Báo cáo Y học: Inhibition of the SERCA Ca21 pumps by curcumin Curcumin putatively stabilizes the interaction between the nucleotide-binding and phosphorylation domains in the absence of ATP pot

For the skeletal muscle isoform of the Ca2þ pump SERCA1, the inhibition of curcumin is noncompetitive with respect to Ca2þ, and competitive with respect to ATP at high curcumin concentra

Inhibition of the SERCA Ca21 pumps by curcumin

Curcumin putatively stabilizes the interaction between the nucleotide-binding and phosphorylation domains in the absence of ATP

Jonathan G Bilmen1, Shahla Zafar Khan1, Masood-ul-Hassan Javed2and Francesco Michelangeli1

1 School of Biosciences, University of Birmingham, Edgbaston, Birmingham, UK;2Shifa College of Medicine, Islamabad, Pakistan.

Curcumin is a compound derived from the spice, tumeric It

is a potent inhibitor of the SERCA Ca2þ pumps (all

isoforms), inhibiting Ca2þ-dependent ATPase activity with

IC50values of between 7 and 15 mM It also inhibits

ATP-dependent Ca2þ-uptake in a variety of microsomal

membranes, although for cerebellar and platelet

micro-somes, a stimulation in Ca2þ uptake is observed at low

curcumin concentrations (, 10 mM) For the skeletal muscle

isoform of the Ca2þ pump (SERCA1), the inhibition of

curcumin is noncompetitive with respect to Ca2þ, and

competitive with respect to ATP at high curcumin

concentrations (< 10 – 25 mM) This was confirmed by

ATP binding studies that showed inhibition in the presence

of curcumin: ATP-dependent phosphorylation was also

reduced Experiments with fluorescein 50-isothiocyanate

(FITC)-labelled ATPase also suggest that curcumin stabilizes the E1 conformational state The fact that FITC labels the nucleotide binding site of the ATPase (precluding ATP from binding), and the fact that curcumin affects FITC fluorescence indicate that curcumin must be binding to another site within the ATPase that induces a conformational change to prevent ATP from binding This observation is interpreted, with the aid of recent structural information, as curcumin stabilizing the interaction between the nucleotide-binding and phosphorylation domains, precluding ATP binding

Keywords: SERCA; ATP binding; curcumin; phosphoryl-ation; fluorescence

Tumeric is extensively used as a spice in Asian cooking and

as a colouring agent in both the food and cosmetic industries

[1] Curcumin (diferuoylmethane or

1,7-bis(4-hydroxy-3-methoxyphenol)-1,6-heptadiene-3,5-dione) is a compound

found in tumeric that gives it its distinctive yellow colour

[2] Recently it has been shown that curcumin has

anti-carcinogenic effects [3] that may be linked to its antioxidant

properties [4] Studies have shown that curcumin can affect

a number of cellular processes including: activation of

apoptosis in Jurkat T-cells [5], inhibition of platelet

aggregation [6,7] and inhibition of inflammatory cytokine

production in macrophages [8] Curcumin has also been

shown to affect the activity of a number of key enzymes

such as cyclooxygenase [9], protein kinase C [10], protein

tyrosine kinases [11] and a Ca2þ-dependent endonuclease

[12] Many of these processes/enzymes are also known to be

regulated by Ca2þ

Cytosolic free Ca2þ concentration ([Ca2þ]cyt) is tightly

controlled, due its importance in the regulation of many

cellular processes The sarco/endoplasmic reticulum Ca2þ

ATPase (SERCA) is one of the major mechanisms by which

the low levels of [Ca2þ]cytare maintained within cells Three isoforms of the SERCA family of Ca2þpumps have so far been identified [13,14] and these are expressed in a tissue-specific manner [14] SERCA1 is found predominantly in fast-twitch skeletal muscle while SERCA2a is found within cardiac and slow-twitch muscle The splice variant form of SERCA2 (SERCA2b), which has an extended C-terminus is found in most nonmuscle cells and is particularly abundant

in neuronal tissues SERCA3 is less widely distributed in nonmuscle tissues but is relatively abundant in macro-phages, platelets and large intestines

The crystal structure of the Ca2þbound form of the SR

Ca2þ-ATPase (SERCA1) was recently resolved and shown

to contain three domains within the cytoplasmic head region [15] These are: the nucleotide binding domain, which binds ATP; the phosphorylation domain, which can be phos-phorylated on Asp351; and the actuator, which may be involved in anchoring the other two domains together during phosphoryl transfer [15] These domains are attached to the membrane by 10 transmembrane helices, containing the two

Ca2þbinding sites that sit side-by-side [15,16]

Using inhibitors to study the ATPase has proved invaluable in helping to elucidate mechanistic steps within the Ca2þtransport process [16 – 18] These steps and their associated conformational changes now need to be placed in context with changes within the tertiary structure of the

Ca2þ-ATPase

In this study, we show that curcumin is a potent inhibitor

of SERCA Ca2þpumps that affects a number of steps within its mechanism We try to rationalize these effects in terms of domain interactions of the known structure

Correspondence to F Michelangeli, School of Biosciences, University

of Birmingham, Edgbaston, Birmingham, UK.

Fax: þ 44 121 414 5925, Tel.: þ 44 121 414 5398,

E-mail: F.Michelangeli@bham.ac.uk

(Received 11 July 2001, revised 5 October 2001, accepted 11 October

2001)

Abbreviations: FITC, fluorescein 50-isothiocyanate; SR, sarcoplasmic

reticulum; SERCA, sarco/endoplasmic reticulum Ca 2þ ATPase.

