Computational, spectroscopic, and resonant mirror biosensor analysis of the interaction of adrenodoxin with native and tryptophan-modified NADPH-adrenodoxin reductase

Computational, spectroscopic, and resonant mirror biosensor analysis of the interaction of adrenodoxin with native and tryptophan-modified NADPH-adrenodoxin reductase
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Computational, Spectroscopic, and Resonant MirrorBiosensor Analysis of the Interaction of Adrenodoxin WithNative and Tryptophan-Modified NADPH-AdrenodoxinReductase  Yelizaveta Sargisova, 1 Francesco-Maria Pierfederici, 2  Andrea Scire`, 3 Enrico Bertoli, 3 Fabio Tanfani, 3,4 * Ferdinando Febbraio, 5 Raffaella Briante, 5  Yelena Karapetyan, 1 and Sona Mardanyan 11  Institute of Biochemistry of Armenian NAS, Yerevan, Armenia 2  Department of Chemical Science University of Catania, Catania, Italy 3  Institute of Biochemistry, Universita` Politecnica delle Marche, Ancona, Italy 4  INFM Ancona, Universita` Politecnica delle Marche, Ancona, Italy 5  Institute of Protein Biochemistry, CNR, Napoli, Italy  ABSTRACT   Insteroidhydroxylationsysteminadrenal cortex mitochondria, NADPH-adrenodoxinreductase (AR) and adrenodoxin (Adx) form a shortelectron-transport chain that transfers electronsfrom NADPH to cytochromes P-450 through FAD in AR and [2Fe-2S] cluster in Adx. The formation of [AR/Adx]complexisessentialfortheelectrontrans-fer mechanism in which previous studies suggestedthat AR tryptophan (Trp) residue(s) might be impli-cated. In this study, we modified AR Trps by N-bromosuccinimide (NBS) and studied AR binding to Adxbyaresonantmirrorbiosensor.Chemicalmodi-ficationoftryptophanscausedinhibitionofelectrontransport. The modified protein (AR*) retained thenative secondary structure but showed a loweraffinity towards Adx with respect to AR. Activitymeasurements and fluorescence data indicated thatone Trp residue of AR may be involved in theelectrontransferringactivityoftheprotein.Compu-tational analysis of AR and [AR/Adx] complex struc-tures suggested that Trp193 and Trp420 are theresidues with the highest probability to undergoNBS-modification. In particular, the modification of Trp420 hampers the correct reorientation of AR*molecule necessary to form the native [AR/Adx]complex that is catalytically essential for electrontransfer from FAD in AR to [2Fe-2S] cluster in Adx.The data support an incorrect assembly of [AR*/  Adx] complex as the cause of electron transportinhibition.Proteins2004;57:302–310.  ©  2004Wiley-Liss,Inc. Key words: NADPH-adrenodoxinreductase;adreno-doxin;protein–proteininteraction;elec-tron transport; resonant mirror biosen-sor;structuralanalysisINTRODUCTION  A permanent great interest of researchers is directedtowards the biological electron-transfer pathways in- volvedinthevitalimportanceofmetabolicprocesses. 1 Oneof the most studied pathways is represented by the steroidhydroxylation system in the adrenocortical mitochondria.ItconsistsofaFAD-dependentNADPH-adrenodoxinreduc-tase (AR), a ferredoxin of [2Fe-2S] class, adrenodoxin(Adx), and a cytochrome P450 (CYP11A1, CYP11B1, orCYP11B2). 2  AR and Adx form a short electron-transportchain that transfers electrons from NADPH to Cyt. P-450:NADPH 3   AR 3   Adx 3  P-450Site-directed mutation experiments 3 showed that Arg240and Arg244 in AR, and Asp76 and Asp79 in Adx, are thebasic key residues stabilizing electrostatically the complexformation between the two proteins. Recently, the 2.3 Å resolution crystal structure of a cross-linked [AR/Adx]complex has been obtained, 4 and the electron-tunnelingrate between FAD in AR and the [2Fe-2S] cluster in Adxwas evaluated to be 10 8 -10 9 s  1 . On the other hand, theelectron transfer rate between the two redox centers, asdetermined experimentally, resulted in 3–4 s  1 . 