Báo cáo Y học: Thermodynamics and kinetics of the cleavage of DNA catalyzed by bleomycin A5 A microcalorimetric study

doi:10.1046/j.1432-1033.2002.02948.x

Eur. J. Biochem. 269, 2851–2859 (2002) (cid:3) FEBS 2002

Thermodynamics and kinetics of the cleavage of DNA catalyzed by bleomycin A5 A microcalorimetric study

Yi Liang1, Fen Du1, Bing-Rui Zhou1, Hui Zhou1, Guo-Lin Zou1, Cun-Xin Wang2 and Song-Sheng Qu2 1College of Life Sciences and 2College of Chemistry and Molecular Science, Wuhan University, China

the cleavage of calf thymus DNA induced by BLM-A5 with those for the scission of calf thymus DNA mediated by adriamycin and by (1,10-phenanthroline)-copper, it was found that BLM-A5 possessed the highest DNA cleavage efficiency among these DNA-damaging agents. These results suggest that BLM-A5 is not as efficient as a DNA-cleaving enzyme although the cleavage of DNA by BLM-A5 follows Michaelis–Menten kinetics. Binding of BLM-A5 to calf thymus DNA is driven by a favorable entropy increase with a less favorable enthalpy decrease, in line with a partial intercalation mode involved in BLM-catalyzed breakage of DNA.

Keywords: bleomycin; DNA cleavage; kinetics; microcalor- imetry; thermodynamics.

Microcalorimetry and UV-vis spectroscopy were used to conduct thermodynamic and kinetic investigations of the scission of calf thymus DNA catalyzed by bleomycin A5 (BLM-A5) in the presence of ferrous ion and oxygen. The molar reaction enthalpy for the cleavage, the Michaelis– Menten constant for calf thymus DNA and the turnover number of BLM-A5 were calculated by a novel thermoki- netic method for an enzyme-catalyzed reaction to be )577 ± 19 kJÆmol)1, 20.4 ± 3.8 lM and 2.28 ± 0.49 · 10)2 s)1, respectively, at 37.0 (cid:2)C. This DNA cleavage was a largely exothermic reaction. The catalytic efficiency of BLM- A5 is of the same order of magnitude as that of lysozyme but several orders of magnitude lower than those of TaqI restriction endonuclease, NaeI endonuclease and BamHI endonuclease. By comparing the molar enthalpy change for

cleavage by BLM in the past two decades [2–6,11–20], thermodynamic information for the scission, which is necessary for a thorough understanding of the mechanism, is eagerly awaited. The purpose of this investigation is to provide detailed thermodynamic and kinetic data for BLM- mediated DNA degradation to furnish insights into the anticancer mechanism of BLM.

Microcalorimetry is an important tool for the study of both thermodynamic and kinetic properties of biological macromolecules by virtue of its general applicability, high accuracy and precision [21–24]. Recently, this method has yielded a large amount of data on the binding reactions of DNA with DNA-targeting molecules, such as adriamycin (ADM) [25], daunomycin [25,26], Hoechst 33258 [27], ethidium bromide [28], 2,7-diazapyrene [29] and dodecyl trimethylammonium bromide [30]. Only a limited number of authors have, however, paid attention to the energetics of drug-induced cleavage of DNA [31].

The bleomycins (BLMs, Fig. 1) are a family of naturally occurring, structurally related, glycopeptide-derived antitu- mor antibiotics discovered by Umezawa and coworkers from cultures of Streptomyces verticillus in 1966 [1], which have more than 200 members, such as A2, A5 and B2 [2]. BLMs consist of an unusual linear hexapeptide, a disac- charide and a terminal amine (the R group in Fig. 1). Mixtures of BLMs are presently used for the clinical treatment of a variety of cancers, notably squamous cell carcinomas, testicular tumors and nonHodgkin’s lym- phoma [2]. The therapeutic effect of BLM is believed to result from its ability to induce single- and double-strand breakage of DNA molecules by oxidation of the deoxyri- bose moiety in the presence of oxygen and a redox-active metal ion, e.g. Fe and Co [2–6]. On the other hand, RNA is also considered as a therapeutically relevant target for BLM [7,8]. It has been found that BLM-induced autoxidation of ferrous iron follows the Michaelis–Menten kinetics [9,10]. Although a significant number of experimental approaches have been used to elucidate the mechanism of DNA

In a previous publication from this laboratory [31], microcalorimetry and agarose gel electrophoresis were applied to check the oxidative degradation of DNA induced by (1,10-phenanthroline)-copper, a well-known DNA-dam- aging agent [32]. In the present paper, microcalorimetry and UV-vis spectroscopy were combined to study the scission of calf thymus DNA by a mixture of bleomycin A5 (BLM-A5), ferrous iron and oxygen. A novel thermokinetic method for an enzyme-catalyzed reaction was proposed and employed to produce not only the thermodynamic constant (DrHm) but also the kinetic properties (Km and k2) of the cleavage of DNA catalyzed by BLM-A5 with the result that BLM-A5 is not as efficient as a DNA-cleaving enzyme. In order to gain insights into the nucleotide binding interactions of BLM, we

Correspondence to Y. Liang, College of Life Sciences, Wuhan University, Wuhan, 430072, China. Fax: + 86 27 8788 2661, Tel.: + 86 27 8721 4902, E-mail: liangyi@whu.edu.cn Abbreviations: ADM, adriamycin; BLM, bleomycin; BLM-A2, bleomycin A2; BLM-A5, bleomycin A5; BLM-B2, bleomycin B2; BR, batch reactor; ME, 2-mercaptoethanol; OP, 1,10-phenanthroline; Vc, vitamin C; UV-vis, ultraviolet and visible. (Received 17 December 2001, revised 12 April 2002, accepted 22 April 2002)

2852 Y. Liang et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

microbatch reactor with a heat-conduction isothermal calorimeter [31,33–35]. For the experiments on DNA cleavage, compartment I of the reaction cell contained 2 mL of a FeCl2 solution and compartment II of the reaction cell contained 4 mL of a DNA/BLM-O2 mixture. This multicomponent system was prepared by mixing DNA and BLM-A5 solutions and saturated with purified oxygen before calorimetric experiments. To avoid the re-oxidation of FeCl2 solution on exposure to air, purified N2 was passed into one compartment of the cell while sample was added to the other. As soon as samples were added, the source of N2 was removed and the plug for the reaction cell was closed tightly. The same procedure was used for adding samples to the reference cell. To avoid the influence of the heat effects of diluting and mixing, etc. on the results, the contents and quantities in both cells were as close as possible except that DNA was not added to the reference cell. For the experiments on DNA binding, compartment I of the reaction cell contained 2 mL of a DNA solution and compartment II of the reaction cell contained 4 mL of a BLM-A5 solution. The heat released by dilution of DNA is negligible.

m and DbS0

m ,DbG0

UV-vis spectroscopy

have elucidated the binding constant (KB) and the standard thermodynamic parameters (DbH 0 m ) for the binding of BLM-A5 to calf thymus DNA using microcalorimetry. The results help understand the binding mode of BLM-A5 to DNA.