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

Curcumin was purchased from Sigma [g32P]ATP was

obtained from Amersham Magnesium green was purchased

from Molecular Probes All other reagents were of

analytical grade

Membrane and protein preparation

Sarcoplasmic reticulum (SR) and the purified Ca2þATPase

were prepared from rabbit skeletal muscle as described by

Michelangeli & Munkonge [19] Porcine cerebellar and

cardiac microsomes were prepared as described by Sayers

et al [20] Human platelet microsomes were prepared based

on the method as described by Le Peuch et al [21] and

resuspended in a buffer containing 5 mM Hepes, 0.32 M

sucrose, 0.1 mM benzamidine, 0.1 mM

phenylmethanesul-fonyl fluoride and 10 mM leupeptin Curcumin was

dissolved in ethanol to give a stock solution of 10 mM

Ca21-ATPase activity

Ca2þ-ATPase activity determination in microsomes was

performed using the phosphate liberation assay as described

by Longland et al [22] Briefly, microsomal extracts were

resuspended in 1 mL of buffer containing 45 mM Hepes/

KOH (pH 7.0), 6 mMMgCl2, 2 mMNaN3, 0.25Msucrose,

12.5 mg·mL21 A23187 ionophore, and EGTA with CaCl2

added to give a free [Ca2þ] of 1 mM Assays were

pre-incubated at 37 8C for 10 min prior to activation with ATP

(final concentration 6 mM) The reaction was stopped by

addition of 0.25 mL 6.5% (w/v) trichloroacetic acid The

assays were put on ice for 10 min prior to centrifugation for

10 min at 20 000 g: 0.5 mL of the supernatent was added to

1.5 mL buffer containing 11.25% (v/v) acetic acid, 0.25%

(w/v) copper sulfate, and 0.2 M sodium acetate

Two-hundred and fifty microliters of 5% (w/v) ammonium

molybdate was then added and mixed thoroughly and

0.25 mL of ELAN solution was added [2% (w/v)

p-methylaminophenol sulfate and 5% (w/v) sodium sulfite]

The colour intensity was measured after 10 min at 870 nm

absorbance and related to a calibration curve of colour

intensity vs known amounts of phosphate, previously

treated as above Controls were performed in the presence of

ethanol, which at maximal curcumin concentrations was

equal to 0.5% (v/v) and had no effect on the ATPase activity

In experiments where the effects of curcumin were

investigated on Ca2þATPase activity as a function of [Ca2þ]

and [ATP], these were carried out on purified SR

Ca2þ-ATPase using a coupled enzyme assay as previously

described in [16,19] in a buffer containing 40 mM Hepes/

KOH, 5 mM MgSO4, 0.42 mM phosphoenolpyruvate,

0.15 mM NADH, 7.5 U pyruvate kinase, 18 U lactate

dehydrogenase, 1.01 mM EGTA, pH 7.2 Free Ca2þ

con-centrations were calculated based on the method and

binding affinities described by Gould et al [23]

Ca21-uptake measurements

The effects of curcumin on Ca2þuptake into microsomes or

SR was measured as described by Michelangeli [24]

Briefly, microsomes were added to a stirred cuvette

containing 2 mL of 40 mM Tris/phosphate, 100 mM KCl

at pH 7.2 in the presence of 1 mMFluo-3 (except for platelet microsomes where, which due to their high Ca2þcontent, the lower affinity magnesium green indicator was used, 0.625 mM), 10 mg·mL21 creatine kinase and 10 mM

phosphocreatine Ca2þuptake was initiated by the addition

of 1.5 mMMgATP Fluorescence intensity was measured at

506 nm excitation/526 nm emission for Fluo-3 and 506 nm excitation/531 nm emission for Magnesium green The results were calculated from fluorescence intensities using the following equation:

½Ca2þŠ ¼ Kd£ ðF 2 FminÞ/ðFmax2 FÞ Where F is the fluorescence level and the addition of 1.25 mMEGTA and 1.5 mMCaCl2(Fluo-3) or 2 mMCaCl2 (Magnesium green) determines Fminand Fmax, respectively The Kd values used for Fluo-3 and Magnesium green are

900 nM[24] and 6 mM, respectively

Effects of curcumin on fluorescein 50-isothiocyanate FITC-labelled Ca21-ATPase

ATPase from SR was labelled with FITC according to the method described by Michelangeli et al [17], with minor modifications to monitor the E2 to E1 transition The SR ATPase was added in equal volume to the starting buffer (1 mMKCl, 0.25Msucrose and 50 mMpotassium phosphate

pH 8.0) FITC in dimethylformamide was then added at a molar ratio of FITC/ATPase, 0.9 : 1 The reaction was incubated for 1 h at 25 8C and stopped by 0.25 mL of stopping buffer (0.2M sucrose, 50 mM Tris/HCl pH 7.0), incubated for 30 min at 30 8C prior to being placed on ice until required Measurements were undertaken in a buffer containing 50 mM Tris, 50 mMmaleate, 5 mM MgSO4and

100 mM KCl at pH 6.0 Fluorescence was measured on a PerkinElmer LS50B spectrofluorimeter at 25 8C (excitation

495 nm, emission 525 nm) EGTA (100 mM) and then Ca2þ (400 mM) or vanadate (100 mM) were added to measure changes in fluorescence

ATP binding to Ca21-ATPase ATP binding to purified Ca2þATPase was also measured using radiolabelled ATP as described by Champeil et al [26] Briefly, 0.3 mg·mL21of purified ATPase was added to

a buffer containing 150 mMTes/Tris, 2 mMMg2þand 2 mM

EGTA (pH 7.0), to which was added ATP doped with [g32P]ATP to give a final concentration of 20 mM(specific activity 10 Ci·mol21) One milliliter of solution was then rapidly filtered through a 0.45-mm Millipore HA filter, and placed in scintillant for counting This was carried out in the absence and presence of curcumin, and controls to estimate nonspecific binding of ATP to the filter were performed as described previously [16]

Phosphorylation studies Phosphorylation of the ATPase by [g-32P]ATP was performed at 25 8C as described by Michelangeli et al [17] Briefly, SR Ca2þ-ATPase was diluted to 75 mg·mL21

in 20 m Hepes/Tris (pH 7.2) containing 100 m KCl,

5 mMMgSO4,1 mMCaCl2in a total volume of 1 mL The

reaction was initiated by addition of ATP doped with

[g-32P]ATP (specific activity, either 10 or 100 Ci·mol21)

and stopped after 15 s by addition of ice-cold 40% (w/v)

trichloroacetic acid The assay was placed on ice for

30 min subsequent to the addition of BSA (final

concentration 1 mg·mL21) The protein was separated

from solution by filtration through Whatman GF/C filters

The filters were washed with 12% (w/v) trichloroacetic

acid/0.2 M H3PO4, and left to dry, then placed in

scintillant and counted

Membrane permeability studies

To measure the effects of curcumin on membrane permeability, an assay was carried out as described by Longland et al [22] Ten milligrams (12.5 mmol) of egg phosphatidylcholine was dissolved in 0.25 mL of chloro-form and evaporated to dryness under a stream of N2 The phospholipid film was dispersed in 400 mL of 40 mM

