One-Electron Reduction of 17-(Dimethylaminoethylamino)-17

ARTICLE pubs.acs.org/JPCA

One-Electron Reduction of 17-(Dimethylaminoethylamino)-17demethoxygeldanamycin: A Pulse Radiolysis Study Sara Goldstein* Chemistry Institute, The Accelerator Laboratory, The Hebrew University of Jerusalem, Jerusalem 91904, Israel ABSTRACT: Geldanamycin, a benzoquinone ansamycin antibiotic, is a natural product inhibitor of Hsp90 with potent and broad anticancer properties but with unacceptable levels of hepatotoxicity. Consequently, numerous structural analogs, which differ only in their 17-substituent, have been synthesized including the water-soluble and less toxic 17-(dimethylaminoethylamino)-17-demethoxygeldanamycin (17-DMAG, Alvespimycin). It has been suggested that the different hepatotoxicity reflects the redox active properties of the quinone moiety. The present pulse radiolysis study was aimed at studying the one-electron reduction of 17-DMAG. The UV
visible spectrum of the semiquinone radical, its pKa, and the second-order rate constants for the reactions of 17-DMAG with CO2•
and (CH3)2C•OH have been obtained. The reduction potential of 17-DMAG has been determined to be
194 ( 6 mV (vs NHE) using oxygen, 1,4-naphthoquinone, and menadione as electron acceptors. This reduction potential is lower than that of O2 demonstrating that thermodynamically the semiquinone radical can reduce O2 to superoxide, particularly since the concentration of O2 is expected to exceed that of the drug in cells and tissues.

’ INTRODUCTION Geldanamycin (GM, scheme 1), a benzoquinone ansamycin antibiotic, is a natural product inhibitor of Hsp90 with potent and broad anticancer properties. Its mechanism of action was unknown for years until GM was characterized as a direct inhibitor of the ATPase function of Hsp90.1,2 While GM showed promise in preclinical studies, its progression to clinical trials was halted due to unacceptable levels of hepatotoxicity.3 Consequently, numerous structural analogs, which differ only in their 17substituent, have been synthesized including the water-soluble and less toxic 17-(dimethylaminoethylamino)-17-demethoxygeldanamycin (17-DMAG, Alvespimycin, Scheme 1), which has recently entered clinical trials.4
6 Cytotoxicity associated with exposure to quinones has generally been attributed to either redox cycling and/or to their interaction with cellular nucleophiles such as thiols.7,8 The present study is concerned only with the redox cycling of the quinone (Q), which can be reduced by cellular enzymes such as NADPH-cytochrome P450 reductase to the respective semiquinone radical (Q•
) giving rise to O2•
formation in the presence of oxygen (reaction 1) Q •
þ O2 h O•
2 þ Q

ð1Þ

Equilibrium 1 is established rapidly, and its position is governed by the reduction potentials of the Q/Q•
and O2/O2•
redox couples and the relative concentrations of Q and O2. The toxicity of GM and its analogs as well as their antitumor effects have been attributed to the formation of O2•
,9
11 which readily reacts with NO to yield peroxynitrite.12 Consequently, physiological response to endogenously produced NO can be r 2011 American Chemical Society

lost in addition to further generation of reactive oxygen and nitrogen species via the decomposition of peroxynitrite,13 which may stimulate cellular oxidative damage.14,15 However, the basic information on the chemical properties of Q•
derived from these drugs and their reduction potentials in aqueous solutions is not available. The short lifetime of Q•
in water precludes, in most cases, its study by conventional methods. Reported below is a pulse radiolysis study, which is an ideal technique for generating and studying short-lived radicals, aimed at obtaining the reduction potential of the water-soluble 17-DMAG.

