Báo cáo khoa học: Modular kinetic analysis reveals differences in Cd2+ and Cu2+ ion-induced impairment of oxidative phosphorylation in liver pot

In agreement with modular kinetic analysis data, metabolic control analysis showed that Cd2+ and Cu2+ions increased control of the respiratory and phosphorylation flux by the respiratory

Cu2+ ion-induced impairment of oxidative phosphorylation

in liver

Jolita Ciapaite1, Zita Nauciene1,2, Rasa Baniene2, Marijke J Wagner3, Klaas Krab3and

Vida Mildaziene1

1 Centre of Environmental Research, Faculty of Natural Sciences, Vytautas Magnus University, Kaunas, Lithuania

2 Institute for Biomedical Research, Kaunas Medical University, Lithuania

3 Department of Molecular Cell Physiology, Institute for Molecular Cell Biology, VU University, Amsterdam, The Netherlands

Many pollutants, even at low effective concentrations,

can harm living organisms by weakening their ability

to cope with long-term environmental challenges At

excess amounts, the heavy metals cadmium and copper

are toxic and carcinogenic [1] The ability of cadmium

and copper to accumulate in the bones, liver and kid-neys determines their toxicity Their deleterious effects can be ameliorated to some extent by binding to metallothionein [2] Cellular dysfunction induced by cadmium and copper is thought to involve alterations

Keywords

cadmium and copper; lipid peroxidation;

metabolic control analysis; modular kinetic

analysis; oxidative phosphorylation

Correspondence

J Ciapaite, Centre of Environmental

Research, Faculty of Natural Sciences,

Vytautas Magnus University, Vileikos 8,

LT-44404 Kaunas, Lithuania

Fax: +370 37 327916

Tel: +370 37 455193

E-mail: jolita.ciapaite@falw.vu.nl

(Received 29 January 2009, revised 18 April

2009, accepted 5 May 2009)

doi:10.1111/j.1742-4658.2009.07084.x

Impaired mitochondrial function contributes to copper- and cadmium-induced cellular dysfunction In this study, we used modular kinetic analy-sis and metabolic control analyanaly-sis to assess how Cd2+ and Cu2+ ions affect the kinetics and control of oxidative phosphorylation in isolated rat liver mitochondria For the analysis, the system was modularized in two ways: (a) respiratory chain, phosphorylation and proton leak; and (b) coen-zyme Q reduction and oxidation, with the membrane potential (Dw) and fraction of reduced coenzyme Q as the connecting intermediate, respec-tively Modular kinetic analysis results indicate that both Cd2+ and Cu2+ ions inhibited the respiratory chain downstream of coenzyme Q Moreover,

Cu2+, but not Cd2+ ions stimulated proton leak kinetics at high Dw val-ues Further analysis showed that this difference can be explained by Cu2+ ion-induced production of reactive oxygen species and membrane lipid peroxidation In agreement with modular kinetic analysis data, metabolic control analysis showed that Cd2+ and Cu2+ions increased control of the respiratory and phosphorylation flux by the respiratory chain module (mainly because of an increase in the control exerted by cytochrome bc1 and cytochrome c oxidase), decreased control by the phosphorylation module and increased negative control of the phosphorylation flux by the proton leak module In summary, we showed that there is a subtle differ-ence in the mode of action of Cd2+and Cu2+ ions on the mitochondrial function, which is related to the ability of Cu2+ ions to induce reactive oxygen species production and lipid peroxidation

Abbreviations

CiJP, flux control coefficient, quantifying the control of phosphorylation flux J P by module i; CiJR, flux control coefficient, quantifying the control of respiratory flux JRby module i; CoQ, coenzyme Q; COX, cytochrome c oxidase; DCPIP, 2,6-dichlorophenolindophenol; JL,proton leak flux; JP,phosphorylation flux; JR,respiratory flux; MCA, metabolic control analysis; ROS, reactive oxygen species; SDH, succinate dehydrogenase; TBA, 2-thiobarbituric acid; TBARS, thiobarbituric acid reactive substances; X CoQ, fraction of reduced coenzyme Q;

Dp, proton-motive force; Dw, mitochondrial transmembrane electric potential.

in mitochondrial metabolism Initially, the

accumula-tion of metal ions in mitochondria may protect the cell

against metal overload, however, later, their

incorpora-tion may cause complex disturbances in mitochondrial

function [3–5], resulting in severe defects in cellular

metabolism

Cd2+ions are transported into the mitochondria via

a Ca2+ -uniporter [6], whereas the accumulation of

Cu2+in mitochondria proceeds via a different,

energy-independent mechanism [7] Both metal ions interact

with important functional groups (in particular, thiol

groups) in a variety of enzymes in the matrix and inner

mitochondrial membrane [8] At micromolar

concentrations, Cd2+ ions uncouple oxidative

phos-phorylation and inhibit respiration in actively

ADP-phosphorylating (state 3) isolated rat liver

mito-chondria [4] At higher concentrations, Cd2+ions

inhi-bit succinate dehydrogenase (SDH) and H+-ATPase

[5,9] Increasing the amount of Cu2+ions added per mg

of mitochondrial protein has been shown to successively

cause inhibition of phosphate transport, accumulation

of K+ions, membrane aggregation, stimulation of

res-piration in the absence of active ADP phosphorylation

(state 4), an increase in passive membrane permeability

to cations and anions, uncoupling and swelling, and

inhibition of respiration in state 3 [3,4] Furthermore, it

has been suggested that Cu2+ions inhibit SDH [10]

