Probing the Role of Crystallographically Defined/Predicted Hinge-bending Regions in the Substrate-induced Global Conformational Transition and Catalytic Activation of Human Phenylalanine Hydroxylase by Single-site Mutagenesis*

Phenylalanine hydroxylase (PAH) is generally considered to undergo a large and reversible conformational transition upon l-Phe binding, which is closely linked to the substrate-induced catalytic activation of this hysteretic enzyme. Recently, several crystallographically solvent-exposed hinge-bending regions including residues 31-34, 111-117, 218-226, and 425-429 have been defined/predicted to be involved in the intra-protomer propagation of the substrate-triggered molecular motions generated at the active site. On this basis, single-site mutagenesis of key residues in these regions of the human PAH tetramer was performed in the present study, and their functional impact was measured by steady-state kinetics and the global conformational transition as assessed by surface plasmon resonance and intrinsic tryptophan fluorescence spectroscopy. A strong correlation (r2 = 0.93-0.96) was observed between the l-Phe-induced global conformational transition and Vmax values for wild-type human PAH and the mutant forms K113P, N223D, N426D, and N32D, in contrast to the substitution T427P, which resulted in a tetrameric form with no kinetic cooperativity. Furthermore, the flexible intra-domain linker region (residues 31-34) seems to be involved in a more local conformational change, and the biochemical/biophysical properties of the G33A/G33V mutant forms support a key function of this residue in the positioning of the autoregulatory sequence (residues 1-30) and thus in the regulation of the solvent and substrate access to the active site. The mutant forms revealed a variably reduced global conformational stability compared with wild-type human PAH, as measured by thermal denaturation and limited proteolysis.

that catalyzes the hydroxylation of L-Phe to L-Tyr in the presence of the pterin cofactor (6R)-L-erythro-5,6,7,8-tetrahydrobiopterin (BH 4 ) and dioxygen, the rate-limiting step in L-Phe catabolism. Mutations in human PAH (hPAH), leading to altered kinetic properties and/or reduced in vivo stability of the enzyme, are associated with the autosomal recessive metabolic disorder phenylketonuria/hyperphenylalaninemia, and so far 439 putative disease-causing alleles have been identified in the PAH gene coding for a 452-residue polypeptide (1) (www. pahdb.mcgill.ca).
Mammalian PAH is a homotetramer (in equilibrium with a dimeric form) with the protomer consisting of three functional domains, i.e. a regulatory, a catalytic (core), and an oligomerization domain (for review, see Ref. 2). The tetramer is a dimer of two conformationally different dimers, denoted AC and BD, related by a 2-fold noncrystallographic symmetry axis (3). Based on a number of independent biochemical and biophysical criteria, it is generally accepted that the enzyme exists in at least two different activity and conformational states (4 -8), and that it undergoes a large scale but relatively slow (secondsto-minutes) global conformational transition (isomerization) upon binding of its substrate L-Phe, which is characteristic of a hysteretic enzyme (5)(6)(7)(8). Thus, the catalytic activation of tetrameric hPAH by its substrate can be represented by the equilibria in Equation 1 (6,8), Ϫk 2 PAH*⅐Phe (Eq. 1) where PAH*⅐Phe represents the binary complex of the relaxed high activity state of the primed enzyme (PAH⅐Phe), and k 2 represents the slow (hysteretic) conformational transition. Thus, the binding of the substrate (7), as well as the pterin cofactor (9,10), occurs by an induced-fit mechanism, with a perturbation of the structure of both the enzyme protomer and the ligand. Although the specific regions involved in the reversible conformational transition and catalytic activation of the tetramer, related to substrate binding at the active site, are only partly characterized (7,11), the three-dimensional crystal structures of different truncated forms of hPAH (3,12) and rat PAH (rPAH) (13) have given valuable insights into the molecular mechanism of substrate activation. Thus, hinge-bending regions in the catalytic domain have been identified to be involved in this isomerization on the basis of two conformational states (conformers) of the catalytic core structure (7,11), and three additional inter-and intra-domain hinge-bending regions, schematically shown in Fig. 1, have been predicted from the structure of complementary truncated forms of hPAH (3,14) and rPAH (13), for which only one x-ray conformer is yet available. There is, however, no crystal structure for any fulllength tetrameric/dimeric PAH, which leaves unsolved the overall mechanism of the intramolecular propagation of the conformational changes initiated by the cooperative binding of L-Phe at the active site (11). In order to gain further insight into the role of the crystallographically defined/predicted hinge-bending regions in the substrate-induced conformational transition and the related catalytic activation, single amino acid substitutions were made at key residues in full-length hPAH as follows: (i) in the flexible intra-domain linker region (residues 31-34) of the regulatory domain structure that follows the N-terminal autoregulatory sequence (13); (ii) in the regulatory-to-catalytic inter-domain loop structure (residues 111-117) (13); (iii) in an intra-domain hinge-loop region (residues 218 -226) of the catalytic core domain structure (11); and (iv) in the dimerization-to-tetramerization loop structure (residues 425-429) (3) preceding the C-terminal amphipathic ␣-helix involved in tetramerization by three-dimensional domain swapping ( Figs. 1 and 2). A key objective of the present paper was to study the effects of the structural perturbations of hinge flexibility, resulting from specific mutations, on the catalytic and global conformational properties of the tetrameric form of the enzyme. Intrinsic tryptophan fluorescence spectroscopy and surface plasmon resonance (SPR) were used to follow in real time the slow (hysteretic) global conformational transition (k 2 in Equation 1) that is triggered by the reversible binding of L-Phe to hPAH (2,6,8).