Fig. 1. Structure of BLM-A2, A5 and B2.

M A T E R I A L S A N D M E T H O D S

Materials

UV and visible spectra were measured using a Shimadzu UV-2401PC spectrophotometer. A reaction system contain- ing 21.5 lM BLM-A5, 20 lM ferrous iron and 15.2 lM calf thymus DNA was saturated with purified oxygen and incubated in 50 mM Tris/HCl buffer at pH 7.4 and 20 (cid:2)C for 30 min and then scanned from 250 to 500 nm. Five control systems were chosen to investigate the effect of DNA cleavage by BLM-A5 on the spectrum of BLM-A5. The first one was 21.5 lM BLM-A5, the second was a mixture containing 21.5 lM BLM-A5 and 20.0 lM ferric iron, and the third was a mixture containing 21.5 lM BLM- A5 and 20.0 lM ferrous iron saturated with purified nitrogen. The fourth was a solution containing 21.5 lM BLM-A5, 20 lM ferric iron and 15.2 lM calf thymus DNA and the fifth was 15.2 lM calf thymus DNA. These solutions were also incubated in 50 mM Tris/HCl buffer at pH 7.4 and 20 (cid:2)C for 30 min and then scanned from 250 to 500 nm.

R E S U L T S

Novel thermokinetic models for enzyme-catalyzed reactions

For a simple single-substrate, single-intermediate, enzyme- catalyzed reaction occurring in a batch reactor (BR) with negligible mass-transfer limitations and without self-inacti- vation of the enzyme, from the Michaelis–Menten kinetics, it follows that

Calf thymus DNA (Sigma Chemical Co., MI, USA) was purified by ethanol precipitation and centrifugal dialysis and sheared by sonication at ice bath temperatures for 30 min. The absorbances at 260 and 280 nm for purified DNA were measured at room temperature. DNA concentrations were determined spectroscopically at 260 nm using a molar )1Æcm)1 and expressed as extinction coefficient of 13 200 M base pair concentrations throughout this paper. The con- centration of BLM-A5 (Hebei Pharmaceutical Factory, Tianjin, China) was determined at 291 nm using a molar )1Æcm)1 and the concen- extinction coefficient of 15 500 M tration of ADM (Haimen Pharmaceutical Factory, Zhejiang, China) was determined at 480 nm using a molar )1Æcm)1. FeCl2Æ4H2O extinction coefficient of 11 500 M (analytical grade) was purchased from Merck’s reagent Co., Germany. Other chemicals used were made in China and of analytical grade. All reagent solutions were prepared in 10 mM Tris/HCl buffer (pH ¼ 7.4). As the FeCl2 solution is easily oxidized by oxygen, it was placed in a brown bottle and then flushed with purified nitrogen for 10 min, sealed and stored in a refrigerator until use. Moreover, it was freshly prepared on each occasion.

ð1Þ

(cid:2)

(cid:2)

lnð1 (cid:2) xÞ ¼

1 t

(cid:1) (cid:2) x t

k2½E(cid:6)0 Km

½S(cid:6)0 Km

Isothermal microcalorimetry

The cleavage of calf thymus DNA by a mixture of BLM-A5, ferrous ion and oxygen and the binding of BLM-A5 to calf thymus DNA, were studied in 10 mM Tris/HCl buffer at pH 7.4 and 37.0 (cid:2)C. The heat effects of the reactions mentioned above were determined using a LKB-2107 batch microcalorimeter (Stockholm, Sweden), which consists of a

where t is the reaction time, x the fraction of substrate converted into product at time t, which is nondimensional, Km the Michaelis constant, [S]0 and [E]0 the initial concen- trations of substrate and enzyme, respectively, and k2, also known as the turnover number of the enzyme [36], the rate constant of breakdown of the enzyme–substrate complex to product.

(cid:3) FEBS 2002

Thermokinetics of DNA cleavage catalyzed by BLM-A5 (Eur. J. Biochem. 269) 2853

Combining Eqns (6), (7) and (8), we get

D ¼

ð9Þ

Dme1(cid:2)t=tm

t tm

ð10Þ

k ¼ eDm=ðbAÞ

Substituting Eqn (10) in Eqn (5), we obtain

ð11Þ

x ¼

þ

a A

bD eDm

Under certain conditions, the rate of self-inactivation of an enzyme may be sufficiently great that it must be taken into account in the study of the kinetics of the reaction undergoing catalysis [37]. The self-inactivation reactions are sometimes, although not always, of first order kinetics. Even when first order inactivation is taken into account, other kinetics schemes, such as the second- or zero-order self- inactivation, can be accounted for, according to Laidler & Bunting [37]. In the present paper, attention will be confined to first-order self-inactivation, but the methods are readily extended to other cases.

When a single-substrate enzyme-catalyzed reaction is taking place in a conduction calorimeter, the molar reaction enthalpy is:

For a single-substrate enzyme-catalyzed reaction occur- ring in a BR with the first-order self-inactivation of the enzyme, the general rate equation is

ð12Þ

DrHm ¼ Q1;1=ðVT (cid:9) ½S(cid:6)0Þ

(cid:2)

¼

ð2Þ

d½S(cid:6) dt

k2½E(cid:6)½S(cid:6) Km þ ½S(cid:6)

where [S] and [E] are the concentration of substrate and the total concentration of active enzyme at time t, respectively. The decay law for the first-order self-inactivation is

Here, Q1,1 is the total heat effect of the reaction, which can be calculated by the integration type of Tian’s equation from the experimental calorimetric curves. VT is the total volume of the reacting system, 6 mL in the present case.

