Cross-Linking of Dithiols by Mitomycin C - Chemical Research in

Jul 7, 2010 - Upon reduction, the antitumor drug mitomycin C undergoes a cascade of reactions to give a bis-electrophile that alkylates cellular nucle...
0 downloads 0 Views 743KB Size
1384

Chem. Res. Toxicol. 2010, 23, 1384–1392

Cross-Linking of Dithiols by Mitomycin C Manuel M. Paz* Departmento de Quı´mica Orga´nica, UniVersidade de Santiago de Compostela, Facultade de Ciencias, Campus de Lugo, 27002 Lugo, Spain ReceiVed April 9, 2010

Upon reduction, the antitumor drug mitomycin C undergoes a cascade of reactions to give a biselectrophile that alkylates cellular nucleophiles. We recently reported that dithiols activate mitomycin C by reduction, and we report here that dithiols, after executing the reductive activation of mitomycin C, are bis-alkylated by the activated drug to form S,S′-cross-links as the predominant end products. The diastereomeric pair of adducts formed by 1,3-propanedithiol has been fully characterized by UV, HRMS, CD, and NMR experiments. Racemic dithiol (()-dithiothreitol gave four diastereomeric cross-links, and (()-dihydrolipoic acid gave eight cross-links (two regioisomers with four diastereomers each) that were partially characterized by UV and MS. The observed dependence of cross-link formation on dithiol concentration indicated the requirement of a second reduction step by dithiol, prior to the alkylation of the second arm of the dithiol. The existence of unidentified reaction pathways was manifested by the formation of unexpected intermediates during the course of the reaction of mitomycin C with dithiols and by the formation of unsoluble mitosene derivatives in the reaction between equimolar amounts of dithiol and mitomycin C. Mechanistic details of the reaction are addressed in light of these results. Finally, we discuss the potential relevance of our findings for the interaction of mitomycin C with dithiol-containing proteins. Introduction 1

Mitomycin C (MC, 1 ) is an antitumor antibiotic used clinically in cancer chemotherapy (1) and in ophthalmologic procedures (2). There is ample consensus that the cytotoxicity of MC is caused by the formation of covalent DNA adducts, in particular interstrand DNA-DNA cross-links, creating a lesion that inhibits DNA replication leading to cell death (3). MC is a smart prodrug, inert toward nucleophiles in its original structure but transformed into a highly reactive bis-electrophile after reduction (4). The mechanism of activation of MC is initiated by the reduction of the quinone ring of MC to form hydroquinone 2. Subsequent elimination of methanol generates an iminium ion that isomerizes to form the indole ring of leucoaziridinomitosene 3, a bis-electrophile that alkylates cellular nucleophiles such as phosphate (5), bicarbonate (6), glutathione (7), and DNA (8) (Scheme 1). Sequential alkylation of DNA at the 1 and 10 positions of 3 generates a cytotoxic interstrand cross-link (6, NuH ) DNA) that specifically links N-2 of opposite deoxyguanosine residues at CpG sites of double stranded DNA (8). A number of oxidoreductase enzymes have been implicated in the in vivo activation of MC, most notably cytochrome C reductase and DT-diaphorase (9). Recent reports indicate that glucose regulatory protein (GRP58) also plays a significant role in the cellular activation of MC (10), and that the activity of this protein resides in its two thioredoxin-like domains (11). We recently reported a successful activation of MC using simple dithiols as chemical models for the dithiol functional group present in the active site of proteins of the thioredoxin family * To whom correspondence should be addressed. E-mail: manuel.paz@ usc.es. 1 Abbreviations: DHLA, dihydrolipoic acid; DTT, D,L-dithiothreitol; MC mitomycin C; Trx, thioredoxin; TrxR, thioredoxin reductase; mitosene: structure 5, without substituents in the 1-, 2- and 7-positions.

Scheme 1. Mechanism for the Reductive Activation of MC by Dithiolsa

a

NuH represents a nucleophile.

(12). The involvement of proteins containing a dithiol active site in the biological activity of MC is not restricted to the abovementioned activation by GRP58. Thioredoxin (Trx), a ubiquitous oxidoreductase, has been reported to be involved in the modulation of mitomycin C activity: high levels of Trx resulted in the protection of Fanconi anemia cell lines from the cytotoxic effects of MC (13, 14); Trx levels correlated with MC-resistance in human bladder cancer cell lines (15). Continuing with our investigation on the reactions of dithiols with MC, we report here that dithiols, in addition to performing

10.1021/tx100134h  2010 American Chemical Society Published on Web 07/07/2010

Cross-Linking of Dithiols by Mitomycin C

the reductive activation of MC, are alkylated by activated MC to form S,S′-cross-links as the almost exclusive end products. The full characterization of the cross-links formed with 1,3propanedithiol is presented, while the cross-links formed by DTT and DHLA were partially characterized. We also report the results obtained from studies on the identification of intermediates during the reaction of MC with dithiols and on how the concentration of dithiol influences the outcome of the reaction. While these studies shed some light on the mechanism of the reaction, they also revealed the existence of unconventional pathways in the reaction of MC with dithiols.