’ MATERIALS AND METHODS Chemicals. Water for solutions preparation was purified using a Milli-Q system. All chemicals were of the highest available grade and were used as received: 17-DMAG was purchased from LC Laboratories (Woburn, MA), 1,4-naphthoquinone (NQ), 1,4naphthoquinone-2-sulfonate (NQS), and 2-methyl-1,4-naphthoquinone (Menadione, MN) were purchased from Sigma-Aldrich. Stock solution of 2 mM 17-DMAG and NQS were prepared in water and those of NQ and MN in 2-propanol. N2O gas was passed through an oxygen trap (OXY-TRAP, Alltech Associate Inc.). Mass flow controllers (Tylan, Torrance, CA) were used for preparing gas mixtures of N2O and O2. The concentration of oxygen-saturated solutions was taken as 1.2 mM at 22 °C and 690 mmHg. Radiolysis. Pulse radiolysis experiments were carried out using a 5-MeV Varian 7715 linear accelerator (0.1
0.3 μs electron Received: June 1, 2011 Revised: June 30, 2011 Published: July 01, 2011 8928

dx.doi.org/10.1021/jp205161v | J. Phys. Chem. A 2011, 115, 8928–8932

The Journal of Physical Chemistry A

ARTICLE

Scheme 1. Structures of GM (R = CH3O) and 17-DMAG (R = H2NCH2CH2NH)

Figure 1. The UV
visible spectrum of 17-DMAG at pH 1.5
7.0 in aqueous solutions.

pulses, 200 mA current). A 200-W Xe lamp produced the analyzing light. A cutoff filter at 340 nm was used to minimize photochemistry. Measurements were done in 2 cm spectrosil cells with three light passes. The dose was determined using the thiocyanate dosimeter (10 mM KSCN in N2O-saturated water).16 All experiments were done at room temperature.

’ RESULTS Optical Absorption Spectra. The UV
visible spectrum of 17-DMAG is shown in Figure 1, which is independent of the pH 1.5
7.0. The concentration of 17-DMAG was determined before each experiment using ε331 = 18,560 M
1 cm
1 or ε527 = 890 M
1 cm
1 depending on the reaction mixture. On the delivery of an electron pulse to N2O-saturated solutions ([N2O] = 24 mM) containing 17-DMAG, 4 mM phosphate buffer and either 50 mM NaHCO2 or 1 M 2-propanol, the respective Q•
(DMAG•
), was formed by the sequence of reactions 2
7. γ

• • þ H 2 O ^ f e
aq ð2:6Þ, OHð2:7Þ, H ð0:6Þ, H3 O ð2:6Þ, H2 O2 ð0:72Þ

ð2Þ In eq 2, the values in parentheses are G-values represent the yields of the species (in 10
7 M Gy
1), which are lower by about 7% than those formed in the presence of high solute concentrations. •
e
aq þ N2 O þ H2 O f OH þ N2 þ OH

ð3Þ

•

ð4Þ

OH=H• þ HCO2
f CO2 •
þ H2 O=H2

CO2 •
þ 17-DMAG f DMAG•
þ CO2

ð5Þ

or •

OH=H• þ ðCH3 Þ2 CHOH f ðCH3 Þ2 C• OH þ H2 O=H2

ð6Þ

ðCH3 Þ2C• OH þ 17
DMAG f DMAG•
þ ðCH3 Þ2 CO þ Hþ

ð7Þ

DMAG•
þ Hþ h DMAGH•

ð8Þ •


The formation and decay of DMAG could not be studied at λ < 380 nm under our experimental conditions due to the high absorption of 17-DMAG at this region (Figure 1). The

Figure 2. The absorption spectra of DMAG•
at pH 7.0 (b) and its protonated form at pH 1.5 (O). Solutions were saturated with N2O and contained 62 μM 17-DMAG, 50 mM NaHCO2, and 4 mM phosphate buffer (pH 7.0) or 32 μM 17-DMAG and 1 M 2-propanol at pH 1.5. The spectra were corrected for the depletion of 17-DMAG. Inset: The pH dependence of the initial absorption at 450 nm demonstrating pKa = 3.2 ( 0.1. Solutions were saturated with N2O and contained 35 μM 17DMAG, 1 M 2-propanol, and 4 mM phosphate buffer. The dose was 5.0 Gy, and the optical path 6.2 cm.