Cd2+ and Cu2+ ions induce cell death by necrosis

and apoptosis via mechanisms involving opening of

the mitochondrial permeability transition pore and

increased generation of reactive oxygen species (ROS)

[11–13] In turn, metal-induced stimulation of ROS

production has been suggested to stem from both

increased ROS production by the mitochondrial

respi-ratory chain and decreased activity of the antioxidant

enzymes [9,11,14–16]

Although the effects of Cd2+ and Cu2+ ions on

some individual mitochondrial enzymes and processes

have been studied extensively, few attempts have been

made to elucidate the mode of action of Cd2+ and

Cu2+ ions at the system level, which, in turn, would

allow us to understand the complex metabolic effects

of these substances

Metabolic control analysis (MCA) is a useful tool

for studying complex biological systems because it

allows the quantification of the contribution made by

each system component to system behavior (e.g fluxes,

metabolite concentrations) in terms of control

coeffi-cients [17,18] In turn, knowledge of the system’s

con-trol structure is valuable in that it allows identification

of system components that are potentially most

impor-tant in mediating the effects of external effectors on the

system (i.e the component with the highest control

coefficient) [19] A ‘top-down’ elasticity analysis (or modular kinetic analysis) was developed to simplify experimental assessment of the control structure of the complex system via MCA [20], and was initially used to study the control of fluxes and intermediates in the oxi-dative phosphorylation system [21] The method is also valuable in determining the sites of action of external effectors within a system [22–27] In this type of analy-sis, the system of interest is conceptually subdivided into functional modules (reaction blocks) in such a way that the selected modules interact via a single connect-ing intermediate In the further analysis, each module is treated as a single enzyme Figure 1A shows how the oxidative phosphorylation system can be subdivided into three functional modules (respiratory chain, phos-phorylating and proton leak module) with membrane

Fig 1 Modularization of the system Division of the oxidative phos-phorylation system into (A) the respiratory chain, phosphos-phorylation and proton leak modules, with Dw as the connecting intermediate; (B) the CoQ-reducing and CoQ-oxidizing modules with fraction of reduced CoQ (X CoQ ) as the connecting intermediate; and (C) the CoQ-reducing, cytochrome bc1+ COX, phosphorylation and proton leak modules with Dw and X CoQ as the connecting intermediates Succ, succinate; SDH, dicarboxylate carrier and succinate dehydro-genase; cyt bc1, cytochrome bc1, COX, cytochrome c oxidase, XCoQ, fraction of reduced CoQ Arrows marked e and h indicate electron and trans-membrane proton flux, respectively (in this study all fluxes were analyzed in terms of oxygen consumption flux).

potential (Dw) as the connecting intermediate [21] To

obtain module kinetics, titrations with module-specific

inhibitors are performed and the flux and level of the

connecting intermediate measured Elasticity

coeffi-cients, quantifying sensitivity of the flux through the

module to a change in the level of the connecting

inter-mediate, can be calculated from the slope of the

inhibi-tor titration curves at a steady state In turn, elasticity

coefficients and steady-state flux values can be used to

calculate the flux control coefficients of the modules

[21] Repeating the procedure in the presence of a fixed

concentration of external effector reveals how that

effector affects the kinetics of each module and the

magnitude of the control that each module exerts

over system fluxes The drawback of the ‘top-down’

approach to MCA is that it yields a coarse picture of

the control structure of the system Different ways of

modularizing the system of interest may allow a more

resolved picture However, this is often limited by the

feasibility of assigning modules that interact via a single

connecting intermediate [28]

In this study, we used modular kinetic analysis and

MCA to determine the effects of low concentrations

(5 lm) of CdCl2and CuCl2on the kinetics and control

of oxidative phosphorylation in isolated rat liver

mito-chondria respiring on succinate To obtain a more

resolved picture of the effects of Cd2+ and Cu2+ ions

on the system, we subdivided the oxidative

phosphory-lation system into modules in different ways (Fig 1)

We showed that at the concentration tested, both

metal ions inhibited respiratory chain module

compo-nents downstream of coenzyme Q (CoQ) In addition,

Cu2+ ions increased the permeability of the inner

membrane to ions at high Dw levels We tested a

hypothesis that the latter effect resulted from Cu2+

ion-induced formation of ROS and lipid peroxidation

Results

Three-modular kinetic analysis of effects of Cd2+

and Cu2+ions on oxidative phosphorylation

To determine which oxidative phosphorylation

compo-nents were affected by Cd2+ and Cu2+ ions in liver

mitochondria oxidizing succinate, we first used

three-modular kinetic analysis with Dw as the connecting

intermediate (Fig 1A) We assessed the effects of a

low metal ion concentration, which did not induce

mitochondrial swelling (results not shown)