Hepes/KOH (pH 7.2), 100 mM KCl buffer containing

100 mM calcein To this was then added 35 mL of 10.5% (w/v) potassium cholate in 40 mM Hepes/KOH (pH 7.2)

Fig 1 Inhibition of Ca 21 -ATPase activity by curcumin Graphs representing both Ca2þ-dependent ATPase activities (X) and Ca2þuptake (B) in

a variety of microsomes are presented: (A) and (B), skeletal muscle SR; (C) and (D), cardiac SR; (E) and (F), platelet microsomes; and (G) and (H), cerebellar microsomes All experiments were performed at 37 8C, pH 7.2 Each data point represents the mean ^ SD of three determinations IC 50 for inhibition compared to controls (i.e in the absence of curcumin) are as follows: (A) 15.0 ^ 0.8 m M (B) 5.0 ^ 0.3 m M (C) 7.4 ^ 0.4 m M (D) 20.3 ^ 2.2 m (E) 8.8 ^ 1.3 m (F) 34.3 ^ 1.5 m (G) 13.7 ^ 4.2 m (H) 50 ^ 2 m curcumin.

buffer) The suspension was sonicated to clarity at room

temperature Excess detergent was then removed by passing

the suspension through a pre-equilibrated Sephadex G-25

column (with 40 mM Hepes/KOH (pH 7.2), 100 mM KCl

buffer at room temperature), followed by 200 mL of Hepes

buffer prior to centrifugation at 200 g for 20 s into a clean

conical centrifuge tube The resulting column eluate was

passed through a second column as before, providing a

suspension of reconstituted lipid vesicles The dye filled

vesicles were diluted in 1.8 mL of Hepes buffer and

fluorescence intensity was measured at excitation and

emission wavelengths of 490 nm and 520 nm, respectively

Ten microliters of 3 mMCoCl2was then added to the vesicle

suspension and the rate of fluorescence quenching was

monitored in the absence or presence of various

concentrations of curcumin

Fluorescence studies

Experiments to investigate the fluorescence of curcumin

bound to the ATPase were performed in a buffer containing

20 mM Mes, 20 mM Mops, 80 mM KCl, 1 mM EGTA

(pH 6.0 or 7.0) In Ca2þbinding experiments to cerebellar

microsomes, 12.5 mg·mL21Ca2þionophore (A23187) was

also added All experiments were performed with 1 mM

curcumin unless otherwise stated Fluorescence was

measured at 25 8C (excitation 411 nm, emission 500 nm)

Ca2þ and EGTA were added to measure changes in

fluorescence at appropriate free Ca2þconcentrations, using

the constants given previously [23]

Ca2þ binding to the Ca2þ-ATPase was measured by

monitoring the change in tryptophan fluorescence [17,27]

Purified ATPase was used at 2 mM in a buffer containing

20 mM Mes, 20 mM Mops, 80 mM KCl, 1 mM EGTA

(pH 6.0 or 7.0) Ca2þ binding was measured as percent

increase in initial fluorescence, over a range of free Ca2þ

concentrations as described in [27] Fluorescence was

monitored at 25 8C (excitation 295 nm, emission

340 nm)

R E S U L T S

Figure 1 shows the Ca2þ-dependent ATPase activity and

Ca2þ uptake in microsomes from various tissue extracts

The tissues were selected for their differential expression of

SERCA subtypes: Skeletal SR membranes (Fig 1A,B)

express predominantly SERCA 1; cardiac SR (Figs 1C,D)

express predominantly SERCA 2a; cerebellar microsomes

(Fig 1E,F) express mostly SERCA 2b and platelet

microsomes (Fig 1G,H) express a mixture of SERCA 2b

and SERCA 3 The activities were measured at various

curcumin concentrations, using the phosphate liberation

assay in the presence of A23187 ionophore, and so were fully uncoupled The Ca2þ ATPase activity in all of the microsomes showed a high degree of inhibition, with half-maximal inhibition (IC50) values ranging from 7.4 ^ 0.4 mM (platelets) to 15.0 ^ 0.8 mM (SR), with almost complete inhibition occurring at about 50 mM in all membranes It was found that the IC50values for curcumin inhibition of SR and the purified Ca2þ-ATPase varied from 7

to 17 mM dependent upon the preparation and conditions used In addition, curcumin was tested to see if it was reversible with respects to inhibition of the ATPase This was performed by initially preincubating the ATPase with

60 mM curcumin for 10 min followed by dilution in assay buffer to 0.06 mM, where the activity was found to be similar

to controls

From Fig 1B,D,F,H where the effects of curcumin on

Ca2þ uptake were monitored, the IC50 values were calculated (compared to control) Skeletal muscle SR appears to be most sensitive to curcumin inhibition (IC50¼ 5 ^ 0.3 mM), whilst cerebellar microsomes was least affected (IC50¼ 50 ^ 1.7 mM) Cardiac and platelet microsomes were inhibited at intermediate concentrations (IC50¼ 20 ^ 2.2 mMand 34 ^ 1.5 mM, respectively) ATP-dependent Ca2þuptake was stimulated in cerebellar microsomes, and to a lesser extent platelets, upon addition

of low concentrations of curcumin Maximal stimulation in platelets occurred at approximately 5 mMcurcumin, with an increase in uptake of 12% (P , 0.01, students t-test, when compared with control) Maximal stimulation in cerebellar microsomes occurred at approximately 10 mMwith a 76% increase in stimulation (P , 0.001)

To measure the effects of curcumin on membrane permeability to cations, reconstituted liposomes were loaded with calcein, a fluorescent dye, and exposed to

Co2þ The rate of quenching was then monitored (Fig 2) After addition of curcumin, the rate of quenching was increased, showing that curcumin permeabilizes the membrane to metal ions The increase in membrane permeability was seen to be dependent on the amount of curcumin that is added This result indicates that the stimulation of Ca2þ uptake in some of the microsome preparations is unlikely to be due to a decrease in ion leakage through the phospholipid membrane

Figure 3 illustrates the dependence of purified Ca2þ -ATPase activity on both Ca2þ(Fig 3A) and ATP (Fig 3B)

in the absence and presence of curcumin, using the coupled enzyme assay In Fig 3A, the half-maximal activation of the ATPase by Ca2þwas measured in the absence of and in the presence of 10 and 25 mM curcumin The EC50 was found to change insignificantly, from 0.52 ^ 0.12 mM to 0.63 ^ 0.30 mM Ca2þ, although the maximal activity decreased from 18.95 IU·mg21 (control) to 4.33 IU·mg21

Fig 2 Curcumin increases membrane

permeability The traces represent experiments of

Co2þquenching calcein trapped within liposomes.

