Unraveling multistate unfolding of pig kidney fructose-1,6- bisphosphatase using single tryptophan mutants Heide C. Ludwig, Fabian N. Pardo*, Joel L. Asenjo*, Marco A. Maureira, Alejandro J. Yan ˜ ez and Juan C. Slebe Instituto de Bioquı ´ mica, Facultad de Ciencias, Universidad Austral de Chile, Valdivia, Chile Fructose-1,6-bisphosphatase (EC 3.1.3.11, FBPase) cat- alyzes a control step in the gluconeogenic pathway, the hydrolysis of fructose-1,6-bisphosphate [Fru(1,6)P 2 ]to fructose-6-phosphate and inorganic phosphate. Diva- lent metal ions such as Mg +2 ,Mn +2 or Zn +2 are required for catalytic activity [1,2]. FBPase is inhibited synergistically by AMP and fructose-2,6-bisphosphate [Fru(2,6)P 2 ]. AMP binds to an allosteric site and its inhibition is cooperative, whereas Fru(2,6)P 2 is a competitive inhibitor, that binds to the active site, according to structural and kinetic evidence [3,4]. The pig kidney FBPase is a homotetramer having D 2 symmetry with a relative molecular mass of 146 000 [5]. The crystal structures of this enzyme com- plexed with various ligands have been solved [4,6–8] (pdb: 1FPB; 1FRP; 1FBF). The four subunits of FBPase are designated C1, C2, C3 and C4 and are labeled clockwise. The C1 and C2 subunits correspond Keywords fructose-1,6-bisphosphatase; protein unfolding; single tryptophan mutants; tetrameric intermediate; phase diagram Correspondence J. C. Slebe, Instituto de Bioquı ´ mica, Universidad Austral de Chile, Campus Isla Teja, Valdivia, Chile Fax: +56 63 221406 Tel: +56 63 221797 E-mail: jslebe@uach.cl *These authors contributed equally to this work (Received 18 June 2007, revised 14 August 2007, accepted 21 August 2007) doi:10.1111/j.1742-4658.2007.06059.x Pig kidney fructose-1,6-bisphosphatase is a homotetrameric enzyme which does not contain tryptophan. In a previous report the guanidine hydrochlo- ride-induced unfolding of the enzyme has been described as a multistate process [Reyes, A. M., Ludwig, H. C., Yan ˜ ez, A. J., Rodriguez, P. H and Slebe, J. C. (2003) Biochemistry 42, 6956–6964]. To monitor spectroscopi- cally the unfolding transitions, four mutants were constructed containing a single tryptophan residue either near the C1–C2 or the C1–C4 intersubunit interface of the tetramer. The mutants were shown to retain essentially all of the structural and kinetic properties of the enzyme isolated from pig kid- ney. The enzymatic activity, intrinsic fluorescence, size-exclusion chromato- graphic profiles and 1-anilinonaphthalene-8-sulfonate binding by the mutants were studied under unfolding equilibrium conditions. The unfold- ing profiles were multisteps, and formation of hydrophobic structures was detected. The enzymatic activity of wild-type and mutant FBPases as a function of guanidine hydrochloride concentration showed an initial enhancement (maximum  30%) followed by a biphasic decay. The activity and fluorescence results indicate that these transitions involve conforma- tional changes in the fructose-1,6-bisphosphate and AMP domains. The representation of intrinsic fluorescence data as a ‘phase diagram’ reveals the existence of five intermediates, including two catalytically active inter- mediates that have not been previously described, and provides the first spectroscopic evidence for the formation of dimers. The intrinsic fluores- cence unfolding profiles indicate that the dimers are formed by selective disruption of the C1–C2 interface. Abbreviations ANS, 8-anilinonaphthalene-1-sulfonate; FBPase, fructose-1,6-bisphosphatase; Fru(1,6)P 2 , fructose-1,6-bisphosphate; Fru(2,6)P 2 , fructose-2,6- bisphosphate; GdmCl, guanidinium chloride. FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS 5337 to the upper dimer and the C3 and C4 subunits to the lower dimer. Each subunit of the enzyme can be divided into two folding domains: residues 1–200 con- stitute the AMP domain, and residues 201–337 the Fru(1,6)P 2 domain. The AMP domain has the AMP binding site at the C1–C4 interface and the Fru(1,6)P 2 domain contains the active site at the C1–C2 interface. Two quaternary conformations have been established, the R- and the T-forms, that differ by a 17° rotation of the lower dimer C3C4 relative to the upper dimer C1C2 [9,10] (pdb: 1FBP; 4FBP). AMP induces the transition from the active R-form to the inactive (or less active) T-form. Understanding the folding ⁄ unfolding and self-assem- bly processes of oligomeric proteins remains a major problem. Equilibrium denaturation studies of such proteins provide important information on the rela- tionship of folding and oligomerization processes and on the influence of quaternary structure on protein sta- bility [11,12]. In a previous publication from this labo- ratory [13] the unfolding of pig kidney FBPase induced by GdmCl was investigated. In contrast to an earlier study [14] that suggested that inactivation and dissociation occur simultaneously, we demonstrated the existence of an inactive tetrameric intermediate. Furthermore, it was shown that the equilibrium unfolding pathway is characterized by the presence of three intermediate states. In these studies, fluorescent reporter groups (2-(4 ¢-maleimidylanilino)naphthalene- 6-sulfonic acid and N-(acetylaminoethyl)-5-naphthyl- amine-1-sulfonic acid) were chemically attached to Cys128, a reactive thiol group located near to the active site to monitor conformational changes and enzyme dissociation. However, the introduction of these fluorescent groups caused a destabilization of the active site region. Furthermore, at high protein con- centration (1 mgÆmL )1 ) the aggregation of dimeric and monomeric unfolding intermediates masked the