Figure 2 shows the effect of 5 lm CdCl2 on the

kinetics of the three modules The plots indicate that

Cd2+ ions inhibited the respiratory chain module

because the respiratory flux (JR) is lower in the

pres-ence of Cd2+ ions than in their absence, when com-pared at the same Dw value (Fig 2B) The kinetics of the proton leak and phosphorylation modules were not significantly affected by Cd2+ions (Fig 2A,C), as indi-cated by similar values for the proton leak (JL) and phosphorylation (JP) flux in the presence and absence

of CdCl2, when the fluxes are compared at the same

Dw value Inhibition of the respiratory chain module

by Cd2+ions resulted in a decrease in Dw by 6 mV in state 3 (i.e the state of maximal ADP phosphorylation) (Table 1) Furthermore, Cd2+ ions decreased JR and

JP by 23 and 25%, respectively However, Cd2+ ions had no significant effect on JLin state 3

Three-modular kinetic analysis of the effects of Cu2+ ions is shown in Fig 3 Similar to Cd2+, Cu2+ ions inhibited the respiratory chain module (Fig 3B), although to a lesser extent Cu2+ions had no significant effect on the kinetics of the phosphorylation module (Fig 3C), but clearly stimulated proton leak kinetics (Fig 3A) The increase in JL was more prominent at higher Dw values corresponding to state 4 (i.e the state with no ADP phosphorylation) (Fig 3A) In state 4,

Cu2+ions stimulated JLby 42% and caused a decrease

of 11 mV in Dw In state 3, Cu2+ ions inhibited JRby 16% and JPby 17%, respectively, but had no significant effect on JL(Table 1) Despite moderate effects on the fluxes, in state 3 Cu2+ions had a strong effect on Dw, which decreased by 12 mV (Table 1)

It should be noted that the effects of Cd2+and Cu2+ ions were determined in two separate series of experi-ments (performed in spring and autumn, respectively) resulting in two sets of flux and Dw values under control conditions (K1 and K2; Table 1) The difference between the two data sets may have been caused by hormone-related seasonal variations in mitochondrial properties (e.g the expression levels of the enzymes involved in the process of oxidative phosphorylation),

as observed in different tissues in rodents [29,30]

Bimodulular kinetic analysis of the effects of Cd2+ and Cu2+ions on oxidative phosphorylation The data in Figs 2 and 3 indicate that Cd2+and Cu2+ ions affected the respiratory chain module Therefore,

as a next step, we set out to pinpoint the components

of the respiratory chain module affected by Cd2+and

Cu2+ ions To achieve this, we conceptually subdi-vided the oxidative phosphorylation system into two modules: (a) CoQ reducing, comprising dicarboxylate carrier, fumarase and succinate dehydrogenase; and (b) CoQ oxidizing, comprising cytochrome bc1, cyto-chrome c oxidase (COX) and the rest of the oxidative phosphorylation system, including proton leak and

enzymes involved in ATP synthesis We used the

frac-tion of CoQ (XCoQ) as the connecting intermediate

(Fig 1B)

The results of bimodular kinetic analysis with XCoQ

as the connecting intermediate are presented in Fig 4

A similar JR value at any given XCoQ indicates that

the kinetics of the CoQ-reducing module was not

sig-nificantly affected by either Cd2+ (Fig 4B) or Cu2+

ions (Fig 4D) Both metal ions inhibited the

CoQ-oxi-dazing module (Fig 4A,C) because lower JR values

were observed when comparison was made at the same

XCoQ level Because three-modular analysis showed

that neither of the ions had any effect on the enzymes

involved in ATP synthesis (Figs 2C and 3C) or on the

proton leak kinetics close to state 3 (Figs 2A and 3A),

we can conclude that the site of action of Cd2+ and

Cu2+ions must be cytochrome bc1and⁄ or COX

Effects of Cd2+and Cu2+ions on the activity of succinate dehydrogenase

Bimodular kinetic analysis showed that SDH (a com-ponent of the CoQ-reducing module) was not signifi-cantly affected by either Cd2+or Cu2+ions However, literature reports suggest that SDH is the target of both metal ions [5,9,10] To check whether data obtained using modular kinetic analysis were correct,

we determined the effect of Cd2+ and Cu2+ ions on SDH activity in isolated rat liver mitochondria The dependence of SDH activity on the concentration of CdCl2 and CuCl2 is shown in Fig 5A and B, respec-tively At 5 lm neither CdCl2 nor CuCl2 had any significant effect on SDH activity, which was

51 ± 8 nmol 2,6-dichlorophenolindophenol (DCPIP)Æ min)1Æmg protein)1 under control conditions and

52 ± 7 and 48 ± 9 nmol DCPIPÆmin)1Æmg protein)1

in the presence of 5 lm CdCl2and 5 lm CuCl2, respec-tively A significant effect on SDH activity was observed only at CdCl2 and CuCl2 concentrations exceeding 10 lm (Fig 5)

Effects of Cd2+and Cu2+ions on H2O2production and lipid peroxidation

We hypothesized that stronger stimulation of proton leak kinetics by Cu2+ ions (Fig 3A) compared with

Cd2+ions (Fig 2A) may be explained by the ability of

Cu2+ ions to stimulate ROS production and induce peroxidation of the membrane lipids Therefore, we

0

50

100

150

200

100 120 140 160 180

J L

J R

–1 )

J P

–1 ·mg

Δψ

Δψ (mV)