absorption spectra of DMAG•
at pH 7.0 (ε450 = 7,200 M
1 cm
1) and DMAGH• at pH 1.5 (ε400 = 4,300 M
1 cm
1) are shown in Figure 2. The absorption at 450 nm decreases upon decreasing the pH demonstrating a pKa = 3.2 ( 0.1 (Figure 2, inset). Kinetics. The formation of DMAG•
via the reduction of 17-DMAG by CO2•
or (CH3)2C•OH was followed at 450 nm. The rates obeyed first-order kinetics and were linearly dependent on [17-DMAG] (Figure 3). The rate constants were determined from the slope of the lines such as in Figure 3 to be k5 = (4.3 ( 0.1)  109 M
1 s
1 and k7 = (1.9 ( 0.1)  109 M
1 s
1 independent of pH 3.0 - 8.0. At pH < 3, kobs is affected by the reaction of 17-DMAG with H 3 formed via H+
reaction with eaq . Electron Transfer from DMAG•
to O2. On saturating solutions of 50
93 μM 17-DMAG, 50 mM NaHCO2, and 4 mM phosphate buffer (pH 7.0) with different mixtures of N2O/O2 (1.8
20% O2), the absorption of DMAG•
at 450 nm decayed in a first-order reaction, which was accelerated upon 8929

dx.doi.org/10.1021/jp205161v |J. Phys. Chem. A 2011, 115, 8928–8932

The Journal of Physical Chemistry A

ARTICLE

Figure 3. Dependence of the observed first order rate constant of the formation of the absorption at 450 nm as a function of [17-DMAG]. Solutions were saturated with N2O and contained 4 mM phosphate buffer (pH 7.0) and 50 mM NaHCO2 (b) or 1 M 2-propanol (9). The dose was 2.95 Gy.

Figure 5. Dependence of the rate of approach to equilibrium on the concentrations of 17-DMAG and O2. Solutions were saturated with mixtures of N2O/O2 and contained 50
93 μM, 50 mM NaHCO2, and 4 mM phosphate buffer (pH 7.0). The dose was 2.95 or 5.0 Gy.

E7(DMAG/DMAG 3
) was calculated using eq 11 and E(O2/O2•
) =
180 mV (1 M O2, NHE)17,18 to be
191 ( 2 mV. •
EðO2 =O•
2 Þ
E7 ðDMAG=DMAG Þ ¼ 0:06 log K1

ð11Þ

Figure 4. Kinetic traces monitored at 450 nm obtained upon pulse irradiation of N2O-saturated solutions with and without 5% O2. The solutions contained 63 μM 17-DMAG, 50 mM NaHCO2, and 4 mM phosphate buffer (pH 7.0). The dose was 2.95 Gy, and the optical path 6.2 cm.

increasing [O2] or [17-DMAG]. Figure 4 shows typical kinetic traces with and without O2 at constant [17-DMAG]. The competition of O2 with 17-DMAG for CO2•
is reflected in the decrease of the initial absorption of DMAG•
as compared with that observed in the absence of O2 (e.g., Figure 4). Since the reduction of O2 and 17-DMAG by CO2•
is completed within less than 10 μs before any appreciable decay of DMAG•
takes place, the observed decay in the 200
400 μs range is attributed to the relaxation of reaction 1 toward equilibrium (eq 9). kobs ¼ k1 ½O2  þ k
1 ½17-DMAG

ð9Þ

A better estimate of the equilibrium constant from kinetics and the individual rate constants is obtained by rearranging eq 9 (eq 10). kobs =½O2  ¼ k1 þ k
1 ½17-DMAG=½O2 

ð10Þ

Figure 5 shows a plot of kobs/[O2] vs [17-DMAG]/[O2] demonstrating a straight line with an intercept = k1 = (1.2 ( 0.1)  108 M
1s
1, a slope = k
1 = (7.8 ( 0.2)  107 M
1 s
1, and intercept/slope = K1 = 1.54 ( 0.17.

Note that E(O2/O2•
) was initially determined to be
160 mV (1.0 M O2).19,20 However, a careful re-examination of this couple indicates that a better value is
180 mV.17,18 Electron Transfer from DMAG•
to other quinones. To verify the reduction potential determined using O2 as an electron acceptor, the latter was replaced with various quinones (Qref): 1, 4-naphthoquinone-2-sulfonate (E7(NQS/NQS•
) =
80 mV), 1,4-naphthoquinone (E7(NQ/NQ•
) =
165 mV), and 2-methyl1,4-naphthoquinone (E7(MN/MN•
) =
228 mV). The reduction potentials in parentheses were corrected to E (O2/O2•
) =
180 mV since the original ones were determined using O2 as a reference with E (O2/O2•
) =
155 or 160 mV.21 DMAG•
þ Q ref h Q ref •
þ 17-DMAG

