lable at ScienceDirect

Biochimie 92 (2010) 178e186

Contents lists avai

Biochimie

journal homepage: www.elsevier .com/locate/biochi

Research paper

Kinetic properties and specificity of trimeric Plasmodium falciparumand human dUTPases

Indalecio Quesada-Soriano a, Juan M. Casas-Solvas b, Eliseo Recio c, Luis M. Ruiz-P�erez c,Antonio Vargas-Berenguel b, Dolores Gonz�alez-Pacanowska c, Luis García-Fuentes a,*

a �Area de Química Física, Facultad de Ciencias Experimentales, Universidad de Almería, La Ca~nada de San Urbano, 04120 Almería, Spainb �Area de Química Org�anica, Facultad de Ciencias Experimentales, Universidad de Almería, La Ca~nada de San Urbano, 04120 Almería, Spainc Instituto de Parasitología y Biomedicina “L�opez-Neyra”, Consejo Superior de Investigaciones Científicas, Avda. del Conocimiento s/n, Parque Tecnol�ogicode Ciencias de la Salud, 18100 Granada, Spain

a r t i c l e i n f o

Article history:Received 5 May 2009Accepted 21 October 2009Available online 29 October 2009

Keywords:KineticCalorimetrydUTPasePlasmodium falciparumInhibitionSpecificityBinding

Abbreviations: dUTPase, dUTP pyrophosphatase (titration calorimetry; TS, thymidylate synthase; FdU50-triphosphate; IdUTP, 5-iode-20-deoxyuridine 50-trip* Corresponding author. Tel.: þ34 950 015618; fax:

E-mail address: lgarcia@ual.es (L. García-Fuentes).

0300-9084/$ e see front matter � 2009 Elsevier Masdoi:10.1016/j.biochi.2009.10.008

a b s t r a c t

Deoxyuridine 50-triphosphate nucleotidohydrolase (dUTPase, EC 3.6.1.23) catalyzes the hydrolysis ofdUTP to dUMP and pyrophosphate, and plays important roles in nucleotide metabolism and DNAreplication. Hydrolysis of other nucleotides similar in structure to dUTP would be physiologicallynegative and therefore high substrate specificity is essential. Binding and hydrolysis of nucleotidesdifferent to dUTP by the dUTPases from Plasmodium falciparum (PfdUTPase) and human (hdUTPase) wasevaluated by applying isothermal titration calorimetry (ITC). The ribo and deoxyribonucleosidetriphosphates dGTP, dATP, dCTP, dTTP, UTP, FdUTP and IdUTP have been analysed. dUTP and FdUTP werethe most specific substrates for both enzymes. The specificity constants (kcat/Km) for the remaining ones,except for the IdUTP, were very similar for both enzymes, although PfdUTPase showed a slightly higherspecificity for dCTP and UTP and the human enzyme for dTTP and dCTP. PfdUTPase was very efficient inusing FdUTP as substrate indicating that small size substituents in the 50 position are well tolerated. Inaddition product inhibition was assessed by binding studies with the nucleoside monophosphatederivatives and thermodynamic parameters were established. When FdUTP hydrolysis was monitored,Plasmodium dUTPase was more sensitive to end-product inhibition by FdUMP than the human enzyme.Taken together these results highlight further significant differences between the human and Plasmo-dium enzymes that may be exploitable in selective inhibitor design.

� 2009 Elsevier Masson SAS. All rights reserved.

1. Introduction

The enzyme deoxyuridine 50-triphosphate diphosphohydrolase(dUTPase, EC 3.6.1.23) is involved in deoxynucleotide metabolismand catalyzes the hydrolysis of the a-b-diphosphate bond of dUTPto yield dUMP and inorganic pyrophosphate (PPi). dUTPase activityis dependent on the presence of the divalent ion Mg2þ and theMichaelis complex is assumed to involve Mg2þ and dUTP [1e10].Other divalent metal ions such as Co2þ or Mn2þ can substitute thephysiological cofactor. Although for the majority of dUTPases thekcat, in the presence of those metal ions, is significantly reduced

EC 3.6.1.23); ITC, isothermalTP, 5-fluoro-20-deoxyuridinehosphate.þ34 950 015008.

son SAS. All rights reserved.

compared to dUTP$Mg2þ, some studies with Escherichia coli dUT-Pase showed a kcat value for dUTP$Mn2þ similar to that calculated inthe presence of its physiological cofactor [11]. The hydrolysisproduct, dUMP, is transformed to dTMP by thymidylate synthase(TS). Although dUTP is a normal intermediate in DNA synthesis incertain cell types, its accumulation and misincorporation into DNAas uracil is lethal. Thus, by dUTP hydrolysis, the enzyme increasesthe ratio dTTP/dUTP, minimizing misincorporation of uracil intoDNA during replication and repair. A solid understanding of thecontrol of the action mechanisms of these two enzymes is impor-tant both for maintaining DNA integrity in dividing cells [12] andfor the prevention of uracil incorporation into DNA.

dUTPases from prokaryotic and eukaryotic sources are known tobe very specific. Compared to dUTP, hydrolysis of other naturallyoccurring pyrimidine nucleotides appears to be negligible. Hydro-lysis of other nucleotides similar in structure to dUTP wouldconceivably be physiologically detrimental and high specificity ofthe reaction is necessary. For instance, no hydrolysis of dTTP, dCTP

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186 179

or UTP has been detected on behalf of the human enzyme isolatedfrom lymphoid cells [13]. In the case of E. coli dUTPase, discrimi-nation is mainly achieved by differences in binding [1]. Thus,whereas the Km value for dUTP$Mg2þ is 0.2 mM, Km values for UTPand dCTP are approximately 2.5 and 4 mM, respectively [1].

5-fluorouracil (5FU) is an anticancer agent metabolized withinthe cell to FdUMP, which acts as a mechanism-based inhibitor of TS.After exposure of cells to 5FU, potential intracellular derivativesinclude both FdUMP and the 5-fluorouridine di- and triphosphates[14]. However, the triphosphate derivatives of 5FU are not usuallyfound, since the FdUTP synthesized is largely converted intracel-lularly to the monophosphate by dUTPase [15,16].

