Logo of prosciprotein sciencecshl presssubscriptionsetoc alertsthe protein societyjournal home
Protein Sci. 2003 May; 12(5): 1018–1029.
PMCID: PMC2323872
PMID: 12717024

Contribution of the dimeric state to the thermal stability of the flavoprotein D-amino acid oxidase

Loredano Pollegioni,1 Stefania Iametti,2 Dimitrios Fessas,3 Laura Caldinelli,1 Luciano Piubelli,1 Alberto Barbiroli,3 Mirella S. Pilone,1 and Francesco Bonomi2
Author information Article notes Copyright and License information Disclaimer

Abstract

The flavoenzyme DAAO from Rhodotorula gracilis, a structural paradigm of the glutathione-reductase family of flavoproteins, is a stable homodimer with a flavin adenine dinucleotide (FAD) molecule tightly bound to each 40-kD subunit. In this work, the thermal unfolding of dimeric DAAO was compared with that of two monomeric forms of the same protein: a Δloop mutant, in which 14 residues belonging to a loop connecting strands βF5–βF6 have been deleted, and a monomer obtained by treating the native holoenzyme with 0.5 M NH4SCN. Thiocyanate specifically and reversibly affects monomer association in wild-type DAAO by acting on hydrophobic residues and on ionic pairs between the βF5–βF6 loop of one monomer and the αI3′ and αI3″ helices of the symmetry-related monomer. By using circular dichroism spectroscopy, protein and flavin fluorescence, activity assays, and DSC, we demonstrated that thermal unfolding involves (in order of increasing temperatures) loss of tertiary structure, followed by loss of some elements of secondary structure, and by general unfolding of the protein structure that was concomitant to FAD release. Temperature stability of wild-type DAAO is related to the presence of a dimeric structure that affects the stability of independent structural domains. The monomeric Δloop mutant is thermodynamically less stable than dimeric wild-type DAAO (with melting temperatures (Tms) of 48°C and 54°C, respectively). The absence of complications ensuing from association equilibria in the mutant Δloop DAAO allowed identification of two energetic domains: a low-temperature energetic domain related to unfolding of tertiary structure, and a high-temperature energetic domain related to loss of secondary structure elements and to flavin release.

Keywords: Flavoprotein, lipophilic ions, folding, dimerization, energetic domains, structural domains, quaternary structure, thermal stability

Flavoproteins are a group of enzymes that catalyze a variety of reactions and that are characterized by the presence of a flavin cofactor. Concerning the coenzyme binding, flavoproteins divide into three classes: those possessing the coenzyme covalently linked through the isoalloxazine ring to a histidine, cysteine, or tyrosine residue of the polypeptide chain (Decker 1991); those in which the flavin is noncovalently bound; and those that use flavin adenine dinucleotide (FAD) or flavin mononucleotide (FMN) as a substrate. In the first two flavoprotein classes, attainment of the catalytic activity requires binding of the cofactor to the apoprotein, which generally has a different conformation from the holoenzyme (van Mierlo and Steensma 2000).

Cofactor binding is of paramount importance in vivo, because many flavoenzymes are nuclear encoded, synthesized as larger precursors on free cytosolic ribosomes, and posttranslationally imported into the final organelle. Many peroxisomal enzymes (such as DAAO) generate toxic hydrogen peroxide, and their activity before targeting could be extremely dangerous for the cell. On the other hand, loss of the coenzyme turns off the enzyme activity and, because of the concomitant formation of an apoprotein form that frequently possesses a looser conformation and thus a lower stability, triggers its degradation.

A second important feature of flavoproteins concerns their quaternary structure. They are often multisubunit proteins constituted either by identical or by different polypeptide chains. In some cases the multimeric organization is linked to the enzyme function. Detailed information on the relationships between cofactor uptake, folding, and intracellular trafficking is available only for a handful of well-characterized systems, such as yeast alcohol oxidase (Waterham et al. 1997). In general, less is known on the role of the stable interaction between identical subunits and on its significance for the apoprotein–FAD interactions that are relevant to biogenesis of a functional protein and to its turnover in the cell.

To gain insights into these important topics, we undertook a study of the stability of structural elements in the peroxisomal flavoenzyme DAAO (EC 1.4.3.3) from the yeast Rhodotorula gracilis (RgDAAO). DAAO is considered the paradigm of the dehydrogenase-oxidase class of flavoproteins (Massey and Hemmerich 1980), is a component of the GR folding family, and is a peroxisomal marker (for review, see Curti et al. 1992; Pilone 2000). DAAO catalyzes the oxidative dehydrogenation of D-isomers of amino acids to give α-keto acids, ammonia, and hydrogen peroxide. Over the years, DAAOs have been the object of extensive investigation (Curti et al. 1992; Pilone 2000). RgDAAO in solution is a stable 80-kD homodimer, with a molecule of FAD tightly—but noncovalently—bound to each 40-kD subunit. The stable dimeric aggregation state, as well as the high catalytic activity and tight binding of FAD to the apoprotein polypeptide, distinguishes the yeast enzyme from mammalian DAAO. These features probably are the evolutionary answer of the different role of this flavoenzyme in higher and lower eukaryotes (Pilone 2000).

The tridimensional (3D) structure of RgDAAO has been resolved at very high resolution, allowing researchers to find the rationale for its high catalytic efficiency (Umhau et al. 2000; Pollegioni et al. 2002). Between the possible modes of monomer–monomer interaction in the crystal, the one having the highest buried surface area (3049 Å2) is the “side-to-tail” model (Pollegioni et al. 2002). A large contribution to the interaction between monomers in RgDAAO is given by a long (21 amino acids) loop connecting β-strands F5 and F6, which is unique to RgDAAO. Recently, by rational design and site-directed mutagenesis, a stable monomeric holoenzyme form of RgDAAO has been obtained by partial elimination of the βF5-βF6 loop (Piubelli et al. 2002). The deleted portion (from Ser 308 to Lys 321: SPLSLGRGSARAAK) contains three positively charged residues and no aromatics.

The Δloop mutant RgDAAO shows slightly altered spectral and kinetic properties. The main differences with respect to the wild-type enzyme are a lower stability at temperatures higher than 35°C and a fivefold increase in the Kd for FAD binding. Conversion to a monomeric form from native RgDAAO can also be achieved by removal of the coenzyme to yield the corresponding apoprotein (Casalin et al. 1991). This indicates a structural relationship between the FAD-harboring domain and the regions involved in dimerization.