ð12Þ

Equation 12 was studied in N2O-saturated solutions containing 4 mM phosphate buffer (pH 7.0) and 50 mM NaHCO2 in the case of NQS or 1 M 2-propanol in the case of NQ and MN due to their poor solubility in water. The absorption changes were followed at 390 nm where Qref•
absorbs (ε390 ≈ 12,000 M
1 cm
1) or at 450 where the absorption of DMAG•
is higher than that of Qref•
.19,22,23 In the presence of 10
40 μM NQS, the initial absorption of NQS•
at 390 nm was unaffected by the addition of 30 μM 17DMAG though most of CO2•
radicals reduce 17-DMAG (k5 = (4.3 ( 0.1)  109 M
1 s
1 vs (1.5 ( 0.1)  109 M
1 s
1, which has been determined in the present study). In addition, kobs of the formation of the absorption at 390 nm increased linearly with [NQS] (k12 = (9.7 ( 0.3)  108 M
1 s
1, results not shown) demonstrating that DMAG•
effectively reduces NQS. Hence, k12 . k
11 and E7(DMAG/DMAG•
) , E7(NQS/NQS•
) =
80 mV. In the case of MN and NQ the initial absorption at 390 nm decreased upon increasing [17-DMAG], and the fast observed first order decay, which increased upon increasing [17-DMAG], is attributed to relaxation of reaction 12 toward equilibrium. Figure 6 shows plots of kobs/[Qref] vs [17-DMAG]/[Qref] 8930

dx.doi.org/10.1021/jp205161v |J. Phys. Chem. A 2011, 115, 8928–8932

The Journal of Physical Chemistry A

ARTICLE

17-DMAG as is the case with OH-linkage, e.g., 2-hydroxy-1,4naphthoquinone (E7(Q/Q•
) =
118 mV, pKa = 3.65)27 and 5,8dihydroxy-1,4-naphthoquinone (E7(Q/Q•
) =
110 mV, pKa = 2.7).28 1,4-Benzoquinone (BQ) is a typical high electron affinity quinone, E7(BQ/BQ•
) = +0.099 V.29 Substitution into the ring by electron-donating or -withdrawing groups decreases or increases, respectively, the one-electron reduction potential of the quinone in a predictable manner. In the case of BQ, an approximate predictor is given by eq 13, where σp is the Hammett sigma value of the substituents.30 E7 ðDMAG=DMAG•
Þ ¼ E7 ðBQ =BQ •
Þ þ 0:62 Σσp

Figure 6. Dependence of the rate of approach to equilibrium on the concentrations of 17-DMAG and MN followed at 390 nm (9) or NQ followed at 450 nm (b). Solutions were saturated with N2O and contained 1 M 2-propanol and 4 mM phosphate buffer (pH 7.0). The dose was 2.95 or 5.0 Gy.

yielding in the case of MN k12 = (6.7 ( 2.6)  107 M
1 s
1, k
12 = (2.1 ( 0.2)  108 M
1 s
1 and K11 = 0.32 ( 0.17. In the case of NQ one obtains k12 = (4.5 ( 0.6)  108 M
1 s
1, k
12 = (1.5 ( 0.2)  108 M
1 s
1, K12 = 3.0 ( 0.9. Hence, E7(DMAG/ DMAG•
) =
198 ( 11 and
193 ( 7 mV, respectively, in excellent agreement with the value determined using O2 as a reference.