Different methods have been frequently employed for moni-toring the dUTPase enzymatic reaction. In the past, discontinuousand slowmethods have been used [17,18]. In most studies dUTPaseactivity wasmeasured by a spectrophotometric continuousmethodbased on the use of the stopped-flow pH indicator technique[1,19e21]. Recently, we have demonstrated the applicability andvalidity of isothermal titration calorimetry (ITC) for monitoring andmeasuring the rate of the enzymatic reaction of dUTPase, particu-larly Plasmodium falciparum dUTPase (PfdUTPase) [22]. At present,this homotrimeric enzyme,which shows structural differenceswithother trimeric dUTPases [23] such as E. coli [24] and human [3], isbeing explored as a potential target for the development of drugs forthe treatment of malaria in humans and certain tritylated deriva-tives that specifically inhibit the parasite enzyme versus the humanprotein have been identified that present antiplasmodial activity[20,25].

In this report, we have extended the use of ITC to analyze indetail the interaction of two trimeric dUTPases (P. falciparumand human, hdUTPase) with different triphosphate derivatives.We also have evaluated both FdUTP and IdUTP hydrolysis andproduct inhibition in several experimental conditions. The resultsreported herein contribute to a better understanding of dUTPasebinding requirements and reveal certain differences that may befurther useful in the design of specific inhibitors of the plasmo-dial enzyme.

2. Material and methods

2.1. Chemicals

UTP, dTMP, dCMP, dGMP and dAMPwere purchased from Sigma.dATP, dUTP, dTTP, dCTP, dGTP, FdUTP, FdUMP and IdUTP werepurchased from Jena Biosciences, Germany. 2-glycerophosphate,Pipes, MES and TES buffers were purchased fromMerck and Sigma.Other chemicals used in this investigation were of the highestavailable quality. All solutions were made with distilled anddeionized (Milli Q) water, and they were degassed and filteredthrough a 0.45-mm Millipore filter before use. The concentration ofnucleotides was determined spectrophotometrically.

2.2. Enzymes

Recombinant P. falciparum was expressed and purified asdescribed elsewhere [23]. The human dUTPase sequence waskindly provided by Per Olof Nyman (Lund University, Sweden). Thefull length coding sequencewas cloned in pET11, expressed in E. coliBL21/DE3 cells and purified using the same methodology as indi-cated for Plasmodium dUTPase. After purification, the enzymeswere homogeneous as judged by SDS-PAGE. Protein concentrationswere determined by Bradford assay [26]. Before use, the purifiedenzymes were concentrated and dialyzed at 4 �C against therespective buffers. Moreover, it's very important to note that

the enzyme concentrations, in the kinetic assays, were given as theconcentrations of active sites. However, in the binding experiments,those were expressed as molecular concentrations (i.e. per trimer).

2.3. Colorimetric enzyme activity assay

FdUTP hydrolysis by these enzymes was monitored by mixingthe enzyme and FdUTP with a rapid kinetic accessory (Hi-TechScientific) attached to a Cary BIO50 spectrophotometer (Varian).Activity assays were performed at 25 �C and the same experimentalconditions to those described for the physiological substrate[19,22]. Due to product inhibition, the real Km for substrate hydro-lysis was determined by assays performed at FdUTP concentrationsranging between 10 and 1200 mM. The plot of apparent Km versussubstrate concentration allows the calculation of the real Km forFdUTP hydrolysis as well as the Kip for the product, FdUMP [19,22].

2.4. NMR measurements

31P-NMR spectra were recorded on a Bruker Avance DPX spec-trometer at 121.50 MHz. Chemical shifts are given relative to 85%H3PO4 (d 0 ppm). A capillary insert filled with D2O was used asa field frequency lock. All experiments were carried out in glycer-ophosphate buffer (25 mM glycerophosphate, 5 mM NaCl, 25 mMMgCl2, and 1 mM b-mercaptoethanol at pH 7.0) aqueous solutionsat room temperature.

2.5. Isothermal titration calorimetry (ITC)

Calorimetric experiments were conducted using an ITC instru-ment (VP-ITC, MicroCal Inc., Northampton, USA). Reaction cellswere filled with enzyme solutions and equilibrated at the experi-mental temperature. After this equilibration, an additional delayperiod was allowed to generate the baseline used in the subsequentdata analyses. ITC measurements were routinely performed in25 mM MES, 5 mM NaCl, 1 mM b-mercaptoethanol and 25 mMMgCl2 at pH 7.0.

2.6. Kinetic data analysis

Enzyme kinetics were investigated using the so called multipleinjections assay as described by Todd and Gomez [27]. Briefly, thismethod measures the rate of enzyme catalysis during stepwiseincrease of substrate concentration in an enzyme solution. A simpleexplanation to the calculations is as follows. The primary experi-mental observable is heat-flow (power) measured in mcal s�1. Thispower (dQ/dt) is directly proportional to the reaction rate (v),expressed as d[p]t/dt, and where [P]t, is the concentration ofproduct formed at reaction time, t.

v ¼ d½P�tdt

¼ 1DHappV0

dQdt

(1)

However, calculation of the reaction rate requires knowledge ofthe apparent enthalpy change of the reaction (DHapp) and thevolume (V0) of the reaction cell (effective volume 1.42 mL). Theenthalpy change is calculated by measurements of the total heat,which is consumed or released and the total number of moleswhich are converted from substrate to product (nconv):

DHapp ¼ 1nconv

24Z N

0

�dQdt

�dt

35 (2)

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186180

The substrate concentration, [S]t, as a function of time can becalculated using the integrated heat as a function of time:

½S�t ¼ ½S�0�R t0ðdQ=dtÞdtV0$DHapp

(3)

Combining eqs. (1) and (3), it is possible to calculate both thereaction rate and substrate concentration as a function of time andfinding the kinetic parameters kcat and Km by fitting the experi-mental data to the MichaeliseMenten equation. In some particularcases, the simple injection assay was also applied in conditionssimilar to those described for dUTP [22]. A more detailed discussionof these analyses can be found elsewhere [27e29].

