Intersubunit Binding DomainsWithin Tyrosine Hydroxylase andTryptophan Hydroxylase

G.J. Yohrling IV,1 G.C.-T. Jiang,2 S.M. Mockus,3 and K.E. Vrana1*1Department of Physiology and Pharmacology, Wake Forest University School of Medicine, Winston-Salem,North Carolina2Molecular Genetics Program, Wake Forest University School of Medicine, Winston-Salem, North Carolina3Neuroscience Program, Wake Forest University School of Medicine, Winston-Salem, North Carolina

Tryptophan hydroxylase (TPH), the rate-limiting enzymein the biosynthesis of the neurotransmitter serotonin(5-HT) belongs to the aromatic amino acid hydroxylasesuperfamily, which includes phenylalanine hydroxylase(PAH) and tyrosine hydroxylase (TH). The crystal struc-tures for both PAH and TH have been reported, but acrystallographic model of TPH remains elusive. For thisreason, we have utilized the information presented in theTH crystal structure in combination with primary se-quence alignments to design point mutations in potentialstructural domains of the TPH protein. Mutation of a THsalt bridge (K170E) was sufficient to alter enzyme mac-romolecular assembly. We found that the disruption ofthe cognate intersubunit dimerization salt bridge (K111–E223) in TPH, however, did not affect the macromolec-ular assembly of TPH. Enzyme peaks representing onlytetramers were observed with size exclusion chromatog-raphy. By contrast, a single-point mutation within thetetramerization domain of TPH (L435A) was sufficient todisrupt the normal homotetrameric assembly of TPH.These studies indicate that, although the proposed saltbridge dimerization interface of TH is conserved in TPH,this hypothetical TPH intersubunit binding domain,K111–E223, is not required for the proper macromolec-ular assembly of the protein. However, leucine 435 withinthe tetramerization domain is necessary for the propermacromolecular assembly of TPH. J. Neurosci. Res. 61:313–320, 2000. © 2000 Wiley-Liss, Inc.

Key words: catecholamines; leucine-zipper; mutagene-sis; salt bridge; serotonin

Tryptophan hydroxylase (TPH; EC 1.14.16.4) is amember of the aromatic amino acid hydroxylase super-family. The other members of this family include tyrosinehydroxylase (TH; EC 1.14.16.2) and phenylalanine hy-droxylase (PAH; EC 1.14.16.1). TPH serves as the rate-limiting enzyme in the biosynthesis of the neurotransmit-ter serotonin. Serotonin (5-HT) is an importantmonoamine that has been implicated in a wide variety ofcentral nervous system functions, including temperature

control, aggression, pain, and memory (Kandasamy andWilliams, 1984; Cornwell-Jones et al., 1989; for reviewsee Mockus and Vrana, 1998). Aberrations in 5-HT syn-thesis and regulation have been linked to a diverse groupof neurological and psychiatric disorders such as Parkin-son’s disease, Gilles de la Tourette’s syndrome, multiplesclerosis, depression, and obsessive compulsive disorder(Volicer et al., 1985; Schauenburg and Dressler, 1992;Tabaddor et al., 1978; Mann et al., 1997; for review seeDelgado and Moreno, 1998).

A potential target in the etiology of many of thesediseases may be TPH. However, research on TPH hasfallen behind that for the other aromatic amino acid hy-droxylases because of its extreme instability and the inabil-ity to obtain large quantities of pure protein (Kuhn et al.,1980; D’Sa et al., 1996a,b; Moran et al., 1998). For thisreason, it has become important to use the information wehave regarding the other related hydroxylases to aid inelucidating the structure and regulation of TPH. Therecent crystallizations of both PAH and TH have yieldedmuch of this information (PAH: Erlandsen et al., 1997;Kobe et al., 1997, 1999; Fusetti et al., 1998; TH: Goodwillet al., 1997, 1998).

It has long been known that point mutations within thePAH gene lead to insufficient PAH activity, thus leading tothe metabolic imbalances and mental retardation characteris-tic of phenylketonuria (PKU). Recently, point mutations inthe catalytic domain of TH have been linked to Segawa’ssyndrome and L-dopa-responsive Parkinsonism (Ludecke etal., 1995, 1996; Knappskog et al., 1995). In addition, anintronic polymorphism in the human TPH gene has beenidentified (Nielsen et al., 1992) that is associated with suicide

S.M. Mockus’s present address is Department of Pharmacology, Universityof Washington, Seattle, WA 98195-7280.