To provide a basis for understanding the structure–function relationships in flavoproteins and the determinants of their stability, we have attempted to compare the unfolding of dimeric DAAO with that of monomeric forms obtained either by site-directed mutagenesis or by chemical treatment. This was done with the ultimate ambitious goal to improve current understanding of the correlations between protein folding, coenzyme binding, subunit interaction, and protein targeting. In particular, by using a combination of calorimetric and spectroscopic methods, we aimed at discriminating among the steps involved in dimer dissociation and those related to the loss of the polypeptide structure and of the FAD cofactor during thermal treatment of the different proteins. Direct comparison of the thermal stability of apo- versus holo- forms of RgDAAO was not feasible, because of the marked thermal instability of the apoproteins that aggregate in the course of prolonged exposure at high temperature, as required for temperature-ramp experiments and calorimetric studies.

Results

Effect of thiocyanate on the aggregation state of wild-type RgDAAO

Thiocyanate is a lipophilic ion (Ptistyn and Semisotnov 1991) that has the capacity to interfere with ionic pairs located either on the protein surface or in the hydrophobic interior of folded proteins. In such a way, lipophilic ions affect the tertiary and quaternary structures of proteins more selectively than most generic chaotropes, such as urea or guanidinium ions, which affect the hydrogen bond network in the solvent water and in the proteins themselves.

The effect of NH4SCN on the quaternary structure of wild-type RgDAAO was investigated by means of size-exclusion chromatography on a Superdex 200 column. The elution profile of wild-type RgDAAO (1 mg/mL) at different NH4SCN concentrations is shown in Figure 1 ▶. RgDAAO was present in a dimeric form in the absence of NH4SCN, but at 0.5 M NH4SCN it eluted as a single peak with a retention time corresponding to a monomeric form (44.2 ± 3.9 kD). This effect is peculiar to thiocyanate, because the elution volume of RgDAAO is not modified in the presence of 0.5 M NaCl or NH4Cl (not shown). The absorption spectrum of the enzyme eluted at 0.5 M NH4SCN showed the visible bands at 450 nm and 380 nm typical of FAD-containing enzymes. Its specific activity was 95 U/mg protein (~110 U/mg for native RgDAAO) and was not affected by the addition of exogenous FAD in the activity assay. However, prolonged incubation of the protein with NH4SCN at a concentration ≥0.5 M prior to the chromatographic separation resulted in decreased recovery of the protein and of its activity. Also, the electronic spectrum of RgDAAO was altered at thiocyanate concentrations higher than 0.5 M, likely because of partial denaturation (vide infra).

An external file that holds a picture, illustration, etc. Object name is 61237-18f1_L1TT.jpg

Effect of NH4SCN concentration on the quaternary structure of RgDAAO. RgDAAO (0.2 mL, 1 mg/mL) was incubated on ice in the presence of 0, 0.4, and 0.5 M NH4SCN before gel permeation chromatography on a Superdex 200 column in 50 mM potassium phosphate (pH 7.5), 150 mM potassium chloride, and 10% glycerol. The elution buffer contained 0 (full line), 0.4 (dashes and dots), or 0.5 M (dots) NH4SCN, according to the concentrations used in the treatment before the separation step. (DB) Dextran Blue.

Under appropriate conditions, the dimer–monomer conversion observed at 0.5 M NH4SCN was fully reversible both structurally and functionally. The dimeric holoenzyme was fully recovered (as judged by gel permeation chromatography) when the NH4SCN concentration was slowly decreased by dilution of the monomeric enzyme solution (from 0.5 M to 25 mM in about 16 h at 4°C). Protein recovery paralleled the activity recovery, that is, ~65%. The addition of a competitive inhibitor such as crotonate (Kd = 0.35 mM; Pilone 2000) did not improve the recovery figures. The absorbance spectrum of the recovered dimeric DAAO was indistinguishable from that of the native enzyme. However, fast removal of NH4SCN from the buffer (i.e., by desalting on a G25 column or by dialysis) resulted in protein and activity loss. This indicates that reacquiring a competent structure after treatment with NH4SCN is not limited thermodynamically but kinetically, and that the presence of residual NH4SCN in the course of slow dilution favors the structural rearrangements required to drive the refolding process toward recovery of a native-like structure.

Effect of thiocyanate on the secondary and tertiary structure of wild-type and Δloop RgDAAO mutant

Far-UV circular dichroism (CD) spectra of the wild-type and Δloop mutant RgDAAO did not reveal any major difference in the features related to the secondary structure of the two proteins (not shown), indicating that the loop deletion leading to loss of the dimeric state did not alter the content and the distribution of secondary structure elements. The very high background absorption of 0.5 M NH4SCN below 230 nm (higher than 3 absorbance units even with a 0.1-mm light path) prevented analysis of the secondary structure modifications induced by NH4SCN.

As shown in Figure 2 ▶, the near-UV CD spectra of the wild-type and Δloop mutant RgDAAO in the absence of NH4SCN were different. The differences may be ascribed to a different contribution from aromatic amino acid residues, which are responsible for most transitions in the near-UV spectral region. Because the deleted portion does not contain aromatic residues, the different spectroscopic features of the two proteins can only be explained in terms of altered mutual relationships between nearby structural elements. The addition of thiocyanate affected the tertiary structure of both wild-type and Δloop mutant RgDAAO, although dissociating concentrations of NH4SCN did not have a marked effect on wild-type DAAO. This indicates that dimer dissociation occurred without any sensible alteration of the tertiary structure in each monomer. Indeed, the spectra reported in Figure 2 ▶ clearly indicate a much higher sensitivity to NH4SCN for the Δloop RgDAAO mutant with respect to the wild-type protein. It is also interesting that the spectra of wild-type DAAO (both native and NH4SCN-dissociated) were isochroic with that of the NH4SCN-treated Δloop RgDAAO mutant at 272 nm. This indicates that the addition of NH4SCN may have eliminated some of the structural perturbation introduced in the wild-type protein by removal of the βF5–βF6 loop.

An external file that holds a picture, illustration, etc. Object name is 61237-18f2_L1TT.jpg

Effect of NH4SCN on the near-UV CD spectra of wild-type (thin lines) and Δloop (thick lines) RgDAAO in the absence (continuous line) and in the presence (dashed line) of 0.5 M NH4SCN. Proteins were 0.5 mg/mL in 50 mM potassium phosphate (pH 7.5), containing 10% glycerol and 2 mM EDTA. Measurements were performed at 15°C.