EXPERIMENTAL PROCEDURES
Materials-The primers for site-directed mutagenesis were obtained from MWG Biotech AG (Ebersberg, Germany). The QuikChange TM site-directed mutagenesis kit was from Stratagene. The Big Dye TM terminator ready reaction mix used to prepare DNA for sequencing was delivered by Applied Biosystems, and the DNA sequencing was carried out on an ABI Prism TM 377 DNA sequencer (Applied Biosystems). The restriction proteases factor Xa and enterokinase were obtained from Protein Engineering Technology ApS (Aarhus, Denmark) and Invitrogen, respectively. The reagents and the sensor chip CM5 used in the SPR analyses were purchased from BiaCore AB (Uppsala, Sweden). (6R)-L-erythro-5,6,7,8-Tetrahydrobiopterin was from the laboratory of Dr. B. Schircks (Jona, Switzerland). Trypsin and trypsin inhibitor were obtained from Sigma.
Site-specific Mutagenesis-The specific mutations addressed in the present study were selected based on predictive protein mutation stability changes (babylone.ulb.ac.be/popmusic/). The mutation K113P was introduced into the cDNA of the wt-hPAH by using the QuikChange TM site-directed mutagenesis kit. The pMAL-hPAH plasmid containing a cleavage site for factor Xa was used as the template, and the following primers were used for mutagenesis (mutated nucleotides are shown in boldface type): GCTTTCACGAGATCCGAAGAAAGACACAGTGCC-CTGG (forward) and CCAGGGCACTGTGTCTTTCTTCGGATCTCGT-GAAAGC (inverse). The N-terminal mutations (N32D and G33A/G33V) were introduced into the cDNA of the pMal vector containing the enterokinase cleavage site (D 4 K), and the remaining mutations (N223D, N426D, and T427P) were introduced into the pMal-hPAH expression system containing the factor Xa cleavage site (IEGR) (14,15). DNA sequencing was used to verify the introduction of the desired mutations as described previously (16).
Expression and Purification of hPAH-The wild-type and mutant fusion proteins were expressed in Escherichia coli with maltose-binding protein as the fusion partner and harvested after 2 h of induction with isopropylthio-␤-D-galactoside, and the tetrameric fusion proteins were purified to homogeneity as described previously (17). The fusion proteins were cleaved by the restriction protease factor Xa or enterokinase, followed by size-exclusion chromatography at 4°C (17) using a HiLoad Superdex 200 HR column (1.6 ϫ 60 cm), prepacked from Amersham Biosciences. The isolated hPAH tetramers were collected, and the concentration was measured spectrophotometrically by using the absorption coefficient A 280 (1 mg⅐ml Ϫ1 ⅐cm Ϫ1 ) ϭ 1 (17).
Assay of Enzymatic Activity-The PAH activity was measured at 25°C in a standard reaction mixture containing 0.4 M subunit (unless otherwise stated) of tetrameric wild-type/mutant forms of hPAH, 1 mM L-Phe, 0.5 mg/ml bovine serum albumin, 1 mM dithiothreitol, 0.04 mg/ml catalase, 100 M ferrous ammonium sulfate in 100 mM Hepes buffer, pH 7. After preincubation (5 min) with L-Phe the reaction was started by adding 75 M (unless otherwise stated) BH 4 to the reaction mixture. The amount of L-Tyr formed after 1 min was measured by high pressure liquid chromatography and fluorimetric detection, and the steady-state kinetic parameters were calculated by nonlinear regression analysis where both the cooperativity and substrate inhibition were taken into consideration, using the Jandel Scientific SigmaPlot® technical graphing software (18).
SPR Analyses-Real time interaction analyses between immobilized enzyme (wt-hPAH or its different mutant forms) and L-Phe was measured by SPR analyses using the BiaCore X instrument (BiaCore AB, Uppsala, Sweden) as described previously (6,8). Full-length hPAH diluted in 10 mM sodium acetate buffer, pH 5.4, to a final concentration of 0.23 mg/ml was covalently immobilized to the carboxymethylated dextran matrix of the sensor chip (CM5) with the amine coupling procedure (6,8). The amount of immobilized protein was estimated from the increased ⌬RU values by assuming that 1000 RU corresponds to ϳ1 ng of immobilized protein/mm 2 (19). The wt-hPAH and its mutant forms were immobilized at high ⌬RU values in the range of 26,000 -32,000 RU, corresponding to 26 -32 ng of bound protein/mm 2 . The double-truncated form ⌬N102/⌬C24-hPAH (0.37 mg/ml in sodium acetate buffer, pH 5.4) was immobilized to the reference cell by the same procedure in order to correct for the changes in the bulk refractive index together with the minor ⌬RU response associated with binding at the active site (without any measurable conformational change) of the low molecular mass substrate L-Phe (165 Da) (8). The system was equilibrated as described (6) using HBS-EP buffer (10 mM Hepes, 150 mM NaCl, 3.4 mM EDTA, and 0.005% (v/v) of the surfactant P-20, pH 7.4) at a flow rate of 5 l/min at 25°C. Various concentrations of L-Phe in HBS buffer were injected over immobilized hPAH, and equilibrium SPR measurements were taken once the ⌬RU reached a stable plateau value after ϳ3 min (6,8). The time-dependent positive SPR response, expressed as ⌬RU/(ng protein/mm 2 ), has been shown previously (6,8) to measure in real time the global conformational transition triggered by L-Phe binding. The L-Phe concentration at half-maximum response ([S] 0.5 ) was calculated from the equilibrium binding isotherm using the Jandel Scientific SigmaPlot® technical graphing software for nonlinear regression analysis.