*Correspondence to: Kent E. Vrana, Ph.D., Department of Physiology andPharmacology, Wake Forest University School of Medicine, Medical Cen-ter Blvd., Winston-Salem, NC 27157-1083.E-mail: kvrana@wfubmc.edu

Received 16 February 2000; Revised 20 April 2000; Accepted 21 April2000

Journal of Neuroscience Research 61:313–320 (2000)

© 2000 Wiley-Liss, Inc.

and variable 5-hydroxyindolacetic acid concentrations inFinnish Caucasians (Nielsen et al., 1994) and depressed pa-tients (Mann et al., 1997). It seems probable that furtherinvestigation of TPH will implicate point mutations in crit-ical regulatory and structural domains of the protein in dis-orders associated with defective serotonergic function.

Recently, this laboratory has identified intersubunitbinding domains in both the regulatory and the catalyticdomains of the rabbit TPH protein (Mockus et al., 1997a;Yohrling et al., 1999) and the rat TH protein (Walker etal., 1994; Vrana et al., 1994b; Yohrling and Vrana, un-published results). These interactions are thought to behydrophobic interactions that are necessary for the propermacromolecular assembly of TPH and TH. It is hypoth-esized that there are additional structural domains withinTPH that work in tandem with those already identified, toallow for the regulation of the enzyme.

Goodwill and colleagues (1997) identified a saltbridge and surrounding hydrogen bonds in TH (Lys 170–Glu 282) as a dimerization interface between monomers Iand II of the tetrameric protein. Sequence alignments haveshown this salt bridge to be conserved in rabbit TPH (Fig.1A,B). A goal of this study was to test the hypothesis thatthis conserved salt bridge interaction (Lys 111–Glu 223) isinvolved in the structure and function of the TPH protein.Additionally, we wished to determine the contribution ofindividual leucine residues within the C-terminal tet-ramerization domain to the macromolecular assembly ofTPH (Fig. 1C; Vrana et al., 1994b; Mockus et al., 1997a).It was hypothesized that disruption of the stronger, hy-drophobic interactions between tetramerization domainswill result in a higher degree of structural dissociationcompared to the disruption of the proposed ionic inter-action. The results suggest that hydrophobic interactionswithin the C-terminus play a more important role than theionic interaction (K111–E223) in the subunit assembly ofTPH.

MATERIALS AND METHODS

All materials were obtained from the Sigma Chemical Co.(St. Louis, MO) with the following exceptions; restriction andDNA modifying enzymes were supplied by Promega Corpora-tion (Madison, WI) and New England Biolabs (NEB; Beverly,MA); L-[5-3H] tryptophan and L-[3,5-3H] tyrosine were fromDu Pont NEN Research Products (Boston, MA); and isopropylthio-b-D-galactoside (IPTG), ampicillin, and activated charcoal(Darco G-60) were from Fisher Scientific (Pittsburgh, PA). ThepET-21c vector and BL21[DE3](F2 OmpTrB

2mB2)-

competent cells were acquired from Novagen Corporation(Madison, WI). Reinforced nitrocellulose (Duralose-UV) andthe QuikChange Site-Directed Mutagenesis kit were purchasedfrom Stratagene (LaJolla, CA). Prestained protein ladder waspurchased from Life Technologies (Gaithersburg, MD). Chemi-luminescent reagents (Renaissance) were purchased from NENResearch Products (Boston, MA). The mouse anti-TPH mono-clonal antibody (WH3) was obtained from RBI (Natick, MA).The rabbit anti-TH monoclonal antibody was obtained fromPel-Freez Biologicals (Rogers, AK). The sheep anti-mouse IgGHRP-coupled and rabbit anti-donkey IgG HRP-coupled sec-

ondary antibodies were purchased from Amersham Life Sciences(Arlington Heights, IL). The Sephacryl 200 high-resolutionsize-exclusion chromatography resin was purchased from Phar-macia Biotech (Upsala, Sweden). Finally, all sequencing wasperformed by the DNA Sequencing and Gene Analysis Facilityof the Molecular Genetics Program (Wake Forest UniversitySchool of Medicine) using a Perkin Elmer/Applied Biosystems377 Prism automated DNA sequencer.