The drop in fluorescence intensity represents

quenching of the fluorescent dye by Co2þions.

Upon addition of curcumin, the rate of quenching

is substantially increased and dependent upon the

concentration of curcumin The traces are

representative of three or more experiments.

(25 mM drug) The inset on Fig 3A shows a double reciprocal (Lineweaver – Burk) plot for activity against [Ca2þ]free As can be seen from the plots, the lines converge

at a single point on the 1/[Ca2þ] axis, indicating noncompetitive inhibition with respect to Ca2þ

Figure 3B shows a complex stimulation of the Ca2þ -ATPase with increasing concentrations of ATP [28] This data could be fitted to a bi-Michaelis – Menton equation assuming two sites, designated the high affinity catalytic site and the lower affinity regulatory site [28,29] The kinetic parameters that describe the data are given in Table 1 A range of values for the kinetic parameters could be used to define the data profiles These fits suggest that the Vmax values for both the catalytic and regulatory sites are reduced

by curcumin, while at higher concentrations of curcumin (25 mM) the Km for the catalytic site is also possibly increased

In order to further assess the possibility of curcumin affecting the interaction of ATP binding to the ATPase, this was directly measured using [32P]ATP in the absence of

Ca2þ(Fig 4A) The data showed that the amount of ATP bound to the ATPase was reduced by curcumin The binding inhibition had a apparent Ki(IC50) of about 9 mM Reversing the order of addition to the ATPase (i.e curcumin then ATP) had a similar effect on the extent of ATP binding, i.e in both cases, the amount of ATP binding to the ATPase was significantly decreased

Experiments to assess the effects of curcumin on the ATP-dependent phosphorylation of the ATPase, were also undertaken Figure 4B shows that ATP-dependent phos-phorylation was inhibited by the presence of 50 mM

curcumin In the absence of curcumin maximal phosphoryl-ation occurred at around 10 mMATP where 1.7 ^ 0.3 nmol E-P per mg ATPase was phosphorylated At 50 mM

curcumin, the maximum level of phosphorylation was reduced by about 80% to 0.40 ^ 0.1 nmol E-P per mg ATPase

Figure 5A shows the traces of experiments obtained with FITC-labelled Ca2þATPase in SR at pH 6, upon addition of

Ca2þ and vanadate These changes have been used to monitor the transition between the E2 and the E1 step [28,30] In the absence of curcumin, a 9.5% change in fluorescence is observed upon addition of Ca2þ This is believed to occur as at pH 6 the ATPase is essentially all in

an E2 conformation (high fluorescence state) while the

Fig 3 The effects of Ca21and ATP dependence of Ca21ATPase

activity by curcumin (A) Ca2þ-ATPase activity was measured as a

function of [Ca 2þ ] free in the absence (B) and presence of 10 m M (W) and

25 m M (X) curcumin (B) shows ATPase activity as a function of [ATP]

in the absence (B) and presence of 10 m M (W) or 25 m M (X) curcumin.

The kinetic parameters are given in the text or Table 1 The experiments

performed at 37 8C pH 7.2 Each data point is the mean ^ SD of three

to five determinations.

Table 1 Kinetic parameters of Purified Ca21ATPase activity as a function of ATP concentration in the presence of curcumin Note: these values are calculated from the best fits to the data in Fig 3B The numbers in brackets correspond to the range of values for each parameter which could also give adequate fits to the experimental data (i.e where chi2for the fits are # 0.7).

Curcumin

concentration

(m M )

Catalytic K m (m M )

Catalytic V max (IU·mg21)

Regulatory K m (m M )

Regulatory V max (IU·mg21)

(2.7 – 6.6) (6.3 – 9.3) (3.8 – 1.0) (13.4 – 14.2)

(1.9 – 3.6) (1.3 – 2.5) (0.24 – 0.42) (11.9 – 12.2)

(4.9 – 8.0) (1.8 – 2.4) (0.23 – 0.41) (4.7 – 5.2)

addition of Ca2þ shifts it to an E1 conformation (low

fluorescence state) [28,30] Vanadate, on the other hand,

would shift the ATPase towards E2 and therefore increase

the FITC-ATPase fluorescence if it were in an E1

conformational state [30] The fluorescence change due to

addition of Ca2þwas shown to decrease upon preincubation

with curcumin (5 mM), as well as slowing down the rate of

this transition In order to establish that the decrease in

Ca2þ-induced fluorescence change was due to a shift in the

E1 to E2 step towards E1, the experiments were repeated

with vanadate At pH 6, vanadate induces little change in the

FITC-ATPase fluorescence, indicative of it already being in

an E2 state However, in the presence of curcumin (5 mM),

vanadate caused a rise in fluorescence suggesting that

curcumin had shifted the equilibrium towards E1 Figure 5B

shows the concentration effects of curcumin on both the

Ca2þ-induced fluorescence decrease and vanadate-induced

fluorescence increase of FITC-labelled SR The data show

that the concentration of curcumin inducing half-maximal

fluorescence changes in both cases were similar (< 5 –

6 mM)

It was found that curcumin strongly fluoresces in the presence of ATPase (excitation 411 nm, emission 500 nm), but little in its absence This observation was used to assess curcumin binding to the Ca2þ-ATPase Titrations were performed by addition of either curcumin or ATPase and interpreted using Langmuir isotherms

Binding can be described by the following equation:

½EŠ þ ½LŠ $ ½ELŠ

Where [E] and [L] are the concentrations of free sites and ligands, respectively, and [EL] is the concentration of bound ligand The total concentration of sites can be expressed as N[E]0, the product of total protein concentration ([E]0) and the number of binding sites per protein molecule (N ) The concentration of bound ligand [EL] can be derived in terms

of the dissociation constant Kd, defined as;

Kd¼ ½EŠ·½LŠ/½ELŠ ¼ ½N·E02 ELŠ·½L02 ELŠ/½ELŠ where [L0] and [E0] are the total ligand and protein concentrations This equation can then be rearranged to give

Fig 4 Effects of curcumin on ATP binding and ATP-dependent

phosphorylation (A) Displacement of [32P]ATP (20 m M ) bound to the

Ca2þ-ATPase by curcumin, measured at pH 7.2, 25 8C Values

represent the mean ^ SD of eight determinations (B) Phosphorylation

of SR Ca2þ-ATPase by [32gP]ATP (0 – 100 m M ) in the absence (B) and

presence of 50 m M curcumin (W), measured at pH 7.2, 25 8C Values

represent the mean ^ SD of three to five determinations.