5 Thisdiscrepancy indicates that some structural peculiaritieswere not taken into account in the evaluation of theelectron-tunneling rate. It is possible that the presence of any group(s), in the space between electron donor andacceptor centers, might influence the rate of electrontransport. Aromatic amino acids are known to be potentially activein intramolecular charge transfer. 6–8 In AR, six Trpresidues with different solvent accessibility are present.Early fluorimetric investigations demonstrated the close-  Abbreviations:  Adx, adrenodoxin; AR, NADPH-adrenodoxin reduc-tase; NBS, N-bromosuccinimide; AR*, AR treated with NBS; Cyt c,cytochrome  c ; EDC, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide;NHS, N-hydroxysuccinimide; KPi, potassium phosphate buffer.Grant sponsor: NATO; Grant number: LST.CLG.977079; Grantsponsor: Universita` Politecnica delle Marche; Grant sponsor: Ministryof Education and Science of Armenia; Grant number: 2000-257.*Correspondence to: Prof. Fabio Tanfani, Institute of Biochemistry,Faculty of Science, Universita` Politecnica delle Marche, Via Ranieri,60131 Ancona, Italy. E-mail: f.tanfani@univpm.itReceived 25 August 2003; Accepted 4 March 2004Published online 11 June 2004 in Wiley InterScience(www.interscience.wiley.com). DOI: 10.1002/prot.20174PROTEINS: Structure, Function, and Bioinformatics 57:302–310 (2004)  ©   2004 WILEY-LISS, INC.  ness of Trp(s) in AR to FAD and Adx binding sites, 9 andtheARcrystalstructurestudysuggestedthedisplacementof Trp367 in the Adx binding interface of AR. 10 To studythe possible participation of Trp in AR activity, the use of chemical modifications of amino acid residue(s) is of greatrelevance.Inourearlyexperiments, 11 thechemicalmodifi-cation of AR with N-bromosuccinimide (NBS) resulted inTrp residues oxidation, accompanied by inhibition of elec-trontransferfromARtoAdx.Thepossiblewaysofelectrontransfer inhibition due to Trp(s) oxidation may be due tothe involvement of these amino acid residues in theintra-complex electron transfer or to AR conformationalchanges affecting [AR/Adx] complex formation and/or theFADmicroenvironment.Wemayexcludethedirectinvolve-ment of Trp residues in the intermolecular [AR/Adx]complex formation, because the electrostatic nature of interaction between the two proteins, involving specificamino acids, has been clearly demonstrated. 3,4 In this study, a controlled chemical modification of Trpsin AR was performed, and the number of Trp residuesessential for AR electron transferring activity was deter-mined. The most probable NBS-modifiable Trp residue(s)was identified by computational analysis of AR and [AR/  Adx] structures. The parameters of [AR*/Adx] and [AR/  Adx]complexformationwereobtainedbyusingaresonantmirror biosensor. CD and fluorescence spectroscopy wereused to analyze the secondary structure of AR and AR*and the fluorescent properties of the [AR*/Adx] and [AR/  Adx] complexes, respectively. From this study, we couldconclude that one Trp residue in AR is important for theelectron transferring activity of the protein and that theelectron transfer inhibition due to oxidation of solvent-exposed Trp(s) is caused by an incorrect assembly of [AR*/Adx] complex. MATERIALS AND METHODSMaterials Carboxymethyl dextran dual well cuvettes and thesuitable coupling kit, consisting of 1-ethyl-3-(3-dimethyl-aminopropyl)carbodiimide(EDC)andofN-hydroxysuccin-imide (NHS), were from Affinity Sensors (Thermo Lab-systems, Cambridge, UK). BrCN-Sepharose, used for Adx-Sepharose preparation, was from Sigma (St. Louis, MO).Cytochrome  c  (Cyt c) was purified from bovine heart, andNBS was synthesized from succinimide and Br 2  as previ-ously reported. 11  All the other chemicals used were commercial samplesof the purest quality. PreparationofAdxandAR  Adx and AR from adrenal cortex were isolated andpurified as previously described. 