transi- tions occurring at GdmCl concentrations above 1.2 m. However, no large aggregates have been detected by light scattering measurements at 50 lgÆmL )1 [13]. Finally, a main unresolved question is which of the FBPase interfaces is broken first by the GdmCl treat- ment. As FBPase does not contain tryptophan, introduc- tion of this fluorescent amino acid by site-directed mutagenesis as nonperturbing probe is an attractive experimental approach to examine the unfolding of the enzyme at low protein concentration. The tryptophan probe, which is very sensitive to a variety of environ- mental conditions, yields structural and dynamic infor- mation about its surroundings [15]. In the present report, FBPase mutants carrying a single replacement of a Phe at position 16, 89, 219 or 232 by Trp were engineered (Fig. 1). Phe16 and Phe89 are residues of the AMP domain located near the C1–C4 interface, whereas Phe219 and Phe232 are in the Fru(1,6)P 2 domain and near the C1–C2 interface. The single-Trp mutants were shown to retain essentially all of the structural and kinetic properties of the enzyme isolated from pig kidney. The GdmCl-induced unfolding transi- tions studied by fluorescence spectroscopy provide evi- dence for the existence of five unfolding intermediates and indicate that the loss of quaternary structure begins by disruption of the C1–C2 interface. Results Catalytic and spectroscopic properties of tryptophan mutants of FBPase The single tryptophan mutants, Phe16Trp, Phe89Trp, Phe219Trp and Phe232Trp FBPases exhibited identical electrophoretic mobility ( 37 kDa) as FBPase isolated from pig kidney and were at least 96% pure using SDS ⁄ PAGE as a criterion (data not shown). As seen in Table 1, the mutations in general do not affect cata- lytic properties significantly, except the loss of AMP cooperativity (h value ¼ 1) observed for Phe16Trp FBPase. The other kinetic parameters only demon- strate slight differences with respect to the recombinant wild-type FBPase and are similar to those published Fig. 1. Schematic of FBPase showing the location of the trypto- phan residues. Active sites and AMP binding sites are labeled FBP and AMP, respectively. Dotted ovals represent ligand binding sites on faces of the tetramer hidden from view. The FBPase tetramer is in the T-state conformation. The location of the phenylalanine resi- dues which were mutated is shown. Selective GdmCl disruption of FBPase C1–C2 interface H. C. Ludwig et al. 5338 FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS elsewhere for nonrecombinant FBPase [16–18]. The CD spectra of the nonrecombinant, recombinant wild- type and mutant FBPases were essentially superimpos- able from 200 to 250 nm (data not shown). Emission spectra of equimolar amounts of the Trp mutants, when excited at 295 nm, are shown in Fig. 2. The emission maxima of the mutants are summarized in Table 2. Phe16Trp and Phe219Trp FBPases have emission maxima at 338 and 335 nm, respectively, indi- cating that these tryptophan residues are located in a nonpolar environment [15]. In contrast, the emission maxima of Phe89Trp and Phe232Trp FBPases are at 356 nm and 352 nm, respectively, indicating that these tryptophan residues are in a polar environment, exposed to the solvent. Figure 2 also shows that Phe232Trp FBPase presents the highest fluorescence quantum yield, whereas the quantum yield of Phe16Trp FBPase is considerably lower than those of the other mutants. The environment of a specific tryptophan residue can also be evaluated by its accessibility to a collisional fluorescence quencher, as acrylamide [19,20]. Table 2 presents the results of the Stern–Volmer analysis of the quenching data of the tryptophan mutants by acrylam- ide. The values of the Stern–Volmer constants (K SV ) indicate that Trp219 is shielded from the solvent (K SV ¼ 3.19 m )1 ), Trp16 (K SV ¼ 5.81 m )1 ) and Trp89 (K SV ¼ 6.28 m )1 ) are moderately accessible and Trp232 (K SV ¼ 11.8 m )1 ) is almost completely solvent exposed. These results agree with the crystallographic structure of the enzyme [21]. Examination of protein unfolding by catalytic activity, size-exclusion high-performance liquid chromatography and 8-anilinonaphthalene- 1-sulfonate (ANS) binding Enzyme activity can be regarded as the most sensitive probe for studying protein unfolding, as it reflects sub- tle readjustments of the active site and detects very small conformational variations of an enzyme struc- ture. Figure 3 shows the changes in enzymatic activity of the nonrecombinant, recombinant wild-type and the mutant pig kidney FBPases as a function of GdmCl Table 1. Kinetic parameters for wild-type and single tryptophan mutants of pig kidney FBPase. Enzyme k cat K m Fru(1,6)P 2 I 50 Fru(2,6)P 2 I 50 AMP h AMP K a Mg +2 s )1 lM lM lM mM Nonrecombinant 20.7 ± 1.0 5.9 ± 0.6 0.9 ± 0.2 10.2 ± 0.2 2.2 ± 0.1 0.16 ± 0.01 Wild-type 19.7 ± 0.9 4.8 ± 0.6 1.0 ± 0.2 7.1 ± 0.2 2.1 ± 0.1 0.28 ± 0.01 Phe16Trp 18.1 ± 1.2 5.0 ± 0.9 0.7 ± 0.3 5.7 ± 0.4 1.0 ± 0.1 0.42 ± 0.03 Phe89Trp 19.2 ± 0.8 4.3 ± 0.8 0.7 ± 0.2 3.1 ± 0.2 1.8 ± 0.2 0.87 ± 0.04 Phe219Trp 18.7 ± 1.3 6.0 ± 1.9 1.6 ± 0.3 3.4 ± 0.3 1.5 ± 0.3 0.64 ± 0.04 Phe232Trp 14.2 ± 0.9 4.6 ± 1.2 1.9 ± 0.1 4.6 ± 0.1 2.2 ± 0.2 0.66 ± 0.06 Fig. 2. Fluorescence emission spectra of FBPase mutants. Each enzyme was 60 lgÆmL )1 in 20 mM Tris ⁄ HCl buffer, pH 7.5, contain- ing 0.1 m M EDTA. The excitation wavelength was 295 nm. The var- ious traces correspond to the following samples: – Æ –, Phe16Trp; – – –, Phe89Trp; ÆÆÆÆ, Phe 219Trp; – ÆÆ – ÆÆ , Phe232 Trp; –––, recombi- nant wild-type FBPase. Table 2. Fluorescence properties of the single tryptophan mutants of pig kidney