A further sensitive marker of protein conformation, and in particular of changes in hydrophobic regions of the structure, is tryptophan emission fluorescence. Tryptophan emission at 345 nm (following excitation at 280 nm) was threefold higher for Δloop RgDAAO than for wild-type (Fig. 3A ▶). This result points to a lower relevance of quenching interactions between tryptophan side chains (in particular of Trp 243, vide infra) and nearby side chains in the monomeric form of RgDAAO with respect to the dimeric form. Increasing the NH4SCN concentration up to 0.5 M increased the intensity of the 345-nm emission for both proteins. At an NH4SCN concentration of ≥0.5 M, the two DAAO forms had similar emission values (Fig. 3A ▶). In the presence of thiocyanate, the tryptophan emission of wild-type and Δloop DAAOs was shifted to a longer wavelength than in the untreated proteins (from 340 to 348 nm), indicating exposure of tryptophan residues to the solvent (Fig. 3B ▶). The shift in the wavelength of maximal emission was complete at 1.5 M NH4SCN.

An external file that holds a picture, illustration, etc. Object name is 61237-18f3_L1TT.jpg

Effect of NH4SCN concentration on tryptophan fluorescence of wild-type (filled symbols) and Δloop (open symbols) RgDAAO. DAAO (0.1 mg/mL) was incubated for 15 min at 15°C in 50 mM potassium phosphate (pH 7.5), 2 mM EDTA, and 10% glycerol containing the given concentration of NH4SCN. Emission spectra were recorded from 300 to 400 nm with excitation at 280 nm. The reported values have been corrected for the emission of the solution before protein addition. (A) Tryptophan fluorescence intensity. (B) Tryptophan fluorescence emission maximum.

Effects of thiocyanate on cofactor binding and on enzymatic activity in wild-type and Δloop RgDAAO mutant

A sensitive marker of the folding of a flavoenzyme is represented by the fluorescence of the FAD cofactor, which is much lower in the holoenzyme-bound form with respect to free FAD (Casalin et al. 1991). The emission at 520 nm, following excitation at 450 nm, was higher for Δloop than for wild-type RgDAAO (in the 0.015–0.5 mg/mL protein concentration range), pointing to a different microenvironment surrounding the FAD coenzyme in the two enzyme forms (see Fig. 4 ▶).

An external file that holds a picture, illustration, etc. Object name is 61237-18f4_L1TT.jpg

Effect of NH4SCN on the flavin fluorescence of wild-type (filled symbols) and Δloop (open symbols) RgDAAO. DAAO (0.1 mg/mL) was incubated for 15 min at 15°C in 50 mM potassium phosphate (pH 7.5), 2 mM EDTA, and 10% glycerol in the presence of various NH4SCN concentrations, and the flavin emission spectrum was recorded in the 460–600-nm wavelength range (excitation at 450 nm). The reported values have been corrected for the background emission of the buffer prior to protein addition.

An increase in NH4SCN concentration had a completely different effect on the two proteins. Flavin fluorescence of the Δloop mutant decreased with increasing NH4SCN concentration. Wild-type DAAO showed a small increase in fluorescence up to 0.5 M NH4SCN, followed by a decrease at higher NH4SCN concentration, where fluorescence attained values similar to those determined for the Δloop mutant (Fig. 4 ▶). Interestingly, the flavin fluorescence at 0.5 M NH4SCN, corresponding to conversion of the dimeric enzyme to the monomeric form (see earlier), was similar for wild-type and Δloop RgDAAO. The decrease in flavin fluorescence at increasing concentration of NH4SCN is mostly due to chemical quenching on direct interaction of free FAD and NH4SCN. In fact, a strong decrease of the 520-nm emission at increasing NH4SCN concentration was also observed for free FAD (from 195 to 8 arbitrary units using 2.5 μM purified FAD, not shown). This behavior is qualitatively similar to that observed for Δloop DAAO, and is indicative of a higher solvent exposure of the isoalloxazine ring of FAD in the Δloop mutant than in wild-type RgDAAO. Circumstantially, this confirms also the weaker binding of the coenzyme to the apoprotein moiety in the monomeric mutant DAAO than in the dimeric wild-type one (Casalin et al. 1991; Piubelli et al. 2002).

The effect of NH4SCN on the activity of RgDAAO was investigated by measuring the enzyme activity in the presence of different thiocyanate concentrations in the assay mixture. At NH4SCN concentrations higher than 0.5 M, the oxygen-consumption trace was clearly biphasic: a first short and fast phase (~1–3 min) was followed by a second phase in which the slope was significantly lower. The activity value corresponding to the first phase was completely lost at NH4SCN concentrations higher than 1 M (not shown). The activity was assayed also by using a standard assay mixture (that is, in the absence of NH4SCN) and DAAO samples previously incubated for 15 min at 15°C with different concentrations of NH4SCN. As reported in Figure 5 ▶, the enzyme activity measured in these conditions started to decline at ≈0.5 M NH4SCN and was lost at 1 M NH4SCN, concomitant to irreversible alteration of the protein conformation, as discussed earlier (see also Fig. 3 ▶). Activity of the Δloop mutant protein was also lost after preincubation with 1 M NH4SCN. As shown in Figure 5 ▶, incubation at dissociating NH4SCN concentration (0.5 M) decreased the activity of wild-type RgDAAO activity to 81% of that measured for the untreated protein. This is close to the 86% decrease in specific activity measured on the monomer recovered from size-exclusion chromatography in the presence of 0.5 M thiocyanate.

An external file that holds a picture, illustration, etc. Object name is 61237-18f5_L1TT.jpg

Effect of NH4SCN concentration on enzyme activity of wild-type RgDAAO. The activity was determined using the oxygen-electrode assay on protein samples incubated at 15°C for 15 min in the presence of different concentrations of NH4SCN. The carryover of thiocyanate in the activity assay was negligible. Data were fitted to the equation y = a/{1 + exp[−(x−xm)/b]}, where a = activity of the enzyme incubated in the absence of thiocyanate (79.0 U/mg protein), b = residual activity of the enzyme incubated at high thiocyanate concentration (≈0.08 U/mg protein at 1 M NH4SCN), and xm = thiocyanate concentration at which 50% of the initial activity is lost (0.61 M).