12,13 The last stage of ARpurification included the affinity chromatography on Adx-Sepharose. The purity indexes of oxidized proteins were A  415  /A  280  0.8 for Adx, and A  280  /A  450  7.0 for AR. Theirconcentrations were determined from the extinction coeffi-cients of 9.8 mM  1 cm  1 at 414 nm for Adx, and 10.9mM  1 cm  1 at 450 nm for AR. 14 EnzymeAssay The electron transferring activity of AR was evaluatedin a reconstituted electron carrier chain consisting of NADPH as an electron source of the electron carriers ARand Adx, and Cyt. c as a model of terminal acceptor of electrons. The incubation mixture in 0.4 ml of 20 mMpotassium phosphate (KPi) buffer, pH 7.4, contained: 10  M Cyt c, 0.2  M AR or AR*, 0.2  M Adx; the reaction wasinitiated by NADPH addition up to 10 mM. The time-dependent increase of reduced Cyt c absorption at 550 nmwas registered. The amount of reduced Cyt c was calcu-lated using the differential extinction coefficient (reduced-oxidized) 19.0 mM  1 cm  1 at 550 nm. 14 SpectralMeasurements UV-VIS spectrophotometers (2401PC Shimadzu, Japanor Specord M-40, Germany) and quartz cuvettes of 0.5 or 1cm light path, in a thermostated holder, were used forspectral and kinetic measurements. Fluorimetric and CDdata were obtained using a MPF-44 spectrofluorimeter(Perkin-Elmer) and a Mark M-III dichrograph (Jobin- Yvon, France), respectively. The optical density of fluori-metric samples did not exceed 0.1 to exclude the internalquenching. TreatmentofARWithNBS The treatment of AR with NBS was described ear-lier 11,15 as follows: small aliquots of 2.25 mM NBS stocksolution in water were added to 0.4 ml of 2–5   M ARsolution in 10 mM KPi buffer, pH 7.4, and the absorptionat 280 nm was registered. The measured value wascorrected,accountingforthedilutionoftheproteinsampledue to the reagent addition. The number of modifiedresidues per mol of protein, n, was estimated from theabsorption drop at 280 nm using the equation of Spandeand Witkop. 16 n  1.31  ε 280    A  280  /5500   A  280  (1)where   280  and A  280  are the molar extinction and theinitialintensityofproteinsolutionat280nm,respectively;   A  280  is the decrease of absorption at 280 nm in responseto NBS treatment; 1.31 is an empirical coefficient account-ing the oxyindol absorption at 280 nm; 5,500 is the molarextinction of Trp at 280 nm. BindingStudies Binding of AR or AR* to Adx was carried out using theresonant mirror biosensor “IAsys plus” (Thermo Lab-systems, Cambridge, UK). A two-well cuvette of the car-boxymethyldextran(CMD)typewasused.Immobilizationof Adx to CMD was performed as described by Davies etal. 17 Carboxyl groups of CMD were activated with EDC/ NHS and then Adx was added in one well at pH 4.6(working well), while the second well was used as a control(control well) to check unspecific binding. Unreacted car-boxyl groups were blocked by 1M ethanolamine, pH 8.5.Finally, the wells were washed several times with 20 mMKPi buffer, pH 7.4. The amount of Adx covalently linked tothe CMD matrix was 8 ng/mm 2 (5.7    10  13 mol/mm 2 ). INTERACTION OF AR AND Adx  303  The control well was subjected to the same treatment asthe working well, but without the step of Adx immobiliza-tion. All the experiments were carried out at 25°C. Run-ning and regeneration buffers were 20 mM KPi, pH 7.4,without and with 300 mM KCl, respectively. The bindingstudieswerecarriedoutbyaddingintothetwowellsofthecuvettesolutionsofARorAR*inthe52–560or52–300nMconcentration ranges, respectively. Higher concentrationsdetermined some aggregate formation and the consequentimpossibility of appraisal of the real binding.It must be emphasized that we chose the pH of mediumfor Adx immobilization, taking into account that the activecenter of Adx consists of [2Fe-2S] cluster, where 2S atomsareacidlabile. 