FBPase. The Stern–Volmer quenching constants for acrylamide (K SV ) were determined in 20 mM Tris ⁄ HCl buffer, pH 7.5, containing 0.1 m M EDTA as described under Experimental procedures. Enzyme k max nm K SV M )1 Phe16Trp 338 5.81 ± 0.16 Phe89Trp 356 6.28 ± 0.20 Phe219Trp 335 3.19 ± 0.18 Phe232Trp 352 11.84 ± 0.54 H. C. Ludwig et al. Selective GdmCl disruption of FBPase C1–C2 interface FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS 5339 concentration at 15 °C. A similar behavior can be observed for the six enzymes: after an enhancement in enzymatic activity (maximum  30%) a decrease of the activity occurs. According to control experiments the residual GdmCl concentrations (2–20 lm) in the assay medium do not affect the enzymatic activity. The maximum activity is observed at 0.2 m (Phe89Trp, Phe219Trp and Phe232Trp FBPases), 0.3 m (Phe16Trp and recombinant wild-type FBPases) and 0.4 m GdmCl (nonrecombinant FBPase). Two phases can be distinguished in the activity decrease, an initial phase of slight decay followed by a sharp decrease. In accor- dance with previous data [13] the midpoint for GdmCl-based inactivation for the nonrecombinant enzyme is 0.75 m. The recombinant enzymes are less resistant to GdmCl inactivation than the nonrecombi- nant FBPase, as indicated by the lower denaturant concentration required for half-maximum inactivation: recombinant wild-type and Phe219Trp FBPases, 0.70 m GdmCl; Phe16Trp and Phe232Trp FBPases, 0.64 m GdmCl and Phe89Trp FBPase, 0.58 m GdmCl. It has been described that the inactivation of non- recombinant FBPase takes place without dissociation of the tetramer, and therefore the enzyme at 0.9 m GdmCl elutes as a single peak from a size-exclusion column pre-equilibrated with the same solvent [13]. The elution profiles of the tryptophan mutants of FBPase at various concentrations of GdmCl were obtained (data not shown). Between 0 and 0.9 m GdmCl the enzymes elute as a single peak centered at 7.5 min, indicating that the mutants maintain their tet- rameric structure. A shoulder at a higher elution time (aproximately 8.0 min) is observed in the elution pat- terns at 1.0 and 1.2 m GdmCl, indicating the presence of dimers (relative molecular mass  70 000), as has been described for the nonrecombinant enzyme [13]. ANS, a hydrophobic fluorophore, can be used as an external probe for the unfolding of proteins [22]. This fluorophore has a low emission in aqueous solutions, but its fluorescence is increased in nonpolar environ- ments in such a way that the changes in ANS fluores- cence are related to the increase in accessible hydrophobic surface upon protein unfolding. As shown in Fig. 4, there is a sharp rise in ANS fluores- cence and thus in ANS binding to Phe89Trp FBPase between 0.4 m and 0.6 m GdmCl. This transition is coincident with the loss of catalytic activity. Beyond 0.7 m GdmCl the ANS emission shows a gradual decrease, reflecting the disappearance of the hydropho- bic patches where ANS binds. In the case of the Phe16Trp, Phe219Trp and Phe232Trp mutants the increase in ANS binding takes place approximately between 0.5 m and 0.9 m GdmCl, and is also coinci- dent with the loss of catalytic activity. These results are similar to those described for nonrecombinant pig kidney FBPase [13]. Monitoring changes of the intrinsic tryptophan fluorescence The fluorescence of the indole ring is highly sensitive to its environment; this makes tryptophan an ideal res- idue to detect conformational changes of protein mole- cules [15]. GdmCl-induced denaturation of the tryptophan mutants was monitored by the change in fluorescence emission spectra at an excitation wave- length of 295 nm. The results were plotted by taking the average emission wavelength [23] and the fluores- cence intensity at the emission maximum of each mutant in the native state versus GdmCl concentra- tion. The average emission wavelength was used instead of k max because it is a more sensitive value as it reflects changes in the shape of the spectrum as well as in position. The unfolding curves (Fig. 5) are mostly biphasic or triphasic and differ greatly in shape. All of the tryptophan residues detect the transition by which enzymatic activity is lost. When enzymatic activity and fluorescence were measured for the same samples of an unfolding experiment, a perfect coincidence between catalytic activity loss, change of average emission GdmCl, M 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Enzymatic activity, % 0 20 40 60 80 100 120 140 Fig. 3. Enzyme activity of wild-type and mutant FBPases, as a func- tion of GdmCl concentrations. Samples of nonrecombinant FBPase (d), recombinant wild-type (h), or the mutant enzymes Phe16Trp (s), Phe89Trp (m ), Phe219Trp (n) or Phe232Trp (j) (50 lgÆmL )1 ) were incubated at different concentrations of GdmCl. The denatur- ant effect was then evaluated measuring enzyme activity, as described in Experimental procedures. Selective GdmCl disruption of FBPase C1–C2 interface H. C. Ludwig et al. 5340 FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS wavelength and emission intensity was obtained for Phe16Trp and Phe89Trp FBPases. The emission intensity of Phe16Trp FBPase increases at low GdmCl concentrations in a biphasic way (Fig. 5). The first phase, between 0 and 0.2 m GdmCl, correlates with the increase in enzymatic activ- ity (Fig. 3), and the second phase, between 0.55 and 0.8 m GdmCl, correlates with the activity loss. At denaturant concentrations higher than 0.8 m GdmCl the emission intensity decreases. The average emission wavelength is shifted gradually towards higher values, and a pronounced increase of this parameter is