In addition, if proton release occurs during the enzymaticreaction, the ionization enthalpy of the buffer, DHion, has to be

A

B

C

D

Fig. 1. Raw data for thermal power change as a function of time in the multiple injectiondUTPase, and dCTP (C) and dTTP (D) by human dUTPase. The non-linear least squares fits ofparameters are shown by the right side of each thermogram. Kinetic assays were performedmaking 5 mL injections of 15 mM substrate into calorimetric cell containing 15e25 mM enz

taken into account to calculate the true reaction enthalpy, DHrxn,according to the relation:

DHapp ¼ DHrxn þ n$DHion (4)

where n is the number of protons released. Proton release duringFdUTP hydrolysis was evaluated by performing the ITC kineticstudy in buffers which have different ionization enthalpies (inkcal$mol�1 at 25 �C): 2-glycerophosphate (�0.17), PIPES (2.74), MES(3.71) and TES (7.83) [30].

2.7. Binding experiments

The procedure and data analysis that were applied were similarto those described elsewhere [31,32]. Briefly, the inhibitor solution

s assay for the hydrolysis process of dCTP (A) and dTTP (B) by Plasmodium falciparumthe experimental data to the MichaeliseMenten equation to determine the hydrolysisin 25 mM MES, 5 mM NaCl, 1 mM b-mercaptoethanol and 25 mM MgCl2 (pH 7, 25 �C),yme.

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186 181

was loaded in the injection syringe and added stepwise to thecalorimetric cell containing the enzyme solution. As controlexperiments (dilution experiments) the inhibitor was injected intothe buffer under identical experimental conditions. Raw data werecollected, corrected for the ligand heats of dilution, and integratedusing the Microcal Origin software supplied with the instrument.A non-cooperativemodel was used to fit the experimental data. Thedata analysis produced three parameters, viz., stoichiometry (n),association constant (K) and standard enthalpy change (DH) for thebinding of the ligand to enzyme. The binding Gibbs energy change(DG0 ¼�RT ln K) was calculated from the binding constant, and theassociation entropy change (TDS0 ¼ DH�DG0) was also estimated.

3. Results and discussion

3.1. dNTPs and UTP hydrolysis

The specific role of dUTPase is to remove dUTP and to bevirtually inactive with the other nucleotides [13]. dUTP hydrolysisand product inhibition by PfdUTPase have been previously char-acterized [22,23]. Here, PfdUTPase was tested for its ability tohydrolyse other purine and pyrimidine nucleotides. Fig. 1 showstwo representative thermograms corresponding to two ad hocdesigned ITC experiments to obtain the kinetic data for hydrolysisof dCTP (Fig. 1A) and dTTP (Fig. 1B) by PfdUTPase. Significanthydrolytic activity was detected and kinetic parameters wereestimated from the data as described in the Materials and methodssection. In both cases, as a result of the deoxynucleoside triphos-phate injections, an initial endothermic peak corresponding to theheat dilution was generated. The baseline then stabilises at a lowerpower level than those of the precedent injections as consequenceof the heat generated by the enzymatic reaction and due to the factthat the higher the substrate accumulation, the faster the reactionoccurs. The drop in the baseline indicates that these particularreactions proceed with a negative (exothermic) enthalpy change.Thus, the enzyme-generated thermal power (dQ/dt) shouldincrease after the substrate injection, which occurs with dCTP(Fig. 1A). However, if there is product inhibition, this does nothappen [33,34], like for the hydrolysis of dTTP where the generatedthermal power diminishes after a time of 35 min (Fig. 1B). There-fore, product inhibition is detected in the hydrolysis reaction ofdTTP and Plasmodium enzyme, and so, for the kinetic analysis, wetruncated the rate measurements at w35 min. Thus, within thistime range (0e35 min), a normal plateau for the steady state wasobtained (Fig. 1, right panel). Moreover, in the case of dTTP andPfdUTPase (Fig. 1B), different values of Km were obtained when

Table 1Substrate specificity of Plasmodium falciparum and human dUTPases.a

Nucleotide Km (mM) kcat/Km (s�1 M�1)

P. f. human P. f. huma

dUTP 1.87 1.1 7.1$106 6.6$10FdUTP 1.73 6.1 1.1$105 6$105

dCTP 465 687 21.72 161.5UTP 571 540 11.21 12.78dTTP 290 761 9.20 19.71IdUTP 438 523 5.25 6.10dATP 1130 n.d. 2.65 e

dGTP 1960 n.d. 0.13 e

n.d. not determined.a Parameters values were the mean of at least three experiments.b Discrimination is expressed as kcat/Km for dUTP divided by the corresponding value

initial concentration of dTTP was varied, confirming the productinhibition. A plot of apparent Km versus dTTP concentration allowsfor the calculation of the real Km for dTTP hydrolysis as well as theKi for the product, dTMP. This determination gave values forKm ¼ 290 mM and Ki ¼ 470 mM. For comparative purposes, thehydrolysis of these two nucleotides (dCTP and dTTP) was alsoexamined with the human enzyme (Fig. 1C, D). In these cases, thecalorimetric thermograms for did not evidence significant productinhibition as occurs for PfdUTPase and dTTP. Similar calorimetricexperiments were performed with dATP and dGTP, which were lessefficiently hydrolysed by both the Plasmodium and the humanenzymes (data not shown). Significant hydrolytic activity wasobtained for UTP, yet product inhibitionwas not observed (data notshown). The raw data of the thermograms for these nucleotideswere transformed into plots of reaction rates versus nucleotideconcentration (Fig. 1, right panels) and the kinetic parameters, kcatand Km, were obtained by fitting the data to MichaeliseMentenequation using non-linear regression (Table 1). However, as indi-cated in Eqs. (1) and (3), to obtain the these kinetic parameters itwas necessary to determine previously the molar enthalpy (DHapp)values for these biochemical reactions. Values of DHapp for thedNTPs or UTP hydrolysis by both enzymes were determined fromother ITC experiments designed ad hoc, following the singleinjection assay and under conditions of total substrate conversion[27]. Using this single injection procedure DHapp values aremeasured when the titrating enzyme is at a much higher concen-tration than when multiple injections are employed. At highenzyme concentration the reaction time is shorten. Fig. 2 displaysa representative ITC experiment to determine the hydrolysisapparent enthalpy for dCTP by PfdUTPase in MES buffer at 25 �Cand pH 7.0. In this case, 66 mM enzyme contained in the calori-metric cell was titrated with 8.5 mM dCTP by 5 mL injections.The ligand dilution was determined independently. A value of�9.2 � 0.2 kcal mol�1 of DHapp was calculated for dCTP hydrolysisby subtracting the heat associated with the dilution event from thetotal heat evolved in one injection (area) and dividing by the molesof dCTP hydrolysed. The similar size/shape for all injectionsconfirmed that no appreciable product inhibition occurs in agree-ment with the ITC experiments of multiple injections describedpreviously. The hydrolysis molar enthalpies for the other nucleo-tides were determined analogously. The molar enthalpy data foreach nucleotide were calculated from the average of first threeinjections. The kinetic parameters for all assayed nucleotides areshown in Table 1.