12  AtpHlowerthan7.0,theclustermaylosethe sulfur atoms resulting in distortion of the activecenter. To prevent this, we checked the Adx immobiliza-tion to CMD matrix of biosensor cuvette at different pHsand tested its activity. The best conditions that promoted Adx linking to the CMD matrix and maintained the Adxactivity were at pH 4.6.Kinetics of Adx-AR or Adx-AR* interactions at thesurface of the cuvette were analyzed using FASTfit andGRAfitsoftwaressuppliedwiththeIAsysplusinstrument.The apparent on-rate constants, k on , was calculated foreach association curve related to different AR or AR*concentrations using the monophasic kinetics of binding.Monophasic equation contains one exponential term 18,19 :R t   R   R 0  1  exp   k on t   R 0  (2)In this equation, R t  is the response at time t, R 0  is theinitial response, and R   is the maximal response. The value of k on  varies with ligate concentration [L] (in ourcase AR or AR*) as described by the equation 18,19 :k on  k diss  k ass [L] (3)Thus, a plot of k on  against [L] allows the kinetic parame-ters of complex formation, the association (k ass ) and disso-ciation (k diss ) rate constants, to be determined from theslope and the intercept with the ordinate axis, respec-tively. The dissociation equilibrium constant K  D , can beevaluated using the ratio 18,19 :K  D  k diss  /k ass  (4)The K  D  and k diss  constants can be obtained also as follows.The FASTfit software, extrapolating the data concerningthe association curves at different AR or AR* concentra-tions, calculates the equilibrium response (R eq ). PlottingR eq  versus AR or AR* concentration, a curve, described bythe following equation, is obtained 19 :R eq   R max [L]   /   K  D   L   (5)In this equation, R max  is the response when the immobi-lized ligand (Adx) is saturated with ligate L (AR or AR*)and K  D  is the value of [L] at R max  /2. Once the K  D  valueswere obtained from equation [5], the dissociation con-stants, k diss , were determined using the equation [4]. ComputationalAnalysis The3DstructuresofAR(PDBaccessionnumber1CJC) 10 and the [AR/Adx] complex (PDB accession number 1E1E) 4 resolved by X-ray at 1.7 and 2.3 Å, respectively, wereobtainedfromtheProteinDataBank(http://www.rcsb.org/ pdb/). As reported by Mu¨ller et al., 4 two [AR/Adx] com-plexes, related by a non-crystallographic screw rotation,are present in the asymmetric unit. Complex I containsresidues 5–117 of Adx and 4–460 of AR (AR1), andcomplex II contains residues 5–110 of Adx and 5–460 of  AR (AR2). All the AR structures were superimposed and surfaces,accessibility, volumes, and cavities were determined usingthe program Swisspdb 3.7. 20 Maximum accessibility, in the program Swisspdb, isdefined as being the accessible surface of an amino acid X in a pentapeptide GGXGG in extended conformation. Thisis only an approximate scale, but perfectly sufficient todifferentiate core amino acids from surface ones.The temperature factor (B-factor) of Trp residues hasbeen determined as a media of B-factors of any side-chainsatom.The program molmol 21 was used to calculate the ex-posed solvent surface (ESS) for each amino acid residues,using a solvent radius 1.4 Å and a precision of 3 (good ratiobetween data accuracy and analysis rate). RESULTS AND DISCUSSIONNumberofChemicallyModifiedTrpResiduesandTheirInvolvementinARActivity ThenumberofmodifiedTrpresidueswasobtainedusingequation 1, obtaining a degree of modification of 0.3, 0.5,0.8, 1.5, 2.3, and 2.7 Trp residues for six different samplesof AR treated as described in Materials and Methods. Trpmodification in AR inhibits the electron transfer in thereconstituted model electron carrier chain (NADPH 3   AR 3   Adx 3  Cyt. c ). The number of Trp residues involved inthe AR activity were evaluated measuring the residualactivity at different degrees of Trp(s) modification (Fig. 1).Theextrapolationoftheinitialslopeindicatesthatoneoutof 3 modifiable Trps is important for the electron transfer-ring activity of AR from NADPH to Adx. SecondaryStructureofNBS-TreatedAR The treatment of AR with NBS led to the spectrallyobservable chemical modification of Trp residues, 11,22 ex-tent of which could be evaluated using equation [1]. Theintegrity of possible NBS target amino acids (Cys, Tyr,Lys, Arg) in the conditions of our experiments had beenprovedpreviously. 