A

State 4

0

50

100

150

200

100 120 140 160 180

Δψ

Δψ (mV)

B

State 3

State 4

0

50

100

150

200

100 120 140 160 180

Δψ

Δψ (mV)

C

State 3

Fig 2 Effect of Cd 2+ ions on the kinetics of the proton leak module (A), the respiratory chain module (B) and the phosphorylation module (C) The kinetics of the proton leak module were obtained by titrating with a specific inhibitor of the respiratory chain module, malonate (0–12.5 l M ), when phosphorylation module activity is fully blocked with oligomycin (0.7 lgÆmL)1) The kinetics of the respiratory chain module were obtained by titrating with a specific inhibitor of the phosphorylation module, carboxyatractyloside (0–0.5 l M ) The kinetics of the phosphorylation module were obtained by titrating with a specific inhibitor of the respiratory chain module, malonate (0–3.125 l M ), and subsequently calculating JPby subtracting JLfrom JRat the same value of Dw [21] JR, respiratory flux; JP, phosphorylation flux; JL, proton leak flux Open symbols, no CdCl 2 added; closed symbols, plus 5 l M CdCl 2 Average of n = 6 independent experiments ± SEM.

Table 1 Effect of Cd 2+ and Cu 2+ ions on system properties in

state 3 Average of n = 4 independent experiments ± SEM K1 and

K2, control experiments with no CdCl 2 or CuCl 2 added; J R ,

respira-tory flux; JP, phosphorylation flux; JL, proton leak flux.

K1 5 l M CdCl2 K2 5 l M CuCl2

J R (nmol OÆmin)1

Æmg protein)1)

171 ± 14 131 ± 16* 140 ± 7 118 ± 1*

JP(nmol OÆmin)1

Æmg protein)1)

162 ± 14 122 ± 15* 131 ± 6 109 ± 2*

JL(nmol OÆmin)1

Æmg protein)1)

Dw (mV) 143 ± 2 137 ± 2* 140 ± 3 128 ± 1*

*P < 0.05 versus the condition with no CdCl 2 or CuCl 2 added.

assessed how Cd2+ and Cu2+ ions affect overall ROS

production in isolated mitochondria oxidizing succinate

in state 2 (i.e the resting state with no ADP

phosphory-lation) Figure 6A shows that 5 lm CdCl2 (i.e the

concentration used for modular kinetic analysis) had no

significant effect on overall H2O2 production, as

indi-cated by the unchanged oxidation rate of

2¢,7¢-dichloro-fluorescin (DCF) In turn, 5 lm CuCl2 stimulated the

rate of DCF oxidation by 43% (Fig 6A) Increasing the concentration of CdCl2and CuCl2to 10 lm resulted in DCF oxidation rates that were 1.7 (P < 0.01) and 2.1 (P < 0.01) times higher, respectively

Next, we assessed the ability of both metal ions to induce lipid peroxidation in isolated rat liver mitochon-dria respiring on succinate in state 2 Figure 6B shows the effects of Cd2+and Cu2+ions on the formation of

0

50

100

150

200

Δψ

Δψ (mV)

A

State 4

0

50

100

150

200

Δψ

Δψ (mV)

B

State 3

State 4

0

50

100

150

200

Δψ

Δψ (mV)

C

State 3

J L

J R

–1 )

J P

–1 ·mg

Fig 3 Effect of Cu 2+ ions on the kinetics of the proton leak module (A), the respiratory chain module (B) and the phosphorylation module (C) The kinetics of the modules were obtained as described in the legend for Fig 2 JR, respiratory flux; JP, phosphorylation flux; JL, proton leak flux Open symbols, no CuCl 2 added; closed symbols, plus 5 l M CuCl 2 Average of n = 5 independent experiments ± SEM.

0

50

100

150

200

250

Fraction of reduced CoQ (%)

D

State 3

0

50

100

150

200

250

Fraction of reduced CoQ (%)

C

State 3

0

50

100

150

200

250

20 30 40 50 60 70 80

20 30 40 50 60 70 80

Fraction of reduced CoQ (%)

B

State 3

0

50

100

150

200

250

Fraction of reduced CoQ (%)

A

State 3

J R

J R

J R

J R

Fig 4 Effect of Cd 2+ and Cu 2+ ions on the kinetics of the reducing and CoQ-oxidizing modules Effect of Cd2+on the kinetics of the CoQ-oxidizing module (A) and CoQ-reducing module (B) Effect of Cu 2+ on the kinetics of the CoQ-oxidizing module (C) and CoQ-reducing module (D) The kinetics

of the CoQ-oxidizing module were obtained

by titrating with a specific inhibitor of the CoQ-reducing module, malonate (0–3.125 l M ) The kinetics of the CoQ-reducing module were obtained by titrating with a specific inhibitor of the CoQ-oxidizing module, myxothiazol (0–80 n M ) JR, respira-tory flux Open symbols, no CdCl2or CuCl2 added; closed symbols, plus 5 l M CdCl 2

(A,B) or 5 l M CuCl2(C,D) Average of n = 4 (A,B) and n = 2 (C,D) independent experiments ± SEM.

