cytochrome c reductase with NADH in anaerobic solution
Toxicities of DNA Cross-Links and Monoadducts
Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1403
Table 2. Frequencies of DNA Adducts of MC and DMC Formed in EMT6 Cells
drug DMC MC
a
treatment concentration (µm)
total adducta (×105)
2 10 20 2b 10 20
1.9 5.4 ( 2.1 11 0.067 ( 0.1b 0.26 ( 0.05 0.60
monoadducta (×105)
bisadduct (interstrand cross-link 3a) (×105)
bisadduct (interstrand cross-link 3b) (×105)
bisadduct (intrastrand cross-link 4) (×105)
total bisadduct (3a + 3b + 4) (×105)
1.9 5.0 ( 1.8 10 0.03 ( 0.00b 0.13 ( 0.04 0.35
c 0.03 ( 0.01 d 0.012 ( 0.00b 0.048 ( 0.026 0.089
c 0.33 1.2 undetectableb