ð3Þ

½E(cid:6) ¼ ½E(cid:6)0e(cid:2)k1t where k1 is the first-order rate constant for self-inactivation of the enzyme.

Substituting Eqn (3) in Eqn (2) and performing the integration between limits [S]0 to [S] and 0 to t, we obtain

x (cid:2)

lnð1 (cid:2) xÞ ¼

ð1 (cid:2) e(cid:2)k1tÞ

ð4Þ

Km ½S(cid:6)0

k2½E(cid:6)0 k1½S(cid:6)0

Eqns (1), (11) and (12) are called the analog calorimetric curve model of a single-substrate enzyme-catalyzed reaction without taking self-inactivation of the enzyme into account. It is a novel application of the thermo-analytical analog curve method and suitable to both fast and slow enzyme- catalyzed reactions. A plot of –ln(1 ) x)/t against x/t is linear with an axis intercept of k2[E]0/Km and a slope of –[S]0/Km. The values of Km and k2 can be calculated from the slope and intercept, respectively, using the calorimetric data from only a single experiment.

If the heat-transfer process in a BR obeys Tian’s equation [21,33–35], the substrate conversion in a BR may be written as

x ¼ ðD þ kaÞ=kA

ð5Þ

Eqns (4) and (11) are called the analog calorimetric curve model of a single-substrate enzyme-catalyzed reaction with the first-order self-inactivation of the enzyme. The values of Km, k2 and k1 were obtained from the equations by substituting in at least three sets of experimental data (x and t) and using the MATHSOFT MATHCAD software (version 2001). The value of s, the lifetime of self-inactivation, was calculated using the rate constant k1.

A thermodynamic model for the binding of small molecules to DNA

Here, D is the calorimetric height at time t (i.e. the thermopile potential difference at times t and 0), a is the area under the calorimetric curve and the time-axis over the interval (t – 0), A is the total area under the calorimetric curve and k is the Newton cooling constant of the calorimeter system which can be easily determined by electric calibration [34].

According to the thermo-analytical analog curve method [38], the calorimetric curve for a reaction occurring in a conduction calorimeter can be approximately simulated by the following relationship:

Understanding the thermodynamics of the binding of small molecules to DNA is of practical interest, because many small molecules that bind to DNA are effective pharmaceutical agents, especially in cancer chemotherapy [25].

D ¼ ate(cid:2)kbt

ð6Þ

From these experiments, it is found that the interactions of DNA with many small molecules, such as BLM and ADM, are at rapid equilibrium:

At t ¼ tm, dD/dt ¼ 0 and D ¼ Dm, substituting in Eqn (3), we get:

DNA þ L Ð DNA (cid:9) L

ð13Þ

ð7Þ

a ¼ eDm=tm

ð8Þ

b ¼ 1=ktm

where L is a small molecule that binds to DNA and DNAÆL the complex between DNA and L. The intrinsic binding constant, KB, is defined by the equation [24,28,29]:

ð14Þ

KB ¼

y ð1 (cid:2) yÞð½DNA(cid:6)0 (cid:2) ny½L(cid:6)0Þ

where a and b are the analog parameters related to the thermokinetic system, Dm and tm are the calorimetric curve characteristic data representing the maximum calorimetric height and time corresponding to Dm, respectively. For a fast reaction, the value of b turns out to be 1. For a slow reaction, however, the value of b is 2/3 [38].

Here, [DNA]0 and [L]0 are the initial concentrations of DNA and L, respectively, n is the exclusion parameter which presents the number of base pairs covered by each L.

2854 Y. Liang et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

The degree of L binding to DNA, y, can be determined by the formula:

ð15Þ

y ¼ DbHm;a=DbH0 m

where DbH 0 m is the standard binding enthalpy per mole of L and DbHm,a is the apparent molar binding enthalpy which can be calculated using the equation:

ð16Þ

DbHm;a ¼ Q2;1=ð½L(cid:6)0 (cid:9) VTÞ

Here Q2,1 is the total heat effect of L binding to DNA, which can be calculated by the integration type of Tian’s equation from the experimental calorimetric curves. The molar ratio, r, of DNA to L is defined as

ð17Þ

r ¼ nDNA;0=nL;0 ¼ ½DNA(cid:6)0=½L(cid:6)0

where nDNA,0 and nL,0 are the initial amounts of DNA and L, respectively. Substituting Eqns (15) and (17) into Eqn (14), we get

m (cid:2) DbHm;aÞ

½DNA(cid:6)0KBnDbHm;aðDbH0

r ¼

½DNA(cid:6)0KBðDbH0

mÞ2 (cid:2) ð½DNA(cid:6)0KB þ 1ÞDbHm;aDbH0 m ð18Þ

solutions taken from both the reaction cell and the reference one were brownish yellow.

This thermodynamic model was used to perform a nonlinear least-squares analysis of the apparent molar binding enthalpy, DbHm,a, as an explicit function of the ratio r using the MICROCAL ORIGIN software molar (ver. 6.0) and the values for three unknown binding parameters, KB, DbH 0 m and n, were thus obtained. The v2 test was used to examine the appropriateness of the model statistically.