Materials and Methods General Procedures. Mitomycin C was a gift from Kyowa Hakko Kogyo Co., Ltd., Japan.. DTT, 1,3-propanedithiol, dihydrolipoic acid, and other reagents were from Sigma-Aldrich. HPLC was performed with a HP Agilent Series 1100 diode array system. LCMS was performed with an HP Agilent Series 1100 diode array HPLC system connected to an HP Agilent Series 1100 MSD mass spectrometer. A Macherey Nagel (Nucleodur C18, 4.6 × 125 mm, 5 µm) column was used. The elution system was 5% B to 30% B in 20 min, 30% B to 55% B in 10 min (A ) 10 mM ammonium acetate, pH 5.5; B ) acetonitrile), with a flow rate of 1 mL/min. LC/HRMS experiments were performed with a Bruker Microtof mass spectrometer, using a Phenomenex (Kinetex C18, 2.1 × 100 mm, 2.6 µm) column. The elution system was 5% B to 30% B in 20 min, 30% B to 55% B in 10 min (A ) 10 mM ammonium acetate, pH 5.5; B ) acetonitrile, containing 0.1% formic acid), with a flow rate of 0.4 mL/min. Circular dichroism spectra were obtained with a Jasco 715 spectropolarimeter using a 1 cm quartz cell in MeOH. NMR experiments were performed with a Varian Mercury 300 spectrometer or in a Varian VNMRS-500-WB. UV spectra were obtained with a CARY 300 UV-visible spectrometer. Analysis of the Products from the Reaction of MC and DTT by LC/MS. MC in H2O (100 µL of a 5 mM solution) was treated with DTT (25 µL of a 100 mM solution in H2O, pH 9.5). DTT (pKa1 ) 9.2) was used as the internal buffer by adjusting the pH with NaOH. The reaction was allowed to stand at room temperature for 8 h, quenched by the addition of 0.500 mL of 0.2 M phosphate buffer (pH 6.5), and analyzed by LC/MS. Compound 7: retention times (Nucleodur column), 8.4, 10.0, 11.8, and 12.2 min. UV (CH3CN/10 mM NH4OAc pH 5.5) (λmax) 314 nm; 258 nm. MS (ES+): 396 (M + H+), 418 (M + Na+), 434 (M + K+). HRMS (HPLC-ESI-TOF): retention times, 9.4, 10.5, 11.4, and 11.9 min. Calculated for C16H20N3O2S2 (M + H+): 396.1046, found, 396.1051 (9.4 min); 396.1050 (10.5 min); 396.1055 (11.4 min), 396.1050 (11.9 min). Analysis of the Products from the Reaction of MC with 1,3-Propanedithiol or DHLA by LC/MS. MC (100 µL of a 5 mM solution in 100 mM Na2CO3 pH 9.5) was treated with 1,3propanedithiol or DHLA (25 µL of a 100 mM solution in CH3CN/ H2O 1:1). The reaction was allowed to stand at room temperature for 8 h, quenched by the addition of 0.200 mL of 0.2 M potassium phosphate (pH 6.5), and analyzed by LC/MS. Compound 8: retention times (Nucleodur column), 20.1 and 20.8 min. UV (CH3CN/10 mM NH4OAc pH 5.5) (λmax) 314 nm; 258 nm. MS (ES+): 350 (M + H+), 372 (M + Na+), 388 (M + K+). 9: retention times (Nucleodur column), 21.7, 22.1, 22.2 (sh), 22.7, 23.0, 23.5, and 23.8 min. UV (CH3CN/10 mM NH4OAc pH 5.5) (λmax) 312 nm; 258 nm. MS (ES+): 450 (M + H+), 472 (M + Na+), 488 (M + K+). HRMS (HPLC-ESI-TOF): retention times, 13.8, 14.3, 14.8, 15.1, 15.4, 15.9, and 16.6 min. Calculated for C21H28N3O4S2 (M + H+), 450.1516; found, 450.1509 (13.8 min); 450.1511 (14.3 min); 450.1517 (14.8 min), 450.1499 (15.1 min), 450.1505 (15.4 min), 450.1015 (15.9 min). Isolation and Characterization of MC-1,3-Propanedithiol Cross-Links. A solution of MC (20 mg, 0.060 mmol) in water (12 mL) was treated with a solution of 1,3-propanedithiol (16 mg, 0.15 mmol) in acetonitrile (3 mL) and NaOH (0.50 mL of a 0.10 M

Chem. Res. Toxicol., Vol. 23, No. 8, 2010 1385 solution in H2O, 0.05 mmol) and stirred at room temperature for 12 h. After this period, the resulting suspension was partially concentrated to remove acetonitrile, extracted with CH2Cl2/i-PrOH 4:1 (3 × 15 mL), dried (Na2SO4), and concentrated in vacuum. The residue was purified by preparative TLC using repeated elutions on an analytical 20 × 20 plate using 5% MeOH in CH2Cl2 as eluant to give 3 mg of cis-8 and about 0.5 mg of trans-8 (yields calculated from the UV absorption at 312 nm, using an extinction coefficient of 11500 M-1 cm-1) (16). 1,2-Cis -8: 1H NMR (CDCl3) δ (ppm) 4.90 (s, 2H, NH2-7); 4.53 (dd, 1H, J ) 7.1, 12.6 Hz, H-3a); 4.38 (d, 1H, J ) 6.2 Hz, H-1); 4.23 (d, 1H, J ) 14.4 Hz, H-10a); 4.20-4.17 (m, 1H, H-2); 3.98 (dd, 1H, J ) 9.0 Hz, 12.6 Hz, H-3b); 3.90-3.88 (m, 1H, H-2); 3.74 (d, 1H, J ) 14.4 Hz, H-10b); 2.82-2.65 (m, 4H, H-11 and H-13); 1.99-1.91 (m, 1H, H-12a); 1.83 (s, 3H, CH3); 1.82-1.76 (m, 1H, H-12b). 13C NMR (CDCl3) δ (ppm) 178.4 (C8); 177.5 (C5); 145.8 (C6); 139.5 (C9a); 129.4 (C4a); 120.6 (C9); 116.8 (C8a); 107.4 (C7); 59.0 (C2); 51.8 (C3); 48.4 (C1); 32.3 (C12); 29.8 (C13); 29.3 (C11); 28.0 (C10); 8.0 (CH3). MS (ES+): 350 (M + H+), 372 (M + Na+), 388 (M + K+). HRMS: calculated for C16H20N3O2S2, 350.0991; found, 350.0996. UV (MeOH) (λmax) 530 nm; 352 nm (sh); 312 nm; 252 nm. 1,2trans-8: 1H NMR (CDCl3) δ (ppm) 4.90 (s, 2H, NH2-7); 4.53-4.49 (m, 1 H, H-3a); 4.28 (d, 1H, J ) 14.2 Hz, H-10a); 3.97-3.91 (m, 3H, H-1, H-2, H-3b); 3.69 (d, J ) 14.2 Hz, H-10b); 2.81-2.74 (m, 4H, H-11 and H-13); 1.98-1.91 (m, 1H, H-12a); 1.83 (s, 3H, CH3); 1.80-1.73 (m, 1H, H-12b). 1H NMR (CD3OD) δ (ppm) 4.59 (s, 2H, NH2-7); 4.44 (dd, 1H, J ) 6.1, 12.9 Hz, H-3a); 4.19 (d, 1H, J ) 14.2 Hz, H-10a); 4.11 (d, 1H, J ) 3.7 Hz, H-1); 3.98 (dd, 1H, J ) 4.0 Hz, 12.9 Hz, H-3b); 3.90-3.88 (m, 1H, H-2); 3.74 (d, 1H, J ) 14.2 Hz, H-10b); 3.90-3.88 (m, 1H, H-2); 2.85-2.70 (m, 4H, H-11 and H-13); 1.97-1.90 (m, 1H, H-12a); 1.80 (s, 3H, CH3); 1.78-1.72 (m, 1H, H-12b). 13C NMR (CDCl3) δ (ppm) 178.3 (C8); 177.5 (C5); 145.8(C6); 137.9 (C9a); 129.3 (C4a); 121.2 (C9); 117.3 (C8a); 107.3 (C7); 63.1 (C2); 53.7(C3); 50.6(C1); 31.5 (C12); 30.5 (C13); 29.0 (C11); 28.1 (C10); 8.0 (CH3). MS (ES+): 350 (M + H+), 372 (M + Na+), 388 (M + K+). HRMS: calculated for C16H20N3O2S2, 350.0991; found, 350.0997. UV (MeOH) (λmax) 534 nm; 357 nm (sh); 312 nm; 253 nm. Reaction of MC with 1,3-Propanedithiol: Influence of the Concentration of 1,3-Propanedithiol. Reaction mixtures were prepared by admixing 50 µL of 5 mM MC in 100 mM sodium carbonate (pH 9.5) with 50 µL of propanedithiol in 1:1 CH3CN/ H2O (propanedithiol stock solutions were 2.0, 5.0, 10.0, 15.0, 20.0, and 50.0 mM). The reactions were maintained at 25 °C for 6 h, then quenched by the addition of 0.200 mL of 0.2 M sodium phosphate (pH 6.5), and analyzed by LC/MS (injection of 100 µL). No internal standard was included due to the inherent complexity of the mixtures. Peaks were identified by their MS and UV, and their abundance was determined by integration of the peaks observed in the HPLC trace recorded at 310 nm. Reaction of MC with DTT: Influence of the Concentration of DTT. Reaction mixtures were prepared by admixing 60 µL of 5 mM MC, 30 mL of H2O, and 10 µL of DTT (pH 9.5, DTT as internal buffer). DTT stock solutions were 15, 20, 30, 60, 90, 120, and 150 mM. The reactions were kept at 25 °C for 6 h, then quenched by the addition of 0.200 mL of 0.20 M sodium phosphate (pH 6.5); the UV spectrum (λ ) 200-400 nm) was recorded, and the mixtures were analyzed by HPLC as above. Reaction of MC with an Equimolar Amount of DTT. A solution of MC (25 mg, 0.075 mmol) in 15 mL of water was treated with a solution of DTT adjusted to pH 9.5 (0.750 mL of a 100 mM solution, 0.075 mmol). The reaction mixture was allowed to stand at room temperature overnight. The resulting suspension was centrifuged, and the supernatant was removed. The solid residue was resuspended and centrifuged successively with MeOH (2 × 5 mL) and acetonitrile (2 × 5 mL). The resulting material was dried in vacuum to give 19 mg of a dark red solid. The combined mother liquors were diluted to 50 mL, and the concentration of mitosenes was determined by UV assuming an extinction coefficient ε312 ) 11500 M-1 cm-1. The total amount of soluble mitosenes was 0.011