observed between 2.0 and 2.7 m GdmCl. A probable cause for this pronounced increase is the disruption of the C1–C4 interface next to Trp16, which exposes the tryptophan residue to the solvent. For Phe89Trp FBPase a large decrease of the aver- age emission wavelength is observed between 0.4 and 0.7 m GdmCl (Fig. 5) correlated with a decrease in the fluorescence intensity. Notably, the maximum emission wavelength (k max ) value of the emission spectrum at 0.8 m GdmCl is 339 nm, characteristic for a nonpolar environment. A shift of the average emission wave- length in the opposite direction between 1.8 and 2.5 m GdmCl indicates that the tryptophan residue now moves into a polar environment. In accordance with the results obtained for Phe16Trp FBPase, this red shift probably corresponds to the disruption of the C1–C4 interface next to Trp89. The unfolding curves of Phe219Trp FBPase, which contains a tryptophan residue located near the C1– C2 interface, have certain features differing from those of the mutants with a tryptophan residue near the C1–C4 interface. The fluorescence intensity of Phe219Trp FBPase decreases in a transition that extends beyond 0.9 m GdmCl (Fig. 5), a concentra- tion at which the catalytic activity is completely lost. The intensity decrease and the increase of the average emission wavelength between 0.9 and 1.4 m GdmCl probably is caused by the disruption of the C1–C2 interface next to Trp219. Moreover, for Phe219Trp FBPase only a modest increase in the average emis- sion wavelength (less than 30% of the total increase) and no change of the emission intensity is detected between 1.8 and 2.7 m GdmCl. The effect of GdmCl on the fluorescence intensity of Phe232Trp FBPase is similar to that of the Phe219Trp mutant (Fig. 5). This tryptophan residue, also located near the C1–C2 GdmCl, M 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Enzymatic activity, % 0 20 40 60 80 100 120 140 Relative fluorescence at 480 nm 6 8 10 12 14 16 18 20 Phe89Trp GdmCl, M 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Enzymatic activity, % 0 20 40 60 80 100 120 140 Relative fluorescence at 480 nm 6 8 10 12 14 16 18 Phe232Trp GdmCl, M 0,0 0,5 1,0 1 ,5 2,0 2,5 3,0 Enzymatic activity, % 0 20 40 60 80 100 120 140 Relative fluorescence at 480 nm 6 8 10 12 14 16 18 Phe219Trp GdmHCl, M 0,0 0,5 1,0 1,5 2,0 2,5 3,0 Enzymatic activity, % 0 20 40 60 80 100 120 140 6 8 10 12 14 16 18 Relative fluorescence at 480 nm Phe16Trp Fig. 4. ANS fluorescence and catalytic activity of FBPase mutants at different concentrations of GdmCl. Samples of Phe16Trp, Phe89Trp Phe219Trp and Phe232Trp FBPases (50 lgÆmL )1 ) were denatured by GdmCl. Catalytic activity (d) and ANS emission (s) were measured as described in Experimental procedures. The final ANS concentration was 40 l M. H. C. Ludwig et al. Selective GdmCl disruption of FBPase C1–C2 interface FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS 5341 interface, is already in a polar environment in the native state (Table 2), and therefore only minor changes of the average emission wavelength are observed. Acrylamide quenching of Phe89Trp FBPase intrinsic fluorescence As the blue shift of the emission spectrum of Phe89Trp FBPase during denaturation is rather unusual, quench- ing studies were performed. The Stern–Volmer plots for acrylamide quenching are shown in Fig. 6 for the mutant in the native state and after denaturation by different GdmCl concentrations. The quenching plots are linear within the concentration range used. Consis- tent with the changes of the average emission wave- length (Fig. 5), at 0.7 m and at 1.2 m GdmCl the tryptophan residue is considerably more shielded from the solvent (K SV ¼ 2.90 ± 0.10 m )1 and K SV ¼ 2.69 ± 0.12 m )1 , respectively) than in the native state (K SV ¼ 6.28 ± 0.20 m )1 ). At 2.4 m GdmCl, an increase of the Stern–Volmer constant to Acrylamide, M 0.0 0.1 0.2 0.3 0.4 0.5 0.6 F 0 / F 0 1 2 3 4 5 6 7 Fig. 6. Stern-Volmer–plots of acrylamide quenching of Phe89Trp FBPase denatured by different GdmCl concentrations. Phe89Trp FBPase in 0.1 M Hepes-NaOH buffer, pH 7.5, containing 0.1 mM EDTA, 5 mM dithiothreitol and 2 mM MgSO 4 was incubated in the absence (d) or in the presence of GdmCl 0.7 M (s), 1.2 M (j)or 2.4 M (n). Quenching experiments were conducted as described in Experimental procedures. The lines were obtained by fitting the data to the Stern–Volmer equation. GdmCl, M 01234 Average emission wavelength, nm 344 346 348 350 352 354 356 358 360 Emission intensity at 338 nm 20 40 60 80 100 120 140 160 GdmCl, M 01234 Average emission wavelength, nm 344 346 348 350 352 354 356 358 360 Emission intensity at 335 nm 20 40 60 80 100 120 140 160 GdmCl, M 01 Average emission wavelength, nm 344 346 348 350 352 354 356 358 360 Emission intensity at 350 nm 20 40 60 80 100 120 140 160 Phe16Trp Phe219Trp Phe232Trp GdmCl, M 01234 234 Average emission wavelength, nm 344 346 348 350 352 354 356 358 360 Emission intensit y at 356 nm 20 40 60 80 100 120 140 160 Phe89Trp Fig. 5. Unfolding curves of FBPase mutants monitored by tryptophan fluorescence. The fluorescence emission spectra of the mutants dena- tured by GdmCl were obtained at 15 °C. The intensity-averaged emission wavelength (d) and the fluorescence intensity measured at the k max of emission of each mutant in the native form (s) are plotted as a function of GdmCl concentration. The excitation wavelength was 295 nm and the protein concentrations were: 60 lgÆmL )1 for Phe16Trp and Phe89Trp FBPases and 30 lgÆmL )1 for Phe219Trp and Phe232Trp FBPases. Selective GdmCl disruption of FBPase C1–C2 interface H. C. Ludwig et al. 5342 FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS 8.72 ± 0.22 m )1 indicates an increased accessibility to the solvent. Phase diagram analysis of tryptophan fluorescence data The method of ‘phase diagrams’ has been elaborated by Burstein [24] for the analysis of fluorescence data. It has been shown that this method is extremely sensi- tive for the detection of unfolding ⁄ refolding intermedi- ates of proteins [24–26]. Figure 7 shows the phase diagrams representing the unfolding of Phe16Trp FBPase, Phe89Trp FBPase, Phe219Trp FBPase and Phe232Trp FBPase. Four independent experiments performed with each mutant gave similar results. The phase diagram plotted for the Phe16Trp mutant con- sists of six linear parts, corresponding to 0–0.3, 0.3– 0.5, 0.5–0.8, 0.8–1.4, 1.4–2.3 and 2.3–2.7 m GdmCl. This suggests the existence of six independent transi- tions during unfolding. The intermediate that accumu- lates at 0.8 m GdmCl is the catalytically inactive tetramer, whereas the first intermediate must be a tet- rameric species of enhanced catalytic activity, as can be deduced from Fig. 3. The second intermediate that accumulates at 0.5 m GdmCl is an enzyme having approximately the same activity as the native FBPase. The intermediates formed at 1.4 m GdmCl and at 2.3 m GdmCl should be dimeric and monomeric species, respectively. Interestingly, the phase diagram plotted for Phe89Trp FBPase detects only one interme- diate at 0.7 m GdmCl, corresponding to the inactive tetramer. Concerning these results it must be pointed out that the linearity of the parametric relationship found in a phase diagram does not necessarily indicate that the transition is of a one-step character [27]. This is highlighted by the results obtained for the Fig. 7. Phase diagrams representing the unfolding of FBPase mutants induced by an increase in GdmCl concentrations. The data correspond to two independent sets of experiments performed with each mutant. Denaturant concentration values are indicated in the vicinity of the corresponding symbol. The fluorescence intensities of the native enzymes were taken as unity. The excitation wavelength was 295 nm. H. C. Ludwig et al. Selective GdmCl disruption of FBPase C1–C2 interface FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS 5343 Phe219Trp and the Phe232Trp mutants. The phase diagrams plotted for these enzymes do not detect unfolding intermediates. Reactivation of FBPase upon dilution of GdmCl When samples of the unfolded FBPase mutants in 3.5 m GdmCl were diluted to a concentration of 0.1 m GdmCl the recoveries of enzymatic activity were as follows: Phe16Trp FBPase, 60.9%; Phe89Trp FBPase, 57.2%; Phe219Trp FBPase, 59.8%; and Phe232Trp FBPase, 63.6%. These results indicate that the unfold- ing process is not completely reversible. The reduced reversibility is similar to that observed for nonrecombi- nant FBPase (65.8%), a value that is comparable to previous data [13]. The reduced reversibility can be attributed to an aggregation of intermediates. For this reason the unfolding data shown in Fig. 5 are only qualitatively discussed. No quantitative analysis of the unfolding thermodynamics was attempted. Discussion The guanidine-induced unfolding of pig kidney FBPase has been previously studied in this laboratory using enzyme activity, intrinsic (tyrosine) protein fluores- cence, fluorescence of extrinsic probes and size-exclu- sion chromatography [13]. It has been shown that the unfolding is a multistate process, involving as interme- diates a catalytically inactive tetramer, compact dimers and monomers. As the dimeric and monomeric inter- mediates tend to associate at the relatively high protein concentrations (1 mgÆmL )1 ) used for size-exclusion chromatography, the coexistence of aggregates with intermediates complicates the analysis. The introduc- tion of tryptophan residues in different parts of the protein (present work) provided us with the possibility to further characterize the unfolding process at low protein concentrations, detecting specific transitions. Phenylalanine and tryptophan are both neutral non- polar aromatic amino acids, and usually substitution of Phe for Trp does not cause large changes in the whole protein structure. As expected, the mutants pre- sented almost the same catalytic and regulatory prop- erties as wild-type FBPase and the CD spectra are about the same. Clearly, the structural integrity of the enzyme was not affected. The selective loss of AMP cooperativity without loss of AMP sensitivity observed for Phe16Trp FBPase is an effect that has been described previously for the enzyme as a result of chemical modification [28] or replacement by site-direc- ted mutagenesis [16] of Lys50. The AMP cooperativity is based on a specific signal transmission between FBPase subunits that is lost without loss of the quater- nary structure and without loss of the cooperativity for the cofactor Mg +2 , therefore it is reasonable to assume that the unfolding mechanism for Phe16Trp is the same as for wild-type FBPase. The Phe16Trp mutant has a considerably lower quantum yield than the other tryptophan mutants. The local environment in protein structure can result in either very large or very small quantum yields of Trp residues [15]. Examination of the three-dimensional structure of FBPase [21] reveals that the side chain of Phe16 is at distances of less than 4 A ˚ from the side chains of Gln20, Asn182 and Arg198, residues that have been described as tryptophan quenchers [29,30]. The quenching is partially relieved upon the first steps of unfolding, probably because conformational changes at the tetramer level decrease the efficiency of the quenching. Interestingly, the