Spectroscopic and calorimetric studies on the thermal stability of the different aggregation states of RgDAAO

Preliminary studies indicated a significant lower stability of the Δloop mutant with respect to wild-type RgDAAO. After 30 min at 40°C, the Δloop mutant retained <5% of its initial activity, as compared with 60% for wild-type RgDAAO (Piubelli et al. 2002). To compare the temperature sensitivity of specific structural features in the two proteins, and to assess the nature and extent of spectroscopic modifications ensuing by progressive heating of wild-type and Δloop mutant RgDAAO, we performed temperature ramp experiments. Conditions used when monitoring temperature-dependent changes of the various spectroscopic features of the protein were identical to those used in DSC experiments (vide infra). In order to achieve information on the dependence of thermal stability on the protein aggregation state, we also performed these experiments (when feasible) at various protein concentration, with the assumption that this parameter would affect the relative concentration of differently associated species.

Table 1​1 reports the midpoint transition temperatures for the two DAAO forms as detected by the various experimental approaches, and using different sets of experimental conditions.

Table 1.

Comparison of Tm as determined by different approaches on the untreated wild-type and Δloop mutant DAAO

Wild typeΔLoop mutant
Method3.5 mg/mL1.5 mg/mL0.5 mg/mL0.1 mg/mL3.5 mg/mL1.5 mg/mL0.5 mg/mL0.1 mg/mL
DSC56.955.9——48.148.0——
Far UV CD——50.248.0——44.843.8
Trp fluorescence———46.4———42.7
FAD fluorescence——52.050.3——47.046.2

All experiments were performed in 50 mM potassium phosphate (pH 7.5), 2 mM EDTA, and 10% glycerol, at an identical heating rate (0.5°C/min). Tms were obtained by derivatization of the spectroscopic signals, and are uncorrected for delay effects. Standard deviation was within 0.2°C for all determined values.

The DSC tracings reported in Figure 6 ▶ clearly indicate that Δloop RgDAAO was thermodynamically much less stable than wild type. The wild-type protein had a denaturation Tms of 55.9°C and a denaturation enthalpy of 1030 kJ/mole at a concentration of 1.5 mg/mL, compared with 48.0°C and 775 kJ/mole for the Δloop RgDAAO at the same concentration. All Tm figures reported in Table 1​1 have an average standard error ≤0.2°C, whereas a 3% average standard error affects the enthalpy estimates. The DSC signals of wild-type RgDAAO were sensitive to protein concentration in the 1.5–3.5 mg/mL range, showing an increase in Tm and a more pronounced aggregation at high temperature with increasing protein concentration. Concentration effects on the denaturation enthalpies were within the experimental error in our experiments. Because concentration effects on Tm were not observed for Δloop RgDAAO, they cannot be ascribed to a generic “structural protection” exerted by higher protein concentration, but likely are linked to the association/dissociation equilibrium in the dimeric wild-type RgDAAO (D’Auria et al. 1997). The sharp fall of the DSC signal at temperatures above Tm in the case of wild-type RgDAAO also supports this hypothesis (D’Auria et al. 1997). These effects (Tm shift and asymmetry of DSC thermographs) were much less pronounced in the case of Δloop RgDAAO, which did not aggregate above Tm. All of the modifications observed by DSC on either protein were irreversible.

An external file that holds a picture, illustration, etc. Object name is 61237-18f6_L1TT.jpg

DSC tracings for wild-type (thin line) and Δloop (thick line) RgDAAO using protein concentrations of 1.5 (dotted line) or 3.5 mg/mL (continuous line). The heating rate was 0.5°C/min.

Taking into account these evidences, thermodynamic analysis of the modifications detected in the DSC experiments was limited to the Δloop RgDAAO mutant, that is, in the absence of possible complications ensuing from association equilibria in the native or in the denatured forms. It has been proven that even irreversible processes (as those observed here) can be in some cases split into two steps: achievement of a thermodynamic equilibrium, which is responsible for the protein unfolding enthalpy, followed by an irreversible step that does not affect the measured enthalpy. For example, it has been shown that the irreversible thermal denaturation of the lac repressor quantitatively obeys equilibrium thermodynamics (Manly et al. 1985), as reported also for the thermal denaturation of the subunits of Escherichia coli aspartate transcarbamoylase (Edge et al. 1985) and of bovine serum albumin (Giancola et al. 1997).

As shown in Figure 7 ▶, application of a thermodynamic model considering two independent energetic domains (Barone et al. 1994; Fessas et al. 2001) provided a good fit of the DSC signal for the Δloop mutant, that is, in the absence of aggregation at high temperature. A first energetic domain unfolded with a Tm of 45.1°C and contributed ΔdH = 230 kJ mole−1, ΔdS = 723 J mole−1 K−1, whereas a more thermostable domain unfolded with a Tm of 47.8°C and contributed ΔdH = 555 kJ mole−1, ΔdS = 1730 J mole−1 K−1. These figures indicate that modifications affecting the high-temperature energetic domain may involve a larger portion of the protein structure, at least assuming that the unfolding process does not involve the breakage of bonds of unusual strength (Dill 1990).

An external file that holds a picture, illustration, etc. Object name is 61237-18f7_L1TT.jpg

Best fit of the experimental DSC tracing obtained for Δloop RgDAAO using a 1.5 mg/mL protein concentration. The original trace is given as a thin solid line. The thick solid line is a fit performed according to a two-state model assuming two independent domains (see text). Dashed and dotted lines correspond to the contributions attributed to each thermodynamic domain.

When temperature-induced loss of secondary structure was monitored by following the loss of the CD signal at 220 nm in temperature-ramp experiments using the same temperature gradient applied in DSC studies (Fig. 8A ▶), slightly lower Tm values were found for both protein forms with respect to those detected by DSC (although at 0.5 mg protein/mL).

An external file that holds a picture, illustration, etc. Object name is 61237-18f8_L1TT.jpg

Temperature dependence of different spectroscopic signals for wild-type (circles) and Δloop (squares) RgDAAO. Proteins were in 50 mM potassium phosphate (pH 7.5), 2 mM EDTA, and 10% glycerol. Spectral signals were monitored continuously during progressive heating from 25°C to 75°C at a heating rate of 0.5°C/min, and are given as percent of the total observed change. (A) CD signal at 220 nm. (open symbols) 0.1 mg /mL, (filled symbols) 0.5 mg protein/mL. (B) Fluorescence intensities at a protein concentration of 0.5 mg/mL. (open symbols) Tryptophan fluorescence, (filled symbols) flavin fluorescence.