In order to discard a potential spontaneous hydrolysis of thenucleoside triphosphate during the ITC experiments 31P-NMRmeasurements were performed as detailed in Ref. [22]. Nucleotides

Discrimination[dUTP/(d)NTP]b

DHapp (kcal$mol�1)

n P. f. human P. f. human6 e e �10.01 �10.40

65 11 �12.50 �11.507 3.3$105 4.1$104 �9.20 �12.20

6.3$105 5.2$105 �9.21 �11.027.8$105 3.3$105 �10.64 �14.801.4$106 1.1$106 �2.13 �3.412.7$106 e �2.15 �3.275.3$107 e �12.46 �14.03

for (d)NTP.

Fig. 2. Representative ITC experiment for the determination of the hydrolysis apparententhalpy for dCTP by PfdUTPase. (A) Isothermal titration calorimetry thermogram ofPfdUTPase (66 mM) catalyzed hydrolysis of dCTP (5 mL injections of 8.5 mM substrate)in 25 mM MES, 5 mM NaCl, 1 mM b-mercaptoethanol and 25 mM MgCl2 at pH 7 and25 �C. The change in instrumental thermal power was monitored until substratehydrolysis was complete, when it returned to the original baseline. (B) Subtraction ofthe dilution heat (area under dotted line) from first peak of titration (dashed line),contained within the broken line in the main plot, to obtain heat resulting (area undersolid line) for PfdUTPase-dCTP hydrolysis process is shown as inset.

A

B

cal

cal

cal

Fig. 3. 31P-NMR measurements of nucleotides. (A) 31P-NMR spectra of dTTP in a glyc-erophosphate buffer (pH 7) after being incubated for 24 h in the absence (bottom) andin the presence (up) of PfdUTPase. The signals at d �9.71, �18.25 and �4.86 ppm wereassigned to the a-, b- and g-P of the triphosphate moiety, respectively. The incubationwith the enzyme resulted in the disappearance of these signals and the appearance oftwo singlets at d 3.67 and �4.46 ppm corresponding to the a-P of dTMP and the twoequivalents P of the pyrophosphate anion (PPi), respectively. Similar spectra werefound for dCTP and dUTP (not shown). (B) 31P-NMR spectrum of dATP in glycer-ophosphate buffer (pH 7) after the incubation in the presence of PfdUTPase for 24 h.The signals at d �8.26, �16.99 and �3.70 ppmwere assigned to the a-, b- and g-P of thetriphosphate moiety, respectively. No hydrolysis products were detected within thesensitivity of the NMR measurement (5%) although a slightly hydrolysis was noticed bycalorimetric experiment (inset). Similar results were found for dGTP (not shown).

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186182

were incubated in glycerophosphate buffered (pH 7.0) solutions for24 h at room temperature both in the absence and in the presenceof the PfdUTPase protein (Fig. 3). The NMR data were consistentwith the observations obtained from calorimetry. In the case of thepyrimidine derivatives dTTP, dCTP and dUTP, hydrolysis was onlyobserved in the presence of enzyme. The recorded 31P-NMR spectraof the nucleotides after incubation with PfdUTPase showed thedisappearance of the three signals (two doublets and a triplet) thatcorrespond to the ligand triphosphate a-, b- and g-P and theappearance of two singlets assigned to the monophosphate P of thenucleotides (dTMP, dCMP and dUMP) and the pyrophosphate anion(PPi) that results from the hydrolysis. In the case of the purinederivatives dATP and dGTP, the NMR measurements did not showchanges in the structure of the substrate indicating that duringthe time of incubation these nucleotides either do not undergohydrolysis (spontaneous or enzymatic) or the hydrolysis occursbelow the detection limit of the NMR measurements (5%). Theseresults are not in contradiction with those obtained from a moresensitive calorimetric technique which detected a small kineticprocess for those ligands (Table 1).

In summary the results obtained (Table 1) show that the dUT-Pase from P. falciparum is highly specific for dUTP and discriminatesboth the base and sugar as well as the phosphate moiety, asreported for other dUTPases [1e3,6,9,19,21,35,36]. This specificity isvery similar to that obtained for human dUTPase and thosedescribed for E. coli [1,37] and Drosophila dUTPase [37], while EIAV,HSV-I, MMTV and Leishmania major dUTPases exhibit slightly lessrestrictive substrate specificity [19,38,39]. The second bestsubstrate among the nucleotides serving as building blocks for DNAwas dCTP. This nucleotide was hydrolysed w 3$105 and w4$104

times less efficiently than dUTP by PfdUTPase and hdUTPase,respectively (Table 1) although the human enzyme was w7 timesmore efficient than PfdUTPase.

The pyrimidine nucleotides UTP and dTTP were found to behydrolysed by PfdUTPase w 1.5 and 3 times less efficiently thandCTP, respectively. Clearly the methyl group of thymine seems to be

poorly tolerated. A structural basis for the discrimination betweenthese nucleotides is deduced from structure of the hdUTPase:a,b-imido-dUTP complex [7], the E. coli enzyme in complex with thesubstrate analogue dUDP [2] and the Plasmodium enzyme incomplex with a trityl derivative (PDB code: 1VYQ) [23]. In thiscomplex, the carbonyl at position 4 of the uracil ring is hydrogenbonded directing to atom ND2 of residue Asn103 and also formsa hydrogen bond with a conserved water molecule nearby. Thiswater molecule is in turn hydrogen bonded to atom OD1 of Asn103and the main chain N of residue Ala 111. The carbon atom at posi-tion C-5 of the uracil is in van der Waals' contact with a watermolecule only. When a methyl group is modelled in the 5 position,it appears to undergo steric clashes with the main chain atoms ofresidues Gly106 and Leu 107 and the side chain of residue Asn103.An amino group in position 4 is also difficult to accommodate thusexplaining the lowing binding affinity for these nucleotides.