Construction of the TH and TPH Point Mutations

The QuikChange Site-Directed Mutagenesis kit fromStratagene was utilized to create point mutations to disruptpotential dimerization interfaces in both TH and TPH and theleucine zipper tetramerization domain within the carboxyl ter-minus of TPH. Synthetic oligonucleotide primers (sense andantisense) were synthesized to alter the appropriate amino acids.The specific primers used are presented in Table I.

The point mutants were generated in a Perkin ElmerDNA Thermal Cycler (Gene Amp PCR System 2400) using25 ng of rabbit TPH DNA or rat TH DNA, 125 ng of eachprimer, 25 mM of each dNTP, and 13 reaction buffer (con-

Fig. 1. Sequence alignments (single-letter amino acid code) of therabbit TPH (Grennet et al., 1987; SWISS-PROT accession P17290),rat TH (Grima et al., 1985; SWISS-PROT accession P04177), andhuman PAH (Kwok et al., 1985; SWISS-PROT accession P00439).A,B: Salt bridge residues that contribute to a hypothetical dimerizationinterface within TH. The corresponding residues in TPH (K111 andE223) are indicated above the alignments. The boldface residues indi-cate 100% conservation between the three proteins surrounding thetheoretical dimerization interface. The VPWFP amino acid sequence isconserved in all amino acid hydroxylases and delineates the boundarybetween the regulatory and catalytic domain (Grenett et al., 1987;Daubner et al., 1993). C: Sequence alignments of TPH, TH, and PAHshowing the carboxy terminal tetramerization domain (leucine zipper).The leucine zipper motif is completely conserved between TH andTPH. The italized Ls indicate leucine residues 435 and 442, whichwere mutated to alanines (A).

314 Yohrling et al.

taining 100 mM KCl, 100 mM (NH4)2SO4, 200 mM Tris-HCl(pH 8.8), 20 mM MgSO4, 1% Triton X-100, and 1 mg/mlnuclease-free bovine serum albumin). Following the addition of2.5 U of Pfu turbo DNA polymerase, the reaction was subjectedto a 30 sec denaturing step of 95°C, followed by 16 cycles of95°C denaturation for 30 sec, 55°C annealing for 1 min, and68°C extension for 12 min 36 sec. After the cycling, sampleswere subjected to a final extension step at 68°C for 15 min. Thetemplate was digested with 10 U of DpnI (specific for themethylated parental DNA) for 1 hr at 37°C and then used totransform Epicurian Coli XL-1 Blue Supercompetent cells(Stratagene, LaJolla, CA). DNA from XL-1 colonies was puri-fied and digested with NdeI and BamHI to identify clonescontaining the TPH or TH insert. After sequence analysis con-firmed positive point mutations (and eliminated extraneous ar-tifactual mutations), the clones were then transformed intoBL21[DE3] cells for expression of the recombinant protein.

Bacterial Expression of Rabbit TPH and Rat TH

A full-length cDNA clone for rabbit TPH (Grenett et al.,1987; American Type Culture Collection, Laurel, MD) and afull-length cDNA clone for rat TH (Grima et al., 1985) havepreviously been developed for prokaryotic expression (TPH:Vrana et al., 1994a; Kumer et al., 1997; Mockus et al., 1997a,b;Yohrling et al., 1999; TH: Vrana et al., 1994b; Walker et al.,1994). All recombinant proteins were expressed using the ap-proach detailed by Yohrling et al. (1999).

TH and TPH Protein Analysis and Enzyme Activity

Denaturing polyacrylamide gel electrophoresis (Laemmli,1970; 10% acrylamide, 0.27% bis- acrylamide) was performed

on equal total protein aliquots (25 mg) of high-speed supernatantfor seven representative mutant TPH proteins and two THproteins (Fig. 2). The proteins were transferred to Duralose-UV(Stratagene) membrane using a semidry electroblot apparatus setat 400 mA for 1.5 hr (Owl Scientific, Cambridge, MA). Bench-mark prestained protein ladder was resolved with the samples(Life Technologies, Gaithersburg, MD). After transfer, TPHproteins were detected by probing with a 1:1,000 dilution ofaffinity-purified mouse anti-TPH monoclonal antibody (WH3;RBI, Natick, MA) followed by a 1:1,500 dilution (1.5 mg/ml)of sheep anti-mouse IgG coupled with HRP. The mouse