Tm figures for FAD release, as monitored in independent temperature-ramp fluorescence measurements (Fig. 8B ▶), were close to those obtained by DSC, and thus only slightly higher than those obtained when monitoring the loss of secondary structure elements. It is worth observing that a fivefold decrease in the protein concentration (from 0.5 to 0.1 mg protein/mL) only marginally affected Tm for the Δloop RgDAAO (see Table 1​1),), but caused a sensible decrease in Tm for the wild-type protein in terms of loss of secondary structure elements and of flavin release. This supports the higher stability of associated forms of dimeric wild-type RgDAAO, as inferred by the concentration dependence of thermal stability observed in DSC studies.

Studies on the temperature sensitivity of tryptophan fluorescence (taken as a reporter of tertiary structure modifications) confirmed the different inherent stability of the two proteins. These data also indicated that loss of tertiary structure elements involving hydrophobic regions of the protein occurs at much lower temperatures than all other modifications discussed earlier. This is made evident by the various tracings in Figure 8B ▶, which compares the temperature dependence of tryptophan and FAD fluorescence and provides an immediate visual appreciation of the behavior of the different signals in the two proteins. The corresponding Tm values are reported in Table 1​1.. (See Electronic Supplemental Material.) We attempted to calculate transition enthalpies from our spectroscopic data according to independent two-state transition models (van’t Hoff). Results of our calculations, when compared with the calorimetric measurements, indicated that the condition of independence between the spectroscopic variables is not completely satisfied.

In order to understand the relationships between thermal stability and presence of a dimeric structure in DAAO, we also performed some of the experiments described earlier in the presence of 0.5 M NH4SCN, that is, under conditions in which the wild-type protein is completely dissociated into monomers. Temperature-sensitivity of the spectroscopic features of Δloop RgDAAO mutant was also studied in the presence of dissociating concentrations of NH4SCN (0.5 M), for the purpose of (i) establishing some sort of reference with respect to the wild-type monomer and (ii) assessing the effects of NH4SCN on the βF5–βF6 loop region itself.

A summary of the Tm values obtained in the presence of 0.5 M NH4SCN for wild-type and Δloop DAAO, is reported in Table 2​2.. The two panels of Figure 9 ▶ show the original tracings for temperature-induced irreversible changes in tryptophan and flavin fluorescence in the presence and in the absence of 0.5 M NH4SCN. It has to be noted that the experiments presented in Table 2​2 and in the related Figure 9 ▶ were performed at a higher heating rate (2°C/min) with respect to those presented in Table 1​1 and Figure 8 ▶. The higher heating rate was necessary in order to minimize artefacts due to protein aggregation on prolonged exposure at high temperature in the presence of 0.5 M NH4SCN.

Table 2.

Comparison of Tm as determined by different approaches on wild type and Δloop mutant DAAO in the absence and in the presence of 0.5 M NH4SCN

Wild typeΔLoop mutant
Method−NH4SCN+NH4SCN−NH4SCN+NH4SCN
Trp fluorescence55.039.049.041.3
FAD fluorescence56.739.748.443.8

All experiments were performed in 50 mM potassium phosphate (pH 7.5), 2 mM EDTA, and 10% glycerol, at a heating rate of 2°C/min and using a protein concentration of 0.5 mg protein/mL. Tms were obtained by derivatization of the spectroscopic signals, and are uncorrected for delay effects. Standard deviation was within 0.2°C for all determined values.

An external file that holds a picture, illustration, etc. Object name is 61237-18f9_L1TT.jpg

Temperature dependence of different spectroscopic signals for wild-type (circles) and Δloop (squares) RgDAAO, in the absence (filled symbols) and in the presence (open symbols) of 0.5 M NH4SCN. (A) Tryptophan fluorescence, (B) flavin fluorescence. Proteins were 0.5 mg/mL in 50 mM potassium phosphate (pH 7.5), 2 mM EDTA, and 10% glycerol, and were heated from 25°C to 65°C at a heating rate of 2°C/min.

The presence of NH4SCN strongly affected the thermostability parameters for wild-type DAAO, whereas the Δloop RgDAAO mutant was affected to a lesser extent. In the presence of 0.5 M NH4SCN, both proteins were in the monomeric state at the beginning of the thermal treatment, but the wild-type RgDAAO monomer lost its tertiary structure and released flavin at lower temperatures than the Δloop mutant monomer. These observations clearly indicate that the effect of NH4SCN is not restricted to the dimerization interface. The shape of the temperature-dependence curves obtained in the presence of 0.5 M NH4SCN with either protein also indicates a modified cooperativity of the processes leading to irreversible thermal denaturation.

Discussion

We demonstrated here the possibility of obtaining a monomeric form of RgDAAO by treatment with 0.5 M NH4SCN without deletion of the βF5-βF6 loop. Such a conversion is reversible but hysteretic for kinetic reasons. Both protein forms denature at >1M NH4SCN. Scheme 1 ▶ summarizes in a pictorial form the events ensuing from thermal treatment of monomeric and dimeric RgDAAO, either produced by mutagenesis or by dissociation with lipophilic ions of the native dimer.

An external file that holds a picture, illustration, etc. Object name is 61237-18s1_L1TT.jpg

A schematic representation of the structural changes ensuing from progressive heating of dimeric wild-type and monomeric Δloop RgDAAO in the absence and in the presence of 0.5 M NH4SCN. The triangular protrusion identifies the βF5-βF6 loop. Round or square corners in the main rectangle (or in boxes around the cofactor) indicate changes in local structures, as detected by spectroscopy. (PF) Protein fluorescence; (FF) flavin fluorescence. T1–T4 indicate the “midpoint melting temperature” (Tm) for individual protein forms.

The monomeric and dimeric untreated DAAO differ in tryptophan and flavin fluorescence, as well as in their CD spectral properties in the near-UV region. The monomeric forms of DAAO obtained by site-directed mutagenesis and NH4SCN treatment are somewhat different, showing different CD spectra and different fluorescence/exposure of the FAD coenzyme (see Figs. 2 ▶, 4 ▶). However, flavin and protein fluorescence reach similar values at 0.5 M NH4SCN in both wild-type and Δloop DAAO (see Fig. 4 ▶). The change in flavin fluorescence is due to two different events: loss of the dimeric aggregation state (as indicated by the different flavin fluorescence of wild-type and Δloop DAAO in the absence of NH4SCN), and alteration of the tertiary structure specifically consequent to the addition of NH4SCN.

Inspection of the buried surface on each DAAO monomer for the “head-to-tail” model of RgDAAO dimer formation (Pollegioni et al. 2002) shows that Trp 243 in each subunit is placed in strict contact with the corresponding residue in the other subunit at the dimer interface (Fig. 10A ▶). Therefore, the spectral differences in the untreated proteins (Figs. 2 ▶, 3 ▶) may be indicative of changes in the environment of Trp 243 following the deletion of the βF5–βF6 loop and conversion to the monomeric state.