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186 183

The catalytic efficiency of Plasmodium dUTPase was much lowerfor UTP (w6$105 times) than that for dUTP, as a consequence ofboth a lower catalytic constant and a higher Km value indicatingexquisite specificity with regard to the pentose moiety of thesubstrate. As previously observed in other dUTPases [3,35,36], thespecific recognition of the deoxyribose is achieved by a b-turn(b-hairpin), where two highly conserved residues (Tyr and Ile), (Tyr112 and Ile 109 in PfdUTPase), make a sandwich with the deoxy-ribose ring [23]. This tight packing sterically precludes ribosebinding to dUTPase, as the 20-hydroxyl group of ribose wouldpenetrate the van der Waals' radii of the tyrosine side-chain atoms.

The purine nucleotides (dATP and dGTP) bind weakly (higherKm) and the bound nucleotide is more slowly hydrolysed (lowerkcat) and the degree of discrimination is very high. Thus, the size ofthe purine ring prevents both a correct accommodation of thenucleotide and the favourable interactions which are needed forefficient catalytic activity.

3.2. FdUTP and IdUTP hydrolysis

The hydrolysis of FdUTP to FdUMP by Plasmodium and humandUTPases was studied both calorimetrically and spectrophoto-metrically at 25 �C. Fig. 4A shows an example of a calorimetric tracecorresponding to 15 ml injection of 8 mM FdUTP into a 12 nMPfdUTPase solution in 25 mM buffer MES, 5 mM NaCl, 25 mMMgCl2, 1 mM b-mercaptoethanol at pH 7.0. The kinetic plotemerging from ITC after analysing the thermogram is shown inFig. 4B. Deconvolution of the calorimetric traces was carried out asdescribed in Ref. [22]. Thus, the parameters Km and kcat werecalculated by fitting the curve to the MichaeliseMenten equationusing non-linear regression. However, although the experimentwas performed under conditions where product inhibition can beconsidered negligible for the first injection, different values of Km

(apparent Km) were obtained when the initial concentration ofFdUTP was varied. Thus, the determination of the real Km for FdUTPhydrolysis was obtained in ITC assays performed at FdUTPconcentrations ranging between 10 and 1200 mM. The plot of

A B

C

Fig. 4. Calorimetric determination of the PfdUTPase-catalyzed hydrolysis of FdUTP (pH 7.0injection of 8 mM FdUTP into calorimetric cell containing 12 nM enzyme (25 mM MES, 5 mMpower was converted to rate and fit to MichaeliseMenten equation to determine the kineticafter reacting 1 mM FdUTP with 1 mM PfdUTPase. (D) Linear transformation of the data ac

apparent Km versus FdUTP concentration allowed the calculation ofthe real Km. On the other hand, Vmax remained practically constantat different FdUTP concentrations, obtaining amean value of kcat forthe FdUTP concentrations assayed (Table 1).The same procedurewas used for FdUTP hydrolysis by the human enzyme. In order tocompare the kinetic parameters obtained by ITC, we further ana-lysed the hydrolysis using the spectrophotometric method. The Kmand kcat values obtained this way for FdUTP hydrolysis were:PfdUTPase: Km ¼ 3.1 mM, kcat ¼ 4 s�1; hdUTPase: Km ¼ 4.2 mM,kcat ¼ 3.2 s�1. These values are of the same order than thoseobtained by ITC (Table 1).

To estimate both the proton release and the reaction enthalpyfor this hydrolysis, the calorimetric study was carried out in severalbuffers (2-glycerophosphate, PIPES, MES and TES) with differentionization enthalpies, DHion. DHrxn and n were obtained from theintercept (DHion ¼ 0) and the slope of a plot of DHapp vs. DHion,respectively (Eq. (4)). Approximately, 1.6 protons were releasedduring FdUTP hydrolysis by PfdUTPase at 25 �C (data not shown).The hydrolysis enthalpy of FdUTP by this enzyme, DHrxn, taking thenumber of protons exchanged, n, into account was calculated as�6.7 � 0.8 kcal$mol�1 from Equation (4). These values are verysimilar to those obtained for dUTP hydrolysis (n ¼ 1.48;DHrxn ¼ �5.6 kcal$mol�1, Ref. [22]). Therefore, the replacement atC-5 of the uracil base of a hydrogen atom by a fluorine, the latterbeing a less bulky group than the dTTPmethyl group, does not seemto significantly affect the relative heat content of reactants andproducts, or the pKa of their functional groups.

IdUTP hydrolysis was further examined calorimetrically in thesame experimental conditions than FdUTP and dTTP. In agreementwith the previous observations an insignificant hydrolysis wasdetected (PfdUTPase: Km w 440 mM and kcat ¼ 0.002 s�1). Theseresults suggest that indeed the size of substituent at the C-5 posi-tion of the uracil base is a limiting factor for efficient binding.

3.2.1. End-product inhibitionIn order to examine further the inhibition of P. falciparum and

human dUTPases by dTMP and FdUMP, a calorimetric study was

, 25 �C) by the single injection assay. (A) Calorimetric trace obtained making a 15 mLNaCl, 1 mM b-mercaptoethanol and 25 mM MgCl2 at pH 7 and 25 �C). (B) Net thermal

parameters of the hydrolysis process. (C) Spectrophotometric trace recorded at 573 nmcording to the method described in Ref. [1] and the corresponding regression line.