TABLE I Site-Directed Mutagenesis Primers*

Construct Sequencea

K111A senseb 59-CCTTGGTTTCCAAAGGCGCGATAATCAGACCTGGACCATTGTGC-39K111A antisense 59-GCACAATGGTCCAGGTCTGATTATCGCGCCTTTGGAAACCAAGG-39K111E sense 59-CCTTGGTTTCCAAAGGGAGATTTCAGACCTGGAC-39K111E antisense 59-GTCCAGGTCTGAAATCTCCCTTTGGAAACCAAGG-39K111R sense 59-CCTTGGTTTCCAAAGCCGCGGATTTCAGACCTGGAC-39K111R antisense 59-GTCCAGGTCTGAAATCCGCGCTTTGGAAACCAAGG-39E223A sensec 59-GATAATTCAAACTTTTTGGAAAGCCGCGCACAGGTTTTTCC-39E223A antisense 59-GGAAAAACCTGTGCGCGGCTTTCCAAAAAGTTTGAAATATC-39E223K sense 59-CAAACTTTTTAAAAAAAAGCGCACAGGTTTTTCC-39E223K antisense 59-GGAAAAACCTGTGCGCTTTTTTTTAAAAAGTTTG-39E223L sense 59-CAAACTTTTTAAAAACTGCGCACAGGTTTTTCC-39E223L antisense 59-GGAAAAACCTGTGCGCAGTTTTTAAAAAGTTTG-39L435A sensed 59-GTTGTCAGCGACGCCGCGCTGGGAAGGTCAGCAGG-39L435A antisense 59-CCTGCTGACCTTCCCAGCGCGGCGTCGCTGACAAC-39L442A sense 59-GGAAGGTCAGCAGGCAGGCGCGAGTGTCTGACGGCTGCC-39L442A antisense 59-GGCAGCCGTCAGACACTCGCGCCCTGCCTGCTGACCTTCC-39TH K170E sense 59-CCCTGGTTCCCAAGAGGAAGTGTCGGAATTGGAC-39TH K170E antisense 59-GTCCAATTCCGACACTTCTCTTGGGAACCAGGG-39

* Tryptophan hydroxylase (TPH) and tyrosine hydroxylase (TH) site-directed mutagenesis primers. Primers weremade use of with the Quik-Change Mutagenesis Kit from Stratagene (La Jolla, CA).a The altered nucleotides are presented in boldface.b The K111A primers created an EcoRV site for rapid identification of positive clones.c The E223A primers destroy a DraI site to allow for rapid identification of positive clones.d L435A was created first, sequenced, and then used as a template in the mutagenesis reaction with the L442Aprimers to construct the double-point mutation, L435A/L442A within the tetramerization domain of TPH.

Fig. 2. Western analysis of recombinant TPH and TH proteins. Left:Wild-type TPH, K111A, K111E, E223A, E223K, E223L, and L435A.TPH proteins were detected with a monoclonal antibody (WH3) forTPH as outlined in Materials and Methods. All TPH proteins wereobserved to migrate at the appropriate molecular weight of 51 kDa.Right: Wild-type TH and TH K170E. TH proteins were detectedwith an anti-TH monoclonal antibody from rabbit. TH and its mutantmigrated at the predicted molecular weight of 61 kDa.

Mutagenesis of TH and TPH Binding Domains 315

monoclonal antibody recognizes an epitope between residues103 and 109. Similarly, TH protein was detected with a 1:1,000dilution of rabbit anti-TH monoclonal antibody (Pel-Freez Bio-logicals, Rogers, AK) followed by a 1:1,500 dilution (1.5 mg/ml) of donkey anti-rabbit IgG-HRP secondary antibody. Theincubation and wash conditions for western analysis were asoutlined by Mockus et al. (1997b). Immune complexes werevisualized using enhanced chemiluminescence (NEN ResearchProducts, Boston, MA) and exposed to X-ray film (KodakBiomax MR2).

TPH activity was determined using a radioenzymatic 3H2Orelease assay as previously described (Beevers et al., 1983; Vrana etal., 1993). The only significant variation to this method was that afinal concentration of 50 mM BH4 was utilized to minimize sub-strate inhibition of TPH. TH activity was determined at 100 mMtyrosine and 50 mM BH4 as described by Vrana et al. (1993) usingthe method of Reinhard et al. (1986). Activity values derived fromeach assay are expressed as nmol hr21 ml21 in the size-exclusionchromatography elution profiles.