thiobarbituric acid reactive substances (TBARS), which

indicate levels of the lipid peroxidation product

mal-ondialdehyde At the amounts tested (5 and 10 nmol

Cd2+ÆmgÆmitochondrial protein)1), Cd2+ ions had no

significant effect on TBARS formation However, for

Cu2+ions, addition of 5 nmolÆmg protein)1significantly

increased (by 26%) the amount of TBARS per mg of

mitochondrial protein Increasing the amount of added

Cu2+ ions to 10 nmolÆmg protein)1 did not further

increase the amount of TBARS formed (Fig 6B)

Effects of Cd2+and Cu2+ions on the control of

fluxes in oxidative phosphorylation

Using MCA, we assessed the contribution made by

each oxidative phosphorylation module to the control

of JR and JP (Table 2 and Fig 7) Three-modular

analysis (Fig 1A) revealed that control of JR and JP

was shared between the respiratory chain and

phos-phorylation modules, with the former exerting

some-what more control As expected for state 3 conditions,

the proton leak module exerted low levels of positive

control over JRand low levels of negative control over

JP (the latter is because stimulation of this module decreases flux through the phosphorylation branch of oxidative phosphorylation) A somewhat different flux-control pattern was obtained from two separate sets of experiments performed in the absence of CdCl2 and CuCl2 (K1 and K2; Table 2) It has previously been shown that the flux-control structure of the oxidative phosphorylation system is influenced by hormones [31] Therefore, the observed difference may have been caused by seasonal variations in the hormonal state of the animals Bimodular analysis (Fig 1B) showed that the CoQ-reducing module (comprising enzymes involved in substrate transport and SDH) exerted rela-tively low levels of control over JR and JP compared with the control exerted by the remaining system components Combining the results of three- and bimodular MCA made it possible to deduce the con-trol of fluxes exerted by the respiratory chain com-plexes downstream of the CoQ (i.e cytochrome bc1 and COX) (Fig 1C) The data obtained showed that cytochrome bc1 and COX together exerted stronger control over fluxes than the components of the respira-tory chain upstream of CoQ (Table 2)

0

20

40

60

80

100

120

0 20 40 60 80 100 120

A

0 20 40 60 80 100 120 140

0

20

40

60

80

100

120

B

Fig 5 Effect of Cd2+(A) and Cu2+(B) ions on succinate

dehydro-genase (SDH) activity SDH activity under control conditions (i.e.

without CdCl2 or CuCl2) was 50.8 ± 7.8 and 50.8 ± 8.3 nmol

DCPIPÆ min)1Æmg protein)1 (set to 100%) in the experiments

shown in (A) and (B), respectively Average of n = 2 independent

experiments ± SEM.

0 0.05 0.1 0.15

0 nmol·mg protein –1

5 nmol·mg protein–1

10 nmol·mg protein–1

B

* *

0 500 1000 1500

0 μ M

5 μ M

10 μ M

2',7'-dichlorofluorescin oxidation rate (RFU·min

**

**

A

*

Fig 6 Effect of Cd 2+ and Cu 2+ ions on H2O2production (A) and lipid peroxidation (B) Average of n = 3 (A) and n = 4 (B) indepen-dent experiments ± SEM *P < 0.05 and **P < 0.01 versus the condition with no CdCl 2 or CuCl 2 added, respectively RFU, relative fluorescence units; TBARS, thiobarbituric acid reactive substances.

Cd2+ions induced a redistribution in the control over

fluxes through the system (Table 2 and Fig 7) Cd2+

ions tended to increase the control over JRexerted by

the respiratory chain module and significantly decreased the control over JR exerted by the phosphorylation module A similar trend was observed for control over

JP; control by the respiratory chain module increased significantly, whereas control by the phosphorylating module decreased significantly In addition, negative control of JP by the proton leak module increased slightly, but significantly Analysis of the control exerted

by the respiratory chain components revealed that con-trol of JRand JPexerted by the CoQ-reducing module was not affected by Cd2+ions Meanwhile, control of

JR and JP by cytochrome bc1 and COX tended to increase in the presence of Cd2+ions

The effects of Cu2+ions on the control pattern fol-lowed the same trend as the effects of Cd2+ ions but were more pronounced (Table 2 and Fig 7) Three-modular MCA showed that in the presence of Cu2+ ions the respiratory chain module acquired almost complete control of JRand JP Combination of three-and bimodular analysis revealed that this resulted from

a dramatic increase in the control of fluxes by cyto-chrome bc1 and COX Moreover, in agreement with the observation that Cu2+ions are more potent stimu-lators of proton leak kinetics, we showed that the effect of Cu2+ions on the control of fluxes by the proton leak module was stronger than the effect of

Cd2+ions

Taken together, the changes in flux control distribu-tion were consistent with the results of modular kinetic analysis, which revealed that both Cd2+ and Cd2+ ions interfere with oxidative phosphorylation

function-Table 2 Effect of Cd 2+ and Cu 2+ ions on the metabolic control of fluxes Average of n = 4 independent experiments ± SEM K1 and K2, control experiments with no CdCl 2 or CuCl 2 added; C, flux control coefficient; J R , respiratory flux; J P , phosphorylation flux.

Module, i

CiJR

Module, i

CiJP

*P < 0.05 versus condition with no CdCl2or CuCl2added.

