Báo cáo khoa học: Tryptophan 243 affects interprotein contacts, cofactor binding and stability in D-amino acid oxidase from Rhodotorula gracilis ppt

Tryptophan 243 affects interprotein contacts, cofactor binding and stability in D-amino acid oxidase from Rhodotorula gracilis Laura Caldinelli, Gianluca Molla, Mirella S. Pilone and Loredano Pollegioni Department of Biotechnology and Molecular Sciences, University of Insubria, Varese, Italy Oligomeric proteins are the rule rather than the excep- tion. As recently reviewed by Mei et al. [1], a surpris- ingly high number of proteins are made up of two subunits and, in most of these ( 80%), the two sub- units are identical. Recently, we investigated the role of subunit interaction in homodimeric proteins by using the enzyme d-amino acid oxidase (DAAO) (EC 1.4.3.3), from yeast, as a model tool. DAAO contains, as a coenzyme, a noncovalently bound molecule of FAD per 40 kDa protein monomer. DAAO is a com- ponent of the glutathione reductase family, the FAD- containing protein family that has been studied in greatest detail. The protein members of this populous class catalyze diverse reactions and adopt the Ross- mann fold [2]. DAAO belongs to the second sub- family, GR 2 , whose members only align well at their N terminus ( 30 residues) [3]. Two major structural features have been proposed to be responsible for the ‘head-to-tail’ mode of dimeriza- tion of yeast DAAO (Fig. 1A–C), namely (a) electro- static interactions between the positively charged residues of the long loop (from Pro302 to Glu322), con- necting b-strands F5 and F6 and the negatively charged residues belonging to the a-helices I3¢ and I3¢¢ of the other monomer and (b) the interactions in the core of the dimer (where Trp243 is located, see below) [4–6]. The importance of the bF5–bF6 loop for the mono- mer–monomer interaction has been validated by site- directed mutagenesis, when a stable monomeric holoenzyme was obtained by removing part of this loop Keywords cofactor binding; flavoprotein; oligomerization state; rational design; structural stability Correspondence L. Pollegioni, Dipartimento di Biotecnologie e Scienze Molecolari, Universita ` degli Studi dell’Insubria, Via J. H. Dunant, 3–21100 Varese, Italy Fax: +39 0332 421500 Tel: +39 0332 421506 E-mail: loredano.pollegioni@uninsubria.it (Received 10 October 2005, revised 18 November 2005, accepted 2 December 2005) doi:10.1111/j.1742-4658.2005.05083.x The flavoenzyme d-amino acid oxidase from Rhodotorula gracilis is a homodimeric protein whose dimeric state has been proposed to occur as a result of (a) the electrostatic interactions between positively charged resi- dues of the bF5–bF6 loop of one monomer and negatively charged residues belonging to the a-helices I3¢ and I3¢¢ of the other monomer, and (b) the interaction of residues (e.g. Trp243) belonging to the two monomers at the mixed interface region. The role of Trp243 was investigated by substituting it with either tyrosine or isoleucine: both substitutions were nondisruptive, as confirmed by the absence of significant changes in catalytic activity, but altered the tertiary structure (yielding a looser conformation) and decreased the stability towards temperature and denaturants. The change in conformation interferes both with the interaction of the coenzyme to the apoprotein moiety (although the kinetics of the apoprotein–FAD complex reconstitution process are similar between wild-type and mutant d-amino acid oxidases) and with the interaction between monomers. Our results indicate that, in the folded holoenzyme, Trp243 is situated at a position optimal for increasing the interactions between monomers by maximizing van der Waals interactions and by efficiently excluding solvent. Abbreviations ANS, 1,8-anilinonaphtalene sulfonic acid; C m , midpoint concentration of urea required for unfolding; DAAO, D-amino acid oxidase; T m , melting temperature. 504 FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS [7,8]. Thermodynamic studies performed on monomeric and dimeric DAAO forms showed that the shift to such a monomeric form resulted in an enzyme with increased sensitivity to thermal denaturation [6]. The importance of Trp243 in the monomer–mono- mer interaction was envisaged by comparing the near- UV CD spectra and the tryptophan fluorescence between dimeric wild-type and monomeric DLOOP proteins [6]: the observed differences may be indicative of changes in the environment of Trp243 following deletion of the bF5–bF6 loop and conversion to the monomeric state because no other tryptophan residues are buried on the surface of each monomer of the DAAO dimer (Fig. 1B,C). Although the atoms of the Trp243 side chain in each subunit are too remote to allow a direct and strict contact between the two residues at the dimer interface (the distance between the a-carbons is  3.7 A ˚ and between the