An external file that holds a picture, illustration, etc. Object name is 61237-18f10_C4TT.jpg

A schematic representation of the intermonomer interactions in dimeric wild-type RgDAAO. (A) Stacking of Trp-243 (red) in the buried region at the dimer interface (no other tryptophan residues are located in the buried surface between monomers). The individual peptide chains are in blue and green. (B) Left: Specific ion pairing involving the βF5-βF6 loop region of one monomer and the αI3′-αI3″ region of the symmetry-related monomer. The different peptide chains are in green and light blue, respectively. Right: Molecular surface representation showing positively charged residues in the βF5-βF6 loop region (blue) and negatively charged residues in the αI3′- αI3″ region (red).

When monomerization of wild-type RgDAAO is induced by the addition of modest amounts of thiocyanate, other factors have to be taken into consideration. The βF5–βF6 loop in wild-type RgDAAO contains three positively charged residues (Arg 314, Arg 318, and Lys 321) that interact with negatively charged residues belonging to αI3′ and αI3″ helices on the symmetry-related molecule (Asp 269, Glu 273, Glu 276; Fig. 10B ▶). Among all DAAOs, these two regions are unique to this protein (Pollegioni et al. 2002). Because nonlipophilic anions do not promote dissociation, thiocyanate is effective in that it affects both hydrophobic and charge contacts between monomers. By acting on these structural elements, thiocyanate may also affect the conformation of the flavin-binding domain (vide infra).

A second relevant finding is the identification of a clear sequence of events in the course of thermal unfolding of both protein forms that involves (in order of increasing temperature) (1) loss of tertiary structure, (2) loss of some elements of secondary structure, and (3) general unfolding of the protein structure, concomitant to FAD release.

Indeed, by comparing the spectroscopic evidence with the deconvolution of the DSC tracings (Fig. 11 ▶), one can observe that the signal originating from modification of the tertiary structure in the Δloop RgDAAO overlaps changes in the low-temperature energetic domain, whereas the high-temperature energetic domain unfolded with a temperature dependence almost identical to the one detected for loss of secondary structure and release of the cofactor. Apparently, the first low-temperature energetic domain relates to the unfolding of tertiary structure regions, whereas the second energetic domain relates to the loss of secondary structure elements and to the release of the cofactor at higher temperatures. The sequence of events leading to irreversible general unfolding of the protein and loss of the flavin cofactor have Tms that cover a >10°C temperature span (46.4°C to 56.9°C, with appreciable concentration effects) for the wild-type protein, but a much narrower temperature span (<6°C, from 42.7°C to 48.1°C, with only minor concentration effects) for the Δloop RgDAAO.

An external file that holds a picture, illustration, etc. Object name is 61237-18f11_L1TT.jpg

Comparison of the experimental DSC tracing (full line) with the first derivative of the tryptophan (dots) and flavin (dashes) fluorescence signals obtained for Δloop RgDAAO. The fluorescence data used for derivative calculation are those reported in Figure 8B ▶. The DSC data are the same as those reported in Figure 7 ▶.

We also confirm here the reported higher thermostability of dimeric, wild-type RgDAAO (Piubelli et al. 2002). In this frame, it is worth noting that DSC experiments gave a significantly higher denaturation enthalpy for wild-type than for mutant DAAO (1030 vs. 775 kJ/mole), concentration independent. Because the denaturation enthalpy is contributed mostly by the high-temperature energetic domain (at least in the Δloop RgDAAO, see Fig. 7 ▶), it is possible to envision a fundamental role of the factors affecting stability of the flavin-harboring domain in the overall stability of the protein. Indeed, DSC and temperature-dependent spectroscopic studies at various protein concentration (Table 1​1)) are supportive of a definite role of dimerization in promoting thermostability: the dimeric, untreated wild-type DAAO possesses the highest Tm (T1 > T2, see Scheme 1 ▶). As for the possible molecular determinants of this property, it is worth noting that the hydrophobic regions including reporter tryptophans in the thiocyanate-treated proteins (most likely Trp 243, see Fig. 10A ▶) attained a similar degree of unfolding at room temperature and had almost identical temperature sensitivity (Fig. 9A ▶). However, this does not hold true for the flavin-binding domain, that—although similar in both proteins at 15°C in 0.5 M NH4SCN (Fig. 2 ▶)—showed a much higher temperature sensitivity in the thiocyanate-treated wild-type RgDAAO (Fig. 9B ▶). This observation, along with the independence of the energetic domains made evident by DSC studies on the monomeric Δloop mutant RgDAAO (see Fig. 11 ▶ and related comments), may be taken as an evidence of the absence of a strict structural relationship among the two regions in each DAAO monomer, but indicates that the relationships between the two regions are remarkably different for the wild-type protein dimer. Indeed, because this latter in 0.5 M thiocyanate dissociates into a monomer, it is easy to envision a primary role of dimerization in the stabilization of the flavin-harboring regions of the structure. In particular, thiocyanate acts on the wild-type dimeric RgDAAO by affecting ionic pairs between secondary elements of the FAD-binding domain of one monomer and those of the intersubunit domain of the symmetry-related monomer (Fig. 10B ▶). Thus, the effects of thiocyanate on the stability of RgDAAO are indicative of a correlation between the presence or the conformation of the βF5–βF6 loop in RgDAAO (not conserved in other known DAAO sequences and allowing stabilization of the dimeric state) with the tight binding of the flavin coenzyme (10-fold and 5-fold tighter in yeast dimeric DAAO than in the monomeric mammalian enzyme and in the Δloop monomeric RgDAAO, respectively; Casalin et al. 1991; Curti et al. 1992; Piubelli et al. 2002).