The standard molar binding free energy (DbG0

m ) and the standard molar binding entropy (DbS0 m ) for the binding reaction were calculated by the fundamental equations of thermodynamics [28]:

ð19Þ

DbG0

ð20Þ

DbS0

m ¼ (cid:2)RT (cid:9) ln KB m (cid:2) DbG0

m ¼ ðDbH 0

mÞ=T

Tables 1 and 2 summarize the molar reaction enthalpies and the kinetic parameters for the cleavage of calf thymus DNA by a mixture of BLM-A5, Fe2+ and O2 at different DNA concentrations and at 37.0 (cid:2)C obtained from the analog calorimetric curve models of a single-substrate enzyme-catalyzed reaction without taking self-inactivation of BLM-A5 into account and with the first-order self- inactivation of BLM-A5. It should be pointed out that Fe2+ is used in about 30-fold molar excess relative to BLM despite the fact that only 2.28 turnovers (presumably corresponding to DNA cleavage events) per 100 second are observed (Table 1). From Fig. 2 and Table 1, it can also be seen that this DNA cleavage was a largely exothermic reaction and followed Michaelis–Menten kinetics. Thus, the observed rate law for the cleavage of DNA catalyzed by BLM-A5 at excessive ferrous ion and oxygen concentrations can be expressed as

ð21Þ

t0 ¼

k2½BLM(cid:6)0½DNA(cid:6)0 Km þ ½DNA(cid:6)0

Thermodynamics and kinetics of the cleavage of DNA catalyzed by BLM-A5

Here t0 is the initial rate for the DNA cleavage by BLM-A5. It should be pointed out that the fact that a reaction can be simulated using the Michaelis–Menten theory kinetics does not per se imply that a reaction is enzymatic.

UV-vis spectra of BLM-A5

value of

y-axis

Figure 3A compares the UV and visible spectrum from 250 to 500 nm of BLM-A5 after the cleavage of calf thymus DNA by a mixture of BLM-A5, Fe2+ and O2 with those of the five control systems mentioned in the (cid:2)Materials and methods(cid:3) and Fig. 3B shows those between 350 and 500 nm. It can be seen from Fig. 3A that the large underlying peak at 291 nm for BLM, which has been ascribed to the bithiazole p ) p* and n ) p* transitions [19], does not shift after this scission, provided that the absorb- ance for calf thymus DNA has been subtracted from the total absorbance for the reaction system after the cleavage.

From the spectroscopic results, the ratio of the absorbance at 260 nm to that at 280 nm for purified DNA used in the present study is about 2.07. As shown in Fig. 2, the calorimetric curve for the cleavage of calf thymus DNA by a mixture of BLM-A5, Fe2+ and O2 returned to the baseline within 10 min, under the experimental conditions used. The experimental calorimetric curve can be reasonably well fitted by the simulated analog calorimetric curve between 75 and 210 s at 37.0 (cid:2)C. The substrate conversion x at time t in one experiment on the DNA cleavage by BLM-A5 can be calculated using Eqn (11) from the calorimetric data shown in Fig. 2. A plot of –ln(1 ) x)/t against x/t in this range is linear with the intercept being 1.323 · 10)2 s)1, the value of slope being )0.7342 and the linear correlation coefficient being )0.9967. Then, the values of Km and k2 can be calculated from the slope and intercept to be 23.6 lM and 2.90 · 10)2 s)1, respectively. After the calorimetric experiment on DNA cleavage, the residual

Fig. 2. Experimental calorimetric curve (a) and the corresponding simulated analog calorimetric curve (b) of the scission of calf thymus DNA by a mixture of BLM-A5, Fe2+ and O2 at 37.0 (cid:1)C. For curve b, D ¼ 0.03653 te1–t/150 and b ¼ 1. The initial concentrations of calf thymus DNA, BLM-A5, Fe2+ and O2 are 17.3, 10.8, 340 and 650 lM, respectively.

(cid:3) FEBS 2002

Thermokinetics of DNA cleavage catalyzed by BLM-A5 (Eur. J. Biochem. 269) 2855

Table 1. Thermodynamic and kinetic data of the cleavage of calf thymus DNA by a mixture of BLM-A5, Fe2+ and O2 at 37.0 (cid:1)C. The thermodynamic and kinetic data were obtained from the analog calorimetric curve model of a single-substrate enzyme-catalyzed reaction without taking self-inactivation of BLM-A5 into account. The molar enthalpy change for the reaction of BLM-A5, Fe2+ and O2 has been determined by microcalorimetry to be )34.4 ± 3.2 kJÆmol)1. Data are expressed as mean ± SD (n ¼ 8). Here, [BLM-A5]0 ¼ 10.8 lM and R is the correlation coefficient of –ln(1 ) x)/t correlating with x/t.

Rb No. [DNA]0 (lM) [FeCl2]0 (mM) [O2]0 (mM) –Q1,1 (mJ) –DrHm (kJÆmol)1) Km (lM) k2 · 102 (s)1)

1

2

3

4 3.47 6.93 6.93 17.3 17.3 17.3 34.7 34.7 0.34 0.34 0.68 0.34 0.34 0.68 0.34 0.68 0.66 0.66 0.66 0.65 0.65 0.65 0.65 0.65 12.00 25.23 23.28 59.65 62.62 59.50 115.5 117.3 ) 0.9916 ) 0.9968 ) 0.9931 ) 0.9967 ) 0.9912 ) 0.9970 ) 0.9957 ) 0.9989 577 607 560 574 602 572 556 564 577 ± 19 24.1 21.9 15.7 23.6 23.3 14.8 22.4 17.0 20.4 ± 3.8 1.77 1.83 1.75 2.90 2.21 2.30 2.49 2.98 2.28 ± 0.49

Table 2. Comparison of the kinetic parameters for BLM-A5 without taking its self-inactivation into account and those for BLM-A5 with first- order self-inactivation. The kinetic data were obtained from the analog calorimetric curve models of a single-substrate enzyme-catalyzed reaction without taking self-inactivation of BLM-A5 into account and with the first-order self-inactivation of BLM-A5.

Km (lM) k2 · 102 (s)1) s (s) Inactivation type of BLM-A5

This result indicates that the absorbing group is unchanged after this scission. The flat peak at 384 nm for curves a and c in Fig. 3B may result from charge transfer transitions between ferric iron and BLM-A5. The reason why we do not observe the charge transfer band in curve e is unknown.

Thermodynamics of the binding of BLM-A5 to DNA

Figure 4 shows two of the calorimetric curves of BLM-A5 binding to calf thymus DNA at different molar ratios of DNA/BLM-A5. The experimental apparent molar enthalpy changes for these reactions can be calculated using the integration type of Tian’s equation and Eqn (16) from the calorimetric curves. The thermodynamic data for the bind- ing, in which the values of KB, DbH 0 m and n are obtained by fitting the apparent molar enthalpy changes to Eqn (18), are summarized in Table 3. The v2 value of Eqn (18) used to perform a nonlinear least-squares analysis for the binding of BLM-A5 to DNA is 0.0268, indicating a good appropriate- ness of the model proposed. The remaining standard thermodynamic parameters for the binding, DbG0 m and DbS0

m , are calculated by Eqns (19) and (20), respectively.