Tryptophan Fluorescence Measurements-Intrinsic tryptophan fluorescence measurements of the tetrameric forms of wt-hPAH and its different mutant forms were performed on a PerkinElmer Life Sciences LS-50B luminescence spectrometer at 25°C in a buffer consisting of 200 mM NaCl and 20 mM Hepes, pH 7, and at a protein concentration of 0.02 mg/ml. Increasing concentrations of substrate (L-Phe or D-Phe) were added to the cuvette every 3rd min, and equilibrium fluorescence measurements were taken once the ⌬F reached a stable plateau after ϳ3 min. The excitation wavelength was 295 nm, and the L-Phe-specific increase in fluorescence at 358 nm (increased solvent exposure of Trp 120 (20)) was obtained by subtracting the response to added D-Phe (which binds with very low affinity) from added L-Phe in order to correct for a minor nonspecific signal related to each addition and the base-line drift. The binding parameters were calculated by nonlinear regression analyses as described above for the SPR measurements.
Limited Proteolysis with Trypsin-The rate of proteolysis was measured at 25°C in 20 mM Hepes and 200 mM NaCl, pH 7, and a hPAH to trypsin ratio of 200:1 (by mass) in the absence or presence of 1 mM L-Phe. The final protein concentration was 0.3 mg/ml. At timed intervals, 10-l aliquots were mixed with soybean trypsin inhibitor and subjected to SDS-PAGE analyses. Quantification of the protein bands was obtained by using the Quantity One software (Bio-Rad).
Circular Dichroism-CD analyses were performed at 25°C on a Jasco J-810 spectropolarimeter equipped with a Peltier temperature control unit. 5-7 M subunits of wild-type and mutant tetrameric hPAH were diluted in 20 mM potassium phosphate buffer containing 50 mM KF (pH 7) and a stoichiometric amount of ferrous ammonium sulfate. Wavelength scans were taken between 200 and 260 nm by using 0.1-cm path length quartz cells, and the spectra (run in duplicates or triplicates) were the average of four scans; buffer scans were subtracted from the protein spectra. The global conformational stability, as measured by thermal denaturation (range 25-70°C), was monitored by following the changes in the ellipticity at 222 nm at a constant heating rate of 1°C/min. The apparent melting temperature (T m ) values were determined from the first/second derivatives of the denaturation curves after noise reduction by using the standard analysis program provided with the instrument. The first derivative profile was further resolved (deconvoluted) into individual overlapping T m transitions, using a Gaussian distribution function in the PeakFit software program (SPSS Inc., Chicago).

Rationale for Choosing Specific Residues for Mutagenesis-
The schematic presentation of the domain structure of the wt-hPAH protomer shown in Fig. 1 also highlights the defined/ predicted inter-and intra-domain solvent-exposed hinge-loop regions where single-site mutations were introduced. First, the hinge-bending residue Asn 223 , at the center of the crystallographically defined intra-domain hinge-loop region (residues 218 -226) in the catalytic core domain structure (residues 118 -410) (11), was mutated to an aspartate. This substitution is predicted to result in a loss of a stabilizing H-bond to Gly 218 ( Fig. 2A) and thus a local destabilization. Second, the 111 RD-KKKDT 117 loop structure, connecting the regulatory domain with the catalytic domain, in rPAH has been proposed to represent another center of hinge-bending motion as a result of substrate binding (13). On the basis of the prediction of protein mutation stability changes, residue 113 was selected for mutagenesis. Lys 113 was substituted with a proline, as found in the structurally and functionally closely related enzyme tyrosine hydroxylase (2), which is predicted to result in a loss of the H-bonds to Asp 27 (in the autoregulatory sequence) and Asp 315 (in the catalytic core domain) of the same protomer ( Fig. 2A) and thus made this linker region more flexible despite a more rigid backbone conformation. Moreover, in the three-dimensional structure of the tetrameric N-terminally truncated hPAH, a short loop (residues 425-429) was observed in the C-terminal oligomerization domain (residues 411-452) connecting the dimerization and the tetramerization motifs (3). Mutation of residue 427 was predicted to result in a marked destabilization, and a T427P substitution (as in human tyrosine hydroxylase) had revealed previously that the backbone conformation of this loop was indeed involved in the control of the L-Phe-promoted conformational transition in addition to its effects on the tetramer 7 dimer equilibrium and the cooperative binding of substrate (14). For comparison, the functional effects of substituting the neighbor residue Asn 426 with an Asp (as in human tryptophan hydroxylase), predicted to be a harmless substitution, were also studied. Finally, the buried Gly 33 ( Fig. 2A), which represents the center of a proposed flexible intra-domain linker region (residues 31-34) (13), has been proposed also to have a key function in the flexible orientation of the autoregulatory sequence (residues 1-30), and thus it contributes to the regulation of the access of solvent and presumably substrates to the active site (13). Several mutations at Gly 33 are expected to result in a destabilization of hPAH, and here we have selected the two mutations G33A/G33V predicted to have high ⌬⌬G values. For comparison, the functional effects of the N32D mutation were also studied.