CoQ reduction bc1 + COX Phosphorylation Proton leak

0

0.2

0.4

0.6

0.8

1

J R

A

J P

0

0.2

0.4

0.6

0.8

1

B

Fig 7 Control distribution of the respiratory (A) and phosphorylation

flux (B) among the CoQ-reducing, cytochrome bc 1 + COX,

phosphor-ylation and proton leak modules Division of the system of oxidative

phosphorylation into modules is depicted schematically in Fig 1C.

Average of n = 4 data sets J R , respiratory flux; J P , phosphorylation

flux; K1 and K2, control experiments with no CdCl2or CuCl2added.

ing by inhibiting the respiratory chain downstream of

CoQ (i.e cytochrome bc1and⁄ or COX)

Discussion

Living systems are continuously exposed to low levels

of multi-component pollution, which may affect many

cellular processes simultaneously Although no cellular

process is hampered severely, there may be a

cumula-tive effect on the functioning of various metabolic

pathways and this may ultimately challenge cellular

metabolism as a whole Many pollutants, including the

heavy metal ions Cd2+ and Cu2+ are expected to

interfere with several enzymes at the same time because

of a rather nonspecific interaction with their functional

groups In this study, we used modular kinetic analysis

and MCA to elucidate the molecular mechanisms

underlying Cd2+and Cu2+ion-induced impairment of

the main aerobic energy production pathway in the

cell, i.e oxidative phosphorylation By subdividing

oxi-dative phosphorylation into modules in different ways

we were able to obtain a detailed picture of the effects

of Cd2+ and Cu2+ ions We showed that Cd2+ ions

interfere with oxidative phosphorylation solely through

their inhibitory effect on respiratory chain complexes

downstream of CoQ This resulted in lower respiratory

and phosphorylation fluxes and lower Dw values, as

well as an increase in flux control by the respiratory

chain module The overall effect of Cu2+ions on

oxi-dative phosphorylation functioning was similar to that

seen with Cd2+ions, however, it was caused not only

by inhibition of respiratory chain, but also by

stimula-tion of proton leak module activity at high Dw The

latter effect was caused, at least in part, by stimulation

of ROS production and the subsequent peroxidation

of membrane lipids

Modular kinetic analysis uses natural properties of

metabolism, i.e its organization into recognizable

func-tional units, simplifying the analysis and making cellular

complexity manageable [20,32,33] We demonstrated

how different modularization of the system of interest

and subsequent application of modular kinetic analysis

allows identification of molecular targets of Cd2+ and

Cu2+ ions In the first application of the analysis, we

conceptually divided oxidative phosphorylation into

three functional modules based on the production and

consumption of Dw (Fig 1A) We showed that Cd2+

and Cu2+ ions inhibited the respiratory chain module

(Figs 2B and 3B) but had no significant effect on the

phosphorylation module (Figs 2C and 3C)

Interest-ingly, Cu2+ions had a stronger ability to uncouple

oxi-dative phosphorylation than did Cd2+ions, as indicated

by stimulation of proton leak module kinetics at high

Dw values (Figs 2A and 3A) Our data are in agreement with earlier findings that Cu2+ ions can effectively increase membrane permeability, whereas Cd2+ions are much less effective; comparable swelling of isolated rat liver mitochondria was obtained at 5 lm Cu2+ and

40 lm Cd2+[34] It has been shown that the ability of

Cd2+ions to increase membrane permeability is depen-dent on inorganic phosphate transport and increases with increasing phosphate concentration in the medium [35] Therefore, under certain experimental conditions characterized by a high inorganic phosphate concentra-tion, low Cd2+ concentrations can induce uncoupling [35,36] However, at a low inorganic phosphate concen-tration (as in this study) a much higher Cd2+ concentra-tion is needed to stimulate membrane ion permeability and induce uncoupling [37]

Three-modular kinetic analysis has been used previ-ously to localize the sites of action of Cd2+ions in iso-lated potato tuber mitochondria respiring on succinate, and has shown that different concentrations of Cd2+ ions inhibited the respiratory chain module, had no significant effect on the phosphorylation module and stimulated proton leak kinetics at 3.5 lm free Cd2+ [22] The apparent contradiction between the latter finding and our observation that 5 lm CdCl2 has no significant effect on the proton leak kinetics is explained by the fact that we used a lower amount

of Cd2+ ions per mg of mitochondrial protein, and therefore uncoupling did not occur

After establishing that the respiratory chain module is the target of both metal ions, we conceptually divided the system into CoQ-reducing and CoQ-oxidizing mod-ules with XCoQas the connecting intermediate (Fig 1B) This approach revealed that 5 lm CdCl2or CuCl2 inter-fered with the respiratory chain complexes downstream

of CoQ and had no significant effect on the dicarboxy-late carrier and SDH (Fig 4) By contrast, earlier studies identified SDH as the target of both metal ions [5,9,10]

We investigated this and found that 5 lm CdCl2 or CuCl2did not affect SDH activity under our experimen-tal conditions (Fig 5), confirming that modular kinetic analysis yielded the correct results