The relevance of our results may extend beyond mere comparison among different DAAOs. From a structural standpoint, RgDAAO is a component of the large GR family. Based on both structural and sequence homologies (Pollegioni et al. 2002), all the family members adopt the Rossmann fold (Rossmann et al. 1974). In particular, DAAOs belong to the GR2 subgroup, characterized by sequence similarity mainly within 30 residues in the N-terminal region, which represents the dinucleotide-binding motif (Dym and Eisenberg 2001). Interestingly, high topological similarity is found within the GR2 subfamily in the “flavin-binding domain”, whereas only parts of the “interface domain” can be superimposed in pair-wise fashion (Pollegioni et al. 2002). The combination of such structural considerations and our results indicates that the molecular mechanisms regulating dimeric DAAO holoenzyme formation may require two events: (1) binding of the FAD cofactor to the monomeric apoprotein, allowing the formation of the monomeric holoenzyme and (2) electrostatic interaction between positively charged residues of the βF5–βF6 loop and the negatively charged residues of the αI3′ and αI3″ secondary elements belonging to a different monomer, and hydrophobic interaction between Trp 243 residues belonging to two different subunits, leading to dimer stabilization. Whereas cofactor binding, which involves a large portion of the protein, may be common to all members of this class, the second event should be specific of yeast DAAO because it involves structural elements not present in other flavoenzymes (Dym and Eisenberg 2001; Pollegioni et al. 2002).

Besides the academic interest concerning the structure–function relationships in flavoenzymes, knowledge about protein stability is nowadays being exploited in many practical applications in biotechnology and is thus also of industrial importance. In this context, the yeast DAAO is extensively used in the bioconversion of cephalosporin C to 7-aminocephalosporanic acid, a two-step route that uses DAAO and glutaryl-7-ACA acylase in sequence. The two activities require two distinct reactors because the operational stability of DAAO is significantly lower than that of acylase (for a recent review, see Pilone and Pollegioni 2002). The ensuing complications make it worthwhile to take a closer look at the structural determinants of DAAO thermal stability, also from an economic standpoint.

Materials and methods

Purification of wild-type and Δloop RgDAAO, and DAAO activity assay

Recombinant RgDAAO was expressed and purified from E. coli cells as described previously (Molla et al. 1998). Production and properties of DAAO Δloop mutant were described elsewhere (Piubelli et al. 2002). Starting from a 10-L fermentation broth, 180 mg and 80 mg of pure enzyme with a specific activity of 110 and 86 U/mg protein were obtained for wild-type and Δloop DAAO, respectively. The enzyme concentration was determined by using an extinction coefficient of 12.6 mM−1cm−1 for wild-type and 11.3 mM−1cm−1 for Δloop DAAOs (Pilone 2000; Piubelli et al. 2002). DAAO activity was assayed with an oxygen electrode at pH 8.5, air saturation, and 25°C, using 28 mM D-alanine as substrate in the presence of 0.2 mM FAD (Pilone 2000).

Size-exclusion chromatography

Size-exclusion chromatography was performed at room temperature on a Superdex 200 column using an Äkta chromatographic system (Amersham Pharmacia Biotech), at a flow rate of 0.5 mL/min. The eluant was 50 mM potassium phosphate buffer (pH 7.5) containing 0.25 M sodium chloride, 10% glycerol, 2 mM EDTA, and the appropriate concentration of NH4SCN. The column was calibrated with suitable standard proteins.

Spectroscopy

All fluorescence measurements at fixed temperature were performed in a Jasco FP250 or FP750 instrument equipped with a thermostated cell holder by using a 1-mL cell. Tryptophan emission spectra were recorded using an excitation wavelength of 280 nm and flavin emission spectra were recorded using an excitation wavelength of 450 nm. Emission and excitation bandwidths were set at 10 nm. Temperature-ramp experiments were performed in a software-driven, Peltier-equipped Perkin-Elmer LS 50 fluorometer that allowed reproduction of the same temperature gradient (0.5°C/min) used in DSC studies. Emission and excitation bandwidths were set at 5 nm. When necessary, heating rates of 2 °C/min were also used. Tryptophan emission spectra were taken from 300 to 450 nm using an excitation wavelength of 298 nm. Flavin emission spectra were recorded from 460 to 600 nm using an excitation wavelength of 455 nm. Fixed wavelength measurements were taken with emission wavelengths of 340 and 526 nm for tryptophan and flavin fluorescence, respectively. CD spectra were recorded on a J-810 Jasco spectropolarimeter, also fitted with a software-driven Peltier-based temperature controller, and analyzed by means of Jasco software. The cell path was 1 cm for measurements above 250 nm, and 0.1 cm for measurements in the 190- to 250-nm region. Proteins were in 50 mM potassium phosphate buffer (pH 7.5) containing 10% glycerol and 2 mM EDTA.

Calorimetry

Calorimetric measurements were carried out on solutions containing 0.5–3.5 mg protein/mL in the same buffers used for spectroscopic studies. A third-generation Setaram Micro-DSC apparatus was used, which is suitable for dilute solutions of biological macromolecules in the temperature range from −20°C to 120°C. Scan rate was 0.5°C/min. Data were analyzed by means of the THESEUS software (Barone et al. 1992). The excess molar heat capacity <ΔCp> or equation M1, i.e., the difference between the apparent molar heat capacity Cp(T) of the sample and the molar heat capacity of the “native state,” Cp,N(T), was recorded across the scanned temperature range. Cp,N(T) was obtained according to published procedures (Freire and Biltonen 1978) by linear regression of experimental Cp(T) data in the predenaturation region (namely, for T < Ti, where Ti is the highest temperature at which only the native protein is present). The overall calorimetric denaturation enthalpy ΔdH was determined by integration of Cp(T) across the denaturation (Ti ÷ Tf) range, where Tf is the final temperature of the denaturation process, taking into account that the baseline has a sigmoidal trend across the same temperature range. Such a trend can be obtained by assigning a weight factor equal to the ratio incremental area/total area, drawn from the earlier integration, and iterating the overall process until convergence (Barone et al. 1992). As a first step of the iteration procedure, a straight line passing through Cp(Ti) and Cp(Tf) was used as the tentative baseline. All the thermograms reported in this paper were accordingly scaled. Because of the short extension of the baseline before the peaks and, in some cases, of the presence of aggregation phenomena, the heat capacity drop ΔdCp that corresponds to the Cp change on passing from native to the denatured state across the signal was affected by a rather large error and was therefore not taken into account. The procedures described earlier leveled off any uncertainty about ΔdCp and allowed a reliable evaluation of Cp(T) and of the corresponding enthalpy. The theoretical models used to fit the experimental data were tested through the nonlinear Levenberg-Marquardt method (Press et al. 1989).

Acknowledgments

This work was supported by a grant from FAR 2002 to L. Pollegioni and from Fondo Progetti di Eccellenza 2001—University of Insubria to M.S. Pilone and L. Pollegioni. Individual FIRST grants (MiUR, Rome, Italy) supported the other principal investigators (S. Iametti, F. Bonomi). We thank Dr. Gianluca Molla for assistance in preparing Figure 10 ▶.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.