Fig 5 The measurement of E2 – E1 conformational change using FITC-labelled Ca 21 ATPase in SR (A) Effects of curcumin on the fluorescence decrease in FITC-Ca2þATPase induced by either 400 m M

Ca2þ or 100 m M vanadate, initially preincubated in the presence or absence of 5 m M curcumin at pH 6 (B) The effects of curcumin concentration on the fluorescence changes induced by either Ca2þ(B)

or vanadate (W) The experiments were performed at 25 8C and each data point is the mean ^ SD of three determinations.

the following quadratic equation:

½ELŠ22 ½ELŠ·ðKdþ ½NEŠ0þ ½LŠ0Þ þ N½EŠ0½LŠ0¼ 0

Using the formula for the solution of a quadratic equation,

the concentration of bound ligand [EL] is then given by:

½ElŠ ¼ ðA 2 ½A22 4N½EŠ0½LŠ0Š0:5Þ/2

where A ¼ Kdþ N[E]0þ [L]0, N is the number of binding

sites per protein molecule, [E]0is the protein concentration,

[L]0 is the total concentration of ligand, and Kd is the

dissociation constant for binding

Using this equation, curves can be fitted to the

fluorescence data assuming a Kd¼ 0.8 mM and a

stoichi-ometry of 1 (i.e 1 curcumin per ATPase), either by varying

curcumin concentration and keeping ATPase constant

(Fig 6A) or varying ATPase concentration, keeping the

curcumin constant (Fig 6B) The data in Fig 6A could also

be fitted assuming two binding sites for curcumin on the

Ca2þ-ATPase with differing affinities (a high affinity site with a Kd of 0.55 mM and lower affinity site with a Kd of

10 mM)

In addition to enhancement of fluorescence in the presence of ATPase, curcumin bound ATPase also decreased its fluorescence intensity by up to 25% when Ca2þ was added (Fig 7A) To characterize this, the [Ca2þ]free was varied in the presence of ATPase and 1 mMcurcumin at pH 6 and 7 and the fluorescence decrease measured (Fig 7B) In order to assess whether this change is directly monitoring the Ca2þ binding steps, additional experiments were performed to monitor tryptophan fluorescence of the ATPase

as a function of [Ca2þ]freeas this fluorescence change has also

Fig 6 Fluorescence of curcumin bound to the Ca 21 ATPase (A)

Fluorescence change upon addition of curcumin (0 – 8 m M ) to the

Ca2þ-ATPase (2 m M ) ATPase, pH 7.0, 25 8C (B) Addition Ca2þ

ATPase into 1 m M curcumin at 25 8C, pH 7.0 Fluoresence intensity

was measured at 500 nm, and excited at 411 nm All points represent

mean ^ SD of three determinations The curves were fitted assuming a

single binding site for curcumin on the ATPase, with a K d of 0.8 m M

Equally good fits to the data could also be achieved assuming two

curcumin binding sites with K values of 0.55 m and 10 m

Fig 7 Fluorescence of curcumin bound to purified Ca21-ATPase

in the presence Ca21 (A) Spectra of 1 m M curcumin in: (i) buffer alone at 25 8C, pH 7.0; (ii) in the presence of 2 m M purified Ca 2þ ATPase and (iii) after addition of 2.5 m M Ca2þ Results show approximately 25% decrease in fluorescence upon addition of Ca 2þ (B) Fluorescence changes of either curcumin bound to the ATPase or tryptophan residues within the ATPase, upon addition of a range of free

Ca2þconcentrations (3 n M to 100 m M ) Experiments were performed at

25 8C either at pH 6.0 (V, tryptophan, P, curcumin) or at pH 7.0 (B, tryptophan, O, curcumin) (C) Curcumin fluorescence change induced

by Ca2þ, monitored when bound to cerebellar microsomes (200 mg·mL 21 ) Experiments were performed at 25 8C pH 7.0 All data points represent means ^ SD of three determinations Curcumin fluorescence was monitored using the following wavelengths: Excitation l ¼ 411 nm, emission l ¼ 500 nm Tryptophan fluor-escence was monitored by exciting at 295 nm and detecting the emission at 340 nm.

been associated with Ca2þ binding to the ATPase [16] At

pH 7, the curcumin decrease and tryptophan increase in

fluorescence can be superimposed and give a single curve with

the EC50 for Ca2þ occurring at 0.22 ^ 0.11 mM (^ SEM,

n ¼ 9) At pH 6, the curves could again be superimposed but

were shifted to the left (EC50¼ 5.0 ^ 1.0 mM, SEM, n ¼ 7)

as Ca2þbinds more weakly at lower pH Experiments were

performed with cerebellar microsomes (Fig 7C) to see if the

change in curcumin fluorescence with respect to Ca2þcould

be used with crude membrane preparations, where the

presence of Ca2þ ATPase is low and where tryptophan

measurements are impracticable due to abundance of other

proteins It was found that a Ca2þ binding curve could be

derived from the curcumin fluorescence data that gave a EC50

of 1.6 ^ 0.5 mM(^ SEM n ¼ 7) and a maximal decrease of

20%

D I S C U S S I O N

From the Ca2þ-ATPase activities, it can be seen that all

subtypes of SERCA are inhibited to a similar degree by

curcumin suggesting it is not a subtype specific inhibitor of

the Ca2þ-ATPase Interestingly, the corresponding Ca2þ

uptake shows marked differences For platelet and cerebellar

microsomes, an increase in Ca2þ uptake at low

concen-trations of curcumin was observed followed by inhibition at

higher concentrations This biphasic response has been

observed in microsomes upon exposure to ethanol [31,32]