A B

Fig. 5. Representative isothermal titration calorimetry measurements of the binding of dTMP to Plasmodium falciparum (A) and human (B) dUTPases. Titrations were performed in25 mM MES, 5 mM NaCl, 1 mM b-mercaptoethanol and 25 mM MgCl2 at pH 7 and 25 �C. Solutions 30 mM enzyme were titrated with 5 mL injections of 20 mM and 15 mM dTMP forplasmodial and human dUTPases, respectively. The calorimetric thermogram for human dUTPase (B) is similar to that for ligand dilution. The area under each peak for the binding ofdTMP to P. falciparum (A) and human (B) dUTPase was integrated and plotted against the molar ligand/enzyme ratio inside the calorimetric cell. The solid smooth line represents thebest fit of the experimental data to a non-cooperative model (bottom plots).

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186184

carried out using the same experimental conditions as those usedin the above described calorimetric kinetic assays (i.e. MES buffer,pH 7 and 25 �C). Two representative calorimetric experiments forthe binding of dTMP to both enzymes are shown in Fig. 5. Indeedonly PfdUTPase is able to bind dTMP, in agreement with the kineticstudies with dTTP. However, since the affinity of PfdUTPase fordTMP was considerably low, the determination of good thermo-dynamic parameters required an estimation of the number ofbinding sites [40]. Thus, a model with three equal [22,23,32] andnon-interacting binding sites fits adequately the calorimetric dataand the enthalpy change (DH) and binding constant, K, weredirectly obtained from the experimental titration curve (Table 2).The dissociation constant obtained this way was w400 mM, a valuesignificantly higher than that for dUMP (Kd ¼ 77 mM, Ref. [32]). Acomparison of the Km values suggests that binding of dTTP is alsoweaker in the case of hdUTPase than PfdUTPase. As can be seenfromTable 2, the binding of dTMP is both enthalpic and entropicallyfavoured at 25 �C, while dUMP and FdUMP binding were onlyenthalpically driven. However, although the binding of the threemonophosphate derivatives is enthalpically favoured, the entropiccomponent contributes substantially to dTMP binding. Positive

Table 2Apparent thermodynamic parameters from of binding of some monophosphate nucleoti

Nucleotide Kd (mM) DH (kcal$mol�1) DDH (k

P. f. human P. f. human P. f.

dUMP 77 � 12b 366 � 72 �8.8 � 2.0b �22.6 � 4.1 0FdUMP 47 � 5 189 � 21 �12.5 � 1.4 �19.8 � 5.2 �3.7 �dTMP 403 � 32 n.b. �1.1 � 0.2 n.b. 7.7 � 2dCMP, dUMP n.b. n.b. n.b n.b. e

dAMP, dGMP n.b. n.b. n.b n.b. e

n.b.: no binding detected.a The values represent the mean � the S.D. for at least three experiments.b Determined in Ref. [32].

entropy changes during binding are expected to result from theburial of hydrophobic groups from water [41]. Thus, binding ofdTMP and accommodation of the methyl group may requirea higher release of water molecules compared to dUMP and FdUMPgiving rise to a decrease in the number of interactions witha favourable enthalpic contribution.

The binding of FdUMP to PfdUTPase and hdUTPase was alsoinvestigated by ITC at 25 �C and pH 7 (Fig. 6). The enthalpy change,DH, and the binding constant, K, for the enzymeenucleotideinteraction were directly obtained from the experimental titrationcurve shown in the figure. From the fitting parameters a stoichi-ometry of 1:1 (per subunit) was obtained with K ¼ (2.1 � 0.2)$104 M�1 (Kd ¼ 47 mM), and DH ¼ �12.5 � 1.4 kcal mol�1 forPfdUTPase (Table 2). The corresponding values for the humanenzyme were Kd ¼ 189 � 21 mM and DH ¼ �19.8 � 5.2 kcal mol�1

(Table 2). Thus, FdUMP binding affinity for hdUTPase was w4-foldlower than for PfdUTPase. Although only moderate binding occurs,inhibition of human dUTPase by FdUMP an inhibitor of thymidylatesynthase [14], may be taken into account when considering themode of action of 5-fluorouracil (an anticancer agent) and its activemetabolite FdUMPwhich could inhibit both TS and dUTPase further

des to Plasmodium falciparum and human dUTPases at 25 �C and pH 7.0a.

cal$mol�1) TDSo (kcal$mol�1) TDDSo (kcal$mol�1)

human P. f. human P. f. human

0 �2.95 � 0.3b �17.8 � 2.1 0 03.4 2.72 � 9.3 �6.65 � 1.2 �14.8 � 0.5 �3.7 � 1.5 3.0 � 0.3.2 e 3.58 � 0.2 n.b. 6.53 � 0.5 e

e n.b. n.b. e e

e n.b. n.b. e e

A B

Fig. 6. Comparative of the isothermal titration calorimetry thermograms for the binding process of FdUMP to Plasmodium falciparum (A) and human (B) dUTPases. The titrationswere performed in 25 mM MES, 5 mM NaCl, 1 mM b-mercaptoethanol and 25 mM MgCl2 at pH 7 and 25 �C, making 5 mL injections of 4 mM FdUTP into solutions 30 mM enzyme.

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186 185

contributing to modifications in the dTTP/dUTP ratio and cellulartoxicity.

In summary, certain significant differences exist between thePlasmodium and human dUTPases with regard to substrate speci-ficity and product inhibition. These differences are based mainly onthe chemical nature and size of the substituent at the C-5 positionin the uracil base of the nucleotide. Thus, PfdUTPase appears moreefficient than hdUTPase in discriminating against substituents atthe C-5 position. These differences may be further exploitable inthe design of specific inhibitors of the enzyme with medicalapplications.

Acknowledgements

The authors acknowledge the Junta de Andalucía for financialsupport (Grant FQM-3141). This work was supported by researchgrant SAF2007-62596 from DGICYT (co-financed by FEDER), Min-isterio de Ciencia y Tecnología (Spain), the Plan Andaluz de Inves-tigaci�on (PAI groups BIO199 and FQM233), the Plan Propio fromUAL and the FIS Network RICET (RD06/0021). I. Quesada-Sorianowas the recipient of a fellowship from theMinisterio de Educaci�onyCiencia, Spain.

References

[1] G. Larsson, P.O. Nyman, J.O. Kvassman, Kinetic characterization of dUTPasefrom Escherichia coli. J. Biol. Chem. 271 (1996) 24010e24016.