15 Neithertheabsorptionnorthefluores-cence spectra of FAD were influenced by NBS treatment.Should the treatment remove the prosthetic group fromthe protein molecule, an increase of the fluorescenceintensity would be registered. The stability of both thesespectra 22 was evidence of the stability of the FAD attach-ment to the protein molecule during the NBS treatment(spectranotshown).CDspectraofnativeandNBS-treated AR were exactly the same, thus allowing us to rule out the 304  Y. SARGISOVA ET AL.  probability of secondary structure alteration due to Trpsmodification (spectra not shown). Fluorescence Study of the NBS-Treated AR Previously, the quenching of fluorescence both of FADandTrpresiduesinholo-ARhasbeenshown. 9 Thequench-ing of the emission of both fluorophores was also observedin[AR/Adx]complex 9,23 suggestingthatinARTrpsarenotfar from both FAD and Adx binding sites and that someconformational changes took place in AR as a consequenceof [AR/Adx] complex formation. 24 In Figure 2, Stern-Volmer plots are presented for Trpfluorescence quenching by Adx in native and NBS-modified AR. The slope of a linear Stern-Volmer plot isproportional to the constant of bimolecular interaction(K  Q ) between a quencher and a fluorophore. 25 The plots inFigure 2 demonstrate the decreasing of this parameterafter AR treatment by NBS. Moreover, the experimentalpoints for AR modified up to 0.8, 1.5, or 2.7 oxidized Trpsper mol of AR fit the same linear plot. It is likely that thesingle Trp residue modified by a low NBS concentration isthe most solvent-exposed. Hence, the identification of sucharesidueshouldhelptounderstandwhyahigherdegreeof Trp modification does not change K  Q . Binding Studies The binding of AR and AR* to Adx was analyzed by aresonant mirror biosensor. AR* with a mean of 1.5 mol of oxidized Trp per mole of protein was used. The remaining electron-transferring activity of AR* was about 50% withrespect to the native protein.Both AR and AR* bind specifically to Adx covalentlylinked to the CMD matrix of the cuvette. Figure 3 showsthebindingofARtoAdx.AssoonasARwasaddedintothecuvette, a sudden increase of the signal occurred (point 1).Washing the cuvette with running buffer led to a partialdissociation of [AR/Adx] complex (point 2). Washing thecuvette with regenerating buffer (point 3), followed bywashing with running buffer (point 4), led to the completeremoval of AR, as indicated by the same baseline observedbefore the beginning of the experiment. The ability of 300mM KCl containing buffer to disassemble the [AR/Adx]complex proves its stabilization by ionic interactions. Thisis perfectly in agreement with the data reported by otherauthors, evidence of the electrostatic nature of Adx-ARinteraction. 4,26,27 No unspecific binding of AR or AR* tothe CMD matrix occurred, as shown by the flat responseobtained from the control well of the cuvette. Fig. 1. AR activity as a function of the number of NBS-modified Trps.AR activity and number of NBS-modified Trps were calculated asdescribed in Materials and Methods.Fig. 2. Stern-Volmer Plots of AR fluorescence quenching by Adx.NativeAR(solidcircles)andAR*where0.8(squares),1.5(triangles),and2.7 (open circles) Trp residues per mol of protein are modified. TheNBS-induced Trp oxidation is shown. Wavelengths of excitation andemission were, respectively, 295 and 334 nm.Fig. 3. Binding of AR to Adx immobilized on CMD surface. Baselinecorresponds to 20 mM Potassium-Phosphate buffer pH 7.4 (runningbuffer). Dotted line is the response from control well. Continuous linerepresents the response from the cuvette well where Adx was immobi-lized. Point 1: addition of AR to a final concentration of 182 nM in bothwells of cuvette. Point 2: washing