indole rings is ‡ 7.7 A ˚ ; see Fig. 1C), they account for a significant region of the dimerization surface. In order to gain further insight into the role of the structural determinants of this prototypical flavopro- tein, we investigated, by site-directed mutagenesis, the influence of Trp243 on the functionality, the oligo- meric state and the FAD binding of DAAO. Results Spectral properties of Trp243 DAAO mutants Upon purification, the DAAO mutant, W243Y, was found to retain the FAD coenzyme, as confirmed by Fig. 1. Structural features of Rhodotorula gracilis D-amino acid oxidase (DAAO) (PDB code: 1c0p). (A) ‘Head-to-tail’ mode of monomer–mono- mer interaction: the two subunits are presented in different colors (a, green; b, blue). The bF5–bF6 loop is depicted in red, the flavin cofactor is in yellow, and the substrate D-alanine is in purple. (B) Details of the residues involved in the monomer–monomer interaction at the ‘mixed interface’ (blue, positively charged residues; red, negatively charged residues; white, apolar residues). For the sake of clarity, the two sub- units have been separated. (C) Positioning of Trp243 (yellow) in the buried region at the dimer interface (no other tryptophan residues are located in the buried surface between monomers). Amino acids belonging to the A-chain are labeled in black, while residues belonging to the B-chain are labeled in blue. (D) Details of the structure depicting the interactions of strand F5 (blue), which follows the loop containing Arg289 (the site of trypsinolysis in the unfolding intermediate) with the bab elements of the Rossmann fold (green), the C-terminal a-helix containing the SKL sequence for peroxisomal targeting (red), and the region containing Trp243 (the loop following the bI7–aI3, dark blue). L. Caldinelli et al. Structural determinants for dimerization of DAAO FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS 505 the absorption spectrum in the visible region (an E 274 nm ⁄ E 455 nm ratio of  9.6 and an extinction coef- ficient, at 455 nm, of 12 800 m )1 Æcm )1 was deter- mined). In contrast, the W243I mutant was found to exhibit a remarkably higher E 274nm ⁄ E 455nm ratio (ran- ging from 16 to 25, depending on the enzyme pre- paration), indicating a significantly lower content of flavin cofactor. As both mutants showed a purity of ‡ 95% on SDS ⁄ PAGE, the latter result indicates that the W243I mutant is purified as a mixture of holo- and apoprotein. This conclusion is further supported by the observation that a significantly higher enzymatic activity is measured for the W243I mutant after adding exogenous FAD to the assay mixture. The apoprotein forms of W243I and W243Y DAAOs were obtained by dialysis, where a high yield was obtained (‡ 60% of protein recovery). The binding constant for the coenzyme was determined by titrating the apoprotein with increasing amounts of FAD and following the decrease in protein fluorescence: a K d value of 0.37 and 0.035 lm of the apoprotein–FAD complex was obtained for the W243I and W243Y mutants, respectively. The value determined for W243I is 20-fold higher than obtained for the wild-type RgDAAO protein [9] and fourfold higher than those values determined for the monomeric DLOOP DAAO mutant (Table 1) [7,8]. Far-UV CD spectra of the wild-type protein, and of DLOOP and Trp243 DAAO mutants, did not reveal any major change in the features related to the secondary structure of the two mutants. In con- trast, and as shown in Fig. 2, the near-UV CD spec- tra of the different DAAO forms under investigation are different. These alterations may be ascribed to a different contribution from aromatic amino acid resi- dues, which are responsible for most transitions in the near-UV spectral region, in particular in terms of altered mutual relationships between nearby struc- tural elements. Alterations in protein conformation are also made evident by changes in the fluorescence of the indole moiety of tryptophan [10]: the fluores- cence emission at 345 nm (following excitation at 280 nm) was higher for both W243Y and W243I mutants than for dimeric wild-type and monomeric DLOOP DAAOs (Table 1). A sensitive marker of the folding state of a flavo- enzyme is represented by the fluorescence of the FAD cofactor, which is much lower in the holoenzyme- bound form of DAAO with respect to free FAD [9]. Table 1. Comparison of dissociation constants determined for FAD, apparent kinetic parameters and fluorescence spectral properties for wild-type and mutants of D-amino acid oxidase (DAAO). a.u., Arbitrary units determined under the same experimental conditions, at a protein concentration of 0.1 mgÆmL )1 . Enzyme form K FAD d (lM) D-Alanine Fluorescence properties V max (lmol O 2 Æmin )1 Æmg )1 protein) K m (mM) Protein a.u. (nm) Flavin a.u. Wild type 0.02 ± 0.01 a 110 ± 8 b 0.8 ± 0.2 b 135 (334) 12.7 (apoprotein) 940 (342) DLOOP