[2] G. Larsson, L.A. Svensson, P.O. Nyman, Crystal structure of the Escherichia colidUTPase in complex with a substrate analogue (dUDP). Nat. Struct. Biol. 3(1996) 532e538.

[3] C.D. Mol, J.M. Harris, E.M. McIntosh, J.A. Tainer, Human dUTP pyrophospha-tase: uracil recognition by a b-hairpin and active sites formed by threeseparate subunits. Structure 4 (1996) 1077e1092.

[4] J. Kov�ari, O. Barab�as, B. Varga, A. B�ek�esi, F. T€olgyesi, J. Fidy, J. Nagy, B.G. V�ertessy, Methylene substitution at the a-b bridging position within thephosphate chain of dUDP profoundly perturbs ligand accommodation intothe dUTPase active site. Proteins 71 (2008) 308e319.

[5] O. Barab�as, V. Pongracz, J. Kov�ari, M. Wilmanns, B.G. Vertessy, Structuralinsights into the catalytic mechanism of phosphate ester hydrolysis bydUTPase. J. Biol. Chem. 279 (2004) 42907e42915.

[6] S. Chan, B. Segelke, T. Lekin, H. Krupka, U.S. Cho, M.Y. Kim, M. So, C.Y. Kim, C.M. Naranjo, Y.C. Rogers, M.S. Park, G.S. Waldo, I. Pashkov, D. Cascio, J.L. Perry,M.R. Sawaya, Crystal structure of the Mycobacterium tuberculosis dUTPase:insights into the catalytic mechanism. J. Mol. Biol. 341 (2004) 503e517.

[7] B. Varga, O. Barab�as, J. Kov�ari, J. T�oth, E. Hunyadi-Guly�as, E. Klement, K.F. Medzihradszky, F. T€olgyesi, J. Fidy, B.G. V�ertessy, Active site closure facili-tates juxtaposition of reactant atoms for initiation of catalysis by humandUTPase. FEBS Lett. 581 (2007) 4783e4788.

[8] B. Varga, O. Barab�as, E. Takacs, N. Nagy, P. Nagy, B.G. Vertessy, Active site ofmycobacterial dUTPase: structural characteristics and a built-in sensor. Bio-chem. Biophys. Res. Commun. 373 (2008) 8e13.

[9] J. T�oth, B. Varga, M. Kov�acs, A. M�aln�asi-Csizmadia, B.G. V�ertessy, Kineticmechanism of human dUTPase, an essential nucleotide pyrophosphataseenzyme. J. Biol. Chem. 282 (2007) 33572e33582.

[10] B.G. V�ertessy, J. T�oth, Keeping uracil out of DNA: physiological role structureand catalytic mechanism of dUTPases. Acc. Chemical Research 42 (2009)97e106.

[11] D. Mustafi, A. Bekesi, B.G. V�ertessy, M.W. Makinen, Catalytic and structuralrole of the metal ion in dUTP pyrophosphate. Proc. Natl. Acad. Sci. U.S.A 100(2003) 5670e5675.

[12] M.H. Gadsden, E.M. McIntosh, J.C. Game, P.J. Wilson, R.H. Haynes, dUTPpyrophosphatase is an essential enzyme in Saccharomyces cerevisiae. EMBO J.12 (1993) 4425e4431.

[13] H.A. Ingraham, M. Goulian, Deoxyuridine triphosphatase: a potential site ofinteraction with pyrimidine nucleotide analogues. Biochem. Biophys. Res.Commun. 109 (1982) 746e752.

[14] C. Heidelberg, Fluorinated pyrimidines and their nucleoside. in: A. Sartorelli,D.G. Johns (Eds.), Handbook of Experimental Pharmacology. Springer-Verlag,New York, 1974, pp. 193e231.

[15] S.J. Caradonna, Y.C. Cheng, The role of deoxyuridine triphosphate nucleotido-hydrolase, uracil-DNA glycosylase, and DNA polymerase alpha in the metab-olism of FUdR in human tumor cell. Mol. Pharmacol. 18 (1980) 513e520.

[16] H.A. Ingraham, B.Y. Tseng, M. Goulian, Mechanism for exclusion of 5-fluo-rouracil from DNA. Cancer Res. 40 (1980) 998e1001.

[17] B.K. Tye, P.O. Nyman, I.R. Lehman, S. Hochhauser, B. Weiss, Transient accu-mulation of Okazaki fragments as a result of uracil incorporation into nascentDNA. Proc. Natl. Acad. Sci. U.S.A 74 (1977) 154e157.

[18] I. Hoffmann, J. Widstr€om, M. Zeppezauer, P.O. Nyman, Overproduction andlarge-scale preparation of deoxyuridine triphosphate nucleotidohydrolasefrom Escherichia coli. Eur. J. Biochem. 164 (1987) 45e51.

[19] F. Hidalgo-Zarco, A.G. Camacho, V. Bernier-Villamor, J. Nord, L.M. Ruiz-P�erez,D. Gonz�alez-Pacanowska, Kinetic properties and inhibition of thedimeric dUTPase-dUDPase from Leishmania major. Protein Sci. 7 (2001)1426e1433.

I. Quesada-Soriano et al. / Biochimie 92 (2010) 178e186186

[20] C. Nguyen, G. Kasinathan, I. Leal-Cortijo, A. Musso-Buendia, M. Kaiser,R. Brun, L.M. Ruiz-P�erez, N.G. Johansson, D. Gonz�alez-Pacanowska, I.H. Gilbert, Deoxyuridine triphosphate nucleotidohydrolase as a potentialantiparasitic drug target. J. Med. Chem. 48 (2005) 5942e5954.

[21] J.A. Musso-Buendía, A.E. Vidal, G. Kasinthan, C. Nguyen, J. Carrero-L�erida, L.M. Ruiz-P�erez, K. Wilson, N.G. Johansson, I.H. Gilbert, D. Gonz�alez-Paca-nowska, Kinetic properties and inhibition of the dimeric dUTPase-dUDPasefrom Campylobacter jejuni. J. Enzym. Inhib. Med. Chem. 24 (2009) 111e116.