Non-self-inactivation First-order self-inactivation 20.4 4.22 2.28 1.70 188

Thermodynamics of the binding of ADM and (1,10-phenanthroline)-copper to DNA

To establish the action mode of BLM-A5 to DNA, we investigated the energetics for both the binding reactions of ADM and (1,10-phenanthroline)-copper to calf thymus DNA and found that their thermodynamic binding param-

eters were different from those of BLM-A5. ADM is an intercalator, which inserts its aromatic ring between adjacent base pairs of DNA [25,39] and (1,10-phenanthro-

Fig. 3. A comparison of the UV and visible spectrum of BLM-A5 after the cleavage of calf thymus DNA by a mixture of BLM-A5, Fe2+ and O2 with those of five control systems. (a) A reaction system containing 21.5 lM BLM-A5, 20 lM Fe2+ and 15.2 lM calf thymus DNA sat- urated with purified oxygen, after incubation in 50 mM Tris/HCl buffer at pH 7.4 and 20 (cid:2)C for 30 min. (b) 21.5 lM BLM-A5. (c) A solution containing 21.5 lM BLM-A5 and 20 lM Fe3+. (d) A mixture con- taining 21.5 lM BLM-A5 and 20 lM Fe2+ saturated with purified nitrogen. (e) A mixture containing 21.5 lM BLM-A5, 20 lM Fe3+ and 15.2 lM calf thymus DNA. (f ) 15.2 lM calf thymus DNA. (A) shows the optical spectra between 250 and 550 nm and (B) displays those between 350 and 500 nm.

2856 Y. Liang et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

line)-copper binds to either the major or minor groove of the double helix [32]. The thermodynamic data for these binding reactions at 37.0 (cid:2)C are listed in Table 3. The solid lines in Fig. 5B,C are the predicted apparent molar enthalpy changes for these binding reactions as calculated using Eqn (18) and the parameters in Table 3 and in agreement with the experimental data. The v2 value of Eqn (18) used to perform a nonlinear least-squares analysis for the binding of ADM and (1,10-phenanthroline)-copper to DNA are 4.05 and 0.0139, respectively, indicating that the thermodynamic model for the binding of small molecules to DNA proposed in this paper, is reasonable.

Fig. 4. Calorimetric curves of BLM-A5 binding to calf thymus DNA. The initial concentration of calf thymus DNA is 139 lM and the initial concentrations of BLM-A5 are (a) 43.0 lM and (b) 86.0 lM, respect- ively. The experimental temperature is 37.0 (cid:2)C.

D I S C U S S I O N

DNA cleavage efficiency of BLM-A5

In Table 4, we compared the molar enthalpy change for the cleavage of calf thymus DNA induced by BLM-A5 with those for the scission of calf thymus DNA mediated by two well-known DNA-damaging agents, ADM [39,40] and (1,10-phenanthroline)-copper [31,32]. Scission of calf thy- mus DNA induced by BLM in the presence of Fe2+ and O2, converted calf thymus DNA to free nucleic bases [2,5,13,14]. From electrophoresis experiments, it was found that nicking of pBR-322 DNA by a mixture of ADM, Fe3+, Vc and O2 and by a mixture of (1,10-phenanthroline)-copper(II), ME and O2 converted pBR-322 DNA to small DNA fragments [39] and linear DNA [31], respectively. As is seen in Table 4, the higher the degree of DNA strand scission by drugs, the larger the molar enthalpy change for the DNA cleavage.

Fig. 5. Apparent molar enthalpy changes for the binding reactions of (A) BLM-A5 (B) ADM and (C) (OP)2Cu2+, to calf thymus DNA at 37.0 (cid:1)C. The initial concentrations of calf thymus DNA are 139 lM (A,C) and 136 lM (B). Empty circles, experimental data; solid lines, curves predicted by Eqn (18) using the parameters in Table 3.

Table 3. Thermodynamic parameters for the binding reactions of three antitumor drugs, BLM-A5, ADM and (1,10-phenanthroline)-copper, to calf thymus DNA at 37.0 (cid:1)C. These binding reactions were carried out as described in the legend to Fig. 5. Thermodynamic parameters were determined using the thermodynamic model for the binding of small molecules to DNA in the results section. Data are expressed as mean ± SD.

Drug KB · 10)4 )1) (M n (base pairs/drug) DbH 0 m (kJÆmol)1) DbG0 m (kJÆmol)1) DbS0 m (JÆmol)1ÆK)1) Action mode

4.19 ± 0.94 10.9 ± 1.6 21.6 ± 5.7 5.31 ± 0.12 4.83 ± 0.92 3.07 ± 0.10 ) 10.2 ± 0.4 ) 46.3 ± 0.9 16.3 ± 0.2 ) 27.4 ± 0.6 ) 29.9 ± 0.4 ) 31.7 ± 0.7 55.5 ± 3.2 ) 52.9 ± 4.2 155 ± 3 Partial intercalation Intercalation Groove binding BLM-A5 ADM (OP)2Cu2+

(cid:3) FEBS 2002

Thermokinetics of DNA cleavage catalyzed by BLM-A5 (Eur. J. Biochem. 269) 2857

Table 4. Comparison of the molar enthalpy change for the cleavage of calf thymus DNA induced by BLM-A5 and those for the scission of calf thymus DNA mediated by two DNA-damaging agents, ADM and (1,10-phenanthroline)-copper. Data are expressed as mean ± SD (n ¼ 5–8).

a

b

c

Product Reference Cleavage system DrHm (kJÆmol)1)

a T ¼ 37.0 (cid:2)C, pH ¼ 7.4. b T ¼ 25.0 (cid:2)C, pH ¼ 7.4, [ADM]0 ¼ 5.75 lM, [FeCl2]0 ¼ 340 lM, [Vc]0 ¼ 650 lM and oxygen was in excess. c T ¼ 37.0 (cid:2)C, pH ¼ 7.0.

BLM-A5 possessed the highest DNA cleavage efficiency among these DNA-damaging agents.