’ DISCUSSION The reduction potential of 17-DMAG has been determined to be
194 ( 6 mV using O2, NQ, and MN as electron acceptors. The reduction potential of 17-DMAG is lower than that of O2 (
180 mV)17,18 demonstrating that thermodynamically the semiquinone radical can reduce O2 to O2•
. However, the position of eq 1 is governed not only by the reduction potentials of Q/Q•
and O2/O2•
but also by the relative concentrations of Q and O2. By assumption that under physiological conditions [O2] > [Q], eq 1 in the case of 17-DMAG is shifted toward O2•
production. The reduction potential of 17-DMAG can be used to estimate the rate of its reduction by cellular enzymes25,26 and the rate constant of its reaction with thiols.8 For quinones with reduction potential between ∼400 and
165 mV there is a good correlation between the rates of their reduction by purified cytochrome P-450 reductase and E7(Q/Q•
) from which a rate of ∼17 μmol/min per mg is calculated for 17-DMAG.26 The rate constant of the reaction of 17-DMAG with glutathione is estimated to be about 3.5 M
1 s
1 based on the linear correlation demonstrated between log(rate constant) and E7(Q/Q•
) at pH 6.8 At pH 7 the rate constant is about 10-times higher because the reaction proceeds mainly via the deprotonated form of the thiol.8 A correlation between the pKa of the semiquinone radical and the reduction potential of the quinone has been demonstrated for various quinones.24 The pKa of 3.2 for DMAG •
is much lower than expected based on its reduction potential, where quinones with similar negative reduction potentials have pKa values of about 5.24 The relatively low pKa = 3.2 of DMAG•
may be attributed to stabilization of the radical by an intramolecular H-bonding xof the protonated amine-linkage of

ð13Þ

GM and 17-DMAG are BQ substituted by three moieties, two are identical and the third one is methoxy (
OCH3) and dimethylaminoethylamino (
NH(CH2)2N(CH3)2), respectively. Hence, eq 14 is derived from eq 13. E7 ðGM=GM•
Þ ¼ E7 ðDMAG=DMAG•
Þ þ 0:62ðσðOCH3 Þ
σðNHðCH2 Þ2 NðCH3 Þ2 Þ

ð14Þ

The Hammett σp value for
OCH3 is
0.27, and that for
NH(CH2)2N(CH3)2 has not been determined.31 Furthermore,
NH(CH2)2N(CH3)2 (electron-donating) is likely to be protonated at pH 7 (electron-withdrawing), and its σp value should be higher than σ(NHCH2CH3) =
0.61,31 e.g., compare
NH2 (σ =
0.66) with
NH3+ (σ = 0.60).31 An informed guess for σ(NH2+(CH2)2NH+(CH3)2) is obtained using eq 12, E7(DMAG/DMAG•
) =
0.194 V and assuming that the other two substituents are
NHCOCH(CH3)2 (σp =
0.10)31 and
CH2CH2CH2
(σp =
0.26).31 Hence, σ(NH2+(CH2)2NH+(CH3)2) =
0.11, and one obtains E7(GM/GM•
) =
0.293 V. These calculations suggest that E7(GM/GM•
) < E7(DMAG/DMAG•
), the same order observed for E1/2 in DMSO.11 Accordingly, 17-DMAG is more readily reduced by cellular enzymes as has been demonstrated in the case of purified human cytochrome P450 reductase but not in the case of both mouse and human hepatic microsomes.32 In addition, O2 is more readily reduced to O2•
by GM•
compared to DMAG•
as in the case of GM eq 1 is shifted more to the right, i.e., higher K1. Any reagent that removes O2•
will pull the equilibrium in this direction. Potential reagents in cells and tissues are the enzyme superoxide dismutase (SOD) and NO, and the rate of reaction 1 may become the key factor in determining the flux of O2 3
in cells and tissues, which increases with the decrease in E7(Q/Q•
).

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected], Phone 972-2-6586478, Fax: 972-26586925.

’ ACKNOWLEDGMENT This work has been supported by the Israel Science Foundation (Grant No. 1477). ’ REFERENCES (1) Whitesell, L.; Shifrin, S. D.; Schwab, G.; Neckers, L. M. Cancer Res. 1992, 52, 1721–1728. 8931