biphasic increase of fluorescence intensity correlates with the initial increase and the subsequent loss of enzymatic activity. The fluorescence equilibrium unfolding curves of the four single tryptophan mutants are very different (Fig. 5). In general, changes in intrinsic tryptophan flu- orescence intensity upon protein unfolding are com- pletely unpredictable [31]. The only change that can be predicted with confidence is that the spectrum will shift to red upon greater exposure to solvent. Accordingly, we have interpreted the pronounced blue shift observed for the emission of Phe89Trp FBPase between 0.4 and 0.7 m GdmCl as the occurrence of a conformational change that causes a displacement of Trp89 into an apolar environment. This kind of dis- placement is congruent with the reduced degree of exposition detected by acrylamide quenching experi- ments. Concomitantly, hydrophobic patches appear on the surface of the protein, as indicated by the increase in ANS-binding fluorescence and the catalytic activity disappears. It is important to note that the four mutants remain in the tetrameric state and do not aggregate at low concentrations of GdmCl (lower than 0.9 m) as revealed by the size-exclusion experiments. Furthermore, the linearity of the Stern–Volmer plots obtained for Phe89Trp FBPase by acrylamide quench- ing at 0.7 and 1.2 m GdmCl also support the idea that this mutant does not aggregate, as an aggregation should cause heterogeneity and a downward curvature of the plots. We have interpreted the red shift near 2 m GdmCl of the emission of Phe16Trp and Phe89Trp FBPases as a disruption of the C1–C4 interface. A similar situa- tion has been described for the Trp99Phe single trypto- phan mutant of the dimeric Trp aporepressor [23], where the emission of Trp19, a residue that is buried Selective GdmCl disruption of FBPase C1–C2 interface H. C. Ludwig et al. 5344 FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS at the dimer interface, is highly red shifted upon dis- ruption of the interface. On the other hand, the disso- ciation of the catalytically inactive tetrameric FBPase (wild-type and mutants) into dimers begins at a GdmCl concentration around 1 m. Clearly this process does not affect the average emission wavelength of Trp89, which remains constant at between 0.7 and 1.8 m GdmCl, and affects only slightly the average emission wavelength of Trp16. It can be concluded that the interface which is disrupted first during unfolding of FBPase is the C1–C2 interface. The results obtained with Phe219Trp FBPase are in line with this notion. The conclusion that the C1–C2 interface is disrupted before the C1–C4 interface might at first appear to be at odds with the following facts: (a) In FBPase the polypeptide chains of C1 and C2 (or C3 and C4) make up an essential unit for catalytic activity, as they mutu- ally exchange their Arg243 residues at the active sites. Furthermore, both chains are extensively associated through both hydrophilic and hydrophobic interac- tions [32]; (b) In the absence of AMP, the dimers C1C2 and C3C4 associate primarily through interac- tions between the side-chains of residues in two a-heli- ces (H1 and H3) of the AMP domains. When AMP binds to the allosteric site it elicits a 17° rotation between the dimers C1C2 and C3C4, whereas the C1–C2 interface is essentially locked at its existing con- formation in the R state [33]. Nevertheless, the dissocia- tion of the tetramers is preceded by the loss of catalytic activity, and the structural changes that occur at the active site region probably destroy some interac- tions across the C1–C2 interface. Moreover, our results indicate that the transition by which the catalytic activ- ity is lost not only involves conformational changes in the Fru(1,6)P 2 domain, but also at the AMP domain, as it is detected by each of the four tryptophan resi- dues of the mutants. Therefore it is possible that this global change causes the formation of new interactions which stabilize the C1–C4 interface. Interestingly, Nel- son et al. [34] have described a spontaneous subunit exchange between distinct homotetramers of FBPase to form hybrid tetramers at 4 °C that obviously requires the disruption of both interfaces. The phase diagram plotted for Phe16Trp FBPase suggests the existence of five intermediates. Although the difference in the parametric relationship between 0.3 m and 0.5 m GdmCl is moderate, it can not be ignored, as the same change was observed consistently in four independent experiments. Then, according to the phase diagram the first intermediate on unfolding of Phe16Trp FBPase occurs at 0.3 m GdmCl. The exis- tence of this intermediate is also supported by the enhancement of catalytic activity observed at low GdmCl concentrations for wild-type as well as for the mutant enzymes. For the wild-type FBPase the activity enhancement has been interpreted as a local effect, caused by an increased conformational flexibility at the active site [13]. Nevertheless, our present results indi- cate that the effect is not only local, as Trp16 is 30 A ˚ away from the active site. Unfolding intermediates at a low GdmCl concentration (around 0.1 m) have already been described for carbonic anhydrase [35] and for rabbit muscle creatine kinase [25]. The phase diagram for F16W FBPase also reveals the existence of a sec- ond active intermediate at 0.5 m GdmCl. The existence of this intermediate explains the biphasic character of the inactivation of the enzymes. Furthermore, this dia- gram provides the first evidence for the accumulation of an intermediate at 1.4 m GdmCl that corresponds to the dimer. According to the phase