c 0.12 ± 0.01 86 ± 2 1.5 ± 0.2 230 (338) 18.9 (apoprotein) 900 (342) W243I 0.37 ± 0.03 95 ± 2 d 1.6 ± 0.1 d 340 (338) 9.4 (apoprotein) 465 (340) W243Y 0.04 ± 0.01 108 ± 2 1.4 ± 0.1 370 (337) 12.3 (apoprotein) 715 (341) a As described previously [9]. b As described previously [15]. c As described previously [7]. d These values were determined in the presence of a large excess of free FAD (0.2 m M) in the activity assay mixture. Fig. 2. Near-UV CD spectra of wild-type (—), DLOOP (– - –), W243I (- - - -), and W243Y (– – –) D-amino acid oxidases (DAAOs). The pro- tein concentration used was 0.4 mgÆmL )1 in 50 mM potassium phosphate, pH 7.5, containing 10% (v ⁄ v) glycerol and 2 m M EDTA. Structural determinants for dimerization of DAAO L. Caldinelli et al. 506 FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS The emission at 526 nm, following excitation at 450 nm, was higher for the DLOOP mutant than for wild-type DAAO (  19 versus 13 arbitrary units) [6] and was similar among W243Y and wild-type DAAO protein (Table 1). On the other hand, the lower value determined for the purified W243I is caused by the presence of a large amount of apoprotein in the puri- fied sample. In order to investigate the exposure of clusters of buried hydrophobic residues, we studied the binding of the fluorescent probe, 1,8-anilinonaphtalene sulfonic acid (ANS), to hydrophobic side chains in proteins: this binding results in a marked increase in ANS fluor- escence yield for all DAAO forms, accompanied by a blue-shift in the emission fluorescence spectra. An empiric parameter (such as the ratio between the change in fluorescence emission at 500 nm at satur- ating ANS concentration) and the K d for ANS binding were estimated [11]: values of 0.84, 0.78, and 0.21 arbi- trary unitsÆlm )1 for wild-type, W243Y and W243I DAAOs were obtained, whereas the K d for ANS was practically unchanged (140 ± 20 lm for wild-type and Trp243 mutants). The low intensity of ANS fluores- cence observed with the W243I mutant was caused by the presence of 0.1 mm free FAD in the assay mixture, which was added to avoid complications as a result of apoprotein formation. Therefore, it is not possible to compare the exposure of ANS-accessible hydrophobic regions between the W243I mutant and the other DAAO forms. In the apoprotein form of all DAAO forms under investigation, the protein conformation is looser and the tryptophan side chains are transferred to a more polar environment because when the cofactor is removed, tryptophan emission is significantly higher and a shift in the maximum emission is observed (Table 1). The higher fluorescence intensity of the apoprotein forms of wild-type and DLOOP DAAOs compared with those of the W243 mutants probably reflects the contribution of tryptophan exposure at position 243 (see Table 1). Furthermore, similar K d values were determined for the ANS titration of apoprotein forms of wild-type, W243Y and W243I DAAOs (ranging from 24 to 31 lm): these values are fivefold lower than determined for the corres- ponding holoenzymes, and this change is paralleled by a remarkably higher fluorescence emission at sat- uration. The significant increase observed in the empiric DF ⁄ K d ratio (7.2, 7.8 and 7.3 arbitrary unitsÆ lm )1 for W243Y, W243I and wild-type DAAO, respectively) points to a greater exposure of hydro- phobic regions in the apoprotein forms than in the holoenzymes. Kinetic properties of Trp243 DAAO mutants The apparent steady-state kinetic parameters were determined using d-alanine (the most frequently used DAAO substrate) at a fixed (21%) oxygen concentra- tion. In order to avoid problems arising from apopro- tein formation, measurements with the W243I mutant were performed in the presence of 0.2 mm FAD. With both Trp243 mutant DAAOs, the apparent kinetic parameters were modified only slightly: the most signi- ficant change was an increase of approximately two- fold in K m for d-alanine (Table 1). Oligomerization state of Trp243 DAAO mutants The oligomerization state of the Trp243 mutants was investigated by gel-filtration chromatography. The wild-type and W243Y enzymes behaved very similarly during gel-filtration chromatography at 1 mgÆmL )1 protein concentration. However, the elution volume for the mutant enzyme is related to protein concentra- tion: at 0.1 mgÆmL )1 , the wild-type DAAO retained its elution behavior, but the W243Y mutant eluted at a higher elution volume (from a K av of 0.40 at 1mgÆmL )1 to a K av of 0.45 at 0.1 mgÆmL )1 ). On the other hand, the W243I mutant was found to exhibit an elution volume corresponding to a monomeric form of DAAO at all protein concentrations tested (0.2– 16 mgÆmL )1 ). We previously reported that yeast DAAO is a stable dimer and that it converts to a monomeric form in the presence of 0.5 m NH 4 SCN [6]. Analogously, the elu- tion volume of W243Y is also altered