dx.doi.org/10.1021/jp205161v |J. Phys. Chem. A 2011, 115, 8928–8932

The Journal of Physical Chemistry A

ARTICLE

(2) Whitesell, L.; Mimnaugh, E. G.; De Costa, B.; Myers, C. E.; Neckers, L. M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 8324–8328. (3) Supko, J. G.; Hickman, R. L.; Grever, M. R.; Malspeis, L. Cancer Chemother. Pharmacol. 1995, 36, 305–315. (4) Pacey, S. C.; Wilson, R. H.; Walton, M.; Martin, E.; MorenoFarre, J.; Arkenau, H. T.; Steinfeldt, H.; Raynaud, F.; Workman, P.; Judson, I. Mol. Cancer Ther. 2007, 6, 3332S–3332S. (5) Gutierrez, M. E.; Kummar, S.; Gardner, E. R.; Figg, W.; Chen, X.; Zajac-Kaye, M.; Yancey, M. A.; Ivy, P.; Conley, B.; Melillo, G.; Giaccone, G.; Doroshow, J. H.; Murgo, A. J. Ann. Oncol. 2008, 19, 24–24. (6) Lancet, J. E.; Gojo, I.; Burton, M.; Quinn, M.; Tighe, S. M.; Kersey, K.; Zhong, Z.; Albitar, M. X.; Bhalla, K.; Hannah, A. L.; Baer, M. R. Leukemia 2010, 24, 699–705. (7) Monks, T. J.; Lau, S. S. Crit. Rev. Toxicol. 1992, 22, 243–270. (8) Song, Y.; Buettner, G. R. Free Radic. Biol. Med. 2010, 49, 919–962. (9) Dikalov, S.; Landmesser, U.; Harrison, D. G. J. Biol. Chem. 2002, 277, 25480–25485. (10) Fukuyo, Y.; Inoue, M.; Nakajima, T.; Higashikubo, R.; Horikoshi, N. T.; Hunt, C.; Usheva, A.; Freeman, M. L.; Horikoshi, N. Cancer Res. 2008, 68, 6324–6330. (11) Samuni, Y.; Samuni, U.; Ishii, H.; Krishna, M. C.; Goldstein, S.; Mitchell, J. B. Free Radic. Biol. Med. 2010, 48, 1559–1563. (12) Goldstein, S.; Czapski, G. Free Radic. Biol. Med. 1995, 19, 505–510. (13) Goldstein, S.; Lind, J.; Merenyi, G. Chem. Rev. 2005, 105, 2457–2470. (14) Radi, R. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 4003–4008. (15) Dedon, P. C.; Tannenbaum, S. R. Arch. Biochem. Biophys. 2004, 423, 12–22. (16) Buxton, G. V.; Stuart, C. R. J. Chem. Soc. Faraday Trans. I 1995, 91, 279–281. (17) Wardman, P. Free Radic. Res. Commun. 1991, 14, 57–67. (18) Koppenol, W. H.; Stanbury, D. M.; Bounds, P. L. Free Radic. Biol. Med. 2010, 49, 317–322. (19) Meisel, D.; Czapski, G. J. Phys. Chem. 1975, 79, 1503–1509. (20) Ilan, Y. A.; Meisel, D.; Czapski, G. Isr. J. Chem. 1974, 12, 891–895. (21) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637–1755. (22) Patel, K. B.; Willson, R. L. J. Chem. Soc. Faraday Trans. I 1973, 69, 814–825. (23) Meisel, D.; Fessenden, R. W. J. Am. Chem. Soc. 1976, 98, 7505–7510. (24) Rao, P. S.; Hayon, E. J. Phys. Chem. 1973, 77, 2274–2276. (25) Powis, G.; Appel, P. L. Biochem. Pharmacol. 1980, 29, 2567–2572. (26) Butler, J.; Hoey, B. M. Biochim. Biophys. Acta 1993, 1161, 73–78. (27) Mukherjee, T. Radiat. Phys. Chem. 1987, 29, 455–462. (28) Land, E. J.; Mukherjee, T.; Swallow, J. J. Chem. Soc. Faraday Trans. 1 1983, 79, 405–415. (29) Ilan, Y. A.; Czapski, G.; Meisel, D. Biochim. Biophys. Acta 1976, 430, 209–224. (30) Wardman, P. Prediction and measurement of redox properties of drugs and biomolecules. In Selective Activation of Drugs by Redox Processes; Adams, G. E., Breccia, A., Fielden, E. M., Wardman, P., Eds.; Plenum Press: New York, 1990; pp 11
24. (31) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165–195. (32) Guo, W.; Reigan, P.; Siegel, D.; Ross, D. Drug Metab. Dispos. 2008, 36, 2050–2057.

8932

dx.doi.org/10.1021/jp205161v |J. Phys. Chem. A 2011, 115, 8928–8932