Mitidieri & de Meis [31] and Mezna et al [32,33] showed

that at concentrations where ethanol had no effect or

inhibitory effects on Ca2þ-ATPase activity, there was a

significant increase in uptake It is unlikely that the

enhancement of Ca2þuptake is due to curcumin reducing

the permeability of ions through the phospholipid bilayer, as

our data shows that curcumin makes phospholipid

membranes more, not less, leaky (Fig 2) Therefore at

present the simplest explanation would be that curcumin

inhibits Ca2þrelease from these microsomes through a Ca2þ

channel Studies on the inositol-trisphosphate-sensitive

Ca2þ channel, which is abundant in cerebellum and

platelets, have shown it to be inhibited by curcumin and

therefore it seems the most likely target [25]

If the biphasic response of curcumin on Ca2þuptake in

cerebellar and platelet microsomes were to occur in intact

nonmuscle cells, this may well explain many of the reported

cellular effects observed with curcumin For instance high

doses of curcumin can induce apoptosis [34] It is also

known that prolonged elevation of [Ca2þ]cyt induces

apoptosis [35] and agents such as thapsigargin [36] and

alkylphenols [37], which inhibit ER Ca2þpumps can trigger

this process Therefore if high curcumin concentrations act

in a manner similar to thapsigargin and elevates [Ca2þ]cyt

(i.e inhibit Ca2þuptake), this would be the most obvious

mode of action Platelet aggregation, inflammation, and

arachadonic acid production are all processes that have been

shown to be inhibited by curcumin [6,7,38] as well as

requiring Ca2þ[39 – 41] If cells undergoing these processes

were exposed to curcumin concentrations that were able to

stimulate Ca2þ uptake, this would have the effect of

reducing [Ca2þ]cyt, leading to a reduction in stimulation

Therefore the effects of curcumin on these activities could

be explained, at least in part, in terms of its effects on

intracellular Ca2þlevels

The mechanism by which the ATPase transports Ca2þis usually discussed in terms of the model proposed by DeMeis

& Vianna [42], involving two major conformational states defined as E1 and E2 In the E1 form the ATPase is able to bind two Ca2þwith high affinity on the cytoplasmic side of the membrane While in the E2 form the Ca2þbinding sites have translocated across the membrane to the luminal side and are of low affinity Furthermore, in the E1 confor-mational state, the ATPase can be phosphorylated by ATP, which drives the translocation process From the data presented here it appears that curcumin preferentially stabilizes the E1 form of the ATPase as well as acting as a noncompetitive inhibitor with respect to Ca2þ However, the situation is more complex with respect to ATP, with low concentrations of curcumin (up to 5 – 10 mM) acting as a noncompetitive inhibitor with respect to ATP, but higher concentrations appearing to be competitive This is also confirmed from the ATP binding data, which showed that curcumin inhibited ATP binding (IC50 ¼ 9 mM) The hydrophobic nature of curcumin would make it unlikely to compete for the nucleotide binding site by mimicking ATP, and therefore it is more likely to act as a competitive inhibitor by inducing a conformational change, which precludes ATP from binding This suggestion is further supported by the fact that the Ca2þ-induced conformational change monitored by the fluorescence of the FITC-labelled ATPase is affected by curcumin FITC is known to label the ATPase on Lys515 within the ATP binding pocket of the nucleotide binding domain, also precluding ATP binding [15,43] Therefore if curcumin affects this conformationally induced fluorescence change, it must be binding elsewhere

as the ATP binding site is already occupied with FITC

In comparing the 2.6-A˚ resolution crystal structure of the

Ca2þ bound (E1) form of the ATPase with the 8 A˚ low resolution structure of the decavanadate-bound (E2) form of the ATPase, Toyoshima et al [15], have shown by modeling that several major re-arrangements within the three cytoplasmic domains need to occur They predicted that in going from E1 to E2, the nucleotide-binding domain has to move more than 25 A˚ to come into close contact with Asp351 within the phosphorylation domain, for phosphoryl transfer to occur These two domains are linked via a hinge

or bridging region that encompasses amino-acid sequences

355 – 365 and 595 – 605 The actuator domain also under-goes a large motion, moving approximately 30 A˚ and rotating almost 908 to come into closer contact with both the nucleotide-binding and phosphorylation domains In addition, it was suggested that the nucleotide-binding domain is highly mobile in the presence of Ca2þand can come into close contact with the phosphorylation site by simple thermal fluctuation [15]

From this type of analysis of the probable conformational changes that need to occur in the tertiary structure of the

Ca2þ-ATPase in going from E1 to E2, a model can be proposed to explain the competitive nature of curcumin with respect to ATP, without it binding directly to the ATP-binding site (Scheme 1) In this model, we also postulate that the nucleotide-binding and phosphorylation domains are sufficiently mobile to allow them to come into contact with each other in the presence or absence of ATP If both ATP and Ca2þare bound to the ATPase during this contact then phosphorylation can occur and Ca2þcan be transported across them membrane However, if the two domains come

into contact in the absence of ATP, curcumin is then able to

bind to the ATPase (possibly at the hinge region) locking the

two domains together and therefore precluding ATP binding

(i.e inhibiting the ATPase in a ‘competitive manner’) It

would appear unlikely that curcumin can ‘occlude’ ATP

binding, in the same way as chromium-ATP, by trapping the

ATP in the binding site when the two domains come together

[44], as our ATP binding data shows that little ATP is bound

to the ATPase when it is added prior to curcumin

The fluorescence data of curcumin bound to the ATPase

indicates that it may bind with either a high affinity (<

1 mM) to a single site or to two sites of differing affinities

(Kdvalues of 0.55 mMand 10 mM, respectively) However,

to inhibit the ATPase activity by 50% would require between

7 and 15 mM curcumin Therefore this would be more

consistent with the presence of two distinct binding sites for

this molecule of differing affinities Our data would suggest

that the lower affinity binding site (Kd ¼ 10 mM) might

contribute more towards the inhibition on the ATPase, than

the higher affinity one (Kd ¼ 0.55 mM), as this correlates

to the IC50value gained from the activity data (IC50values

of between 7 and 17 mM) In addition, this would also be

consistent with the [32P]ATP binding data (where the IC50

value was 9 mM) and FITC-ATPase data (where the IC50and

EC50values were 5 – 6 mM) As the high affinity binding site

for curcumin can sense Ca2þbinding events, it could also be

speculated that this site is either close to the Ca2þbinding

sites, or at a site which undergoes major changes upon Ca2þ

binding (i.e the actuator domain or transmembrane helices

M1 and M3 which lead into this domain [15])

In conclusion, curcumin inhibits the Ca2þ-ATPase, by

inducing a conformational change, which blocks the ATP

from binding

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

We would like to thank the BBSRC for a PhD studentship to J G B.

and the government of Pakistan for a PhD scholarship to S Z K.