Abbreviations

  • DAAO, D-amino acid oxidase (EC 1.4.3.3)

  • RgDAAO, Rhodotorula gracilis D-amino acid oxidase

  • DSC, differential scanning calorimetry

  • GR, glutathione reductase

Notes

Supplemental material: See www.proteinscience.org.

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.0234603.

References

  • Barone, G., Del Vecchio, P., Fessas, D., Giancola, C., and Graziano, G. 1992. THESEUS: A new software package for the handling and analysis of thermal denaturation data of biological macromolecules. J. Thermal Analysis 38 2779–2790. [Google Scholar]
  • ———. 1994. Denaturation of biological macromolecules: New programs for the deconvolution of DSC measurements. In Chemistry and properties of biomolecular systems (eds. N. Russo, J. Anastassopoulou, et al.), pp. 67–78. Kluwer, Amsterdam, The Netherlands.
  • Casalin, P., Pollegioni, L., Curti, B., and Pilone Simonetta, M. 1991. A study on apoenzyme from Rhodotorula gracilis D-amino acid oxidase. Eur. J. Biochem. 197 513–517. [PubMed] [Google Scholar]
  • Curti, B., Ronchi, S., and Pilone Simonetta, M. 1992. D- and L-amino acid oxidases. In Chemistry and biochemistry of flavoenzymes (ed. F. Müller), Vol. 3, pp. 69–94. CRC, Boca Raton, FL.
  • D’Auria, S., Barone, R., Rossi, M., Nucci, R., Barone, G., Fessas, D., Bertoli, E., and Tanfani, F. 1997. Effects of temperature and SDS on the structure of β-glycosidase from the thermophilic archaeon Sulfolobus solfataricus. Biochem. J. 323 833–840. [PMC free article] [PubMed] [Google Scholar]
  • Decker, K. 1991. Covalent flavoprotein. In Chemistry and biochemistry of flavoenzymes (ed. F. Müller), Vol. 2, pp. 343–393. CRC, Boca Raton, FL.
  • Dill, K.A. 1990. Dominant forces in protein folding. Biochemistry 29 7133–7155. [PubMed] [Google Scholar]
  • Dym, O. and Eisenberg, D. 2001. Sequence-structure analysis of FAD-containing proteins Protein Sci. 10 1712–1728. [PMC free article] [PubMed] [Google Scholar]
  • Edge, V., Allewell, N.M., and Sturtevant, J.M. 1985. High-resolution differential scanning calorimetric analysis of the subunits of Escherichia coli aspartate transcarbamoylase. Biochemistry 24 5899–5906. [PubMed] [Google Scholar]
  • Fessas, D., Iametti, S., Schiraldi, A., and Bonomi, F. 2001. Thermal unfolding of monomeric and dimeric β-lactoglobulins. Eur. J. Biochem. 268 5439–5448. [PubMed] [Google Scholar]
  • Freire, E. and Biltonen, R.L. 1978. Thermodynamics of transfer ribonucleic acids: The effect of sodium on the thermal unfolding of yeast tRNAPhe. Biopolymers 17 463–479. [PubMed] [Google Scholar]
  • Giancola, C., De Sena, C., Fessas, D., Graziano, G., and Barone, G. 1997. DSC studies on bovine serum albumin denaturation. Effects of ionic strength and SDS concentration. Int. J. Biol. Macromol. 20 193–204. [PubMed] [Google Scholar]
  • Manly, S.P., Matthews, K.S., and Sturtevant, J.M. 1985. Thermal denaturation of the core protein of lac repressor. Biochemistry 24 3842–3846. [PubMed] [Google Scholar]
  • Massey, V. and Hemmerich, P. 1980. Active-site probes of flavoproteins. Biochem. Soc. Trans. 8 246–255. [PubMed] [Google Scholar]
  • Molla, G., Vegezzi, C., Pilone, M.S., and Pollegioni, L. 1998. Overexpression in Escherichia coli of a recombinant chimeric Rhodotorula gracilis D-amino acid oxidase. Prot. Expr. Purif. 14 289–294. [PubMed] [Google Scholar]
  • Pilone, M.S. 2000. D-amino acid oxidase: New findings. Cell. Mol. Life Sci. 57 1732–1747. [PubMed] [Google Scholar]
  • Pilone, M.S. and Pollegioni, L. 2002. D-amino acid oxidase as an industrial biocatalyst. Biocatalysis and Biotransformation 20 145–159. [Google Scholar]
  • Piubelli, L., Caldinelli, L., Molla, G., Pilone, M.S., and 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. [PubMed] [Google Scholar]
  • Pollegioni, L., Diederichs, K., Molla, G., Umhau, S., Welte, W., Ghisla, S., and Pilone, M.S. 2002. Yeast D-amino acid oxidase: Structural basis of its catalytic properties. J. Mol. Biol. 324 535–546. [PubMed] [Google Scholar]
  • Press, W.H., Flannery, B.P., Teukolsky, S.A., and Vetterling, W.T. 1989. In Numerical recipes. Cambridge University Press, Cambridge, UK.
  • Ptistyn, O.B. and Semisotnov, G.V. 1991. The mechanism of protein folding. In Conformations and forces in protein folding (eds. B.T. Nall and K.A. Dill), pp. 155–168. American Association for the Advancement of Sciences, Washington, DC.
  • Rossmann, M.G., Moras, D., and Olsen, K.W. 1974. Chemical and biological evolution of nucleotide-binding protein. Nature 250 194–199. [PubMed] [Google Scholar]
  • Umhau, S., Pollegioni, L., Molla, G., Diederichs, K., Welte, W., Pilone, M.S., and Ghisla, S. 2000. The x-ray structure of D-amino acid oxidase at very high resolution identifies the chemical mechanism of flavin-dependent substrate dehydrogenation. Proc. Natl. Acad. Sci. 97 12463–12468. [PMC free article] [PubMed] [Google Scholar]
  • van Mierlo, C.P.M. and Steensma, E. 2000. Protein folding and stability investigated by fluorescence, circular dichroism (CD), and nuclear magnetic resonance (NMR) spectroscopy: The flavodoxin story. J. Biotechnol. 79 281–298. [PubMed] [Google Scholar]
  • Waterham, H.R., Russell, K.A., de Vries, Y., and Cregg, J.M. 1997. Peroxisomal targeting, import, amd assembly of alcohol oxidase in Pichia pastoris. J. Cell Biol. 139 1419–1431. [PMC free article] [PubMed] [Google Scholar]
Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society