BLM-A5 is not as efficient as a DNA-cleaving enzyme

ADM, Fe3+, Vc and O2 and by a mixture of (1,10- phenanthroline)-copper(II), ME and O2 do not, however, follow the Michaelis–Menten kinetics (data not shown), suggesting that ADM and (1,10-phenanthroline)-copper are unlike DNA-cleaving enzymes.

Mode of binding BLM-A5 to DNA

BLM has three functional domains (Fig. 1). The metal binding domain is required for metal complexation, oxygen binding and activation [6] and corresponds to the catalytic site of DNA-cleaving enzymes, e.g. EcoRI endonuclease [15]. The DNA binding domain, encompassing the bithiaz- ole moiety, can be regarded as the substrate binding site [15]. The carbohydrate moiety is involved in cell permeability and selective tumor recognition [6]. Although BLM is much smaller than (cid:2)real(cid:3) DNA-cleaving enzymes, it is comparable, both in size and in domains, to the cleft of the active site of such type of enzymes, e.g. EcoRI endonuclease [15].

As shown in Table 3, the binding of ADM to calf thymus DNA is driven entirely by a large favorable enthalpy reduction but with an unfavorable entropy decrease. In contrast, the binding of (1,10-phenanthroline)-copper to calf thymus DNA shows just an opposite thermodynamics of the reaction driven by a large favorable increase in entropy with an unfavorable raise in enthalpy. Meanwhile, the binding of BLM-A5 to calf thymus DNA seems to be driven by a favorable entropy change with a less favorable enthalpy change. These results indicate that the thermodynamic binding behavior of BLM-A5 ranges between those of ADM and (1,10-phenanthroline)-copper and are in line with a partial intercalation mode involved in BLM-catalyzed breakage of DNA [44,45]. The partial intercalation given here is a threading intercalation mode [6,44,45] in which the bithiazole moiety is partially intercalated between DNA base pairs and the C-terminal substituent has been threaded through the helix to the major groove. The partial interca- lation of BLM induces the relaxation of supercoiled DNA [4], resulting in a moderately favorable increase in entropy.

About the self-inactivation of activated BLM Both Fe2+ and O2 serve as cofactors in DNA cleavage by BLM [2–6]. When ferrous BLM is exposed to O2, a transient complex of drug, iron and oxygen, which is kinetically competent to initiate DNA degradation and commonly termed activated BLM, is formed [2,4,5,13,14,16,18,46].

Table 5 compares the kinetic parameters for BLM-A5 with those for carbonic anhydrase [41], lysozyme [41], TaqI restriction endonuclease [42], NaeI endonuclease [43], BamHI endonuclease [11], blenoxane [11] and BLM-A2 [12]. As shown in Table 5, the catalytic efficiency (repre- sented by k2/Km) of BLM-A5 is of the same order of magnitude as that of lysozyme but several orders of magnitude lower than those of TaqI restriction endonuc- lease, NaeI endonuclease and BamHI endonuclease. As can also be seen from Table 5, the cleavage efficiencies (repre- sented by k2; [11]) of BLM-A5 and of some DNA-cleaving enzymes, such as TaqI restriction endonuclease, NaeI endonuclease and BamHI endonuclease, are of the same order of magnitude but one order of magnitude higher than those of blenoxane and BLM-A2. The catalytic efficiency is a much better measure for the efficiency of an enzyme than k2 (in this case the cleavage efficiency). Therefore, BLM-A5 is not as efficient as a DNA-cleaving enzyme although the cleavage of DNA by BLM-A5 follows Michaelis–Menten kinetics. The cleavage of calf thymus DNA by a mixture of

)577 ± 20 )147.1 ± 6.1 )35.1 ± 1.8 Free nucleic bases Small DNA fragments Linear DNA This work, [2,5,13,14] This work, [39] [31] BLM-A5-Fe2+-O2 ADM-Fe3+-Vc-O2 (OP)2Cu2+-ME-O2

)1Æs)1)

Table 5. Comparison of the kinetic parameters for BLM-A5 and those for carbonic anhydrase, lysozyme, TaqI restriction endonuclease, NaeI endonuclease, BamHI endonuclease, blenoxane and BLM-A2. Here, NAG is N-acetylglucosamine.

–

Reference Enzyme Substrate k2/Km (M Km (M) k2 (s)1)

4 · 105

HCO3 (NAG)2 DNA [41] [41] [42] 9.6 · 10)3 1.75 · 10)4 5.3 · 10)8 0.5 2.2 · 10)2 4.2 · 107 2.9 · 103 4.2 · 106

1.0 · 10)8 8.9 · 10)9 4.5 · 106 7.9 · 105

DNA DNA DNA DNA DNA [43] [11] [11] [12] This work 2.04 · 10)5 4.5 · 10)2 7.0 · 10)3 1 · 10)3 2.39 · 10)3 2.28 · 10)2 1.12 · 103 Carbonic anhydrase Lysozyme TaqI restriction endonuclease NaeI endonuclease BamHI endonuclease Blenoxane BLM-A2 BLM-A5

2858 Y. Liang et al. (Eur. J. Biochem. 269)

(cid:3) FEBS 2002

4. Stubbe, J. & Kozarich, J.W. (1987) Mechanism of bleomycin- induced DNA degradation. Chem. Rev. 87, 1107–1136. 5. Burger, R.M. (1998) Cleavage of nucleic acids by bleomycin. Chem. Rev. 98, 1153–1169.

6. Abraham, A.T., Zhou, X. & Hecht, S.M. (1999) DNA cleavage by Fe (II) bleomycin conjugated to a solid support. J. Am. Chem. Soc. 121, 1982–1983.

7. Holmes, C.E., Carter, B.J. & Hecht, S.M. (1993) Characterization of iron (II) bleomycin-mediated RNA strand scission. Biochem- istry 32, 4293–4307. 8. Hecht, S.M. (1994) RNA degradation by bleomycin, a naturally occurring bioconjugate. Bioconjugate Chem. 5, 513–526.

9. Caspary, W.J., Niziak, C., Lando, D.A., Friedman, R. & Bachur, N.R. (1979) Bleomycin A2: a ferrous oxidase. Mol. Pharmacol. 16, 256–260.