R E F E R E N C E S

1 Srivastava, K.C & Mustafa, T (1993) Pharmacological effects of spices: eicosanoid modulating activities and their significance in human health Biomed Rev 2, 15 – 29.

2 Gescher, A., Pastorino, U., Plummer, S.M & Manson, M.M (1998) Suppression of tumour development by substances derived from the diet – mechanisms and clinical implications Br J Clin Pharmacol 45, 1 – 12.

3 Kelloff, G.J., Boone, C.W., Crowell, J.A., Steele, V.E., Lubet, R.A., Doody, L.A., Malone, W.F., Hawk, E.T & Sigman, C.C (1996) New agents for cancer chemoprevention J Cell Biochem 26,

1 – 28.

4 Xu, Y.X., Pindolia, K.R., Janakiraman, N., Noth, C.J., Chapman, R.A & Gautam, S.C (1997) Curcumin, a compound with anti-inflammatory and anti-oxidant properties, down-regulates chemo-kine expression in bone marrow stromal cells Exp Hematol 25,

413 – 422.

5 Piwocka, K., Zablocki, K., Wieckowski, M.R., Skierski, J., Feiga, I., Szopa, J., Drela, N., Wojtczak, L & Sikora, E (1999) A novel apoptosis-like pathway, independent of mitochondria and caspases, induced by curcumin in human lymphoblastoid T (Jurkat) cells Exp Cell Res 249, 299 – 307.

6 Shah, B.H., Nawaz, Z., Pertani, S.A., Roomi, A., Mahmood, H., Saeed, S.A & Gilani, A.H (1999) Inhibitory effect of curcumin, a food spice from turmeric, on platelet-activating factor- and arachidonic acid-mediated platelet aggregation through inhibition

of thromboxane formation and Ca 2þ signaling Biochem Pharmacol 58, 1167 – 1172.

7 Srivastava, K.C., Bordia, A & Verma, S.K (1995) Curcumin, a major component of food spice turmeric (Curcuma longa ) inhibits aggregation and alters eicosanoid metabolism in human blood platelets Prostaglandins Leukot Essent Fatty Acids 52, 223 – 227.

8 Abe, Y., Hashimoto, S & Horie, T (1999) Curcumin inhibition of inflammatory cytokine production by human peripheral blood monocytes and alveolar macrophages Pharmacol Res 39, 41 – 47.

9 Zhang, F., Altorki, N.K., Mestre, J.R., Subbaramaiah, K & Dannenberg, A.J (1999) Curcumin inhibits cyclooxygenase-2 transcription in bile acid- and phorbol ester-treated human gastrointestinal epithelial cells Carcinogenesis 20, 445 – 451.

10 Liu, J.Y., Lin, S.J & Lin, J.K (1993) Inhibitory effects of curcumin Scheme 1 Proposed model for curcumin inhibition This scheme shows the potential effects of curcumin (I) on the interactions between the nucleotide binding domain (N) and the phosphorylation domain (P), to give competitive inhibition with respect to ATP.

on protein kinase C activity induced by

12-O-tetradecanoyl-phorbol-13-acetate in NIH 3T3 cells Carcinogenesis 14, 857 – 861.

11 Chen, H.W & Huang, H.C (1998) Effect of curcumin on cell cycle

progression and apoptosis in vascular smooth muscle cells Br.

J Pharmacol 124, 1029 – 1040.

12 Chen, Y.C., Kuo, T.C., Lin-Shiau, S.Y & Lin, J.K (1996) Induction

of HSP70 gene expression by modulation of Ca2þion and cellular

p53 protein by curcumin in colorectal carcinoma cells Mol.

Carcinog 17, 224 – 234.

13 MacLennan, D.H., Brandl, C.J., Korczak, B & Green, N.M (1985)

Amino-acid sequence of a Ca2þþ Mg 2þ

-dependent ATPase from rabbit muscle sarcoplasmic reticulum, deduced from its

comp-lementary DNA sequence Nature 316, 696 – 700.

14 Wu, K.D., Lee, W.S., Wey, J., Bungard, D & Lytton, J (1995)

Localization and quantification of endoplasmic reticulum Ca 2þ

-ATPase isoform transcripts Am J Physiol 269, C775 – C784.

15 Toyoshima, C., Nakasako, M., Nomura, H & Ogawa, H (2000)

Crystal structure of the calcium pump of sarcoplasmic reticulum at

2.6A ˚ resolution Nature 405, 647–655.

16 Longland, C.L., Mezna, M & Michelangeli, F (1999) The

mechanism of inhibition of the Ca2þ-ATPase by mastoparan.

Mastoparan abolishes cooperative Ca2þbinding J Biol Chem 274,

14799– 14805.

17 Michelangeli, F., Orlowski, S., Champeil, P., East, J.M & Lee,

A.G (1990) Mechanism of inhibition of the (Ca2þ-Mg2þ)-ATPase

by nonylphenol Biochemistry 29, 3091 – 3101.

18 Wictome, M., Michelangeli, F., Lee, A.G & East, J.M (1992) The

inhibitors thapsigargin and

2,5-di(tert-butyl)-1,4-benzohydroqui-none favour the E2 form of the Ca 2þ ,Mg 2þ -ATPase FEBS Lett.

304, 109 – 113.

19 Michelangeli, F & Munkonge, F.M (1991) Methods of

reconstitution of the purified sarcoplasmic reticulum (Ca 2þ

-Mg 2þ )-ATPase using bile salt detergents to form membranes of

defined lipid to protein ratios or sealed vesicles Anal Biochem.

194, 231 – 236.

20 Sayers, L.G., Brown, G.R., Michell, R.H & Michelangeli, F.

(1993) The effects of thimerosal on calcium uptake and inositol

1,4,5-trisphosphate-induced calcium release in cerebellar

micro-somes Biochem J 289, 883 – 887.