Expression and Isolation of Wild-type hPAH and Its Mutant Forms-Fusion proteins of wt-hPAH and its mutant forms N223D, K113P, N426D, T427P, N32D, and G33A/G33V were all expressed in high yields in the TB1 strain of E. coli. The cells were harvested after 2 h of induction with isopropylthio-␤-D-galactoside at 28°C, i.e. conditions that almost completely prevent the deamidation of three labile Asn residues (at positions 32, 133, and 376), which occurs on longer induction times and higher temperatures (15,18,21) and results in a partly activated wild-type enzyme (characterized by a higher catalytic efficiency, a higher Hill coefficient, and an increased affinity toward L-Phe). Cleavage of the fusion proteins and subsequent isolation of their oligomeric forms by size-exclusion chromatography resulted in a variable recovery of soluble tetrameric forms (data not shown), but for all mutants recovery was in amounts sufficient for the present biochemical and biophysical studies.
Limited Proteolysis and Thermal Stability-Limited tryptic proteolysis is a sensitive probe for conformational changes in hPAH induced by non-enzymatic deamidation of labile asparagine residues (18), phosphorylation of Ser 16 (22), ligand binding (23), as well as in several disease (24) and non-disease (14, 15) related mutant forms. In the present study 78% of the wild-type enzyme was found to be uncleaved after 30 min at the selected experimental conditions in the absence of L-Phe (Table  I). Some of the mutant enzymes revealed an increased susceptibility to limited proteolysis, with a full-length recovery of 67% for N223D and 74% for G33A after 30 min, whereas only 40% of T427P was found to be undigested (Table I). A similar extent of cleavage as wt-hPAH was found for the N426D mutant, whereas the N32D mutant was slightly more resistant to proteolytic degradation, with a full-length recovery of 84% after 30 min (Table I). The hinge-bending region (residues 111-117), representing the linker region between the regulatory and catalytic domains, contains three Lys residues in sequence and is one of the sensitive cleavage sites on limited proteolysis with trypsin. Thus, as expected, the K113P mutant form was more resistant to limited proteolysis than the wild-type, with 88% uncleaved protein after 30 min of proteolysis (Table I), and the disparity was even more pronounced in the presence of substrate (1 mM L-Phe) with 68% full-length enzyme observed for the K113P mutant compared with 29% full-length wt-hPAH (data not shown).
All the mutant forms revealed an ordered secondary structure similar to the wt-hPAH, as accessed by far-UV CD spectroscopy (range 200 -260 nm), and no significant change in the CD spectra (notably in the apparent ␣-helix content) was observed in the presence of L-Phe (data not shown). In the thermal denaturation studies two partially overlapping transitions were observed for the wild-type tetramer, with apparent T m values at ϳ46 and ϳ54°C (Fig. 3), which in the presence of L-Phe shifted to ϳ54 and ϳ59°C, respectively. In general, the tetrameric mutant proteins revealed similar denaturation profiles with two main transitions and similar high temperature T m values as for the wt-hPAH ( Fig. 3 and Table I). The T427P mutant protein was an exception, because its reduced thermal stability revealed only an apparent single transition at ϳ41°C observed in the first derivative profile, but on further deconvolution a minor second transition at ϳ50°C was also indicated ( Fig. 3 and Table I). In the presence of L-Phe the main thermal transition shifted to ϳ45°C. With respect to the first transition, a decrease in the T m values was found for the N223D, K113P, and G33A tetrameric mutant forms as isolated, when compared with the wt-hPAH (Table I).
Catalytic Properties of the Wild-type and Mutant Forms-In the steady-state kinetic analyses of wt-hPAH, the tetrameric form revealed a positive cooperativity of L-Phe binding (n H ϭ 1.5 Ϯ 0.1) with a [S] 0.5 value of 238 Ϯ 18 M and a 4.9-fold activation by preincubation with the substrate at 25°C (Table  II and Fig. 4). The selected mutations were expected to alter one or more kinetic parameters, which was indeed observed (Table II and Fig. 4).
The tetramer of the four mutant forms N223D, N426D, T427P, and N32D revealed a low basal catalytic activity, similar to wt-hPAH (Fig. 4). However, whereas the N426D and N32D mutants gave a close to normal substrate activation (5.1and 4.4-fold, respectively), the N223D and T427P mutant tetramers were activated only 2.7-and 1.2-fold, respectively, on preincubation with L-Phe ( Fig. 4 and Table II). Except for the N32D where a higher catalytic efficiency was observed, the other three mutants revealed a reduced catalytic efficiency either as a result of a reduced V max effect (N223D) or an increase in the [S] 0.5 value for L-Phe (N426D and T427P) ( Table II).
The mutant forms G33A/G33V and K113P were all characterized by an increased basal activity and a reduced enhancement, i.e. 1.4-, 1.5-, and 2.2-fold, respectively, on preincubation with L-Phe (Table II and Fig. 4). Moreover, the K113P mutant tetramer was found to have a slightly increased affinity toward L-Phe ([S] 0.5 ϭ 218 Ϯ 4 M) and a 2-fold increase in the catalytic efficiency (Table II), mainly because of a large increase in V max . The two other mutant forms, G33A/G33V, revealed slightly different kinetic properties (Table II). The largest change was observed for the G33V mutant with a decrease in the [S] 0.5 value for L-Phe together with an increase in V max , leading to a 2-fold increase in the catalytic efficiency as compared with the wild type (Table II).
A positive cooperativity of L-Phe binding is a characteristic feature of the wt-hPAH tetramer (n H ϭ 1.5 Ϯ 0.1). Although a slightly lower Hill coefficient was observed for the mutant    (Table  II). By contrast, the T427P mutant tetramer revealed a noncooperative binding of L-Phe (i.e. n H ϭ 1.0 Ϯ 0.1) ( Table II). Most of the mutants revealed a slightly decreased affinity for the pterin cofactor BH 4 as compared with the wild-type tetramer, with the most dramatic decrease observed for the K113P mutant form (Table II).
Surface Plasmon Resonance and Intrinsic Tryptophan Fluorescence Measurements-SPR spectroscopy has been used successfully to monitor in real time the reversible and large scale global conformational transition that occurred in immobilized wt-hPAH upon L-Phe binding at the active site (6,8) and was positively correlated to the change in intrinsic tryptophan fluorescence in solution (8). The time-dependent increase (within 3 min) in the SPR response to L-Phe injection obtained for the full-length wt-hPAH tetramer (6,8) was also observed for all the mutant forms ( Fig. 5C) with saturation reached at about 2 mM substrate, although the maximum equilibrium SPR response (⌬RU max ) in the binding isotherms varied (Fig. 5, A and  C). For the N223D mutant a ⌬RU max of ϳ60% of the value observed for the wild-type was obtained. By contrast, a significant increase in the ⌬RU max value (ϳ130%) was observed for the K113P tetramer (Fig. 5C). The N426D mutant displayed a similar binding isotherm as the wild type, whereas the T427P mutant revealed an ϳ50% decrease in the ⌬RU max . The Nterminal mutants G33A and G33V also revealed a reduced ⌬RU max response, i.e. ϳ30 and ϳ50%, respectively, of wt-hPAH  in contrast to N32D, which gave a saturating equilibrium binding response similar to the wild type (Fig. 5C) (Table III). K113P demonstrated a slightly higher affinity for L-Phe ([S] 0.5 ϭ 120 Ϯ 3 M), whereas for T427P the affinity was the same as for the wild type, i.e. 135 Ϯ 6 M (Table III). The tryptophan fluorescence spectrum of tetrameric wt-hPAH demonstrated an ϳ10 nm red shift and an increased quantum yield when binding L-Phe at a saturating concentration (20). This response was primarily related to a change in the microenvironment of Trp 120 in each protomer, i.e. from a buried position in the ligand-free enzyme to a more solventexposed position in the substrate-activated form (20), and the tryptophan fluorescence has been used as a valuable conformational probe for both the wild-type and various truncated forms of the enzyme (8). Fig. 5B demonstrates a typical binding isotherm for L-Phe to wt-hPAH tetramer with a [S] 0.5 value of 78 Ϯ 2 M and a Hill coefficient of 2.7 Ϯ 2. For the wild type and four of the mutant tetramers (N223D, K113P, N426D, and N32D) a close positive correlation (r 2 ϭ 0.98, figure not shown) was observed between the ⌬F max value at 358 nm (Fig. 5D) and the ⌬RU max obtained by SPR (Fig. 5C) (see also below). However, the tryptophan fluorescence of the T427P mutation did not change significantly upon the addition of L-Phe. In general, a decrease in the Hill coefficient determined from the binding isotherms was observed for most of the mutant forms (Table III).
The Relationship between the Substrate-triggered Large Scale Global Conformational Transition and Catalytic Activity-Recent x-ray crystallographic analyses of the catalytic core domain structure of hPAH (7,11) and spectroscopic (magnetic circular dichroism and x-ray absorption spectroscopy) studies on wt-rPAH and two disease-associated mutant forms (25) have provided valuable insights into the relation between structure and function of this enzyme. In both studies it was shown that substrate binding results in a change in the binding geometry/coordination of the catalytic iron and a reorientation of residues lining the active site, which represent major contributing factors to catalysis, and new insights into the catalytic mechanism were obtained. Moreover, two distinct confor-mational states (conformers) were observed in the crystal structure of the catalytic core domain representing the unliganded and substrate analog-bound forms of the enzyme (7,11). This supports the original proposal by Shiman and Gray (5) that protein dynamics play a vital role in the PAH catalytic pathway. In the present study the biochemical/biophysical properties observed for the described mutant forms of fulllength tetrameric hPAH have given additional information on the relationship between enzyme dynamics/flexibility and catalytic activity (V max ). The catalytic activity of PAH was measured either by varying the L-Phe concentration or the BH 4 concentration. Whereas varying substrate concentration resulted in sigmoid steady-state kinetics, the BH 4 dependency demonstrated classical Michaelis-Menten kinetics, and a higher V max was also obtained for the latter (Table II). Thus, from Fig. 6 it is seen that the change in catalytic activity for both the L-Phe (Fig. 6A) and the BH 4 (Fig. 6B) concentration curves upon mutation of defined/predicted hinge-bending residues in hPAH correlates strongly with the change in the substrate-induced global conformational transition (linear correlation coefficient r 2 ϭ 0.93-0.96) for N32D, K113P, N223D, N426D, and wt-hPAH tetramers, as measured by SPR and intrinsic tryptophan fluorescence spectroscopy. By contrast, G33A/G33V and T427P did not fit into this correlation (Fig. 6). Although SPR and intrinsic tryptophan fluorescence detect different aspects of the L-Phe-triggered conformational change of the protomer, the ⌬RU and ⌬F responses have complementary time windows and binding isotherms (8). The conformational transition observed by SPR (as a positive ⌬RU value) is compatible with an increase in the hydrodynamic radius as measured by size-exclusion chromatography and dynamic light scattering (26,27). The biphasic increase in refractive index (8) also defines the time scale for the conformational change (i.e. molecular motions) in this hysteretic enzyme. However, the exact molecular basis for the SPR response requires further biophysical studies, including volumetric techniques that may give further insights into hydration and intramolecular packing related to the transition. DISCUSSION The tetrameric form of mammalian PAH displays homotropic cooperative binding of the substrate L-Phe, as observed by steady-state enzyme kinetics and equilibrium binding studies (2,8,28,29). The cooperativity has so far primarily been interpreted within the framework of the classical concerted two-state model of Monod et al. (30), i.e. as a reversible transition between a low affinity/low activity "T-state" and a high affinity/high activity "R-state" of the enzyme tetramer (5, 8, 26 -28). However, the molecular basis for the cooperativity of L-Phe binding and its catalytic activation of the tetramer is only partly understood.
Molecular Motions in the Catalytic Domain Triggered by Substrate Binding-The crystal structures of two different conformational states have been obtained recently for the unliganded and the ternary complex of the dimeric double-truncated form of hPAH, with the substrate analogs 3-(2-thienyl)-L-alanine and L-norleucine (7,11). The structures revealed that binding of substrate at the active site triggered large scale structural changes in the catalytic domain, including the active site crevice structure, representing the "epicenter" of the global conformational transition related to L-Phe activation observed in the full-length protomer and its related oligomeric forms (7,11). The most prominent molecular motions in the structural transition were observed in the loop region comprising residues 134 -139, flanked by two short 3 10 -helices (residues 125-133 and 140 -142), with a maximum backbone displacement of 9.6 Å for Tyr 138 (7,11). In addition, two hinge-bending regions centered at Leu 197 (in a kinked ␣-helix) and Asn 223 (in the 218 -226 loop structure) were defined and found to act in concert upon substrate binding to create a large scale rigid-body movement that displaces the ␤-strands 333-336, 339 -342, 363-366, and 385-389 as well as the ␣-helices 351-356, 369 -372, and 392-403 (7,11). By the substitution of N223D, this hinge region is likely to destabilize the structure by disrupting an H-bond to Gly 218 ( Fig. 2A), resulting in an ϳ50% reduction in the measured global conformational change triggered by L-Phe binding (Fig. 5, C and D). Thus, in this mutant form the affinity for the substrate is not altered (Table II), but the basal catalytic activity and the fold activation by L-Phe preincubation are both significantly reduced ( Fig. 4 and Table II). The results support the involvement of the 218 -226 hinge region in the substrate-induced molecular motions in the catalytic domain as shown in the crystallographic studies (7,11). Since no threedimensional structure of the full-length tetrameric enzyme has yet been obtained, how the conformational changes observed in the catalytic domain are propagated to the N-and C-terminal domains of the protomer and affect the interactions between the subunits in the dimer/tetramer is an open question. Moreover, the molecular dynamics simulations of full-length hPAH have so far not been able to elucidate further the subset of motions and the dynamic pathway of the substrate-induced conformational changes (31). The major motions in this hysteretic enzyme occur on longer time scales, i.e. seconds-to-minutes (2,5,6,8), than are normally covered (nanoseconds) by the molecular dynamics simulations. Here we addressed this issue by studying the functional effects of site-directed mutagenesis of specific residues at the defined (11) and predicted (3,13,14) hinge-bending regions by spectroscopic methods, which allow the global conformational change to be followed in real time (6,8).
The Regulatory-to-Catalytic Interdomain Hinge-bending Region-The functional properties observed here for the K113P mutant tetramer (with the wild-type sequence 111 RD-KKKDT 117 ) further support the structural proposal that the solvent-exposed linker motif 111 RDKEKNT 117 in rPAH, connecting the regulatory domain (residues 2-117) with the catalytic domain (residues 118 -410), represents a center of interdomain movement (13). The Pro substitution will cause not only a change in the backbone conformation but also a loss of two H-bonds to the carboxyl groups of Asp 27 in the autoregulatory sequence and Asp 315 in the catalytic domain ( Fig. 2A). Although both aspartates also form H-bonds to Arg 252 in the catalytic domain ( Fig. 2A), the position of the autoregulatory FIG. 5. The equilibrium L-Phe binding isotherm of wt-hPAH and the mutant forms as measured by SPR and intrinsic tryptophan fluorescence spectroscopy. A, the SPR response isotherm for immobilized wt-hPAH to increasing concentrations of L-Phe. The truncated form ⌬N102/⌬C24-hPAH (catalytic core enzyme) was immobilized to the reference cell. The experiments were performed at 25°C in HBS-EP buffer (10 mM Hepes, 150 mM NaCl, 3.4 mM EDTA, and 0.005% (v/v) of the surfactant P-20, pH 7.4). B, the change in fluorescence intensity of wt-hPAH as a function of added L-Phe; the small, unspecific signal contributed by added D-Phe was subtracted from the L-Phe signal. The measurements were performed at 25°C in a buffer consisting of 20 mM Hepes and 0.2 M NaCl, pH 7, and an enzyme concentration of 0.02 mg/ml. Excitation and emission were at 295 and 358 nm, respectively. C, the maximum SPR response (⌬RU max ) to added L-Phe for selected mutant forms relative to wt-hPAH. D, the maximum change in tryptophan fluorescence (⌬F max ) at 358 nm to added L-Phe for the selected mutant forms relative to wt-hPAH.  (14). b Unfavorable signal-to-noise-ratio in the fluorescence at low L-Phe concentrations.
sequence in relation to the active site crevice structure may still be affected by the mutation through a weakening of the interactions between the regulatory and catalytic domains, and thus facilitate inter-domain movements upon L-Phe binding.
The time-dependent increase in the refractive index of immobilized tetrameric and dimeric wt-hPAH triggered by L-Phe binding has been related to a global hysteretic conformational transition (isomerization), involving both the catalytic and the regulatory domains (8). The K113P mutation per se results in a repositioning (i.e. conformational change) of the flexible Nterminal autoregulatory sequence, as indicated by the 3.5-fold increase in the basal catalytic activity and a correspondingly reduced catalytic activation on preincubation with L-Phe, indicating a more substrate-accessible active site. Thus, the 1.3fold increase in the ⌬RU max response observed here for the K113P tetramer is compatible with a higher mobility of the regulatory domain (relative to the catalytic domain) being responsible for the overall increase in the refractive index of the immobilized enzyme on L-Phe binding.
Hinge-bending Region in the Regulatory Domain That Controls Solvent and Substrate Access to the Active Site-Structural and biochemical evidence has been presented in support of the proposal that a conformational flexibility of the N-terminal autoregulatory sequence (residues 1-30) is involved in the regulation of solvent and substrate access to the active site (13,(32)(33)(34)(35). In the unliganded structure of a dimeric rPAH lacking the C-terminal 24-residue tetramerization motif (⌬C24-rPAH), a segment (residues 19 -29) of the autoregulatory sequence extends over the active site crevice in the catalytic core domain, leaving only a relatively narrow access for the substrate (13). A rotation of the regulatory domain (relative to the catalytic domain) at the linker region 111 RDKEKNT 117 , triggered by L-Phe binding (discussed above), was considered to result in a movement of the autoregulatory sequence away from the active site including a hinge-bending motion centered at Gly 33 (13,32). The biochemical and biophysical results obtained in the present study for the G33A/G33V mutant gives further support for this conclusion. Gly 33 is a buried residue ( Fig. 2A), and its substitution by more bulky hydrophobic amino acids, such as Ala and Val, may introduce a restriction in the conformational space and movement at the hinge. Thus, both mutant forms revealed a markedly increased basal catalytic activity (3.2-fold for both) and a correspondingly reduced catalytic activation on preincubation with L-Phe ( Fig. 4 and Table II), compatible with a more open active site crevice structure in the ligand-free mutant forms, probably because of a more permanent reorientation of the N-terminal arm at this hinge. Although the G33A/ G33V mutations had comparable V max values to the wild type, they showed a reduced tryptophan fluorescence response and a reduced SPR response (⌬RU max ) by ϳ30% (G33A) and ϳ50% (G33V) of the wild-type at 2 mM L-Phe (Fig. 5C) and thus did not fit into the linear activity-conformational change correlation observed for wt-hPAH and the other mutant forms (Fig. 6). The reduced conformational transition may be explained by a more restricted mobility of the N-terminal arm, which does not succeed in the full change of the local environment of the spatially closed Trp 120 , and in the movements related to the N-terminal component of the SPR response. Most interesting, the effects of the mutations at Asn 32 and Gly 33 on the binding of L-Phe at the active site (Table II) support the proposal (31) that the overall conformation of the regulatory domain influences the cooperativity without directly binding the substrate, as compared with previous suggestions of a second regulatory binding site for L-Phe in this domain (5,13,27).
The Dimerization-to-Tetramerization Interdomain Hingebending Region Mediating the Homotropic Cooperative Binding of L-Phe-The crystal structure of an N-terminal truncated tetrameric hPAH (⌬N102-hPAH) has identified a five-residue hinge-bending region ( 425 DNTQQ 429 ) in the dimerization-totetramerization inter-domain structure that links the two antiparallel ␤-strands (residues 411-415 and 420 -424) with the 25-residue C-terminal amphipathic ␣-helix (residues 428 -452), which forms an antiparallel coiled-coil structure by domain swapping in the tetramer (3). Substitution of Asn 426 with an Asp (as in the structurally related human tryptophan hydroxylase) and Thr 427 with a Pro (as in human tyrosine hydroxylase) resulted in mutant forms with quite different properties. The N426D tetramer revealed a V max and a Hill coefficient similar to that of wt-hPAH, but with a reduced catalytic efficiency (mainly a [S] 0.5 effect of L-Phe) (Table II). Moreover, it was activated 5-fold by L-Phe preincubation ( Fig. 4 and Table  II) and revealed a similar substrate-induced global conformational change as wt-hPAH (Fig. 5, C and D, and Fig. 6). By contrast, the T427P tetramer (as isolated) revealed almost no FIG. 6. Secondary plots demonstrating the relationship between the catalytic activity (V max ) and the substrate-induced global conformational change of wt-hPAH and the mutant forms, as measured by SPR and intrinsic tryptophan fluorescence. The V max was measured on the L-Phe preincubated enzymes by varying the L-Phe concentration (A) or by varying the BH 4 concentration (B). The primary data are given in Table II and Fig. 5, C and D. The relative values for the amplitude of the global conformational change are the mean values of the SPR response and the intrinsic tryptophan fluorescence response with the wild-type as a reference of 1.0. The N32D, K113P, N223D, N426D, and wt-hPAH tetramers (q) revealed a strong positive correlation (linear correlation coefficient r 2 ϭ 0.93-0.96) with the catalytic activity (V max ) with both L-Phe (A) and BH 4 (B) as the variable substrate. As discussed in the text, the mutant forms G33A/ G33V and T427P (OE) do not fit into this relationship. activation (1.2-fold) by L-Phe, and its catalytic efficiency (Table  II) was decreased to about 40% of wt-hPAH (mainly a V max effect). A decrease was also observed for the L-Phe-induced global conformational change as measured by SPR spectroscopy, whereas only a small decrease (negative response) in the intrinsic tryptophan fluorescence (of Trp 120 ) was seen (Table  III). It should be noted that the crystal structure of the ligandfree ⌬N102-hPAH tetramer shows a pronounced asymmetry (3), with two alternate conformations in the 425-429 hinge regions of the two adjacent dimers (AC and BD) in the tetramer, related by a 2-fold noncrystallographic symmetry axis (3). In both subunits of the dimer BD, Thr 427 in this loop establishes a stabilizing H-bond (3.2 Å) with the C-terminal carboxyl group of Lys 452 in the subunits of dimer AC (Fig. 2B), respectively. Most interesting, because of the asymmetry of the tetrameric structure, the 425-429 loop regions of dimer AC cannot interact with the C termini of dimer BD because the distance between the two residues is 5.41 Å. This implies that in the ligand-free wild-type tetramer two interactions will occur, one in each end of the coiled-coil motif, involving two monomers in different dimers (AC and BD). Moreover, molecular modeling has suggested that this hinge region also establishes several contacts with the regulatory domain of the same subunit (31) and thus may represent a possible pathway for the propagation of the substrate-induced global conformational changes initiated at the active site (the epicenter). By contrast, in tyrosine hydroxylase Thr 427 is substituted by a Pro residue, and the enzyme exists only as a symmetric tetramer in which the four subunits are related by a crystallographic 222 symmetry axis (36), and the enzyme shows no cooperativity of substrate binding. Thus, the T427P mutation in hPAH resulted in the loss of cooperativity, and instead of a stabilization of the tetrameric form, a shift in the tetramer 7 dimer equilibrium in favor of free dimers (14) was observed, probably because of a loss of the asymmetry and the interactions between the subunits in the two dimers AC and BD (see above).
Global Conformational Stability of Wild-type and Mutant Forms of hPAH-None of the hinge-loop mutations studied here showed any significant change in the ordered secondary structure as accessed by their far-UV CD spectra, but minor conformational changes were suggested on the basis of the four complementary conformational probes used, including the relative thermal stability (Fig. 3). Previous thermal denaturation studies by differential scanning calorimetry and far-UV CD spectroscopy on the wild-type and truncated forms of hPAH (37) have suggested that the thermodynamically irreversible denaturation of the full-length wild-type tetramer occurs through a sequential unfolding of the N-terminal regulatory domains (T m ϳ46°C) and the catalytic domains (T m ϳ54°C) (37). Both transitions are shifted to higher temperatures in the presence of L-Phe (37) (Fig. 3). The observed effect of L-Phe binding on the biphasic thermal denaturation profile of wt-hPAH (37) (Fig. 3) further supports the interpretation of the SPR response in terms of a biphasic time course, possibly contributed by the two domains (8). Most of the tetrameric mutant forms revealed similar thermal denaturation profiles, with two resolved transitions ( Fig. 3 and Table I). Moreover, the conformational changes induced by the mutations G33A, K113P, and N223D affected the T m value for the first transition (related to the regulatory domain), whereas the second transition was relatively unaffected. Most interesting, the thermal unfolding process of the T427P-hPAH tetramer presented essentially a single transition with a low T m value (41°C). The T427P mutation at the 425-429 hinge may thus be responsible for a destabilization of both the regulatory and catalytic domains, which unfold at around the same T m value. Overall, there seems to be a good correlation between the extent of limited proteolysis and the thermal stability of the mutant enzymes when compared with the wild-type, probably related to differences in their intramolecular packing of the structural elements in the tetramers.
Concluding Remarks-Structural dynamics is essential for the biological function of proteins, and hinge-bending motions usually occur in proteins that have two or more domains connected by linkers (i.e. hinges/loops) and are frequently involved in ligand-induced protein conformational changes. PAH is such a protein in which the binding of substrate (L-Phe) at the active site initially triggers local conformational changes at this site (the epicenter), which are propagated globally through several hinge-bending motions, as partly demonstrated by molecular motion analyses of the catalytic core enzyme (7,11; see also under "Experimental Procedures"). The present site-directed mutagenesis study provides further support for the structurebased model in which defined hinge-bending regions have been demonstrated (7,11) or predicted (3,13,14) crystallographically to be involved in the substrate-induced global conformational transition. Most interesting, a strong correlation (r 2 ϭ 0.93-0.96) was observed relating the catalytic activity of the wild-type and the N32D, K113P, N223D, and N426D mutant tetramers to the L-Phe-induced large scale global conformational changes as assessed by both SPR and intrinsic tryptophan fluorescence spectroscopy. This correlation supports previous conclusions, based on crystallographic data (7,11), that substrate binds to the catalytic/active site where it triggers large scale structural changes in the catalytic domain, which are further transmitted to the dimer/tetramer and directly relate the global conformational change to the catalytic activation of the enzyme. The G33A/G33V and T427P mutant tetramers did not fit into this relationship, indicating a more restricted contribution of the related hinges to the global conformational transition, as measured by SPR and fluorescence spectroscopy. Further studies are required to uncover their exact involvement in the activation of PAH. In particular, three-dimensional structural studies of a full-length dimeric/ tetrameric enzyme, in its ligand-free and substrate-bound form, are required to determine how the local conformational changes at the active site are propagated to the entire protomer as well as to the dimer/tetramer leading PAH from its "T-state" to its "R-state."