10. Caspary, W.J., Lando, D.A. & Niziak, C. (1981) Intermediates in the ferrous oxidase cycle of bleomycin. Biochemistry 20, 3868– 3875.

11. Biggins, J.B., Prudent, J.R., Marshall, D.J., Ruppen, M. & Thorson, J.S. (2000) A continuous assay for DNA cleavage: the application of (cid:2)break lights(cid:3) to enediynes, iron-dependent agents, and nucleases. Proc. Natl Acad. Sci. USA 97, 13537–13542. 12. Hashimoto, S., Wang, B. & Hecht, S.M. (2001) Kinetics of DNA cleavage by Fe (II) bleomycins. J. Am. Chem. Soc. 123, 7437–7438. 13. Kozarich, J.W., Worth Jr, L., Frank, B.L., Christner, D.F., Vanderwall, D.E. & Stubbe, J. (1989) Sequence-specific isotope effects on the cleavage of DNA by bleomycin. Science 245, 1396– 1399.

14. Natrajan, A., Hecht, S.M., van der Marel, G.A. & van Boom, J.H. (1990) A study of O2- versus H2O2-supported activation of Fe-bleomycin. J. Am. Chem. Soc. 112, 3997–4002.

Nakamura & Peisach [47] have suggested that the bithiazloe structure of BLM-A2 is altered when it is inactivated. It has also been shown that activated BLM-A2 undergoes self-inactivation to a very substantial extent concomitant with its cleavage of DNA [5,46–49]. As some of the molecules become inactivated and thus are no longer capable of cleaving DNA, the measured kinetics of cleavage will lead to underestimating of the cleaving potential of the remaining molecules. The impact of self-inactivation of activated BLM on the thermodynamics of DNA binding is more complex. As the structural identities of the BLM degradation products are unknown, it is unclear whether those products bind to DNA themselves and with what properties and they might affect BLM binding. Although it is indicated in this paper that the chromophoric group of BLM-A5 is unchanged when it cleaves DNA, activated BLM-A5 could undergo the first-order self-inactivation to some extent. As shown in Table 2, the lifetime of self- inactivation of BLM-A5 obtained from a model with the first-order self-inactivation is close to that reported by Burger and coworkers [46] and the summed v2 of the fit using this model is of the same order of magnitude as that of the model without taking self-inactivation into account (data not shown). A first-order self-inactivation could be due to denaturation of the peptide part of the compound leaving the bithiazloe unit intact but uncoupling the DNA binding part of the metal complexation part (feasible at 37 (cid:2)C). Moreover, in this paper, calf thymus DNA is present when BLM-A5 is mixed with Fe2+ and O2 but not added after drug activation, and it is well known that DNA does protect activated BLM against self-inactivation [5,46– 49]. Activated BLM-A5 may lose its activity slower when complexed with DNA.

15. Owa, T., Haupt, A., Otsuka, M., Kobayashi, S., Tomioka, N., Itai, A. & Ohno, M. (1992) Man-designed bleomycins: significance of the binding sites as enzyme models and of the stereochemistry of the linker moiety. Tetrahedron 48, 1193–1208.

16. Sam, J.W., Tang, X.-J. & Peisach, J. (1994) Electrospray mass spectrometry of iron bleomycin: demonstration that activated bleomycin is a ferric peroxide complex. J. Am. Chem. Soc. 116, 5250–5256.

In this paper, microcalorimetry and UV-vis spectroscopy have been combined to study the scission of calf thymus DNA catalyzed by BLM-A5. A novel thermokinetic method for an enzyme-catalyzed reaction has been pro- posed and employed to produce not only the thermody- namic constant but also the kinetic properties of DNA cleavage by BLM-A5 with the result that BLM-A5 is not as efficient as a DNA-cleaving enzyme. The present thermo- dynamic and kinetic findings have provided further insights into the mechanism with which BLM functions as both a DNA-damaging agent and an antitumor drug.

17. Boger, D.L., Ramsey, T.M., Cai, H., Hoehn, S.T., Kozarich, J.W. & Stubbe, J. (1998) Definition of the effect and role of the bleo- mycin A2 valerate substituents: preorganization of a rigid, com- pact conformation implicated in sequence-selective DNA cleavage. J. Am. Chem. Soc. 120, 9149–9158.

18. Veselov, A., Burger, R.M. & Scholes, C.P. (1998) Q-band electron nuclear double resonance of ferric bleomycin and activated bleo- mycin complexes with DNA: Fe(III) hyperfine interaction with 31P and DNA-induced perturbation to bleomycin structure. J. Am. Chem. Soc. 120, 1030–1033.

A C K N O W L E D G E M E N T S

19. Sam, J.W., Takahashi, S., Lippai, I., Peisach, J. & Rousseau, D.L. (1998) Sequence-specific changes in the metal site of ferric bleo- mycin induced by the binding of DNA. J. Biol. Chem. 273, 16090– 16097.

20. Aso, M., Kondo, M., Suemune, H. & Hecht, S.M. (1999) Chemistry of the bleomycin-induced alkali-labile DNA lesion. J. Am. Chem. Soc. 121, 9023–9033. This work was supported by the 973 Project (G1999075608) from the Chinese Minister of Science and Technology and the grant (39970164) from the National Natural Science Foundation of China. We are also grateful to Prof C. L. Tsou and Prof J. M. Zhou (Institute of Biophysics, Academia Sinica, China) for their critical reading of the manuscript and for their helpful suggestions.

21. Wadso¨ , I. (1997) Isothermal microcalorimetry near ambient temperature: an overview and discussion. Thermochim. Acta 294, 1–11.

R E F E R E N C E S

1. Umezawa, H., Maeda, K., Takeuchi, T. & Okami, Y. (1966) New antibiotics, bleomycin A and B. J. Antibiot. A 19, 200–209. 2. Claussen, C.A. & Long, E.C. (1999) Nucleic acid recognition by metal complexes of bleomycin. Chem. Rev. 99, 2797–2816. 22. Doyle, M.L. (1997) Characterization of binding interactions by isothermal titration calorimetry. Curr. Opin. Biotechnol. 8, 31–35. 23. Ladbury, J.E. & Chowdhry, B.Z. (1998) Biocalorimetry: Appli- cations of Calorimetry in the Biological Sciences. John Wiley & Sons, UK.