by thiocyanate, although the conversion to a monomeric state was achieved at a lower concentration of lipophilic ion (i.e. 0.2 m), further confirming the weaker interaction between monomers. Spectroscopic studies of thermal stability Temperature ramp experiments were performed, as reported previously [6], by following changes of var- ious spectroscopic signals. Following the increase in FAD fluorescence, the two Trp243 DAAO mutants (that at 0.1 mgÆmL )1 protein concentration are both monomeric) show a different increase in flavin fluores- cence and are less thermostable than wild-type DAAO (showing a melting temperature (T m )ofupto 9 °C lower for the W243I mutant than for the dimeric wild- type DAAO, see Table 2). When the temperature- induced loss of the tertiary structure was monitored by following the increase in protein fluorescence and in ANS fluorescence, the Trp243 mutants were still more L. Caldinelli et al. Structural determinants for dimerization of DAAO FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS 507 sensitive to temperature than the wild-type DAAO (Table 2). Taken together, these data show that the loss of the tertiary structure elements and of the flavin cofactor occurs at lower temperatures in the two Trp243 mutants than in the dimeric wild-type DAAO. Urea-induced unfolding The stability towards chemical denaturation of Trp243 mutants was compared with that of wild-type and DLOOP DAAO forms using equilibrium unfolding measurements and by following different spectroscopic signals of these proteins after equilibration in the pres- ence of increasing urea concentration. The change in protein emission intensity (used as a probe of the glo- bal unfolding) was only accurately measured for the W243Y mutant because, for the W243I mutant, a high scattering was observed in both the absence and pres- ence of exogenous FAD. For both mutants, the chan- ges were complete at 2 m urea and suggest a simple two-state transition (see Fig. 3A for W243Y DAAO). The midpoint concentration of urea required for unfolding (C m ) of the W243Y mutant is similar to that determined for the monomeric DLOOP enzyme, which is significantly lower than for wild-type DAAO (Table 2). In contrast, the change in wavelength of maximal emission as a function of urea concentration exhibited significantly higher C m values (2.0 ± 0.2 m for all DAAO forms; Fig. 3A). The lack of correlation between the fluorescence intensity and the maximum emission, two parameters that report on different events (solvent exposition of the indole ring of trypto- phan versus its distance from quenching moieties) sup- ports the hypothesis that the unfolding process of Trp243 mutants is complex. Analogously to that reported previously for wild- type DAAO [11], a biphasic dependence of flavin fluor- escence intensity with increasing urea concentrations was also observed for the W243Y mutant (see Fig. 3B and Table 2), leading to the production of the unfold- ing intermediate at a urea concentration similar to that observed for the monomeric DLOOP mutant (a similar experiment was not performed with the W243I mutant because in the absence of added free flavin it is largely present as apoprotein). At ‡ 7 m urea, the flavin fluor- escence of the W243Y mutant attains values similar to those observed for the DLOOP mutant and wild-type DAAOs under similar conditions and corresponds to that of the free flavin. The binding of ANS was used to study the partial exposure of hydrophobic patches upon loosening of the protein tertiary structure at increasing urea concen- trations. For wild-type and DLOOP DAAOs, two transitions were seen in the intensity of ANS fluores- cence at 500 nm in solutions containing 0.1 mm ANS and increasing denaturant concentration [11]: an increase in ANS fluorescence up to 2 m urea (paral- leled by a change in the maximum ANS emission wavelength) was followed by a quenching of ANS fluorescence along with a shift of  20 nm in the maxi- mum emission to a longer wavelength at higher urea concentrations (Fig. 3C). Using the W243Y DAAO, the ANS fluorescence was high (and the maximum emission wavelength decreased) up to 4.5 m urea (Fig. 3D), indicating that exposure of new hydropho- bic ANS-binding surfaces in the mutant enzyme is observed up to this concentration of urea. In all cases, the surfaces progressively disappeared as urea concen- trations were increased further. Effects of urea on the associative behaviour Upon preincubation with increasing concentrations of urea (0–8 m), both dimeric wild-type and monomeric DLOOP DAAOs eluted from size-exclusion chroma- tography as multiple peaks with retention volumes corresponding to DAAO aggregates comprising 2–10 subunits [11]. Under the same experimental conditions, and at a protein concentration of 1 mgÆmL )1 , the W243I DAAO (which is monomeric in the absence of urea) and the W243Y DAAO (which is dimeric in the absence of urea), converted into aggregates of increas- ing size at urea concentrations from 2 to 4 m, with a Table 2. Comparison of parameters for temperature- and urea- induced unfolding as determined by different approaches on the wild-type and mutant proteins of D-amino acid oxidase (DAAO). Method Wild type a,b DLOOP a,b W243I W243Y Temperature (°C) Trp fluorescence 46.8 39.2 42.4 40.6 FAD fluorescence 47.5 42.3 38.5 41.2 ANS fluorescence 50.0 n.d. 41.7 40.2 Urea C m (M) Trp fluorescence 1.4 (1.8) 1.0 (2.0) n.d. (1.9) 0.8 (2.0) FAD fluorescence c 1.8, 6.2 1.3, 5.6 n.d. 1.1, 5.0 In order to avoid the presence of apoprotein in the assay mixture, the protein- and 1,8-anilinonaphtalene sulfonic acid (ANS) fluores- cence of W243I was determined in the presence of 0.1 m M FAD. Melting temperature (T m ) values were obtained by deriving the spectroscopic signals and not corrected for delay effects. The SD was within 0.2 °C for all values determined. The numbers in paren- theses are the midpoint concentration of urea required for unfolding (C m ) determined from changes in the wavelength of maximum emission as function of urea concentration. a As described previously [6]. b As described previously [11]. c These values were estimated using a three-state model, as des- cribed previously [11,16]. Structural determinants for dimerization of DAAO L. Caldinelli et al. 508 FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS concomitant loss of solubility. These aggregates con- verted into soluble polymeric forms at urea concentra- tions higher than 6 m, as evident by the increase in the total area of the eluted peaks (up to  80% of the value determined for the untreated DAAO). This result is consistent with the surface hydrophobicity data pre- sented in Fig. 3C,D, and indicates a difference in the exposure of hydrophobic regions during the urea- induced unfolding of W243 mutants compared with the wild-type DAAO. Kinetics of reconstitution of apoprotein with FAD Because of the weaker binding of the FAD cofactor to the apoprotein of the W243 mutants, we studied the kinetics of reconstitution of the apoprotein–FAD com- plexes under pseudo-first-order conditions (10- and 20-fold excess of FAD) by following the quenching of protein fluorescence both under steady-state conditions and using a stopped-flow instrument. The reaction course was previously demonstrated to be biphasic using wild-type and DLOOP DAAO apoproteins [12]. The quenching curves obtained with the W243Y apo- protein are similar to those obtained for wild-type DAAO (Fig. 4A). The observed first-order rate con- stant of the fast phase is  1.0 ± 0.1Æs )1 and that of the slow phase is 0.013 ± 0.002Æs )1 for all DAAO forms. Under these experimental conditions, the weak binding of FAD to W243I apoprotein in combination with the small change in protein fluorescence between the corresponding apoprotein and the holoenzyme forms (see Table 1) represents a hindrance in acquiring reasonable time-course fluorescence traces. Under the same experimental conditions, the change in fluores- cence intensity of the first phase corresponds to 40 ± 4% of the total change for the wild-type DAAO and to 32 ± 3% for the W243Y mutant. This result indicates that the quenching of the signal associated with W243 largely belongs to the first phase (i.e. to the rearrangement of the interaction between the flavin Fig. 3. (A) Equilibrium denaturation curves of W243Y D-amino acid oxidase (DAAO) detected by means of tryptophan fluorescence. Samples of DAAO (0.02 mgÆmL )1 ) were equilibrated for 40 min in the presence of increasing urea concentrations, and the fraction of unfolded protein was determined from the fluorescence intensity at  340 nm (s) and emission peak maximum (d). The reported val- ues were corrected for the emission of the solution prior to adding protein. Solid lines represent the best fit obtained using a two-state denaturation model. (B) Comparison of equilibrium urea-denatura- tion profiles of wild-type (d), DLOOP (h), and W243Y (m) DAAOs detected by means of flavin fluorescence (see the legend of panel A for details). Lines represent the best fit obtained using a three- state denaturation model for wild-type (—) and W243Y (- - - -) enzymes [11,16]. (C, D) 1,8-Anilinonaphtalene sulfonic acid (ANS) fluorescence in the presence of wild-type (C) and W243Y (D) DAAOs as a function of urea concentration: fluorescence intensity at 500 nm (d) and wavelength of emission maximum (n). Samples of DAAO (2.5 l M ¼ 0.1 mgÆmL )1 ) were equilibrated for 40 min at 15 °C in the presence of increasing concentrations of urea, and the fluorescence spectra were recorded after the addition of 100 l M ANS. The values reported have been corrected for the emission of the solution prior to adding protein. L. Caldinelli et al. Structural determinants for dimerization