diagram, the Phe89Trp FBPase appears to unfold in a three-state manner (Fig. 7), in which the intermediate is the inactive tetra- mer. However, the existence of linearity of the para- metric relationship in a phase diagram does not necessarily indicate that the transition is of a one-step character [27]. Although the transition between confor- mational states proceeds via an intermediate, the para- metric relationship can be practically linear in the following cases: (a) if the values of the measured charac- teristics of the intermediate state are close to those of the initial or final states; and (b) if the values of the measured characteristics of the intermediate state are somewhat between those of the initial and final states. This is highlighted by the phase diagrams obtained for the Phe219Trp and Phe232Trp FBPases. The multiple probes of the unfolding of these mutants (activity, ANS binding, tryptophan fluorescence and size-exclusion chromatography) indicate the not-one-step character of the process. Nevertheless, the phase diagrams of these mutants clearly do not detect any intermediate. In conclusion, our data are consistent with the fol- lowing scheme of GdmCl-induced unfolding of FBPase: where T N ,T A1 ,T A2 and T I are the native enzyme, the tetrameric intermediate of increased catalytic activity, the second active tetrameric intermediate and the inac- tive tetrameric intermediate, respectively; D are the dimers (C1C4 and C2C3); M and U are a monomeric intermediate and the unfolded monomer, respectively; and A corresponds to aggregates. The existence of M H. C. Ludwig et al. Selective GdmCl disruption of FBPase C1–C2 interface FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS 5345 is supported by the phase diagram of the Phe16Trp mutant and by our previous report [13], in which the fluorescence anisotropy of an N-(acetylaminoethyl)-5- naphthylamine-1-sulfonic acid-labeled FBPase and the average emission wavelength of a 2-(4¢-maleimidylanili- no)naphthalene-6-sulfonic acid-labeled FBPase were measured. On the other hand, the inclusion of aggre- gates in the scheme is based on previous results [13]. These aggregates are formed at high protein concentra- tions. A similar behavior has been described for rabbit muscle creatine kinase [25]. Size-exclusion chromato- graphy studies of this enzyme show the formation of large aggregates at a high (2 mgÆmL )1 ) but not at a low (0.1 mgÆmL )1 ) protein concentration. Interestingly, although the unfolding behavior of FBPase has been studied [13,14,36], the formation of the active tetrameric intermediates T A1 and T A2 and the notion that the loss of quaternary structure begins by disruption of the C1–C2 interface are described here for the first time. Experimental procedures Materials Fructose-1,6-bisphosphatase was purified from pig kidney (nonrecombinant enzyme) as described previously [37]. ANS was obtained from Molecular Probes (Eugene, OR). Auxiliary enzymes were purchased from Sigma (St. Louis, MO) and GdmCl from Merck (Darmstadt, Germany). All other reagents were of analytical grade. Preparation, expression and purification of FBPase mutants Replacement of phenylalanine residues with tryptophan was carried out using the Altered sites II in Vitro Muta- genesis System kit, following the manufacturer’s (Pro- mega, Madison, WI) instructions, as previously described [16]. The following mutagenic oligonucleotides were used (the bases changed appear in bold): 5¢-GCTCACCCTAA CCGCTGGGTCATGGAGGAGGGCAG-3¢ (Phe16Trp); 5¢-GTTAAAGTCATCTTGGGCCACCTGCGTTCTC-3¢ (Phe89Trp); 5¢-GGCTATGCCAGGGAGTGGGACCCTG CCATCACTGAG-3¢ (Phe219Trp); 5¢-CAGAGGAAGAA GTGGCCCCCAGA-3¢ (Phe232Trp). The mutations were confirmed by unique restriction enzyme digestion and by sequence analysis of the mutagenic FBPase plasmids as described earlier [16]. Protein expres- sion and purification were performed as described [16]. For expression, the fragments encoding the wild-type or muta- genic FBPases were excised from the corresponding plasmid and cloned into the vector pET15b (Novagen, San Diego, CA). The purified His-FBPases were subjected to proteolysis with thrombin in order to remove the His-tag. The protein concentration of the samples was measured using the Bio-Rad Protein assay kit with FBPase isolated from pig kidney as standard, or determined by absorbance at 280 nm using a e 1mg⁄ mL value of 0.755 [37] for the enzyme isolated from pig kidney and 0.904 for the single- tryptophan mutants (determined by comparison with the enzyme isolated from pig kidney). Spectrophotometric assay of fructose-1,6- bisphosphatase activity The enzyme activity was determined spectrophotometrically at 30 °C by following the rate of NADH formation at 340 nm in the presence of an excess of both glucose-6-phos- phate dehydrogenase and phosphoglucose isomerase [16,38]. Unless stated otherwise, the reaction system (0.5 mL) con- tained 50 mm Tris ⁄ HCl buffer, pH 7.5, 0.1 mm EDTA, 5mm MgSO 4 ,30lm Fru(1,6)P 2 , 0.3 mm NAD + and 1.2 enzyme units of each auxiliary enzyme. Digital absor- bance values were collected using a Hewlett Packard 8453 spectrophotometer (Hewlett Packard, Palo Alto, CA) and the linear data, from beyond the coupling lag period, were fit to a straight line on a coupled computer using the UV-visible CHEM STATION program. One unit of activity is defined as the amount of enzyme that catalyzes the formation of 1 lmol of fructose-6-phosphate per min at 30 °C under the conditions described [16]. Because the nonrecombinant and mutant FBPases exhibit partial substrate inhibition at high substrate concentrations, substrate saturation curves for all enzymes were fit by nonlinear