[22] I. Quesada-Soriano, I. Leal, J.M. Casas-Solvas, A. Vargas-Berenguel, C. Bar�on, L.M. Ruiz-P�erez, D. Gonz�alez-Pacanowska, L. García-Fuentes, Kinetic andthermodynamic characterization of dUTP hydrolysis by Plasmodium falcipa-rum dUTPase. Biochim. Biophys. Acta 1784 (2008) 1347e1355.

[23] J.L. Whittingham, I. Leal, C. Nguyen, G. Kasinathan, E. Bell, A.F. Jones, C. Berry,A. Benito, J.P. Turkenburg, E.J. Dodson, L.M. Ruiz-Perez, A.J. Wilkinson, G.N. Johansson, R. Brun, I.H. Gilbert, D. Gonz�alez-Pacanowska, K.S. Wilson,dUTPase as a platform for antimalarial drug design: structural basis for theselectivity of a class of nucleoside inhibitors. Structure 13 (2005) 1e10.

[24] E.S. Cedergren-Zeppezauer, G. Larsson, P.O. Nyman, Z. Dauter, K.S. Wilson,Crystal structure of a dUTPase. Nature 355 (1992) 740e743.

[25] C. Nguyen, G.F. Ruda, A. Schipani, G. Kasinathan, I. Leal, A. Musso-Buendia,M. Kaiser, R. Brun, L.M. Ruiz-P�erez, B.L. Sahlberg, N.G. Johansson, D. Gonzalez-Pacanowska, I.H. Gilbert, Acyclic nucleoside analogues as inhibitors of Plas-modium falciparum dUTPase. J. Med. Chem. 49 (2006) 4183e4195.

[26] M.M. Bradford, A rapid and sensitive method for the quantitation of micro-gram quantities of protein utilizing the principle of protein-dye binding.Anal. Biochem. 72 (1976) 248e254.

[27] M.J. Todd, J. Gomez, Enzyme kinetics determined using calorimetry:a general assay for enzyme activity? Anal. Biochem. 296 (2001) 179e187.

[28] T. Lonhienne, E. Baise, G. Feller, V. Bouriotis, C. Gerday, Enzyme activitydetermination on macromolecular substrates by isothermal titration calo-rimetry: application to mesophilic and psychrophilic chitinases. Biochim.Biophys. Acta 1545 (2001) 349e356.

[29] M.L. Bianconi, Calorimetric determination of thermodynamic parameters ofreaction reveals different enthalpic compensations of the yeast hexokinaseisozymes. J. Biol. Chem. 278 (2003) 18709e18713.

[30] H. Fukada, K. Takahashi, Enthalpy and heat capacity changes for the protondissociation of various buffer components in 0.1 M potassium chloride.Proteins 33 (1998) 159e166.

[31] R. T�ellez-Sanz, Z. Yassin, V. Bernier-Villamor, E. Ortiz-Salmer�on, J.A. Musso-Buendia, C. Bar�on, L.M. Ruiz-P�erez, D. Gonz�alez-Pacanowska, L. García-Fuentes, Effect of an Asp80Ala substitution on the binding of dUTP and dUMPto Trypanosoma cruzi dUTPase. Biochimie 89 (2007) 972e980.

[32] I. Quesada-Soriano, J.A. Musso-Buendia, R. Tellez-Sanz, L.M. Ruiz-P�erez,C. Bar�on, D. Gonz�alez-Pacanowska, L. García-Fuentes, Plasmodium falciparumdUTPase: studies on protein stability and binding of deoxyuridine deriva-tives. Biochim. Biophys. Acta 1774 (2007) 936e945.

[33] L. Cai, A. Cao, L. Lai, An isothermal titration calorimetric method to determinethe kinetic parameters of enzyme catalytic reaction by employing theproduct inhibition as probe. Anal. Biochem. 299 (2001) 19e23.

[34] H.J. Wiggers, J. Cheleski, A. Zottis, G. Oliva, A.D. Andricopulo, C.A. Montanari,Effects of organic solvents on the enzyme activity of Trypanosoma cruziglyceraldehyde-3-phosphate dehydrogenase in calorimetric assays. Anal.Biochem. 370 (2007) 107e114.

[35] G.S. Prasad, E.A. Stura, D.E. McRee, G.S. Laco, C. Hasselkus-Light, J.H. Elder, C.D. Stout, Crystal structure of dUTP pyrophosphatase from feline immuno-deficiency virus. Protein Sci. 5 (1996) 2429e2437.

[36] Z. Dauter, R. Persson, A.M. Rosengren, P.O. Nyman, K.S. Wilson, E.S. Cedergren-Zeppezauer, Crystal structure of dUTPase from equine infec-tious anaemia virus; active site metal binding in a substrate analoguecomplex. J. Mol. Biol. 285 (1999) 655e673.

[37] J. Kov�ari, O. Barab�as, E. Takacs, A. B�ek�esi, Z. Dubrovay, V. Pongracz, I. Zagyva,T. Imre, P. Szabo, B.G. Vertessy, Altered active site flexibility and a structuralmetal-binding site in eukaryotic dUTPase: kinetic characterization, foldingand crystallographic studies of the homotrimeric Drosophila enzyme. J. Biol.Chem. 279 (2004) 17932e17944.

[38] J. Nord, G. Larsson, J.O. Kvassman, A.M. Rosengren, P.O. Nyman, dUTPase fromthe retrovirus equine infectious anaemia virus: specificity, turnover andinhibition. FEBS Lett. 414 (1997) 271e274.

[39] O. Bj€ornberg, P.O. Nyman, The dUTPases from herpes simplex virus type 1and mouse mammary tumour virus are less specific than the Escherichia colienzyme. J. Gen. Virol. 77 (1996) 3107e3111.

[40] W.B. Turnbull, A.H. Daranas, On the value of c: can low affinity systems bestudied by isothermal titration calorimetry? J. Am. Chem. Soc. 125 (2003)14859e14866.

[41] J.M. Sturtevant, Heat capacity and entropy changes in processes involvingproteins. Proc. Natl. Acad. Sci. U.S.A. 74 (1977) 2236e2240.