In this study, we showed that at low concentrations,

Cd2+and Cu2+ions affected a rather limited number of components in the oxidative phosphorylation system, i.e the respiratory chain downstream of CoQ (i.e cyto-chrome bc1and COX) and the mitochondrial inner mem-brane This suggests that because these components are affected by a low concentration of Cd2+and Cu2+ions, they are the most sensitive and are therefore responsible for the early response of the system when it is chronically exposed to low levels of pollutants Support for this hypothesis comes from the observation that

mitochon-dria from liver, kidney and muscle of cadmium-treated

rats display a rapid early decrease in COX activity,

fol-lowed by partial restoration after 6 months of treatment,

and a progressive decrease in SDH activity [9]

The variation in the number of molecular targets of

Cd2+ or Cu2+ ions reported in the literature may be

explained by their differing sensitivity to these ions,

resulting in a greater number of affected system

com-ponents at higher ion concentrations For example, the

effects on mitochondrial ATPase were examined at

Cd2+ concentrations in the mm range (up to 10 mm)

[5] It is questionable whether such high concentrations

of Cd2+ions can be achieved even in a heavily

intoxi-cated cell Furthermore, it has been shown that

ATPase activity was induced by very low

concentra-tions of Cu2+ (4–6 nmolÆmg protein)1) However, the

same ion concentration induced maximal uncoupling

under the experimental conditions used by Hwang

et al [34], and the effect might therefore be explained

by uncoupling In contrast to these findings, our data

(Figs 2C and 3C) did not reveal any noticeable effects

of 5 lm CdCl2 and 5 lm CuCl2 on the components of

the ADP phosphorylation machinery

We hypothesized that the different effect of Cd2+and

Cu2+ions on the proton leak kinetics observed in our

study may be determined by the ability of Cu2+ions to

stimulate ROS production, which in turn may lead to

lipid peroxidation and membrane damage We showed

that 5 lm CuCl2, but not 5 lm CdCl2, stimulated the

oxidation of DCF significantly, indicating increased

H2O2 formation in the mitochondrial matrix Because

oxygen favors reduction by one electron at a time, the

primary ROS that is formed by the action of Cu2+ions

must be a superoxide anion radical The latter may

reduce Cu2+to Cu+, leading to formation of the

hydro-xyl radical via the Haber–Weiss cycle [38] In turn, the

hydroxyl radical is a potent inductor of lipid

peroxida-tion In support of this, we showed that Cu2+ions, but

not Cd2+, cause accumulation of TBARS, suggesting

that Cu2+may increase membrane permeability by

stim-ulating lipid peroxidation It has also been shown that,

in intact hepatocytes, Cu2+ions are much more potent

inductors of lipid peroxidation than Cd2+ions [14]

In addition to the evaluation of Cd2+and Cu2+

ion-induced changes in the kinetics of the individual

oxidative phosphorylation components, we were also

interested in how these changes affect the control

struc-ture of the system MCA was designed to deal with the

effects of a small disturbance in enzyme activity on

sys-tem fluxes and is a useful method with which to analyze

and diagnose cell sickliness caused by agents that

simul-taneously affect many enzymes In this study, we

assessed metabolic control of the respiratory (JR) and

phosphorylation (JP) fluxes by the respiratory chain, phosphorylation and proton leak modules, as well as control of the respiratory flux by CoQ-reducing and CoQ-oxidizing modules, and determined how Cd2+and

Cu2+ions affected this control In agreement with pre-viously published data [21,24], we found that control of both fluxes is mainly shared between the respiratory chain and phosphorylation modules, with only slight control being exerted by the proton leak module (posi-tive in the case of JR and negative in the case of JP) (Table 2 and Fig 7) Five micromolar CdCl2and CuCl2 caused an increase in the control of JR and JP by the respiratory chain module, with cytochrome bc1 and COX being the main contributors This is in agreement with our observation that both metal ions interfere with the respiratory chain components downstream of CoQ The effect of Cu2+ ions was stronger than that of

Cd2+ Furthermore, Cu2+ions increased negative con-trol of JPby the proton leak module more than Cd2+ ions, because the former had a stronger effect on the permeability of the inner membrane to ions This obser-vation illustrates how interfering with a system compo-nent that makes a relatively low contribution to the control of system fluxes (i.e proton leak) may com-promise the system via control of other vital system properties (e.g stronger membrane damage by Cu2+ ions may contribute to inhibition of cytochrome bc1 and COX, which are membrane proteins, leading to increased flux control by these enzymes)

Flux control is a property of the whole system rather than of individual components of that system (i.e enzymes, modules), as illustrated by the summa-tion theorem for flux control coefficients [17,18] As a result, individual flux control coefficients for each sys-tem component cannot change independently Conse-quently, we see that an increase in flux control by the respiratory chain and proton leak modules results in decreased control of both fluxes by the phosphory-lation module, although this module is not affected directly by Cd2+and Cu2+ions

Taken together, we have demonstrated how modular kinetic analysis could be gradually applied to identify the sites of action of external effectors such as Cd2+ and Cu2+ions in a multienzyme system like oxidative phosphorylation Although we found that Cd2+ ions affect system behavior by acting on a single target, whereas Cu2+ interferes with two, there is no doubt that this method is valuable, especially when assessing multisite effects of toxic substances on complex meta-bolic systems The next step in this type of analysis may be to elucidate the combined effects of several effectors (e.g a mixture of Cd2+ and Cu2+ions) with individual known sites of action Repeated modular

kinetic analysis using different modularization

approaches might reveal the causes and consequences

of their competition for binding to the same molecular

component or synergism of their action

Materials and methods

Materials

Rotenone, myxothiazol, oligomycin, creatine

phospho-kinase, CoQ1, 2¢,7¢-DCF, 2¢,7¢-dichlorofluorescin diacetate

and 2-thiobarbituric acid (TBA) were from Sigma

(Sigma-Aldrich, St Louis, MO, USA)