regression to a modified form of the Michaelis–Menten equation which incorporated a term for substrate inhibition [16,39]. The K a value and the Hill coefficient (h) for Mg +2 were determined by saturation curves fitting the data to the Hill equation. AMP and Fru(2,6)P 2 inhibition curves were fit to the Taketa and Pogell equation [40]. To prevent FBPase reactivation during the enzyme assay used for the examination of protein unfolding, trypsin (20 lg proteinÆ mL )1 ) was added to the assay mixture [13]. Equilibrium unfolding Equilibrium unfolding of FBPases was performed in 0.1 m Hepes-NaOH buffer, pH 7.5, containing 0.1 mm EDTA, 5mm dithiothreitol, 2 mm MgSO 4 and GdmCl at the desired concentration. The solutions were incubated for 4 h at 15 °C before analysis. The concentration of the GdmCl stock solu- tion was determined by refractometry according to Pace [41]. Reactivation studies The enzymes (900 lgÆmL )1 ) were incubated in 3.5 m GdmCl in 0.1 m Hepes ⁄ NaOH buffer (pH 7.5) containing Selective GdmCl disruption of FBPase C1–C2 interface H. C. Ludwig et al. 5346 FEBS Journal 274 (2007) 5337–5349 ª 2007 The Authors Journal compilation ª 2007 FEBS [...]... different concentrations The intensity of tryptophan fluorescence emission upon excitation at 295 nm was detected between 310 and 400 nm as a function of acrylamide concentration The emission of The ‘phase diagram’ method is a sensitive approach for the detection of unfolding ⁄ refolding intermediates of proteins [25,27,35] The essence of this method is to build up the diagram of I(k1) versus I(k2), where I(k1)... C1–C2 interface residue lysine 50 of pig kidney fructose1,6-bisphosphatase has a crucial role in the cooperative signal transmission of the AMP inhibition, Eur J Biochem 267, 2242–2251 Kelley-Loughnane N & Kantrowitz ER (2001) AMP inhibition of pig kidney fructose-1,6-bisphosphatase, Biochim Biophys Acta 1548, 66–71 Nelson SW, Honzatko RB & Fromm HJ (2002) Hybrid tetramers of porcine liver fructose-1,6-bisphosphatase... Slebe JC (2003) Nativelike intermediate on the unfolding pathway of pig kidney fructose-1,6-bisphosphatase Biochemistry 42, 6956–6964 Yuan C, Xie ZQ, Zhang FW & Xu GJ (2001) Association and activation of fructose 1,6-bisphosphase during unfolding and refolding: spectroscopic and enzymatic studies, J Protein Chem 20, 39–47 Lakowicz JR (1999) Principles of Fluorescence Spectroscopy, 2nd edn Kluver Academic... the same concentration of GdmCl in the same buffer Circular dichroism CD spectra of wild-type and mutant FBPases in Tris ⁄ HCl buffer, pH 7.5, containing 0.1 mm EDTA were recorded at room temperature on a Jasco 600 spectrometer (Jasco, Easton, MD) using the 720 software and a cuvette of 0.1 cm path length Five scans of each spectrum were collected from 250 to 200 nm in increments of 1 nm and averaged... mechanism of catalysis and allosteric inhibition revealed in product complexes Biochemistry 39, 8565–8574 Ptitsyn OB, Pain RH, Semisotnov GV, Zerovnik E & Razgulyaev OI (1990) Evidence for a molten globule state as a general intermediate in protein folding FEBS Lett 262, 20–24 Royer CA, Mann CJ & Matthews CR (1993) Resolution of the fluorescence equilibrium unfolding profile of trp aporepressor using single tryptophan. .. blank-corrected and smoothed by using the software package provided with the instrument ‘Phase diagram’ method Intrinsic fluorescence and ANS-binding experiments Fluorescence spectra were taken at 15 °C on a PerkinElmer LS-50 spectrofluorometer (Perkin-Elmer, Norwalk, CT) using excitation and emission slits of 6 nm The fluorescence was corrected by subtraction of the solvent signal For intrinsic tryptophan fluorescence... folding -unfolding Biochemistry 41, 13127–13132 Ludwig HC, Herrera R, Reyes AM, Hubert E & Slebe JC (1999) Suppression of kinetic AMP cooperativity of fructose-1,6-bisphosphatase by carbamoylation of lysine 50, J Protein Chem 18, 533–545 Chen Y & Barkley MD (1998) Toward understanding tryptophan fluorescence in proteins Biochemistry 37, 9976–9982 Clark PL, Liu ZP, Zhang J & Gierasch LM (1996) Intrinsic tryptophans... concentration of 0.1 m GdmCl and a protein concentration of 25 lgÆmL)1 The recovery of catalytic activity was measured after further incubation at 15 °C for enough time to ascertain that it attained a final stable value Selective GdmCl disruption of FBPase C1–C2 interface blank solutions was subtracted The fluorescence at kmax was corrected for dilution and for the inner filter effect due to the absorbance of acrylamide... relationships of substrate cycle enzymes Annu Rev Nutr 11, 465–515 2 Benkovic SJ & deMaine MM (1982) Mechanism of action of fructose 1,6-bisphosphatase, Adv Enzymol Relat Areas Mol Biol 53, 45–82 3 Pilkis SJ, el-Maghrabi MR & Claus TH (1988) Hormonal regulation of hepatic gluconeogenesis and glycolysis, Annu Rev Biochem 57, 755–783 4 Liang JY, Huang S, Zhang Y, Ke H & Lipscomb WN (1992) Crystal structure of the... Edelstein I, Reardon I & Heinrikson RL (1982) Complete amino acid sequence of pig kidney fructose-1,6-bisphosphatase, Proc Natl Acad Sci USA 79, 7161–7165 Xue Y, Huang S, Liang JY, Zhang Y & Lipscomb WN (1994) Crystal structure of fructose-1,6-bisphosphatase complexed with fructose 2,6-bisphosphate, AMP, and Zn2+ at 2.0-A resolution: aspects of synergism between inhibitors, Proc Natl Acad Sci USA 91, 12482–12486 . Unraveling multistate unfolding of pig kidney fructose-1,6- bisphosphatase using single tryptophan mutants Heide C. Ludwig, Fabian. recombi- nant wild-type FBPase. Table 2. Fluorescence properties of the single tryptophan mutants of pig kidney FBPase. The Stern–Volmer quenching constants for acrylamide