of DAAO FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS 509 and the surrounding amino acid residues to form an holoenzyme intermediate) [12]. In all cases, the enzyme activity was regained during the second phase, as defined by the fluorescence experiments (requiring  95 ± 10 s for all apoproteins; Fig. 4B). The reconstitution process is significantly faster in the presence of 2 m urea (i.e. the denaturant concentra- tion at which the unfolding intermediate is largely produced) (Fig. 3B) [11]. The fluorescence changes obtained for wild-type and W243Y DAAOs at 2 m urea were observed during the dead-time of mixing (£ 10 ms), and also the enzymatic activity is recovered faster than in the absence of urea (requiring £ 40 s for all DAAO forms, Fig. 4B). At 2 m urea,  15% of the activity was recovered, corresponding to the residual activity measured at this urea concentration during chemical unfolding experiments [11]. Therefore, at the urea concentration required to yield the unfolding intermediate, the reconstitution process for the DAAO apoprotein is different from that observed under native conditions and it is not significantly modified by Trp243 substitution. Discussion Both of the Trp243 substitutions introduced in DAAO were nondisruptive, as confirmed by the absence of major changes in catalytic activity, and altered the protein in a different way. Some of the properties of the W243Y mutant are similar to those of the wild-type protein (e.g. the oligomeric state at 1mgÆmL )1 protein concentration, the absorption spectrum, the flavin fluorescence, the far-UV CD spectrum, and the binding with FAD), whereas other properties differ (e.g. the monomeric state at low protein concentration, the protein fluorescence, the near-UV CD spectrum, and the thermal stability). Furthermore, the changes are more evident for the W243I mutant: it has been purified as a monomer whose interaction with the FAD coenzyme is signifi- cantly decreased (Table 1). Substitution of Trp243 modifies the protein confor- mation such that it interferes with both the coenzyme binding to the apoprotein moiety and with the acquisi- tion of the tertiary structure required for the mono- mer–monomer interaction, resulting in a lower stability towards temperature and denaturants. Concerning the monomer–monomer interaction, the structural pertur- bation introduced by substitution of Trp243 with isoleucine is sufficient to prevent dimerization of the enzyme, although the bF5–bF6 dimerization loop is still present in the mutant DAAO (Fig. 1A,B). Simi- larly, a single tryptophan residue (Trp548 in the sub- unit interface region) is also responsible for the integrity of the quaternary structure of the cytosolic malic enzyme [13]. Our results suggest that the ‘hole’ generated by the substitution of Trp243 could be filled with water molecules that, in a hydrophobic environ- ment, might be a major factor that prevents associ- ation (or tight contact) of the two monomers. Such a conclusion is supported by the results obtained for the W243Y mutant, the most conservative substitution. In fact, although the W243Y mutant binds the FAD cofactor tightly and is dimeric, its oligomerization state depends on the protein concentration, and its conver- sion to a monomeric state is obtained at a significantly lower thiocyanate concentration than for wild-type DAAO (0.2 versus 0.5 m, respectively). We hypothesize that Trp243 changes its position during flavin binding Fig. 4. (A) Time course of protein fluorescence change at 340 nm during the binding of FAD to wild-type (d) and W243Y (n) apopro- teins at 15 °C and pH 7.5. Proteins were used at a concentration of 0.1 l M (0.004 mgÆmL )1 )in50mM potassium phosphate, pH 7.5, containing 10% (v ⁄ v) glycerol and 2 m M EDTA, and were reacted with 10-fold excess free FAD. (B) Time course of activity recovery during the binding of FAD to wild-type D-amino acid oxidase (DAAO) apoprotein at 15 °C and pH 7.5 in the presence (s)or absence (d)of2 M urea. The activity was measured using an oxy- gen-consumption assay under the same experimental conditions used for the fluorescence analysis (see Fig. 4A). Structural determinants for dimerization of DAAO L. Caldinelli et al. 510 FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS to the DAAO apoprotein (in particular during the first phase, see Fig. 4A), reaching a location that promotes dimerization. The second structural element fundamental for the stability of yeast DAAO is represented by the coenzyme binding: temperature ramp experiments demonstrated that the apoprotein is less stable than the holoenzyme and that flavin release triggers protein denaturation [6]. Indeed, limited proteolysis experiments have highlighted the looser conformation of the apoprotein form [14]. Recently, we reported that the urea-induced unfolding of DAAO is a three-state process, yielding an