Isolation of mitochondria

The handling of the animals conformed to the rules defined

by the European Convention for the Protection of

Verte-brate Animals Used for Experimental and Other Scientific

Purposes (License No 0006 of State Veterinary Service for

working with laboratory animals) Mitochondria were

iso-lated from the livers of male Wistar rats using a standard

differential centrifugation procedure as described previously

[24], using 250 mm sucrose, 10 mm Tris, 3 mm EGTA and

2 mgÆmL)1 BSA (pH 7.7) as the isolation medium

Mito-chondria were suspended in the medium containing 250 mm

sucrose and 5 mm Tris (pH 7.3) Protein was estimated

according to Bradford [39] using BSA as the standard

Measurement of respiration and membrane

potential

Prior to each measurement, mitochondria (1 mgÆmL)1

mito-chondrial protein) were incubated for 3 min with or without

5 lm CdCl2or 5 lm CuCl2in the assay medium containing

110 mm KCl, 20 mm Tris, 5 mm KH2PO4, 50 mm creatine,

an excess of creatine kinase and 1 mm MgCl2, pH 7.2

Mito-chondrial respiration rate and Dw were measured

simulta-neously at 37C in a closed, stirred and thermostated glass

vessel equipped with a Clark-type oxygen electrode and

TPP+-sensitive electrode, as described previously [40] We

used 5 mm succinate (+ 2 lm rotenone) as the respiratory

substrate ATP (1 mm) was added to initiate state 3

respira-tion Data were processed using the chart program supplied

with MacLab (AD Instruments, Chalgrove, UK)

Determination of CoQ reduction level

CoQ reduction level was determined in mitochondria

(1 mgÆmL)1mitochondrial protein) incubated under the

con-ditions used in Dw measurements in a thermostated (37C)

vessel equipped with platinum and oxygen electrodes, by

polarographically measuring the redox state of exogenous

CoQ1(2 lm) [41] To calibrate the platinum electrode traces,

samples were taken from incubations of mitochondria in standard assay medium without further additions (state 1) and mitochondria were incubated with substrate (5 mm suc-cinate + 2 lm rotenone) One milliliter of sample was quenched with 1 mL of 0.2 m HClO4 in methanol (0C) CoQ was extracted with 3 mL of petroleum ether (40–60C) and determined by HPLC, as described previously [42] Data were processed using the chart and powerchrom programs supplied with MacLab

Determination of SDH activity SDH activity was determined spectrophotometrically at

600 nm from the rate of reduction of DCPIP in the presence

of CoQ1, as described in Ragan et al [43] Briefly, intact mitochondria (1 mgÆmL)1mitochondrial protein) were incu-bated for 3 min with 5 lm CdCl2or 5 lm CuCl2under the experimental conditions used in the respiration and Dw mea-surements Medium was then collected and mitochondrial membranes were ruptured by four freeze–thaw cycles Mea-surement of SDH activity was carried out in the presence of

1 mgÆmL)1CoQ1and 100 lm DCPIP SDH activity was cal-culated using a molar extinction coefficient of 21 mm)1Æcm)1 The dependence of SDH activity on concentrations exceeding

5 lm Cd2+ or Cu2+ ions was determined using a slightly modified protocol: varying amounts of CdCl2or CuCl2were added directly to four-times freeze–thawed mitochondrial suspension (i.e without 3 min preincubation with intact mitochondria) in an assay medium containing 110 mm KCl,

20 mm Tris, 5 mm KH2PO4, 2.24 mm MgCl2, pH 7.2, and SDH activity was determined as described above

Measurement of H2O2production Oxidation of DCF was used as an indicator of H2O2 produc-tion in the mitochondrial matrix Isolated mitochondria were incubated with 5 lm 2¢,7¢-dichlorofluorescin diacetate for

30 min Next, mitochondria were resuspended in medium containing 250 mm sucrose, 5 mm Tris (pH 7.3), centrifuged

at 7300 g for 10 min at 4C, and the supernatant discarded

To assess H2O2production, mitochondria (1 mgÆmL)1 mito-chondrial protein) were incubated at 37C in assay medium (as in SDH activity determination) supplemented with 5 mm succinate and different concentrations of CdCl2and CuCl2 (0, 5 and 10 lm) The oxidation of DCF was measured spectrofluorimetrically (kex= 485 nm, kem= 535 nm) for

3 min using GENios Pro reader (Tecan, Ma¨nnedorf, Swit-zerland) and the DCF oxidation rate was expressed as rela-tive fluorescence unitsÆ min)1Æmg protein)1 To correct for changes in mitochondria-derived background fluorescence, the same measurement was carried out with mitochondria, which were not loaded with 2¢,7¢-dichlorofluorescin diacetate and the rate obtained was subtracted from the DCF oxida-tion rate