3. Goodisman, J., Kirk, C. & Dabrowiak, J.C. (1997) Kinetic ana- lysis of drug cleavage of closed-circular DNA. Biophys. Chem. 69, 249–268. 24. Breslauer, K.J., Freire, E. & Straume, M. (1992) Calorimetry: a tool for DNA and ligand-DNA studies. Methods Enzymol. 211, 533–567.

(cid:3) FEBS 2002

Thermokinetics of DNA cleavage catalyzed by BLM-A5 (Eur. J. Biochem. 269) 2859

25. Chaires, J.B. (1997) Energetics of drug–DNA interactions. Bio- polymers 44, 201–215. 37. Laidler, K.J. & Bunting, P.S. (1973) The Chemical Kinetics of Enzyme Action. pp. 175–180, 413–429. Clarendon Press, Oxford. 26. Roche, C.J., Thomson, J.A. & Crothers, D.M. (1994) Site selectivity of daunomycin. Biochemistry 33, 926–935.

38. Zeng, X.C., Zhang, Y.Q., Meng, X.G. & Tian, A.M. (1997) Thermo-kinetic research methods for simple order reactions: analog curve and characteristic parameter methods. Thermochim. Acta 293, 171–177. 27. Haq, I., Ladbury, J.E., Chowdhry, B.Z., Jenkins, T.C. & Chaires, J.B. (1997) Specific binding of Hoechst 33258 to the d(CGCAAATTTGCG)2 duplex: calorimetric and spectroscopic studies. J. Mol. Biol. 271, 244–257.

39. Wang, H.F., Yang, P., Li, Q.S. & Wu, Q. (1997) Electrochemistry characteristics of adriamycin-Fe (III) complex and its interaction with DNA. Chem. J. Chinese University 18, 671–675.

28. Kagemoto, A., Yoshii, A., Kimura, S. & Baba, Y. (1994) Ther- modynamics of interactions between ethidium bromide and poly(A)-poly(U) mixtures in dilute and concentrated solutions. J. Phys. Chem. 98, 5943–5952. 40. Eliot, H., Gianni, L. & Myers, C. (1984) Oxidative destruction of DNA by the adriamycin–iron complex. Biochemistry 23, 928–936. 41. Britt, B.M. (1997) For enzymes, bigger is better. Biophys. Chem. 69, 63–70.

29. Becker, H.-C. & Norde´ n, B. (1997) DNA binding properties of 2,7-diazapyrene and its N-methylated cations studied by linear and circular dichroism spectroscopy and calorimetry. J. Am. Chem. Soc. 120, 1030–1033. 42. Zebala, J.A., Choi, J. & Barany, F. (1992) Characterization of steady state, single-turnover, and binding kinetics of the TaqI restriction endonuclease. J. Biol. Chem. 267, 8097–8105.

30. Bathaie, S.Z., Moosavi-Movahedi, A.A. & Saboury, A.A. (1999) Energetic and binding properties of DNA upon interaction with dodecyl trimethylammonium bromide. Nucleic Acids Res. 27, 1001–1005. 43. Yang, C.C., Baxter, B.K. & Topal, M.D. (1994) DNA cleavage by NaeI: protein purification, rate-limiting step, and accuracy. Bio- chemistry 33, 14918–14925.

31. Liang, Y., Qu, S.S., Wang, C.X., Liu, Y.W., Wang, Z.Y., Song, Z.H., Zou, G.L. & Ou, R. (1998) Microcalorimetric studies on the DNA scission by (1,10-phenanthroline) copper. Acta Chim. Sin. 56, 1145–1152. 32. Sigman, D.S., Mazamder, A. & Perrin, D.M. (1993) Chemical 44. Wu, W., Vanderwall, D.E., Lui, S.M., Tang, X.J., Turner, C.J., Kozarich, J.W. & Stubbe, J. (1996) Studies of Cobleomycin A2 green: its detailed structural characterization by NMR and mo- lecular modeling and its sequence–specific interaction with DNA oligonucleotides. J. Am. Chem. Soc. 118, 1268–1280. nucleases. Chem. Rev. 93, 2295–2316.

45. Wu, W., Vanderwall, D.E., Turner, C.J., Kozarich, J.W. & Stubbe, J. (1996) Solution structure of Cobleomycin A2 green complexed with d(CCAGGCCTGG). J. Am. Chem. Soc. 118, 1281–1294. 33. Liang, Y., Wu, Y.X., Li, D.H., Wang, C.X., Liu, Y., Qu, S.S. & Zou, G.L. (1997) Thermokinetic models of enzyme-catalyzed reactions in batch and plug-flow reactors. Thermochim. Acta 307, 149–153.

46. Burger, R.M., Peisach, J. & Horwitz, S.B. (1981) Activated bleomycin: a transient complex of drug, iron, and oxygen that degrades DNA. J. Biol. Chem. 256, 11636–11644. 47. Nakamura, M., Peisach, J. & Self-inactivation of Fe (II)-bleo- 34. Liang, Y., Wang, C.X., Qu, S.S., Wu, Y.X., Li, D.H. & Zou, G.L. (1998) Thermokinetic method for faster enzyme-catalyzed reac- tions. Thermochim. Acta 322, 1–7. mycin. (1988) J. Antibiot. 41, 638–647.

48. Burger, R.M., Peisach, J., Blumberg, W.E. & Horwitz, S.B. (1979) Iron–bleomycin interactions with oxygen and oxygen analogues. J. Biol. Chem. 254, 10906–10912. 35. Liang, Y., Li, J., Chen, J. & Wang, C.C. (2001) Thermodynamics of the folding of D-glyceraldehyde-3-phosphate dehydrogenase assisted by protein disulfide isomerase studied by micro- calorimetry. Eur. J. Biochem. 268, 4183–4189.

49. Van Atta, R.B., Long, E.C. & Hecht, S.M. (1989) Electrochemical activation of oxygenated Febleomycin. J. Am. Chem. Soc. 111, 2722–2724. 36. Fersht, A. (1999) Structure and Mechanism in Protein Science: a Guide to Enzyme Catalysis and Protein Folding, pp. 108–111. W.H. Freeman Co, New York.