with running buffer. Point 3: washingwith 20 mM Potassium-Phosphate, 300 mM KCl buffer, pH 7.4 (regenera-tion buffer). Point 4: washing with running buffer. INTERACTION OF AR AND Adx  305  The kinetic and equilibrium constants for the interac-tion of Adx with AR or AR* were calculated as described inMaterials and Methods. In particular, Figure 4 shows theoverlay plots of the association profiles of AR and AR* toCMD-bound Adx. The best fit to these data was obtainedusingthemonophasicequation2,fromwhichtheapparentrate constant, K  on , could be calculated for each AR or AR*concentration.Inadditiontotheoriginalandfittedassocia-tion curves, Figure 4 also reports the error, the low valuesof which indicate the goodness of the fittings. Figure 5A and B reports the plots of K  on  against the AR or AR*concentration. K  ass  could be calculated from the slope andK  diss  from the intercept with the ordinate axis (equation3). The data reported in Figure 4 were used by theIAsys-plus software to calculate also the equilibrium re-sponse as a function of the AR or AR* concentration and togenerate the plot reported in Figure 5C and D. K  D  valueswere obtained from the plot using equation 5, while thedissociation constant K  diss  was obtained from equation 4.Table I reports the values of K  D , K  ass , and K  diss  for theinteraction of AR or AR* to CMD-bound Adx.The K  D  values are consistent with those reported bydifferent scientific groups that show that the order of magnitude, for K  D  in AR-Adx interaction, is in the 10  8 –10  9 M range. 26,28–33 The increase of K  D  up to 2–6  10  6 M was reported in a study concerning the interaction of native AR with apo-adrenodoxin. 5 It is worth noting thatthe same order, 1.1 10  6 M, was obtained in a biosensorstudy of Adx-AR interaction, 34 suggesting that the condi-tions used in that work probably were not ideal to main-tain the native form of Adx. It is possible that the lowering of K  D  value was due to [2Fe-2S] cluster distortion in Adxduring the immobilization of the protein in the biosensorcuvette at pH 4, a more extreme condition than thatchosen in our experiment (pH 4.6).The data reported in Table I show small but significantdifferences in the K  D  values calculated for AR and AR*.Since K  D  depends on K  diss  and K  ass  (equation 4), its higher value in the Adx/AR* system than in the Adx/AR systemcould be due to an increase in K  diss  or to a decrease of K  ass .Table I shows that both K  diss  and K  ass  are lower in the Adx/AR*systembutthatK  ass decreasedmorethank diss asshown by k ass  AR/k ass  AR* and k diss  AR/k diss  AR* ratios.Hence, the data indicate that the modification of Trpsreduces the association rate constant more than the disso-ciation rate constant. In turn, these data suggest that theTrps modification led to some structural changes impor-tant for protein–protein interactions. These indicationswere checked by computational analysis of the structuresof AR and [AR/Adx] complex. Computational Analysis Both Adx and AR are two-domain proteins. Adx consistsof a core domain containing the [2Fe-2S] cluster and asmall interaction domain 35 ; AR contains a FAD domainand a NADPH domain of about equal size. 10 In order to identify the most probable NBS-modifiableTrps in AR, we calculated the exposed solvent surface Fig.4. OverlayofassociationdataforAR( A )orAR*( B )binding to immobilized Adx.Eachassociation curve wasfittedusing amonophasic equation(eq. 1). Bold and thin lines represent experimental and fitted curves, respectively. The AR and AR* concentration used are reported. The residual errorreported in each panel refers to the association curves obtained with 182 nm AR or AR* concentration, and gives an estimate of the fitting goodness.Similar errors were obtained with other protein concentrations. 306  Y. SARGISOVA ET AL.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks