Báo cáo khoa học: Studies on the regulatory properties of the pterin cofactor and dopamine bound at the active site of human phenylalanine hydroxylase pptx

catalytic iron and activation of dioxygen, the natural pterincofactor H4biopterin inhibits the substrate activation of rPAH by forming a binary enzymeÆH4biopterin complex [7,11].The acti

Studies on the regulatory properties of the pterin cofactor

and dopamine bound at the active site of human phenylalanine

hydroxylase

Therese Solstad1, Anne J Stokka1, Ole A Andersen2and Torgeir Flatmark1

1

Department of Biochemistry and Molecular Biology, University of Bergen, Norway;2Department of Chemistry,

University of Tromsø, Norway

The catalytic activity of phenylalanine hydroxylase (PAH,

phenylalanine 4-monooxygenase EC 1.14.16.1) is regulated

by three main mechanisms, i.e substrate (L-phenylalanine,

L-Phe) activation, pterin cofactor inhibition and

phos-phorylation of a single serine (Ser16) residue.To address the

molecular basis for the inhibition by the natural cofactor

(6R)-L-erythro-5,6,7,8-tetrahydrobiopterin, its effects on the

recombinant tetrameric human enzyme (wt-hPAH) was

studied using three different conformational probes, i.e the

limited proteolysis by trypsin, the reversible global

con-formational transition (hysteresis) triggered by L-Phe

bind-ing, as measured in real time by surface plasmon resonance

analysis, and the rate of phosphorylation of Ser16 by

cAMP-dependent protein kinase.Comparison of the inhibitory

properties of the natural cofactor with the available

three-dimensional crystal structure information on the ligand-free,

the binary and the ternary complexes, have provided

important clues concerning the molecular mechanism for

the negative modulatory effects.In the binary complex,

the binding of the cofactor at the active site results in the

formation of stabilizing hydrogen bonds between the dihydroxypropyl side-chain and the carbonyl oxygen of Ser23 in the autoregulatory sequence.L-Phe binding triggers local as well as global conformational changes of the pro-tomer resulting in a displacement of the cofactor bound at the active site by 2.6 A˚ (mean distance) in the direction of the iron and Glu286 which causes a loss of the stabilizing hydrogen bonds present in the binary complex and thereby a complete reversal of the pterin cofactor as a negative effector The negative modulatory properties of the inhibitor dop-amine, bound by bidentate coordination to the active site iron, is explained by a similar molecular mechanism inclu-ding its reversal by substrate bininclu-ding.Although the pterin cofactor and the substrate bind at distinctly different sites, the local conformational changes imposed by their binding

at the active site have a mutual effect on their respective binding affinities

Keywords: tetrahydrobiopterin; dopamine; phosphoryl-ation; surface plasmon resonance; regulation

Mammalian phenylalanine hydroxylase (PAH,

phenylala-nine 4-monooxygenase, EC 1.14.16.1) catalyses the

stereo-specific hydroxylation of L-phenylalanine (L-Phe) to

tyrosine (L-Tyr) in the liver [1], kidney [2,3] and

melano-cytes [4], utilizing (6R)-L

-erythro-5,6,7,8-tetrahydrobiop-terin (H4biopterin) as the physiological electron donor.A

lack or dysfunction of this enzyme in humans is associated with the autosomal recessive disease hyperphenylalanine-mia/phenylketonuria [5] (http://www.mcgill.ca/pahdb) It has been estimated that the liver contains a sufficiently high level of PAH and pterin cofactor to remove all free L-Phe from the blood within a few minutes if all enzyme molecules are fully active [6].However, early on it was recognized that the activity of PAH is effectively controlled by several mechanisms in order to maintain the phenylalanine and tyrosine homeostasis in vivo despite great fluctuations in the dietary intake of L-Phe and the overall rate of protein catabolism.It became apparent that the short-term control

of rat liver PAH (rPAH) is kinetic, primarily through an activation of the enzyme by L-Phe [7,8].This activation is a cooperative, reversible process involving all protomers of the 200-kDa enzyme homotetramer [7,8].rPAH is activated several fold in vitro by preincubation with L-Phe as well as

by some amino acid analogues [9].The recombinant human enzyme (hPAH) has similar regulatory properties in vitro as rPAH, i.e the tetrameric form binds L-Phe with positive cooperativity with a Hill coefficient (h) of 1.6–1.9, and preincubation with substrate results in a sixfold to eightfold activation of the enzyme [10].In addition to its catalytic function as an electron donor in the reduction of the

Correspondence to T.Flatmark, Department of Biochemistry and

Molecular Biology, University of Bergen, A˚rstadveien 19,

N-5009 Bergen, Norway.

Fax: + 47 55586400, Tel.: + 47 55586428,

E-mail: torgeir.flatmark@ibmb.uib.no

Abbreviations: hPAH, human phenylalanine hydroxylase; rPAH,

rat phenylalanine hydroxylase; h, Hill coefficient; H 2 biopterin,

(6R)- L -erythro-7,8-dihydrobiopterin; H 4 biopterin, (6R)- L

-erythro-5,6,7,8-tetrahydrobiopterin; H 4 6-methyl-pterin, 6-methyl-5,6,7,

8-tetrahydropterin; IPTG, isopropyl-thio-b- D -galactoside;

L-Phe, L -phenylalanine; MBP, maltose binding protein;

PKA, cAMP dependent protein kinase; wt, wild-type.

Enzyme: phenylalanine 4-monooxygenase or phenylalanine

hydroxylase (EC 1.14.16.1).

(Received 22 October 2002, revised 15 January 2003,

accepted 20 January 2003)

catalytic iron and activation of dioxygen, the natural pterin

cofactor H4biopterin inhibits the substrate activation of

rPAH by forming a binary enzymeÆH4biopterin complex

[7,11].The activity of rPAH and human phenylalanine

hydroxylase (hPAH) is also regulated by post-translational

mechanisms, notably phosphorylation of Ser16 by

cAMP-dependent protein kinase (PKA), which sensitizes the

enzyme for activation by L-Phe [12,13].Whereas the

binding of substrate slightly increases the rate of

phos-phorylation of rPAH and hPAH by PKA, H4biopterin acts

as a negative effector on the same process [12,13].Finally,

nonenzymatic deamidation of labile Asn residues in hPAH

during its expression in, e.g Eschericia coli, has more

recently been shown to result in a threefold increase in its

catalytic efficiency [10]

The molecular basis for the inhibitory effects observed for

the natural pterin cofactor (H4biopterin/H2biopterin) has

been addressed in a series of studies on rPAH including

direct binding measurements [14], steady-state kinetic

ana-lysis [11] and modulation of Ser16 phosphorylation by PKA

[12].The binding studies by Shiman and collaborators

[11,15] were interpreted to support a working model that

includes three types of binding sites for H4biopterin.The

sites were (a) a redox site, involved in the reduction of the

active site iron [Fe(III)fi Fe(II)], (b) a catalytic site,

involved in the activation of dioxygen and hydroxylation

of L-Phe, and (c) a regulatory site outside the active site,

responsible for its inhibitory properties.However, recent

crystal structure analyses of the ligand-free, the binary and

ternary complexes of hPAH [16–21], as well as the

complementary NMR-molecular modelling structural

stud-ies [22], have not been able to identify more than a single

cofactor binding site, i.e the binding at the active site, with

two alternative orientations in the binary and ternary

complex [20,21].Thus, the molecular basis for the

regula-tory (inhibiregula-tory) properties of H4biopterin is still a matter of

debate, both with respect to the domain localization of the

inhibitory binding site and the essential importance of its

dihydroxypropyl side-chain for inhibition.In the present

study, we address these questions as well as the related

regulatory properties of the catecholamine inhibitor

dop-amine, which is known to bind covalently (bidentate

coordination) to the active site iron [17] with high affinity,

employing recombinant tetrameric wt-hPAH, with the

three-dimensional crystal structural information on this

enzyme [16–21] as a reference

Materials and methods

Materials

The restriction proteases enterokinase and factor Xa were

obtained from Invitrogen (The Netherlands) and Protein

Engineering Technology (ApS, Aarhus, Denmark),

respect-ively.The catalytic subunit of cAMP-dependent protein

kinase was purified to homogeneity from bovine heart and

was a generous gift from S.O.Døskeland, Department of

Anatomy and Cell Biology, University of Bergen.The

pterin cofactors (H4biopterin, H2biopterin and H4

6-methyl-pterin) were purchased from B.Schircks Laboratory

(Joana, Switzerland).Specific chemicals are mentioned in

the text elsewhere

Expression and purification of recombinant hPAH The pMAL expression system was used for the production

of the wild-type fusion proteins MBP-(D4K)ek-hPAH, MBP-(IEGR)Xa-hPAH and its double truncated form MBP-(IEGR)Xa-hPAH(Gly103-Gln428) with maltose binding protein as the fusion partner [23].Cells were grown

at 37C, and expression was induced at 28 C by the addition of 1 mM isopropyl-thio-b-D-galactoside (IPTG); the cells were harvested after 2, 8 or 24 h of induction.Full-length hPAH (residues 1–452) was obtained by enterokinase (at D4K) or factor Xa (at IEGR) cleavage and hPAH(Gly103–Gln428) by factor Xa cleavage, and the tetrameric (full-length form) and dimeric form (the trun-cated catalytic core enzyme) were isolated by size-exclusion chromatography [23]

Protein measurements The concentration of purified enzyme forms was measured

by the absorbance at 280 nm, using the absorption coeffi-cient e280¼ 1.63 for the fusion protein MBP-(pep)-hPAH and 1.0 for the isolated hPAH protein [23]

Protein phosphorylation The tetrameric wt-hPAH was phosphorylated by PKA as described [13].The standard reaction mixture contained

15 mM Na-Hepes (pH 7.0), 0.1 mM ethylene glycol bis (a-amino ether)-N,N,N¢,N¢-tetraacetic acid, 0.03 mMEDTA,

1 mM dithiothreitol, 10 mM magnesium acetate, 60 lM [c-32P]-ATP (Amersham, UK), 25 nM of the catalytic subunit of PKA and 4 lM of hPAH.The reaction was performed at 30C, and at timed intervals, aliquots of the reaction mixture were spotted on phosphocellulose strips [24] to measure the amount of 32P transferred to the substrate.In control experiments, the active-site iron in hPAH was prereduced by the noninhibitory [12] H4 6-methyl-pterin as described [14] prior to initiation of the phosphorylation reaction in the presence of inhibitory

H4biopterin

Limited proteolysis by trypsin Limited tryptic proteolysis of tetrameric wt-hPAH was performed at 25C in 20 mMNa-Hepes, 200 mMNaCl at

pH 7; the ratio of trypsin to hPAH was 1 : 200 (m/m) Aliquots of 3 lg hPAH were removed at timed intervals and mixed with SDS buffer containing soybean trypsin inhibitor.The protein was finally subjected to SDS/PAGE (10%, w/v), stained with Coomassie Brilliant Blue, and the gels were scanned usingDESKSCAN II(Hewlett Packard Co) and further analyzed by using thePHORETIX 1D analysis software from Nonlinear Dynamics Ltd [10]

Reversed-phase chromatography Reversed-phase chromatography of tryptic peptides was performed using a ConstaMetric Gradient System (Laboratory Data Control, USA) and a 4.6 mm· 10 cm Hypersil ODS C18 column (Hewlett Packard, USA) fitted with a 2-cm guard column.Solvent A was 0.1%

(w/v) triflouroacetic acid in water, and solvent B was

0.1% trifluoroacetic acid in 70% (v/v) acetonitrile

A linear gradient of 5–100% solvent B at 1 mLÆmin)1

for 60 min was used for the separation of peptides

Absorbance was monitored at 214 nm using a

Hewlett-Packard model 1040A photodiode array HPLC detector

Surface plasmon resonance analyses

The interactions between tetrameric wt-hPAH and its

substrates (L-Phe and H4biopterin/H2biopterin) were

studied by surface plasmon resonance (SPR) analysis

using the BiaCore X instrument (BiaCore AB, Uppsala,

Sweden).The enzyme, diluted to a final concentration of

0.23 mgÆmL)1in 10 mMsodium acetate buffer, pH 5.4, was

immobilized to the carboxymethylated dextran matrix of a

sensor chip (CM5 from BiaCore AB) by the amine coupling

procedure [25–27].The double truncated dimeric form

hPAH(Gly103-Gln428) was immobilized to the reference

surface by the same procedure [27].Seventy microlitres of

10 mMdithiothreitol in HBS running buffer (0.15MNaCl,

3 mM EDTA, 0.005% surfactant P20 in 0.01M Hepes,

pH 7.4) was allowed to pass through both flow cells (in

series) which empirically decreased the time required to

reach a stable baseline [25].Increasing concentrations of

L-Phe diluted in HBS buffer was injected over the

immo-bilized enzyme in the absence and presence of 100 lM

H4biopterin or 500 lMH2biopterin.The reduced cofactor

(in 1 mMHCl) was kept on ice and the pH adjusted to 7.4

with an accuracy of 0.01 unit immediately before injection

[25].All analyses were carried out at 25C and at a constant

flow of 5 lLÆmin)1.The sensorgrams were obtained as the

difference in SPR (DRU) response between the sample and

the reference cell, which corrects for any changes in the bulk

refractive index, together with the small DRU value

associated with the binding of the 165-Da substrate.The

time-dependent increase in RU (end point at 3 min) was

measured directly from the sensorgrams using the cursor

guided reading of the X- and Y-coordinates, and

was related to the calculated pmol of immobilized

enzyme by assuming that 1000 RU corresponds to 1 ng

protein boundÆmm)2 [28], i.e expressed as DRU/(pmol

subunitÆmm)2)

Structural studies

The recently reported three-dimensional crystal structures

of hPAH and rPAH [16–21] have provided the basis to

further explore the molecular mechanism by which the

pterin cofactor and dopamine function as negative effectors

and how substrate binding completely reverses these effects

The coordinates for the structure of the binary and ternary

complexes of the catalytic core domain

hPAH(Gly103-Gln428), i.e of hPAH-Fe(III)ÆH2biopterin (PDB id codes

1DMW and 1LRM) [18,20], of hPAH-Fe(II)ÆH4biopterin

(PDB id code 1J8U) [20] and hPAH-Fe(II)ÆH4biopterinÆ

3-(2-thienyl)-L-alanine (PDB id code 1KW0) [21], define the

position, orientation, conformation and hydrogen bonding

network of the pterin cofactor in the three structures

Superpositions of the binary and ternary complexes of

hPAH onto the crystal structure of the ligand-free dimeric

rPAH (PDB id code 1PHZ) [19], which contains both the regulatory and the catalytic domains (residues 1–429), were performed to demonstrate the interactions of the dihydroxy-propyl side-chain of the pterin cofactors with the N-terminal autoregulatory sequence.Similarly, the coordinates for the structure of the binary complex with dopamine, i.e hPAH-Fe(III)Ædopamine (PDB id code 5PAH) [17] define the position and orientation of the inhibitor and the interaction of its main-chain with the autoregulatory sequence

Results

Recombinant forms of hPAH were produced at high yields

as fusion proteins in E coli using the pMAL expression vector, and the cleaved tetrameric forms of wt-hPAH and the dimeric truncated form hPAH(Gly103–Gln428) were isolated by size-exclusion chromatography [23].Aliquots (10 mgÆmL)1) were stored in liquid nitrogen and a new aliquot was used for each individual experiment.The conformational differences of tetrameric wt-hPAH resulting from the binding of different ligands (H4biopterin, H2 biop-terin, dopamine and L-Phe) were studied using three different conformational probes, i.e their effects (a) on the limited proteolysis by trypsin, (b) on the reversible con-formational transition (hysteresis) triggered by substrate binding, as followed in real time by surface plasmon resonance (SPR) analyses, and (c) on the rate of phos-phorylation of Ser16 by PKA

Effect of ligands on the limited proteolysis

of recombinant wt-hPAH and its catalytic core Limited proteolysis of rPAH by chymotrypsin has been shown to be a sensitive conformational probe, and among the observed ligand effects was an inhibition of proteolysis

by H4biopterin [29].From Fig.1A it is seen that at saturating concentration (40 lM), H4biopterin has a similar inhibitory effect on the limited proteolysis of tetrameric wt-hPAH by trypsin, which cleaves after Lys/Arg residues

in a putative hinge region (residues 111–117, RDKKKNT) connecting the regulatory domain with the catalytic domain [30].Moreover, the covalently bound inhibitor dopamine (40 lM) is equally efficient in reducing the susceptibility towards proteolysis.However, when 1 mML-Phe was also present during the incubation with trypsin, the inhibitory effects of H4biopterin and dopamine were completely prevented and the rate of proteolysis increased to approxi-mately the same high level as observed in the presence of substrate alone.By contrast, the dimeric catalytic core enzyme hPAH(Gly103–Gln428) [16], lacking both the N-terminal regulatory domain and the C-terminal tetra-merization domain, was as expected [30] found to be more resistant to proteolysis.When this enzyme form was incubated with either H4biopterin, L-Phe, or both ligands combined, no significant effect of the ligands on the rate of proteolysis was observed as compared to the ligand-free enzyme (Fig.1B).In this case, the results from SDS/PAGE analyses were confirmed by reversed-phase chromatogra-phy of the peptides released during incubation (data not shown)

Effect of H4biopterin/H2biopterin on the conformational

transition (hysteresis) triggered by substrate binding

as studied by surface plasmon resonance analysis

The effect of the pterin cofactors on the reversible global

conformational transition induced by L-Phe binding to

tetrameric wt-hPAH was measured in real time by the

time-dependent change in refractive index (i.e as a surface

plasmon resonance (SPR) response) of the immobilized

enzyme [25–27].Approximately 25 ngÆmm)2 (0.48 pmol

subunitÆmm)2) of tetrameric wt-hPAH was immobilized to the dextran matrix of the sample surface, and a slightly lower amount of the dimeric catalytic core enzyme hPAH(Gly103–Gln428) (0.38 pmol subunitÆmm)2) was immobilized to the reference surface [27].The time-dependent conformational change (SPR response) of the full-length wt-hPAH was measured as a function of L-Phe concentration and a steady-state (3 min response) binding isotherm was obtained [26,27].The isotherm observed in the absence of biopterin cofactor was hyperbolic with a concentration of L-Phe at half-maximum saturation ([S]0.5) of 98 ± 7 lM(Fig.2A,B).Saturation was reached

at approximately 2 mMwith a DRU-value of 75 RU/(pmol subunitÆmm)2).Simultaneous injection of L-Phe (variable concentration) and 100 lM of H4biopterin resulted in a

 50% decrease in the maximum SPR response to L-Phe (Fig.2A), and the [S]0.5-value for L-Phe increased to

178 ± 11 lM.The presence of 500 lM of the oxidized cofactor Hbiopterin in the running buffer also lowers the

Fig 2 The effect of pterin cofactor on the global conformational transition of tetrameric wt-hPAH triggered by L-Phe binding as studied

by surface plasmon resonance The effect of increasing L-Phe concen-tration in the absence (d) and presence (s) of the pterin cofactor (A) 100 l M H 4 biopterin was coinjected with L-Phe.(B) 500 l M of the soluble H 2 biopterin was included in the running buffer.The response

in the absence of pterin cofactor represents the average of two separate titration experiments with a basal mean DRU-value of 25090 RU corresponding to 0.12 pmol (0.48 pmol subunit) of immobilized enzyme.The truncated form hPAH(Gly103-Gln428) was present on the reference surface [27].

Fig 1 The effect of ligand binding on the limited proteolysis by trypsin

of full-length wt-hPAH and the truncated form hPAH(Gly103-Gln428).

(A) Tetrameric wt-hPAH (24 induction with IPTG) preincubated with

either no ligand (d), 40 l M H 4 biopterin (m), 40 l M dopamine (n),

40 l M H 4 biopterin and 1 m M L-Phe (j), 40 l M dopamine and 1 m M

L-Phe (h) or 1 m M L-Phe (s) before being subjected to limited

pro-teolysis by trypsin.(B) hPAH(Gly103-Gln428) was preincubated with

either no ligand (d), 40 l M H 4 biopterin and 1 m M L-Phe (j) or 1 m M

L-Phe (s) before being subjected to limited proteolysis by trypsin.The

ratio hPAH : trypsin was 200 : 1 (m/m).The reactions were allowed to

proceed for up to 1 h at 25 C and aliquots were taken at different time

intervals.The reaction was stopped by the addition of soybean trypsin

inhibitor (the ratio trypsin : inhibitor 1 : 1.5, m/m) and finally

sub-jected to SDS/PAGE stained with Coomassie Brilliant Blue.The gels

were scanned using DESKSCAN II (Hewlett Packard Co.); the volume of

the bands was analyzed by using the PHORETIX 1 D analysis software

from Nonlinear Dynamics Ltd, 1996 [10].

maximum SPR response to L-Phe, most significantly

observed at concentrations above 200 lM (Fig.2B), and

in this case, the [S]0.5-value for L-Phe increased to

123 ± 6 lM.It should be noted that in separate binding

experiments with H4biopterin alone the [S]0.5-value for the

cofactor was measured to 5.6 ± 0.8 lMwith a D R-value

of25 RU/(pmol subunitÆmm)2) at saturation [26]

Effect of active site ligands on the rate

of phosphorylation of recombinant hPAH

H4biopterin and L-Phe have been shown to inhibit and

stimulate, respectively, the rate of in vitro phosphorylation

of Ser16 by PKA in rPAH at physiologically relevant

concentrations [12].In the present study, the ligand effects

on the phosphorylation of Ser16 by PKA in tetrameric

wt-hPAH were measured for the reduced cofactor H4

biop-terin, the oxidized cofactor H2biopterin and the

catechol-amine inhibitor dopcatechol-amine.From Fig.3A it is seen that on

preincubation with saturating concentrations (Fig.3B) of

the biopterin cofactors or dopamine the rate of

phosphory-lation was decreased, and at 200 lM of the ligands the

inhibition was more pronounced for H4biopterin ( 26%)

and dopamine ( 26%) than for H2biopterin ( 12%).The

presence of L-Phe during preincubation with H4biopterin or

dopamine completely reversed their inhibitory effects on

phosphorylation (Fig.3A).From Fig.3B it is also seen that

the most potent inhibitor was dopamine, with a

half-maximum inhibition at < 0.1 lM, and H4biopterin ([I ]0.5of

1.4 lM, r¼ 0.96) was more efficient than H2biopterin ([I ]0.5

of 15.8 lM, r¼ 0.86) The oxidized cofactor reached only a

 12% inhibition of phosphorylation compared to the

 26% for H4biopterin and dopamine.Interestingly, the

inhibitory effect of both biopterin cofactors revealed an

apparent negative cooperativity, with a Hill coefficient (h) of

0.46 (r¼ 0.96) for H4biopterin (Fig.3B, insert) and

h¼ 0.75 (r ¼ 0.86) for H2biopterin.Hill coefficients < 1

were also observed for the inhibition of the phosphorylation

of the tetrameric fusion protein MBP-(pep)-hPAH, and

enzyme preparations isolated after a short (2 h) induction

period gave reproducibly a higher Hill coefficient with

H4biopterin than those isolated after a long (24 h) induction

period (data not shown).The relative efficiency of the

inhibition by the two biopterin cofactors is in good

agreement with the previously reported apparent Kdvalues

for H4biopterin (0.09 ± 0.01 lM) and H2biopterin

(1.1 ± 0.02 lM) in their binding to rPAH, as determined

by fluorescence quenching titration [14].The reduction of

the active-site iron [Fe(III)fi Fe(II)] by H46-methyl-pterin

[14] prior to the phosphorylation assay, did not demonstrate

any significant effect on the apparent binding parameters

for H4biopterin in the present study.Moreover, we have

confirmed our previous finding with rPAH as the substrate

[12] that the dihydroxypropyl side-chain is required for the

cofactor inhibition of phosphorylation (data not shown)

The molecular basis for the negative modulatory

effects of the pterin cofactor and dopamine binding

to wt-hPAH and their reversal by substrate

The recently solved high resolution crystal structures of

hPAH and rPAH [16–21] have provided a detailed picture of

the protein contacts involved in the active site binding of the pterin cofactor [18,20,21], dopamine inhibitor [17] and L-Phe [21].Thus, the superposition of the hPAH catalytic core structures of the binary complexes with oxidized [18] or reduced [20] pterin cofactor onto the structure of the ligand-free rPAH (containing the regulatory and catalytic domains) [19] revealed that both the reduced and the oxidized cofactor interact with the N-terminal autoregulatory sequence at Ser23.A close-up of this site of interaction (Fig 4A,B) shows that the O1¢ and O2¢ of the dihydroxypropyl side-chain

Fig 3 The effect of H 4 biopterin, H 2 biopterin, dopamine and L-Phe on the phosphorylation of tetrameric wt-hPAH (A) Time-course for the phosphorylation of wt-hPAH (4 l M ) at standard incubation condi-tions at 30 C, including 60 l M [c-32P]ATP and 25 n M C-subunit of protein kinase A (PKA).At timed intervals, aliquots of the reaction mixture were spotted on phosphocellulose strips [24] to measure the amount of32P transferred to the substrate by scintillation counting The incubations contained no ligand (d), 40 l M H 4 biopterin (m),

40 l M dopamine (n) or 1 m M L-Phe in combination with either 40 l M

H 4 biopterin or 40 l M dopamine (h).(B) The effect of increasing concentrations of H 2 biopterin (s), H 4 biopterin (m) and dopamine (n)

on the rate of phosphorylation (t ¼ 10 min).Each point in the curves represents the average of four measurements.Insert: a conventional Hill plot on the H 4 biopterin data is shown in the main figure.

Y ¼ (v o ) v x )/(v o ) v min ), which is the fractional decrease of phos-phorylation rate seen at the concentration x of H 4 biopterin v o is the rate in the absence of H 4 biopterin and v min is the rate at very high concentrations of H 4 biopterin.The observed Hill coefficient (h) was found to be 0.46 (r ¼ 0.96).

are sufficiently close to form favourable hydrogen bonds to

the carbonyl oxygen of Ser23 (Table 1).However, it is

important to note that the dihydroxypropyl side-chain is

positioned slightly different in the two redox states of the

cofactor and that the distances from Ser23O to O1¢ and O2¢

are slightly different (Table 1) because of the different

positions and hybridizations of the C6 atom of the pyrazine

ring in the two redox states [18,20].This diversity of the

side-chain position is likely to entail different hydrogen-bonding

patterns to Ser23 in the full-length enzyme explaining the

redox state dependent regulatory properties of the pterin

cofactor.Moreover, the superposition of the ternary

struc-ture [21] revealed a similar orientation of the pterin cofactor

as in the binary structures [18,20].However, the reduced

cofactor was found to be displaced by 2.6 A˚ (mean distance)

in the direction of the iron and Glu286 upon substrate binding, and in addition, the hydrogen-bonding network for the cofactor was slightly different when compared to the binary structures [18,20].This displacement of the cofactor results in a loss of stabilizing hydrogen bonds between O1¢ and O2¢ of the dihydroxypropyl side-chain and Ser23O in the autoregulatory sequence (Fig.4C and Table 1) and thus explaining the complete reversal of the pterin cofactor as a negative effector (Figs 1 and 3A,B)

When the crystal structure of the hPAHÆadrenaline/ dopamine binary complex [17] was superimposed onto that

of the ligand-free rPAH containing the regulatory and catalytic domains [19] the catecholamine main-chain is also

Fig 4 Stereo view of the site of interaction between the pterin cofactors and dopamine with the regulatory domain The figure was pro-duced by superimposing the crystal structure

of ligand-free dimeric rPAH (PDB id code 1PHZ), which contains both the regulatory and the catalytic domains (residues 1–429) onto the catalytic core crystal structures of (A) binary hPAH with bound H 2 biopterin (PDB

id code 1LRM) (B) binary hPAH with bound

H 4 biopterin (PDB id code 1J8U) (C) ternary hPAH with bound H 4 biopterin and substrate (PDB id code 1KWO) and (D) binary hPAH with bound dopamine (PDB id code 5PAH) The backbone of rPAH regulatory domain and hPAH catalytic domain are shown in red and green, respectively, while residues Ser23 and Ile25 are shown by ball-and-stick repre-sentation.The figure was prepared using

MOLSCRIPT [42].

seen to interact with the N-terminal autoregulatory

sequence (Fig.4D and Table 1) The dopamine nitrogen

interacts with the carbonyl oxygen of Ser23 and the

dopamine Ca atom interacts with the side-chain of Ile25

Discussion

The catalytic activity of PAH is regulated by four main

mechanisms, i.e by substrate (L-Phe) activation, biopterin

cofactor (H4biopterin/H2biopterin) inhibition, increased

catalytic efficiency on phosphorylation of Ser16 (for review,

see [31]) and activation by spontaneous nonenzymatic

deamidation of specific labile Asn residues [10].Internal

protein dynamics are intimately connected with these

regulatory properties.Based on our previous steady-state

enzyme kinetic and phosphorylation studies on rPAH a

working model was proposed to explain three of these

regulatory properties [12], involving four main

tional states (isomers) of the enzyme.The four

conforma-tional states include a ground state for the ligand-free

enzyme, an activated state with bound substrate (L-Phe), an

inhibited state with bound H4biopterin and finally the state

of catalytic turnover, i.e the ternary enzyme-substrate

complex.Recent crystal structure analyses of the

catalyti-cally active core enzyme in different ligand-bound forms

[16–21] and complementary biophysical studies (reviewed in

[31]) strongly support such a model.Thus, both H4biopterin

and L-Phe bind reversibly at the active site of the core

enzyme by an induced fit mechanism with defined protein

contacts and conformational states [21].Moreover, PAH is

inhibited by the covalent binding of catecholamines, i.e by

bidentate coordination to the active site iron [17].Whereas

the inhibition of the catalytic activity by catecholamines is

well understood at the structural level [17], the molecular

mechanism of the inhibitory properties of the pterin

cofactor (unrelated to its effect as electron donor in the

hydroxylation reaction) has been a controversial issue and is

further discussed below

On the pterin cofactor binding site

Shiman and coworkers [14] have suggested the presence of

several putative binding sites for the natural cofactor

H4biopterin.The proposed binding sites are a regulatory

site (outside the active site) that is responsible for the

observed inhibitory effects of H4biopterin binding and a

redox site responsible for the reduction of the active site iron

in addition to its binding at the catalytic site as part of a

catalytically active ternary complex.Based on the crystal

structure of a ligand-free dimeric C-terminal truncated form

of rPAH [19] and analogies with the structure of pterin-4a-carbinolamine dehydratase (PCD/DCoH) a putative bind-ing site for H4biopterin in the regulatory domain, close to a proposed hinge region (residues 111–117), has been pro-posed [19].This region is also considered as the target for L-Phe induced proteolytic cleavage by trypsin [30].More-over, the binding of L-Phe to a second putative site in the regulatory domain was suggested to induce a conforma-tional transition that modifies the intra-subunit interaction between the regulatory and the catalytic domains.This interaction is followed by the formation of a catalytically activated form of the enzyme and an increased susceptibility

to tryptic cleavage.Thus, binding of H4biopterin to the proposed regulatory site was suggested to prevent the interdomain hinge-bending motion required for activation and susceptibility towards proteolysis [19].However, recent crystal structure analyses of the binary complex with

H2biopterin [18] and the binary and ternary complexes with H4biopterin [20,21] have defined the protein contacts involved in cofactor binding at the active site and their interactions through the dihydroxypropyl side-chain with the autoregulatory sequence in the regulatory domain (Fig.4A,B).Thus, no direct structural evidence (by X-ray

or NMR) has so far been presented in support of a

H4biopterin (and L-Phe) binding site in the regulatory domain [18–21,31,32].Based on this structural information,

we here propose an alternative molecular mechanism for the inhibitory effects of the biopterin cofactor on the rate of phosphorylation of Ser16, the limited proteolysis by trypsin and the substrate induced conformational transition (hys-teresis) related to catalytic activation

The inhibited forms of the enzyme with bound biopterin cofactor or dopamine

The phosphorylation site in PAH (Ser16) is localized in the N-terminal tail (residues 1–18) of the autoregulatory sequence for which no interpretable electron density has been obtained [19], compatible with a rather flexible structure of the N-terminus [33].Our experimental data and the crystal structure analysis discussed above, show the interactions of H4biopterin, H2biopterin and dopamine with the N-terminal autoregulatory sequence at Ser23 and Ile25 (Fig.4A,B,D; Table 1).We can now present an explanation for the inhibitory effect of the biopterin cofactor and dopamine on the rate of phosphorylation (Fig.3), on limited proteolysis (Fig.1A) and on the global conforma-tional transition (hysteresis) related to catalytic activation

Table 1 Comparison of distances (A˚) of the superposition of the crystal structure of ligand-free dimeric rPAH (PDB id code 1PHZ), which contains both the regulatory and the catalytic domains (residues 1–429) onto the catalytic core crystal structures of binary hPAH with bound H 4 biopterin (PDB

id code 1J8U), binary hPAH with bound H 2 biopterin (PDB id code 1LRM), ternary hPAH with bound H 4 biopterin and substrate (PDB id code 1KW0) and binary hPAH with bound dopamine (PDB id code 5PAH).

Distance from Ser23O to H 4 biopterin structure H 2 biopterin structure THA-H 4 biopterin structure Dopamine structure

H 2 biopterin O2¢ 4.0

H 2 biopterin O1¢ 1.3

(Fig 2A,B), i.e.by their direct binding at the active site.That

H4biopterin is a more potent inhibitor than H2biopterin of

the global conformational transition triggered by L-Phe

binding was most directly demonstrated by the SPR

analyses.Whereas 100 lM H4biopterin resulted in a

 50% reduction in the maximum SPR response to L-Phe

binding (Fig.2A), H2biopterin gave only a minor reduction

(maximum 11%) at concentrations higher than 200 lM

(Fig.2B).Interestingly, the inhibitory effect of the biopterin

cofactor on the rate of phosphorylation revealed an

apparent negative cooperativity, with a Hill coefficient (h)

of  0.5 for H4biopterin (Fig.3B, insert) and  0.8 for

H2biopterin.A negative cooperativity has also been reported

for the structurally and functionally related human enzyme

tyrosine hydroxylase in a direct binding assay

[25].Further-more, the Hill coefficient for H4biopterin binding to hPAH

revealed a dependence (both for the isolated tetrameric

wt-hPAH and the tetrameric fusion protein MBP-(D4K)ek

-hPAH) on the induction time with IPTG in E coli, i e a

short induction period (2 h at 28C) gave a slightly higher

Hill coefficient than 24 h induction (data not shown).This

finding may be related to the differences observed in our

steady-state kinetic analyses of the two enzyme forms [10]

Thus, the tetrameric wt-hPAH isolated after 2 h and 24 h of

induction with IPTG in E coli, revealed differences (24 h vs

2 h) in both the affinity for the cofactor H4biopterin

(decreased) and L-Phe (increased), as well as for the catalytic

efficiency (increased) [10].The main physico-chemical

difference between these two enzyme preparations was the

extent of time-dependent nonenzymatic deamidation

(dur-ing expression in E coli) of specific labile Asn residues [10]

Complementary mutagenesis (AsnfiAsp) analyses have

demonstrated that the rate of phosphorylation is indeed

dependent on the extent of deamidation of a very labile Asn

residue (Asn32) in the N-terminal autoregulatory sequence

of wt-hPAH [34]

Dopamine has a dual mechanism for its inhibition of

PAH.First, the bidentate coordination of its catechol

hydroxyl groups to the active site iron [17] results in a

complex with strong inhibition of the catalytic activity [32]

Secondly, on binding at the active site, the dopamine

main-chain interacts with the autoregulatory sequence at Ser23

(by hydrogen bonding) and Ile25 (Fig.4), which results in

an inhibition of Ser16 phosphorylation similar to that of

H4biopterin (Fig.3B) with a half-maximum inhibition at a

concentration < 0.1 lM.This value compares well with the

concentration determined for the half-maximum binding

(0.25 lM) of noradrenaline to rPAH [32].A similar type of

interaction between dopamine and the regulatory domain of

the structurally and functionally closely related enzyme

tyrosine hydroxylase has been reported in experiments using

limited proteolysis as a structural probe [35].In this case, the

affinity of dopamine binding was even higher than in PAH,

with a Kd¼ 1.3 ± 0.6 nM, which is considered to be of

physiological relevance (reviewed in [36])

The interdependent binding of biopterin cofactor

and amino acid substrate

Internal protein dynamics are intimately connected to the

catalytic activity of PAH, and our recent structural studies

on hPAH have revealed some key features of its hysteretic

properties.Thus, the structures of the binary and ternary complexes of hPAH have provided a detailed picture of the protein contacts involved in the binding of biopterin cofactor and amino acid substrate at distinctly different positions in the active site crevice structure [20,21].More-over, the cystal structures have revealed that the binding of both biopterin cofactor and substrate is accompanied by local (active site) conformational changes, most pronounced for the binding of L-Phe, which also triggers global conformational changes (hysteresis) in the protomer as determined by SPR analyses (Fig.2 [26,27]).However, as there is still no crystal structure available for the full-length form of the ligand-free enzyme and the binary substrate complex, it is not known how the observed conformational changes at the active site [21] are propagated to the rest of the full-length protomer in the oligomeric forms.The structural changes observed at the active site of the catalytic domain structure [21] explain why the binding of L-Phe reduces the affinity of pterin cofactor binding (Fig.4C) and vice versa (Fig.2).Thus, the large conformational change observed at the active site on substrate binding changes the position and orientation (relative to the catalytic iron) and the hydrogen-bonding network of the bound H4biopterin.Moreover, the substrate induced repositioning of the cofactor also accounts for the release of its interaction with the autoregulatory sequence (Fig.4C) and the related inhibition of Ser16 phosphorylation (Fig.3) A similar effect of substrate binding was observed for the dopamine inhibition of phosphorylation (Fig.3) and of the limited proteolysis by trypsin (Fig.1), which are both completely reversed by L-Phe binding.The interdependent binding properties of the pterin cofactor and L-Phe have also been observed in kinetic studies on mutant forms of hPAH (point mutations of active site residues), which often have a profound influence on biopterin cofactor and/or substrate binding affinity [18, 37–40].If the affinity for the cofactor is reduced, that of the substrate is often observed to be increased, and vice versa [18,40].It should be noted that enzyme kinetic, three-dimensional structural and biophysical (MCD and EXAFS) studies all support an ordered reaction mechanism wherein both cofactor and substrate must be bound before reaction with dioxygen can occur to generate the active intermediate for the coupled hydroxylation of the cosubstrates [21,41] However, the sequence of binding of cofactor and substrate

is still a matter of discussion and may be different comparing wild-type and mutant forms of the enzyme

Acknowledgements

The study was supported by grants from the Research Council of Norway (NFR), from The Novo Nordisk Foundation, The Nansen Fund, The Blix Family Fund for Advancement of Medical Research and the Norwegian Council on Cardiovascular Diseases.We greatly appreciate the expert technical assistance of Ali Sepulveda Mun˜oz in expression and purification of the recombinant enzymes, and the staff

of the Swiss-Norwegian Beamlines in Grenoble (France).

References

1.Kaufman, S.(1971) The phenylalanine hydroxylating system from mammalian liver Adv Enzymol Relat Areas Mol Biol 35, 245– 319.

2.Richardson, S C.& Fisher, M J.(1993) Characterization of

phenylalanine hydroxylase from rat kidney Int J Biochem 25,

581–588.

3 Lichter-Konecki, U., Hipke, C.M & Konecki, D.S (1999)

Human phenylalanine hydroxylase gene expression in kidney and

other nonhepatic tissues Mol Genet Metab 67, 308–316.

4 Schallreuter, K U , Lemke, K R , Pittelkow, M R , Wood, J M ,

Korner, C.& Malik, R.(1995) Catecholamines in human

kerati-nocyte differentiation J Invest Dermatol 104, 953–957.

5 Scriver, C.R., Waters, P.J., Sarkissian, C., Ryan, S., Prevost, L.,

Cote, D., Novak, J., Teebi, S & Nowacki, P.M (2000) PAHdb: a

locus-specific knowledgebase Hum Mutat 15, 99–104.

6 Shiman, R , Mortimore, G E , Schworer, C M & Gray, D W.

(1982) Regulation of phenylalanine hydroxylase activity by

phenylalanine in vivo, in vitro, and in perfused rat liver J Biol.

Chem 257, 11213–11216.

7 Shiman, R.& Gray, D.W.(1980) Substrate activation of

phenyl-alanine hydroxylase.A kinetic characterization.J Biol Chem.

255, 4793–4800.

8 Shiman, R., Jones, S.H & Gray, D.W (1990) Mechanism of

phenylalanine regulation of phenylalanine hydroxylase J Biol.

Chem 265, 11633–11642.

9.Kaufman, S.(1987) Phenylalanine 4-monooxygenase from rat

liver Methods Enzymol 142, 3–17.

10.Solstad, T.& Flatmark, T.(2000) Microheterogeneity of

recombinant human phenylalanine hydroxylase as a result of

nonenzymatic deamidations of labile amide containing amino

acids.Effects on catalytic and stability properties.Eur J Biochem.

267, 6302–6310.

11 Xia, T., Gray, D.W & Shiman, R (1994) Regulation of rat liver

phenylalanine hydroxylase.III.Control of catalysis by

(6R)-tetrahydrobiopterin and phenylalanine J Biol Chem 269, 24657–

24665.

12 Døskeland, A.P., Døskeland, S.O., Øgreid, D & Flatmark, T.

(1984) The effect of ligands of phenylalanine 4-monooxygenase on

the cAMP-dependent phosphorylation of the enzyme J Biol.

Chem 259, 11242–11248.

13.Døskeland, A P , Martı´nez, A., Knappskog, P.M & Flatmark, T.

(1996) Phosphorylation of recombinant human phenylalanine

hydroxylase: effect on catalytic activity, substrate activation and

protection against non-specific cleavage of the fusion protein by

restriction protease Biochem J 313, 409–414.

14 Shiman, R., Xia, T., Hill, M.A & Gray, D.W (1994) Regulation

of rat liver phenylalanine hydroxylase.II.Substrate binding and

the role of activation in the control of enzymatic activity J Biol.

Chem 269, 24647–24656.

15 Shiman, R., Gray, D.W & Hill, M.A (1994) Regulation of rat

liver phenylalanine hydroxylase.I.Kinetic properties of the

enzyme’s iron and enzyme reduction site J Biol Chem 269,

24637–24646.

16.Erlandsen, H , Fusetti, F , Martı´nez, A., Hough, E., Flatmark, T.

& Stevens, R.C (1997) Crystal structure of the catalytic domain of

human phenylalanine hydroxylase reveals the structural basis for

phenylketonuria Nat Struct Biol 4, 995–1000.

17 Erlandsen, H., Flatmark, T., Stevens, R.C & Hough, E (1998)

Crystallographic analysis of the human phenylalanine hydroxylase

catalytic domain with bound catechol inhibitors at 2.0 A˚

resolu-tion Biochemistry 37, 15638–15646.

18 Erlandsen, H., Bjørgo, E., Flatmark, T & Stevens, R.C (2000)

Crystal structure and site-specific mutagenesis of

pterin-bound human phenylalanine hydroxylase Biochemistry 39, 2208–

2217.

19 Kobe, B., Jennings, I.G., House, C.M., Michell, B.J., Goodwill,

K E , Santarsiero, B D , Stevens, R C , Cotton, R G & Kemp,

B.E (1999) Structural basis of autoregulation of phenylalanine

hydroxylase Nat Struct Biol 6, 442–448.

20 Andersen, O.A., Flatmark, T & Hough, E (2001) High resolution crystal structures of the catalytic domain of human phenylalanine hydroxylase in its catalytically active Fe(II) form and binary complex with tetrahydrobiopterin J Mol Biol 314, 279–291.

21 Andersen, O.A., Flatmark, T & Hough, E (2002) Crystal struc-ture of the ternary complex of the catalytic domain of human phenylalanine hydroxylase with tetrahydrobiopterin and 3-(2-thienyl)- L -alanine, and its implications for the mechanisms of catalysis and substrate activation J Mol Biol 320, 1095–1108.

22 Teigen, K., Frøystein, N.A & Martı´nez, A.(1999) The structural basis of the recognition of phenylalanine and pterin cofactors by phenylalanine hydroxylase: implications for the catalytic mechanism J Mol Biol 294, 807–823.

23.Martı´nez, A., Knappskog, P.M., Olafsdottir, S., Døskeland, A.P., Eiken, H.G., Svebak, R.M., Bozzini, M., Apold, J & Flatmark, T (1995) Expression of recombinant human phenylalanine hydro-xylase as fusion protein in Escherichia coli circumvents proteolytic degradation by host cell proteases.Isolation and characterization

of the wild-type enzyme Biochem J 306, 589–597.

24.Roskoski, R.Jr (1983) Assays of protein kinase.Methods Enzy-mol 99, 3–6.

25.Flatmark, T , Alma˚s, B., Knappskog, P.M., Berge, S.V., Svebak, R.M., Chehin, R., Muga, A & Martı´nez, A.(1999) Tyrosine hydroxylase binds tetrahydrobiopterin cofactor with negative cooperativity, as shown by kinetic analyses and surface plasmon resonance detection Eur J Biochem 262, 840–849.

26 Flatmark, T., Stokka, A.J & Berge, S.V (2001) Use of surface plasmon resonance for real-time measurements of the global conformational transition in human phenylalanine hydroxylase in response to substrate binding and catalytic activation Anal Bio-chem 294, 95–101.

27.Stokka, A J.& Flatmark, T.(2003) Substrate induced con-formational transition in human phenylalanine hydroxylase as studied by surface plasmon resonance analyses.The effect of terminal deletions, substrate analogues and phosphorylation Biochem J 369, 509–518.

28.Jo¨nsson, U., Fa¨gerstam, L., Ivarsson, B., Johnsson, B., Karlsson, R., Lundh, K., Lo¨fas, S , Persson, B , Roos, H , Ro¨nnberg, I., Sjo¨lander, S , Stenberg, E , Sta˚hlberg, R., Urbaniczky, C., O¨stlin, H.& Malmqvist, M.(1991) Real–time biospecific interaction analysis using surface plasmon resonance and a sensor chip technology Bio Techniques 11, 620–627.

29.Phillips, R S , Iwaki, M.& Kaufman, S.(1983) Ligand effects on the limited proteolysis of phenylalanine hydroxylase: evidence for multiple conformational states Biochem Biophys Res Commun.

110, 919–925.

30 Iwaki, M., Phillips, R.S & Kaufman, S (1986) Proteolytic modification of the amino-terminal and carboxyl-terminal regions

of rat hepatic phenylalanine hydroxylase J Biol Chem 261, 2051–2056.

31 Flatmark, T.& Stevens, R.C.(1999) Structural insight into the aromatic amino acid hydroxylases and their disease related mutant forms Chem Rev 99, 2137–2360.

32.Martı´nez, A , Haavik, J.& Flatmark, T.(1990) Cooperative homotropic interaction of 1-noradrenaline with the catalytic site of phenylalanine 4-monooxygenase Eur J Biochem 193, 211–219.

33 Miranda, F.F., Teigen, K., Tho´ro´lfsson, M., Svebak, R.M., Knappskog, P.M., Flatmark, T & Martı´nez, A.(2003) Phos-phorylation and mutations of Ser16 in human phenylalanine hydroxylase.Kinetic and structural effects.J Biol Chem 277, 40937–40943.

34 Carvalho, R.N., Solstad, T., Bjørgo, E., Barroso, J.F & Flatmark, T.(2003) Deamidations in recombinant human phenylalanine hydroxylase.Identification of labile asparagine residues and functional characterization of AsnfiAsp mutant forms J Biol Chem [epub ahead of print on Jan 28].

35 McCulloch, R.I & Fitzpatrick, P.F (1999) Limited proteolysis of

tyrosine hydroxylase identifies residues 33–50 as conformationally

sensitive to phosphorylation state and dopamine binding Arch.

Biochem Biophys 367, 143–145.

36.Flatmark, T.(2000) Catecholamine biosynthesis and physiological

regulation in neuroendocrine cells Acta Physiol Scand 168, 1–17.

37 Knappskog, P.M., Eiken, H.G., Martı´nez, A., Bruland, O.,

Apold, J.& Flatmark, T.(1996) PKU mutation (D143G)

asso-ciated with an apparent high residual enzyme activity: expression

of a kinetic variant form of phenylalanine hydroxylase in three

different systems Hum Mutat 8, 236–246.

38 Bjørgo, E., Knappskog, P.M., Martı´nez, A , Stevens, R C &

Flatmark, T.(1998) Partial characterization and

three-dimen-sional-structural localization of eight mutations in exon 7 of the

human phenylalanine hydroxylase gene associated with

phenyl-ketonuria Eur J Biochem 257, 1–10.

39 Leandro, P , Rivera, I , Lechner, M C , de Almeida, I T & Konecki, D.(2000) The V388M mutation results in a kinetic variant form of phenylalanine hydroxylase Mol Genet Metab 69, 204–212.

40 Jennings, I.G., Cotton, R.G & Kobe, B (2000) Functional ana-lysis, using in vitro mutagenesis, of amino acids located in the phenylalanine hydroxylase active site Arch Biochem Biophys.

384, 238–244.

41 Solomon, E I , Brunold, T C , Davis, M I , Kemsley, J N , Lee, S.-K., Lehnert, N., Neese, F., Skulan, A.J., Yang, Y.-S & Zhou, J (2000) Geometric and electronic structure/function correlations in non-heme iron enzymes Chem Rev 100, 235–350.

42.Kraulis, P J.(1991) MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures J Appl Crys-tallogr 24, 946–950.