intermedi- ate at  2 m urea [11]. The intermediate species lacks the characteristic tertiary structure of native DAAO, but has a significant secondary structure, retains flavin binding, and shows an increased sensitivity to trypsin of a specific site (Arg289) belonging to the loop preceding the b-strand F5 of the FAD-binding domain. Strand F5 has been proposed as a specific connection during holo- enzyme reconstitution between the FAD-binding process and the exposure of Trp243 required for dimeri- zation [11]. In fact, it is connected on one side (and through the C-terminal a-helix containing the peroxi- somal targeting signal) to the bab motif, known as the dinucleotide-binding domain [2], and on the other side to the long loop following the b-strand I7 containing Trp243 (see Fig. 1D). The looser conformation of the mutants DAAO, as evident by the higher protein fluor- escence and by the modification of near-UV CD spec- trum of their holoenzyme forms compared with the wild-type, indicates that the tertiary structure modifica- tion caused by Trp243 substitution (in particular with an isoleucine) prevents the aforementioned interactions from occurring. In conclusion, the site-directed mutagenesis studies on Trp243 in yeast DAAO indicate that (a) the inter- actions at the monomer–monomer interface, which result in tight packing, stabilize the protein by maxim- izing van der Waals interactions and by efficiently excluding solvent and (b) Trp243 substitution alters the protein conformation required for efficient coen- zyme binding. Experimental procedures Mutagenesis, enzyme expression and purification, and apoprotein preparation The mutant DAAO genes were generated by site-directed mutagenesis using the QuikChangeÒ site-directed mutagen- esis kit (Stratagene, LaJolla, CA, USA) and the recombin- ant plasmid, pT7-HisDAAO, as template [15]. Both W243I and W243Y DAAOs are produced as a fusion protein because 13 additional residues are added at the N terminus of the protein before the original methionine: the presence of this short peptide has been demonstrated not to alter the overall properties of the wild-type DAAO [15]. The Trp243 DAAO mutants were expressed using the strain BL21(DE3)pLysS Escherichia coli as host and purified as reported previously [15]; however, the W243I mutant was better expressed when the cells were grown at 25 °C after isopropyl thio-b-d-galactoside induction (up to 50 DAAO unitsÆg )1 cell paste). The overall yield of the purification was  50% for both Trp243 mutants, similar to the yield obtained with other DAAO forms. The apoprotein form of both DAAO mutants was obtained by dialysis in the pres- ence of 2 m KBr and 20% (v ⁄ v) glycerol [9], and can be stored at )20 °C for months without any loss in ability to reconstitute with FAD to give the corresponding holo- enzyme. Activity assay DAAO activity was assayed with an oxygen electrode at 25 °C, as described previously [15], using 28 mmd-alanine as substrate. One DAAO unit corresponds to the amount of enzyme that converts 1 lmol of d-alanine per minute. Size-exclusion chromatography Size-exclusion chromatography was performed on a Super- dex 200 H column (Amersham Biosciences, Piscataway NJ, USA), at a flow rate of 0.5 mLÆmin )1 and at room tempera- ture. The elution buffer used was 50 mm potassium phos- phate, pH 7.5, containing 5% (v ⁄ v) glycerol, 2 mm EDTA, and the appropriate concentration of urea or NH 4 SCN. The column was calibrated with suitable standard proteins. Spectroscopy All experiments were performed in 50 mm potassium phos- phate buffer, pH 7.5, containing 10% (v ⁄ v) glycerol and 2mm EDTA, and at 15 °C. Fluorescence was measured in a Jasco FP-750 instrument and at a protein concentration of 0.1 mgÆmL )1 . For temperature ramp experiments, the instrument was equipped with a software-driven Peltier temperature controller (to produce a 0.5 °CÆmin )1 tempera- ture gradient) [6]. Tryptophan emission spectra were recor- ded from 300 to 400 nm using an excitation wavelength of 280 nm, and flavin emission spectra were recorded from 475 to 600 nm using an excitation wavelength of 450 nm. Emission and excitation bandwidths were set at 10 and 20 nm, respectively. Kinetics of the apoprotein–FAD com- plex formation were measured at a concentration of 0.1 lm apoprotein (0.004 mgÆmL )1 ) and following the emission at 305 nm or at 340 nm (excitation at 280 nm). Reconstitution kinetics were also determined by stopped-flow in a Bio- L. Caldinelli et al. Structural determinants for dimerization of DAAO FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS 511 Logic SFM-300 instrument (BioLogic, Claix, France) inter- faced to the spectrofluorimeter. Reaction rates were calcula- ted by fitting the traces to a sum of exponentials equation using Biokine32 (BioLogic). ANS-binding experiments were carried out at a protein concentration of 2.5 