1386

Chem. Res. Toxicol., Vol. 23, No. 8, 2010

mmol (15%). Elemental analysis of the red solid: C, 53.2; H, 5.7; N, 13.1; S, 8.1. Reaction of MC with an Equimolar Amount of 1,3-Propanedithiol. The reaction was performed in a manner similar to that used in the DTT reaction described above, using MC (17 mg, 0.050 mmol) in H2O (10 mL), 1,3-propanedithiol (0.50 mL of a 100 mM solution in acetonitrile), and NaOH (0.20 mL of a 100 mM solution in H2O) to give 12 mg of a deep red solid. Elemental analysis: C, 51.0; H, 5.8; N, 12.0; S, 16.3. Analysis of Intermediates Formed during the Reaction of MC with Excess Propanedithiol. MC (400 µL of a 5 mM solution in H2O) was treated with 1,3-propanedithiol (100 µL of a 100 mM solution in CH3CN) and NaOH (30 µL of a 100 mM solution in H2O). After 1 min, the autocatalytic reaction that converts MC to mitosenes was observed visually by a sudden change of color from blue to deep purple. Aliquots of the mixture (50 µL) were removed at time intervals, quenched by the addition of phosphate buffer at pH 6.2 (0.300 mL of 0.10 M), and analyzed by LC/MS (100 µL injection). Compound 10: retention times, 20.2 and 23.0 min. UV (CH3CN/10 mM NH4OAc pH 5.5) (λmax) 306 nm; 248 nm. MS (ES+): 843 (2 M + Na+), 449(M + K+), 433 (M + Na+), 350 (M - HOCONH2 + H+). Y: retention times, 13.9 and 17.6 min. UV (CH3CN/10 mM NH4OAc pH 5.5) (λmax) 310 nm; 256 nm. MS (ES+): 693 (M + K+), 677 (M + Na+), 594 (M - HOCONH2 + H+), 576 (M - HOCONH2 - H2O + H+). HRMS (HPLC-ESITOF): retention times, 14.6 and 18.0 min. Found, 677.1837 (14.6 min); 677.1830 (18.0 min). Calculated for C30H34N6O7S2Na: 677.1823. Trapping of the Intermediates Formed during the Reaction of MC with Excess Propanedithiol with Maleimide or Iodoacetic Acid. MC was activated as described above. Aliquots were removed after 10 min and quenched by the addition of maleimide or sodium iodoacetate (50 µL of a 0.25 M solution). After 15 min, the reactions were quenched by the addition of phosphate buffer at pH 6.2 (0.300 mL of a 0.10 M solution) and analyzed by LC/MS. Retention time for Y-MI: 15.64 min. UV (λmax) 306 nm; 248 nm. MS (ES+): 790 (M + K+), 774 (M + Na+), 691(M - HOCONH2 + H+), 673 (M - HOCONH2 - H2O + H+). Retention time for 10-MI: 17.0 min. UV (λmax) 310 nm; 254 nm. MS (ES+): 546 (M + K+), 530 (M + Na+), 447 (M - HOCONH2 + H+). Retention times for Y-IA: 9.4 and 10.9 min. UV (λmax) 304 nm; 250 nm. MS (ES+): 751 (M + K+), 735 (M + Na+), 652 (M - HOCONH2 + H+). Retention times for 10-IA: 12.5 and 16.5 min. UV (λmax) 311 nm; 256 nm. MS (ES+): 507 (M + K+), 491 (M + Na+), 408 (M - HOCONH2 + H+).

Paz

Results

Figure 1. HPLC traces for the reaction of MC with dithiols: (a) DTT; (b) DHLA; (c) 1,3-propanedithiol. Reactions contained a molar ratio dithiol/MC ) 5 and were performed at 20 °C for 8 h, quenched by acidification, and analyzed by LC/MS (λ ) 310 nm).

Products Formed in the Reaction of MC with Excess Thiol. The initial experiments on the reaction of MC with dithiols were performed using DTT. Aqueous solutions of MC were treated with a 5 molar excess of DTT at pH values between 9 and 10. The pH was internally regulated by the thiol/thiolate system. The addition of thiol resulted in an almost immediate reductive activation of MC that could be observed by the naked eye by a change in color from blue to purple. The reactions were allowed to continue for several hours, neutralized, and analyzed by HPLC. A relatively simple chromatogram was observed consisting of 5 major peaks (Figure 1a). One of them was identified as oxidized DTT, while the other 4 peaks presented identical UV and MS spectra. The absorbance maxima (314 and 250 nm) in the UV spectra were characteristic of a mitosene chromophore, and the observed ions in the MS were consistent with a mass of 395 amu, which corresponds to the product resultant from the addition of one molecule of DTT to one molecule of activated MC at its 1- and 10-positions. The presence of four peaks in the HPLC chromatogram is justified by the formation of four diastereomers as the result of the known formation of cis- and trans-isomers at C-1 in the addition of

nucleophiles at activated MC and the use of racemic DTT. The reaction mixture was also analyzed by LC/HRMS. Again, four peaks were observed that fitted the expected exact mass for 7 (Chart 1). DHLA was reacted with MC analogously, resulting in a more complex chromatogram (Figure 1b), where at least 6 peaks with identical UV and MS spectra were observed. The UV spectrum of the six peaks was indicative of a mitosene structure, and the observed MS ions were concordant with a mass of 449 amu, as expected for the addition of one molecule of DHLA to one molecule of activated MC. The observed peaks were assigned to six of the eight expected isomers for 9 (Chart 1). The LC/HRMS revealed seven peaks (and one shoulder) that gave ions within the calculated exact mass for 9 (see Supporting Information for details). The large number of diastereomers formed with racemic DTT and unsymmetrical, racemic DHLA directed us to explore the reaction of MC with 1,3-propanedithiol, anticipating a more manageable mixture, easier to purify and characterize. As expected, the reaction of MC and 1,3-propanedithiol gave two major adducts (Figure 1c) with the distinctive UV of a mitosene structure and mass spectra showing ions corresponding to a mass of 349 amu, consistent

Cross-Linking of Dithiols by Mitomycin C

Chem. Res. Toxicol., Vol. 23, No. 8, 2010 1387

Chart 1. Structures of MC-Dithiol Cross-Links

with the expected structure 8 (Chart 1). With this result in hand, we focused on the preparation of enough quantities of 8 to perform a complete characterization. Isolation and Characterization of MC-1,3-Propanedithiol Cross-Links. Larger scale reactions were performed in 10-mg scale, and they gave the expected products. As 8 was soluble in CH2Cl2 and poorly soluble in water, minor water-soluble mitosenes could be removed by aqueous extraction. HPLC gave a poor and nondependable separation for the two isomers that in many cases resulted in a single peak for the two diastereomers. However, the two compounds could be separated by repeated elutions on TLC, providing enough material for a rigorous characterization. High resolution mass spectra were consistent with the expected molecular formula for both isomers. The 1H NMR of cis-8 (Figure 2A) presented peaks for the expected hydrogen atoms, which were assigned by a COSY experiment. The 1H NMR of trans-8 in CDCl3 showed collapsing peaks for H-1, H-2 and one H-3 (Figure 2B), but the use of CD3OD resulted in well separated peaks that allowed the measurement of coupling constants and the assignment of all peaks by a COSY experiment (see Supporting Information for details). The two doublets corresponding to the AB quartet for H-10 were separated by some 0.5 ppm for both isomers, an indication of a constrained methylene group. By contrast, mitosenes with a linear chain at C-10 showed a separation of less than 0.1 ppm between the two doublets (17–19). The 13C NMR spectrum for cis-8 presented 16 peaks that could be

Figure 2. Fragment of the 1H NMR spectra in CDCl3 for cis-8 (A) and trans-8 (B). See Chart 1 for numbering.

Figure 3. (a) CD spectrum of 1,2-cis-8 (red) and 1,2-trans-8 (blue) in MeOH. (b) UV-vis spectrum of 1,2-cis-8 (red) and 1,2-trans-8 (blue) in MeOH.

unambiguously assigned by a combination of DEPT, HMBC, and HMQC experiments. The 13C NMR of trans-8 also showed the expected 16 peaks, which were assigned by comparison to the 13C NMR spectrum of cis-8. The stereochemistry at C-1 was assigned from two observations: The circular dichroism spectra for the two compounds were almost symmetrical and presented the expected pattern for isomeric mitosenes at C-1, with the diagnostic band at 530 nm (20) (Figure 3). The major isomer bore a cis-configuration, similar to what has been observed for the addition of other nucleophiles at C-1 of activated MC. Additionally, the coupling constants J1-2 were 6.2 Hz (major isomer) and 3.7 Hz (minor isomer). Similarly diverging J1-2 values for cis- and trans-isomers at C-1 of mitosenes have been previously observed (17–19). Influence of the Concentration of Dithiol on the Outcome of the Reaction. The distribution and relative amounts of products formed in the reaction of MC with different concentrations of DTT or 1,3-propanedithiol was studied. MC was reacted with dithiols for 6 h using increasing molar ratios dithiol/MC, and the product composition was analyzed by LC/MS. Because of the inherent complexity of the reaction mixtures, no internal standard was used; instead, precautions were taken to ensure the same final concentration of MC on all samples, and exactly the same volumes of sample were injected for LC/MS analysis. The results of these experiments are shown in Figure 4. The reactions with substoichiometric dithiols resulted in the formation of 1-hydroxymitosenes (5a) as major products, as previously reported (12). The use of molar ratios dithiol/MC higher than two suppressed the formation of 5a and resulted in the formation of dithiol cross-links as major products (Chart 1). When equimolar amounts of dithiol and MC were used, minimal amounts of HPLC-detectable mitosenes were observed (Figure 4), and an insoluble red-brown precipitate was observed for both dithiols assayed. This issue was further investigated using higher scale reactions. Reactions of MC with Equimolar Amounts of Dithiols. The reaction of MC with one molar equivalent of either DTT or 1,3-propanedithiol to give insoluble mitosene-containing materials as major products was reproducibly performed several times at a 50-75 µmol scale. The autocatalytic reaction (12) was observed after about 1 min for 1,3-propanedithiol and after 3-5 min for DTT to give transparent purple solutions. After 1 h, the solutions started to turn turbid, and after a few hours,

1388

Chem. Res. Toxicol., Vol. 23, No. 8, 2010

Paz

Figure 5. Time course of the reaction of MC with 1,3-propanedithiol. Aliquots of a reaction mixture containing 5 mM MC and 25 mM propanedithiol at pH 10.0 (dithiol as internal buffer) were removed at time intervals after the autocatalytic reaction was observed, quenched by the addition of phosphate buffer at pH 6.0, and analyzed by HPLC (λ ) 310 nm). The relative abundance of each compound was determined from the area of the observed peaks: 5a (red squares); transient intermediate Y (blue circles); and cross-links 8 (green triangles). The abundance of each compound was normalized (1 for the sum of mitosenes observed at t ) 0). Figure 4. Dependence of the reaction of MC and dithiols on the concentration of dithiol: (a) DTT; (b) 1,3-propanedithiol. Reaction mixtures containing MC with increasing concentrations of dithiol were analyzed by HPLC (λ ) 310 nm). Lines are hydroxymitosenes 5a (blue circles); cross-links 8 (red squares); and the sum of HPLC-detectable mitosenes (green triangles). The relative abundance of each compound was determined from the area of the observed peaks. The relative abundance of each compound was normalized (1 for the sum of all mitosenes observed with the lowest ratio dithiol/MC).

a significant amount of a purple precipitate could be observed. This material, thereafter termed mitosenes X, was collected by centrifugation and was washed with several solvents, as the solid was insoluble in common organic solvents (halogenated solvents, alcohols, dioxane, EtOAc, acetone, and pyridine). Mitosenes X could be dissolved in DMSO and DMF to give deepred solutions. Mass recoveries for mitosenes X averaged 80% of the original MC mass, and the combined supernatants contained 8-20% of the expected absorbance for full conversion of MC to mitosenes. The HPLC analysis of the supernatant showed a complex mixture of residual soluble mitosenes, containing 5a and dithiol cross-links among other products. UV spectra for mitosenes X were obtained by dilution of an aliquot of a DMSO solution with MeOH, and they showed absorbance maxima at wavelengths characteristic of mitosene structures. Several MS experiments (ES+, ES-, MALDI-TOF) did not provide decipherable peaks. The 1HNMR spectrum gave a complex set of broad peaks at frequencies expected for mitosene compounds (see Supporting Information for details). The 13C NMR spectrum of a sample containing approximately 15 mg/ mL of mitosenes X in DMSO gave no discernible peaks after an overnight acquisition. The derivatization of mitosenes X with Ac2O or Boc2O resulted, in some cases, in the formation of mitosene materials that were soluble in organic solvents, but all attempts to ascertain its identity failed. The elemental analysis of samples of mitosenes X from DTT indicated that on average it contained two mitosene molecules per molecule of DTT, while for the material obtained from 1,3-propanedithiol, the ratio of mitosene/dithiol was 1:1 (see Supporting Information for details). Intermediates Formed during the Reaction of MC with Excess Propanedithiol. We performed a number of experiments aimed at detecting transient intermediates formed during the reaction of MC with 1,3-propanedithiol. Aliquots of a reaction mixture of MC and propanedithiol were removed at time intervals and quenched by lowering the pH to 5-6. LC/MS

analysis showed the formation of 1-hydroxymitosenes in aliquots quenched immediately after the autocatalytic reaction and the formation of cross-links 8 after prolonged times (Figure 5). Aliquots quenched after intermediate times (5-20 min) showed the presence of a transient major product, hereafter termed Y, with an apparent mass of 654 amu. Peaks corresponding to monoadduct 10 were also observed, but they were formed in relatively small quantities (Figure 6A). Intermediates Y were formed independent of the nature of the buffer used in the quenching reaction, and they were also observed when the reaction employed a 10× dilution relative to that described in Materials and Methods, both for the reaction and/or the quenching. When aliquots taken after 5-10 min of the autocatalytic reaction were quenched with the thiol trapping reagents maleimide or iodoacetate, adducts corresponding to maleimide or iodoacetate adducts of Y were observed (termed Y-MI and Y-IA, respectively, Chart 2), together with minor peaks corresponding to the addition of maleimide or iodoacetate on 10 (Figure 6B,C). Several attempts at isolating intermediate Y or its maleimide or iodoacetate derivatives using larger scale reactions failed. An exact mass for Y could be obtained using LC/HRMS, and the only plausible formula fitting the observed ions was C30H34N6O7S2 (see Supporting Information for details), indicating the formation of a dimeric structure containing two mitosene molecules and one molecule of dithiol.2

Discussion We recently reported that simple dithiols, including the oxidoreductase cofactor DHLA, were able to perform the reductive activation of MC (12). This discovery resolved two controversial observations: on the one hand it was reported that MC was inert toward thiols (21, 22), but on the other hand the 2 The intermediate reacted with thiol-trapping reagents to give monoalkylated derivatives, therefore, must contain one free SH group. Hydroxymitosenes 5a were recovered intact after maleimide or iodoacetate treatment. It is therefore concluded that functional groups (other than thiol) present in the mitosene Y are inert toward those electrophiles in the reaction conditions we used. From the fragmetation pattern observed in the EI-MS, we can conclude that Y contains one carbamoyloxy group and one hydroxy group. Considering that the molecule must contain at least one dithiol group, the only plausible formula fitting the observed mass is C30H34N6O7S2 (calculated from the sodium adduct). With the achieved data, we were unable to deduce a structure for Y based on the known reactivity of MC, but we hypothesize that Y is a dimer of 10 with another intermediate mitosene. More details on this issue are discussed in an addendum included in Supporting Information.

Cross-Linking of Dithiols by Mitomycin C

Figure 6. Detection of intermediates formed during the reaction of MC with 1,3-propanedithiol. Aliquots from the reaction of MC (4 mM) with propanedithiol (20 mM) were quenched after 10 min with NH4Ac (A), iodoacetate (B). or maleimide (C) and analyzed by HPLC (λ ) 310 nm). Y represents the unknown intermediates with a mass of 654 amu. Y-IA and Y-MI represent iodoacetate and maleimide adducts of Y. 10-IA and 10-MI represent iodoacetate and maleimide adducts of 10.

Chart 2. Structures for MC-Propanedithiol Monoadducts Identified after Quenching the Reaction of MC and 1,3-Propanedithiol with Acid, Iodoacetate, or Maleimide

dithiol active site of glucose regulatory protein was found to be involved in the reductive activation of MC (10, 11). The activation of MC with dithiols was 2 orders of magnitude faster than that with monothiols, and the pH was identified as the key factor to achieve a successful activation: the activation of MC by dithiols occurred most efficiently at pH values averaging the pKa1 and pKa2 of the dithiol. The activation reaction proceeded by an autocatalytic mechanism, similar to other slowreducing agents for MC (16, 23). As a consequence of the autocatalytic mechanism, the reaction could be performed

Chem. Res. Toxicol., Vol. 23, No. 8, 2010 1389

efficiently using substoichiometric quantities of dithiol. Under those conditions, the formation of aziridinomitosene 4 as the initial product of the reaction was evidenced by trapping 4 with diverse nucleophiles to give mitosenes 5a-d, resulting from alkylation reactions at C-1 of 4 (Scheme 1) (12). Continuing our research on the reactivity of MC with dithiols, we found that the reaction of MC with several dithiols resulted in the formation of covalent bis-adducts of MC with dithiol as the vastly predominant end product. The initial reactions were performed using DTT, and the product profile observed in the HPLC trace (Figure 1a) consisted of four mitosenes with identical UV and MS that we identified as the four possible diastereomeric mitosenes resulting from the alkylation of DTT by a single MC molecule at its 1- and 10-positions. HPLC analysis of reactions between the oxidoreductase cofactor DHLA and MC (Figure 1b) revealed the formation of at least seven of the eight expected isomers, resulting from bis-alkylation of MC with one molecule of DHLA. LC-HRMS gave exact masses that confirmed the molecular formula of the proposed DTT cross-links 7 and DHLA cross-links 9 (Chart 1). For a full characterization of the cross-links of MC with dithiols, we used 1,3-propanedithiol as a means to simplify the composition of the reaction mixture. As expected, the reaction of MC with 1,3propanedithiol (Figure 1c) gave two isomers with UV and MS supporting structure 8. We should note here that the simple chromatograms obtained after the reaction of MC with dithiols (Figure 1) do not imply that all MC is converted to MC-dithiol cross-links. Side reactions do occur, but, as we will discuss later, they result in the formation of unidentified insoluble mitosenes that do not appear in the HPLC trace. The two isomers of 8 could be separated and fully characterized using NMR spectroscopy. The main evidence for the stereochemistry at C-1 was obtained using circular dichroism, a well established technique to assign the stereochemical configuration at C-1 of 2-aminomitosenes (20). In our case, the isomers from the reaction of MC and 1,3-propanedithiol gave CD spectra that were almost symmetrical (Figure 3), and confirmed the expected stereochemistry at C-1, with the major isomer bearing a cis-configuration. A study of the dependence of the reaction on dithiol concentration provided insight in the mechanism for the formation of 8. One observation was that the formation of crosslinks required a ratio thiol/MC higher than two (ideally four to six) for optimum conversion to cross-links. No mitosenes substituted with two molecules of dithiol were detected, even when a large excess of dithiol (10-fold) was used. Also, no products resulting from the alkylation of one molecule of dithiol by two mitosenes were detected, although such products were to be expected when using low dithiol/MC ratios. When the reaction was performed with stoichiometric dithiol (or with small excess) only minor amounts of cross-links were observed (Figure 4). The most intriguing result observed during the dithiol concentration-dependence experiments was that when equimolar concentrations of MC and dithiols were used most of the HPLCdetectable mitosenes disappeared. Larger scale reactions showed that this was caused by the formation of a mitosene-containing material, whose solubility differed dramatically from that for all mitosenes known to us. This intractable material (mitosenes X) could not be identified using available spectroscopic methods (see Supporting Information for details). Elemental analysis indicated that this material contains mitosene and dithiol in ratios ranging from 2:1 to 1:1. As the only reacting species are MC, water, and dithiol, the most plausible interpretation is that the autocataytic activation of MC generates 4 that then reacts with 1,3-propanedithiol to give monoadduct 10 as the initial product. After-

1390

Chem. Res. Toxicol., Vol. 23, No. 8, 2010

Scheme 2. Proposed Mechanism for the Formation of Mitomycin C-1,3-Propanedithiol Cross-Links

wards, compound 10 evolves by an unknown pathway to form the unsoluble material, perhaps a polymer formed as a result of the electrophilic/nucleophilic nature of 10 and other mitosenes. We next performed a number of experiments aimed at detecting intermediates formed during the reaction of MC with 1,3-propanedithiol, with the expectancy of observing the formation of dithiol monoadduct 10 as the major transient adduct. This compound was indeed observed, although in minor proportion compared to other mitosenes (Figure 6A). The major intermediate observed during those experiments was a mitosene with a mass of 654 amu, indicative of a dimeric mitosene. When maleimide or iodoacetate were used to quench aliquots of the reaction, derivatives of Y substituted with a single molecule of alkylating agent were obtained (Chart 2 and Figure 6B,C). All attempts to isolate these intermediates for a rigorous characterization failed, but they can be considered as a footprint for the presence of 10 in the reaction mixture.2 In light of these results, we propose the mechanism shown in Scheme 2, with three dithiol molecules participating at different stages in the cascade of reactions leading to 8. The third dithiol molecule is needed for an additional reductive activation step that activates C-10 for the addition of the second arm of the dithiol, thus explaining that this reaction is favored by a large excess of dithiol. The first step is the autocatalytic reaction of mitomycin C with substoichiometric dithiol to form aziridinomitosene 4; in the second step, a nucleophilic addition of thiolate to C-1 of 4 forms monoadduct 10; the third step involves the reduction of the mitosene-dithiol monoadduct 10 by a third molecule of dithiol to the hydroquinone form 11, with the subsequent elimination of the carbamoyloxy group of the C-10 position to give 12; in the fourth step, a nucleophilic addition of the second arm of the dithiol to the C-10 position generates the cross-link adduct in the reduced form 13; in the final step, oxidation to the quinone form generates the end product 8. Alternatively, 8 could be formed by a direct addition

Paz

of thiolate at C-10 in 10. SN2 substitutions at C-10 of mitosenes have been reported (7, 24), but in this case, the low yields of 8 observed in the reactions with low excess of dithiols argue in favor of the necessity of a second reductive activation step.3 The low yields of 8 obtained in reactions using ratios of dithiol/ MC between 1 and 2 can be justified by the existence of a competing pathway for the evolution of the initially formed monoadduct 10 as discussed above. Another observation arising from the concentration-dependence experiments is that minimal amounts of 5a were formed when molar ratios dithiol/MC higher than 1 were used (Figure 4), reflecting a remarkable preferential reactivity of activated MC toward thiols in aqueous solution, even in the presence of a considerable concentration of hydroxyl anion (reactions were performed at pH 9.5-10). The ease of formation of MC cross-links with simple dithiols raises the question as to whether biological dithiols could be a substrate for MC-induced S,S′-cross-linking. The thiol group of one of the cysteine residues in thioredoxin superfamily is known to be deprotonated at physiological pH and to react as an excellent nucleophile (25), and a similar reactivity has been observed for the selenol group in thioredoxin reductase (TrxR) (26). Alkylation reactions of Trx and TrxR with electrophiles that target DNA such as cisplatin (27) and 4-hydroxy-2-nonenal (28) have been reported, and bis-electrophilic mustards alkylated a single cysteine in the dithiol active site of trypanothione reductase (29).4 The alkylation of TrxR with the bis-electrophile curcumin has been reported to give bis-adducts containing two curcumin molecules, but no cross-links were detected (30). The reactions presented here were performed at pH 9.5-10, and such conditions may seem of no biological significance. However, we must consider that the reactive species in the reactions of thiols as nucleophiles in water is the thiolate (31); therefore the pKa of the thiol dictates the optimum pH value for their reactivity. In the reactions reported here, we employed pH values close to the average pKa of the dithiol (DTT, 9.8; propanedithiol, 10.3; DHLA, 10.7) because that was found to be the optimum pH for the reductive activation of MC by dithiols (12). An extrapolation of the reactivity reported here to dithiol-containing proteins must take into account that the pKa of their dithiol group is abnormally low (about 2 units lower than those of the simple dithiols used here) (32) and that they behave as good nucleophiles at physiological pH (27–30); therefore, the reactions of MC with simple thiols at pH 9.5-10 reported here may be expected to occur with dithiol-containing proteins at physiological pH values. Geometric evaluations indicate that dithiol groups in proteins are well suited to give cross-links with MC: The distance between the two sulfur atoms in reduced human thioredoxin is 3.1 Å, while in E. coli 3 Alternatively, 8 could be formed directly from 3 (bifunctionally activated MC) by sequential alkylation at C-1 and C-10, without the need of a second reducing agent. This would imply that the use of excess dithiol reduces MC fast enough to generate significant amounts of 3, thus reducing the extent of formation of 4 through the autocatalytic path (Scheme 1). Such a scenario does not concord well with the curves obtained during our kinetic studies on the activation of MC with excess dithiols, which indicated a quick autocatalytic reaction at pH 9-10 (12) that gives 4 as the major end product. This was evidenced by the formation of 5a as the only product when reactions are quenched immediately after the autocatalytic process (see Supporting Information, Figure S22a). However, it cannot be rigorously ruled out that 4 is subsequently reduced by excess dithiol to give 3. 4 The work reported in ref 29 indicates that two quinacrine mustard molecules reacted in the active site of trypanothione reductase, an enzyme containing a dithiol active site (Cys-53 and Cys-58). One of the mustard molecules cross-linked the protein by covalently joining the carboxylic side chains from an aspartate and a glutamate residue. Another molecule of mustard alkylated Cys-53 in the dithiol active site. The second arm of the mustard and Cys-58 remained intact.

Cross-Linking of Dithiols by Mitomycin C

thioredoxin, it is 3.8 Å (32). These values are similar to the 3.5 Å distance between the exocyclic 2-amino groups of opposite deoxyguanosine residues at CpG sequences, which are the target for DNA-DNA cross-links generated by MC (33). It is also comparable to the C1-C10 distance calculated for an energyminimized bis-electrophilic mitosene5 (3.38 Å) and to the C1-C10 distance of 3.4 Å measured in the solution structure of an oligonucleotide modified by MC (34). Consequently, we consider that the formation of S,S′-cross-links by MC might occur in proteins containing a dithiol active site. This hypothesis is based on the results presented here, on the known reactivity of dithiols in proteins, and on geometric considerations. The potential formation of dithiol cross-links in proteins by MC constitutes an exciting prospect, as the only reported precedents of chemicals that link two sulfur atoms in dithiol-containing proteins are arsenite derivatives (35).6 We should also remark that prospective studies in the reductive activation of MC by dithiol-containing proteins must bear in mind that the activation reaction might result in inhibition, as a consequence of alkylation reactions in the active site of the enzyme by activated MC.

Conclusions Cross-link adducts formed by MC with three dithiols, including the biological cofactor dihydrolipoic acid, are reported. The formation of the mitomycin-dithiol cross-link requires excess dithiol, and we propose that a second quinone reduction is needed for the activation of the initially formed dithiolmitosene monoadduct for nucleophilic addition at C10. A study of the time course for the reaction of MC with 1,3-propanedithiol showed an initial fast formation of an aziridinomitosene that was slowly converted to dithiol cross-links. The chemical crosslinking of the two sulfur atoms in lipoate cofactors presented here has no reported precedent in the literature to the best of our knowledge. The biological target of MC is considered to be DNA, but the biological activity of electrophiles is also frequently associated with the formation of covalent protein adducts, occurring predominantly at cysteine residues (39). Our work using small dithiols as a chemical model for dithiolcontaining proteins has shown that they are oxidized and crosslinked by MC, and further work should clarify whether proteins with a dithiol active site could be a target for oxidation and/or formation of covalent cross-links by this drug. Acknowledgment. This work was supported by grants from the Spanish Ministry of Education and Science (PS09/00501). We are grateful to Dr. Hitoshi Arai of Kyowa Hakko Kogyo Co., Ltd., for a gift of mitomycin C. We thank Dr. Jose´ Manuel Seco (USC) for performing the CD spectroscopy experiments. We thank Dr. Eugenio Va´zquez Sentı´s and Dr. Jose´ Va´zquez Tato (USC) for the use of their facilities. Supporting Information Available: Spectral data for compounds cis-8 and trans-8; UV and MS spectra for 7, 9, 10, 10IA, 10-MI, Y, Y-IA, and Y-MI; addendum with data on other intermediates formed in the reaction of MC with 1,3-propanedithiol; and a discussion on the possible identitity of 5 The energy-minimized structure for the aziridine-opened electrophilic intermediate derived from 3 was calculated by molecular mechanics (MM2, Chem3D Pro 11.0, Cambridgesoft Corp., Cambridge, MA). 6 The mechanism of inhibition of dithiol-containing enzymes by arsenites has been classically attributed to the formation of complexes with vicinal dithiols (36). Additionally, the formation of dithiol cross-links in Keap1 by some inducers of phase 2 enzymes has been hypothesized (37), and bis-maleimides have been used to cross-link thiol groups in protein disulfide isomerase, but these thiol groups were located at different dithiol sites (38).

Chem. Res. Toxicol., Vol. 23, No. 8, 2010 1391

mitosene Y. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Paz, M. M. (2008) Antitumor Antibiotics, in Anticancer Therapeutics (Missailidis, S., Ed.) pp 112-115, John Wiley and Sons, Chichester, UK. (2) Teus, M. A., de Benito-Llopis, L., and Alio´, J. L. (2009) Mitomycin C in corneal refractive surgery. SurV. Ophthalmol. 54, 487–502. (3) Iyer, V. N., and Szybalsk, W. (1964) Mitomycin and porfiromycin: chemical mechanism of activation and cross-linking of DNA. Science 145, 55–58. (4) Tomasz, M. (1995) Mitomycin C: small, fast and deadly (but very selective). Chem. Biol. 2, 575–579. (5) Tomasz, M., and Lipman, R. (1981) Reductive metabolism and alkylating activity of mitomycin C induced by rat liver microsomes. Biochemistry 20, 5056–5061. (6) Paz, M. M. (2010) Reaction of reductively activated mitomycin C with aqueous bicarbonate: Isolation and characterization of an oxazolidinone derivative of cis-1-hydroxy-2,7-diaminomitosene. Bioorg. Med. Chem. Lett. 20, 31–34. (7) Sharma, M., and Tomasz, M. (1994) Conjugation of glutathione and other thiols with bioreductively activated mitomycin C. Effect of thiols on the reductive activation rate. Chem. Res. Toxicol. 7, 390–400. (8) Tomasz, M., Lipman, R., Chowdary, D., Pawlak, J., Verdine, G. L., and Nakanishi, K. (1987) Isolation and structure of a covalent crosslink adduct between mitomycin C and DNA. Science 235, 1204–1208. (9) Cummings, J., Spanswick, V. J., Tomasz, M., and Smyth, J. F. (1998) Enzymology of mitomycin C metabolic activation in tumour tissue: implications for enzyme-directed bioreductive drug development. Biochem. Pharmacol. 56, 405–414. (10) Celli, M., and Jaiswal, A. K. (2003) Role of GRP58 in mitomycin C-induced DNA cross-Linking. Cancer Res. 63, 6016–6025. (11) Adikesavan, A. K., and Jaiswal, A. K. (2007) Thioredoxin-like domains required for glucose regulatory protein 58 mediated reductive activation of mitomycin C leading to DNA cross-linking. Mol. Cancer Ther. 6, 2719–2727. (12) Paz, M. M. (2009) Reductive activation of mitomycin C by thiols: Kinetics, mechanism, and biological implications. Chem. Res. Toxicol. 22, 1663–1668. (13) Kontou, M., Adelfalk, C., Ramirez, M. H., Ruppitsch, W., HirschKauffmann, M., and Schweiger, M. (2002) Overexpressed thioredoxin compensates Fanconi anemia related chromosomal instability. Oncogene 21, 2406–2412. (14) Ruppitsch, W., Meisslitzer, C., Hirsch-Kauffmann, M., and Schweiger, M. (1998) Overexpression of thioredoxin in Fanconi anemia fibroblasts prevents the cytotoxic and DNA damaging effect of mitomycin C and diepoxybutane. FEBS Lett. 422, 99–102. (15) Yokomizo, A., Ono, M., Nanri, H., Makino, Y., Ohga, T., Wada, M., Okamoto, T., Yodoi, J., Kuwano, M., and Kohno, K. (1995) Cellular levels of thioredoxin associated with drug sensitivity to cisplatin, mitomycin C, doxorubicin, and etoposide. Cancer Res. 55, 4293–4296. (16) Hoey, B. M., Butler, J., and Swallow, A. J. (1988) Reductive activation of mitomycin C. Biochemistry 27, 2608–2614. (17) Schiltz, P., and Kohn, H. (1993) Studies on the reactivity of reductively activated mitomycin C. J. Am. Chem. Soc. 115, 10510–10518. (18) Kohn, H., Zein, N., Lin, X. Q., Ding, J.-Q., and Kadish, K. M. (1987) Mechanistic studies on the mode of reaction of mitomycin C under catalytic and electrochemical reductive conditions. J. Am. Chem. Soc. 109, 1833–1840. (19) Schiltz, P., and Kohn, H. (1993) Studies on the use of Na2S2O4 for the reductive activation of mitomycin C. J. Am. Chem. Soc. 115, 10497–10509. (20) Tomasz, M., Jung, M., Verdine, G., and Nakanishi, K. (1984) Circular dichroism spectroscopy as a probe for the stereochemistry of aziridine cleavage reactions of mitomycin C. Application to adducts ofmitomycin with DNA constituents. J. Am. Chem. Soc. 106, 7367–7370. (21) He, Q.-Y., Maruenda, H., and Tomasz, M. (1994) Novel bioreductive activation mechanism of mitomycin C derivatives bearing a disulfide substituent in their quinone. J. Am. Chem. Soc. 116, 9349–9350. (22) Sharma, M., and Tomasz, M. (1994) Effects of glutathione on alkylation and cross-linking of DNA by mitomycin C. Isolation of a ternary glutathione-mitomycin-DNA adduct. Chem. Res. Toxicol. 6, 390–400. (23) Peterson, D. M., and Fisher, J. (1986) Autocatalytic quinone methide formation from mitomycin C. Biochemistry 25, 4077–4084. (24) Schiltz, P., and Kohn, H. (1992) Sodium dithionite-mediated mitomycin C reductive activation processes. Tetrahedron Lett. 33, 4709–4712. (25) Kallis, G. B., and Holmgren, A. (1980) Differential reactivity of the functional sulfhydryl groups of cysteine- 32 and cysteine-35 present

1392

(26)

(27)

(28) (29)

(30) (31)

Chem. Res. Toxicol., Vol. 23, No. 8, 2010

in the reduced form of thioredoxin from Escherichia coli. J. Biol. Chem. 255, 10261–10265. Zhong, L., Arne´r, E. S. J., and Holmgren, A. (2000) Structure and mechanism of mammalian thioredoxin reductase: The active site is a redox-active selenolthiol/selenenylsulfide formed from the conserved cysteine-selenocysteine sequence. Proc. Natl. Acad. Sci. U.S.A. 97, 5854–5859. Arne´r, E. S. J., Nakamura, H., Sasada, T., Yodoi, J., Holmgren, A., Spyrou, G., Fang, J., Lu, J., and Holmgren, A. (2001) Analysis of the inhibition of mammalian thioredoxin, thioredoxin reductase, and glutaredoxin by cis-diamminedichloroplatinum (II) and its major metabolite, the glutathione-platinum complex. Free Radical Biol. Med. 31, 1170–1178. Fang, J., and Holmgren, A. (2006) Inhibition of thioredoxin and thioredoxin reductase by 4-hydroxy-2-nonenal in vitro and in vivo. J. Am. Chem. Soc. 128, 1879–1885. Saravanamuthu, A., Vickers, T. J., Bond, C. S., Peterson, M. R., Hunter, W. N., and Fairlamb, A. H. J. (2004) Two interacting binding sites for quinacrine derivatives in the active site of trypanothione reductase: a template for drug design. J. Biol. Chem. 279, 29493– 29500. Fang, J., Lu, J., and Holmgren, A. (2005) Thioredoxin reductase is irreversibly modified by curcumin: a novel molecular mechanism for its anticancer activity. J. Biol. Chem. 280, 25284–90. Esterbauer, H. (1970) Kinetik der reaktion von sulfhydrylverbindungen mit R,β-ungesa¨ttigten aldehyden in wa¨βrigem system. Monatsh. Chem. 101, 782–810.

Paz (32) Holmgren, A. (1995) Thioredoxin structure and mechanism: conformational changes on oxidation of the active-site sulfhydryls to a disulfide. Structure 3, 239–243. (33) Tereshko, V., Minasov, G., and Egli, M. (1999) The Dickerson-Drew B-DNA dodecamer revisited at atomic resolution. J. Am. Chem. Soc. 121, 470–471. (34) Sastry, M., Fiala, R., Lipman, R., Tomasz, M., and Patel, D. J. (1995) Solution structure of the monoalkylated mitomycin C-DNA complex. J. Mol. Biol. 247, 338–359. (35) Ramadan, D., Cline, D. J., Bai, S., Thorpe, C., and Schneider, J. P. (2007) Effects of As(III) binding on beta-hairpin structure. J. Am. Chem. Soc. 129, 2981–2988. (36) Aposhian, H. V., and Aposhian, M. M. (2006) Arsenic toxicology: five questions. Chem. Res. Toxicol. 19, 1–15. (37) Dinkova-Kostova, A. T., Holtzclaw, W. D., Cole, R. N., Itoh, K., Wakabayashi, N., Katoh, Y., Yamamoto, M., and Talalay, P. (2002) Direct evidence that sulfhydryl groups of Keap1 are the sensors regulating induction of phase 2 enzymes that protect against carcinogens and oxidants. Proc. Natl. Acad. Sci. U.S.A. 99, 11908–11913. (38) Hawkins, H. C., de Nardi, M., and Freedman, R. B. (1991) Redox properties and cross-linking of the dithiol/disulphide active sites of mammalian protein disulphide-isomerase. Biochem. J. 275, 341–348. (39) Liebler, D. C. (2008) Protein damage by reactive electrophiles: targets and consequences. Chem. Res. Toxicol. 21, 117–128.

TX100134H