21 Le Peuch, C.J., Le Peuch, D.A., Katz, S., Demaille, J.G., Hincke,

M.T., Bredoux, R., Enouf, J., Levy-Toledano, S & Caen, J (1983)

Regulation of calcium accumulation and efflux from platelet

vesicles Possible role for cyclic-AMP-dependent phosphorylation

and calmodulin Biochim Biophys Acta 731, 456 – 464.

22 Longland, C.L., Mezna, M., Langel, U., Hallbrink, M., Soomets,

U., Wheatley, M., Michelangeli, F & Howl, J (1998) Biochemical

mechanisms of calcium mobilisation induced by mastoparan and

chimeric hormone-mastoparan constructs Cell Calcium 24, 27 – 34.

23 Gould, G.W., East, J.M., Froud, R.J., McWhirter, J.M., Stefanova,

H.I & Lee, A.G (1986) A kinetic model for the Ca 2þ þ

Mg 2þ -activated ATPase of sarcoplasmic reticulum Biochem J.

237, 217 – 227.

24 Michelangeli, F (1991) Measuring calcium uptake and inositol

1,4,5-trisphosphate-induced calcium release in cerebellar

micro-somes using Fluo-3 J Fluorescence 1, 203 – 206.

25 Khan, S.Z (2001) The effects of hydrophobic compounds on

intracellular Ca2þchannels and pumps PhD Thesis, University of

Birmingham, UK.

26 Champeil, P & Guillain, F (1986) Rapid filtration study of the

phosphorylation-dependent dissociation of calcium from transport

sites of purified sarcoplasmic reticulum ATPase and ATP

modulation of the catalytic cycle Biochemistry 25, 7623 – 7633.

27 Michelangeli, F., Orlowski, S., Champeil, P., Grimes, E.A., East, J.M & Lee, A.G (1990) Effects of phospholipids on binding of calcium to (Ca2þ-Mg2þ)-ATPase Biochemistry 29, 8307 – 8312.

28 Brown, G.R., Benyon, S.L., Kirk, C.J., Wictome, M., East, J.M., Lee, A.G & Michelangeli, F (1994) Characterisation of a novel

Ca2þpump inhibitor (bis-phenol) and its effects on intracellular

Ca2þmobilization Biochim Biophys Acta 1195, 252 – 258.

29 Coll, R.J & Murphy, A.J (1991) Kinetic evidence for two nucleotide binding sites on the Ca-ATPase of sarcoplasmic reticulum Biochemistry 30, 1456 – 1461.

30 Froud, R.J & Lee, A.G (1986) Conformational transitions in the

Ca 2þ þ Mg 2þ -activated ATPase and the binding of Ca 2þ ions Biochem J 237, 197 – 206.

31 Mitidieri, F & de Meis, L (1995) Ethanol has different effects on

Ca 2þ -transport ATPases of muscle, brain and blood platelets Biochem J 312, 733 – 737.

32 Mezna, M., Patchick, T., Tovey, S & Michelangeli, F (1996) Inhibition of the cerebellar inositol 1,4,5-trisphosphate-sensitive

Ca2þchannel by ethanol and other aliphatic alcohols Biochem J.

314, 175 – 179.

33 Mezna, M., Patchick, T., Tovey, S & Michelangeli, F (1996) Effects of aliphatic alcohols on the ER calcium pumps and InsP 3 receptors from porcine cerebellum Biochem Soc Trans 24, 292S.

34 Sikora, E., Bielak-Zmijewska, A., Piwocka, K., Skierski, J & Radziszewska, E (1997) Inhibition of proliferation and apoptosis

of human and rat T lymphocytes by curcumin, a curry pigment Biochem Pharmacol 54, 899 – 907.

35 Orrenius, S., McConkey, D.J., Bellomo, G & Nicotera, P (1989) Role of Ca 2þ in toxic cell killing Trends Pharmacol Sci 10,

281 – 285.

36 Mukerjee, N., McGinnis, K.M., Park, Y.H., Gnegy, M.E & Wang, K.K (2000) Caspase-mediated proteolytic activation of calcineurin

in thapsigargin-mediated apoptosis in SH-SY5Y neuroblastoma cells Arch Biochem Biophys 379, 337 – 343.

37 Hughes, P.J., McLellan, H., Lowes, D.A., Khan, S.Z., Bilmen, J.G., Tovey, S.C., Godfrey, R.E., Michell, R.H., Kirk, C.J & Michelangeli, F (2000) Estrogenic alkylphenols induce cell death

by inhibiting testis endoplasmic reticulum Ca2þpumps Biochem Biophys Res Commun 277, 568 – 574.

38 Huang, M.T., Lysz, T., Ferraro, T., Abidi, T.F., Laskin, J.D & Conney, A.H (1991) Inhibitory effects of curcumin on in vitro lipoxygenase and cyclooxygenase activities in mouse epidermis Cancer Res 51, 813 – 819.

39 Clark, J.D., Lin, L.L., Kriz, R.W., Ramesha, C.S., Sultzman, L.A., Lin, A.Y., Milona, N & Knopf, J.L (1991) A novel arachidonic acid-selective cytosolic PLA 2 contains a Ca2þ-dependent trans-location domain with homology to PKC and GAP Cell 65,

1043 – 1051.

40 Nishikawa, M & Hidaka, H (1982) Role of calmodulin in platelet aggregation Structure – activity relationship of calmodulin antagonists J Clin Invest 69, 1348 – 1355.

41 Rao, T.S., Shaffer, A.F., Currie, J.L & Isakson, P.C (1994) Role of mast cells in calcium ionophore (A23187)-induced peritoneal inflammation in mice Inflammation 18, 187 – 192.

42 De Meis, L & Vianna, A.L (1979) Energy interconversion by the

Ca2þ-dependent ATPase of the sarcoplasmic reticulum Annu Rev Biochem 48, 275 – 292.

43 Pick, U (1981) Interaction of fluorescein isothiocyanate with nucleotide-binding sites of the Ca-ATPase from sarcoplasmic reticulum Eur J Biochem 121, 187 – 195.

44 McIntosh, D.B (1998) The ATP binding sites of P-type ion transport ATPases Adv Mol Cell Biol 23A, 33 – 99.