lm (0.1 mgÆmL )1 ) [11]. Fluorescence emission spectra were recorded in the 450–600 nm range using an excitation wave- length of 370 nm. In general, the ANS binding was fol- lowed at 500 nm but, because of the FAD-induced quenching of ANS signal, its binding to W243I mutant was estimated using the values at 478 nm, a wavelength at which the change in fluorescence is not affected by the pres- ence of the cofactor. CD spectra were recorded on a Jasco J-810 spectropolarimeter (Jasco Europe, Cremella, Italy) and analysed by means of Jasco software [6,11]. The unfolding equilibrium of DAAO was determined by following the changes in flavin and protein fluorescence, as detailed previously [11]. Protein fluorescence unfolding curves were analyzed, according to a two-state mechanism, using the apparent fraction of the unfolding form, and the flavin fluorescence data were analyzed according to a three- state denaturation pathway (N « I « U) [11,16]. Acknowledgements We thank Dr Stefania Iametti for kindly helping with the CD measurements, and Dr Luciano Piubelli for helpful discussions. This work was supported by grants from Fondazione Cariplo to LP, and from FAR 2004. References 1 Mei G, Di Venere A, Rosato N & Finazzi-Agro A (2005) The importance of being dimeric. FEBS J 272, 16–27. 2 Rossmann MG, Moras D & Olsen KW (1974) Chemical and biological evolution of nucleotide-binding protein. Nature 250, 194–199. 3 Dym O & Eisenberg D (2001) Sequence-structure analy- sis of FAD-containing proteins. Protein Sci 10, 1712– 1728. 4 Umhau S, Pollegioni L, Molla G, Diederichs K, Welte W, Pilone MS & Ghisla S (2000) The x-ray structure of d-amino acid oxidase at very high resolution identifies the chemical mechanism of flavin-dependent substrate dehy- drogenation. Proc Natl Acad Sci USA 97, 12463–12468. 5 Pollegioni L, Diederichs K, Molla G, Umhau S, Welte W, Ghisla S & Pilone MS (2002) Yeast d-amino acid oxidase: structural basis of its catalytic properties. J Mol Biol 324, 535–546. 6 Pollegioni L, Iametti S, Fessas D, Caldinelli L, Piubelli L, Barbiroli A, Pilone MS & Bonomi F (2003) Contri- bution of the dimeric state to the thermal stability of the flavoprotein d-amino acid oxidase. Protein Sci 12, 1018–1029. 7 Piubelli L, Caldinelli L, Molla G, Pilone MS & Pollegioni L (2002) Conversion of the dimeric d-amino acid oxidase from Rhodotorula gracilis to a monomeric form. A rational mutagenesis approach. FEBS Lett 526, 43–48. 8 Piubelli L, Molla G, Caldinelli L, Pilone MS & Polle- gioni L (2003) Dissection of the structural determinants involved in formation of the dimeric form of d-amino acid oxidase from Rhodotorula gracilis: role of the size of the bF5-bF6 loop. Protein Eng 16, 1063–1069. 9 Casalin P, Pollegioni L, Curti B & Pilone Simonetta M (1991) A study on apoenzyme from Rhodotorula gracilis d-amino acid oxidase. Eur J Biochem 197, 513–517. 10 Lakowicz JR (1999) Principles of Fluorescence Spectro- scopy, 2nd edn, pp. 237–266. Kluwer ⁄ Plenum, New York. 11 Caldinelli L, Iametti S, Barbiroli A, Bonomi F, Piubelli L, Ferranti P, Picariello G, Pilone MS & Pollegioni L (2004) Unfolding intermediate in the peroxisomal flavo- protein d-amino acid oxidase. J Biol Chem 279, 28426– 28434. 12 Pollegioni L & Pilone MS (1996) On the holoenzyme reconstitution process in native and truncated Rhodotor- ula gracilis d-amino acid oxidase. Arch Biochem Biophys 332, 58–62. 13 Chang HC & Chang GG (2003) Involvement of single residue tryptophan 548 in the quaternary structural sta- bility of pigeon cytosolic malic enzyme. J Biol Chem 278, 23996–24002. 14 Pollegioni L, Ceciliani F, Curti B, Ronchi S & Pilone MS (1995) Studies on the structural and functional aspects of Rhodotorula gracilis d-amino acid oxidase by limited trypsinolysis. Biochem J 310, 577–583. 15 Fantinato S, Pollegioni L & Pilone MS (2001) Engineer- ing, expression and purification of a His-tagged chimeric d-amino acid oxidase from Rhodotorula gracilis. Enzyme Microbial Technol 29, 407–412. 16 Ayed A & Duckworth HW (1999) A stable intermediate in the equilibrium unfolding of Escherichia coli citrate synthase. Protein Sci 8, 1116–1126. Structural determinants for dimerization of DAAO L. Caldinelli et al. 512 FEBS Journal 273 (2006) 504–512 ª 2006 The Authors Journal compilation ª 2006 FEBS . Tryptophan 243 affects interprotein contacts, cofactor binding and stability in D-amino acid oxidase from Rhodotorula gracilis Laura Caldinelli, Gianluca Molla, Mirella S. Pilone and Loredano. structure, retains flavin binding, and shows an increased sensitivity to trypsin of a specific site (Arg289) belonging to the loop preceding the b-strand F5 of the FAD -binding domain. Strand F5 has. Trp243 changes its position during flavin binding Fig. 4. (A) Time course of protein fluorescence change at 340 nm during the binding of FAD to wild-type (d) and W243Y (n) apopro- teins at 15 °C and