Analysis of TPH and TH Quaternary Structure

Wild-type TPH, wild-type TH, and each of their pointmutants were subjected to size-exclusion chromatography on aSephacryl-200-HR gel filtration column (10 mm i.d. 3 40 cm;Pharmacia Biotech, Upsala, Sweden). Bacterial supernatants(330 ml) were applied directly to the S-200-HR column and theproteins eluted at a rate of 330 ml min21 using a peristaltic pumpwith buffer containing 10 mM Tris-HCl, 200 mM NaCl, 1 mMDTT, and 25 mM tryptophan or tyrosine at pH 7.5. One minutefractions were collected and 25 ml and were then assayed forTPH or TH activity. Elution profiles from individual experi-ments (n 5 3) were averaged to yield composite plots.

The S-200-HR column was calibrated with pure proteinsof known molecular weight as outlined by Yohrling et al.(1999). A standard curve was therefore generated that permittedestimation of the molecular weight of the recombinant proteins(data not shown).

RESULTSTPH and TH Western Analysis

The expression of wild-type TPH and the mutantproteins is depicted in Figure 2 (left panel). Westernanalysis confirmed that all proteins migrated consistentwith their predicted molecular weights of approximately51 kDa. For each recombinant protein, 25 mg of totalprotein (as determined by a Bradford protein assay) wasapplied to the denaturing gel. Also displayed in Figure 2(right panel) is the protein expression of wild-type TH andthe mutated salt bridge construct K170E. Both migratedwith a molecular weight of 61 kDa.

Macromolecular Structure DeterminationsUpon expressing the recombinant proteins in the

pET-21c system, high-speed homogenates were subjectedto nondenaturing size-exclusion chromatography. ASephacryl-200-HR column was utilized because it is ap-propriate for resolving proteins between 5 and 250 kDa.Therefore, tetramers, dimers, and monomers of TPH andTH will all fall within the resolution range for the column.

In Figure 3, the average elution profiles for wild-typeTH and its corresponding salt bridge mutation K170E arepresented. As predicted, wild-type TH migrated as a ho-motetramer with an apparent molecular weight of 206 kDa.The mutant K170E displayed a shift to the right in the elutionprofile for the recombinant protein and displayed an ob-served molecular weight of 172 kDa. This is in the range ofa dimer of TH possessing an extended axial ratio (oval shape)as it migrated through the SEC column. The differences inTH activity reflect variations in protein expression by thebacteria (data not shown).

The average elution profiles for all active TPH pro-teins are depicted in Figure 4. All profiles represent thecomposite pattern for three independent protein induc-tions. In Figure 4A, the SEC profiles for wild-type TPH,K111E, and E223K are shown. These amino acid substi-tutions should place two similarly charged amino acids(within separate subunits) in juxtaposition to each otherand thereby disrupt the hypothetical salt bridge (and poten-tially tetramer assembly). However, each protein migrated asa tetramer with molecular weights of 206–214 kDa. Figure4B compares the observed molecular weight of wild-typeTPH (206 kDa) to four additional mutations at the proposeddimerization interface (K111A, K111R, E223A, and E223L).Once again, these mutations had no effect on the macromo-lecular structure of TPH; all recombinant enzymes continued

Fig. 3. Size-exclusion chromatography elution profiles for TH and itssalt bridge mutant (K170E). TH activity is represented in the unitsnmol hr21 ml21 on the abscissa. Each fraction collected and assayed isrepresented as a Ve/Vo ratio on the ordinate. The average startingactivity loaded onto the column for the mutants is as follows (in nmolhr21 mg total protein21): wt TH 44.9 6 11.1; TH K170E 17.8 6 6.1.The left y-axis corresponds to wild-type TH activity and the righty-axis corresponds to K170E activity. Differences in activity reflectdifferences in protein expression by the bacteria (data not shown). Peakfractions of standard molecular weight markers are indicated [bA,b-amylase (200,000 Da); ADH, alcohol dehydrogenase (150,000 Da);Hb, hemoglobin (66,500 Da)].

316 Yohrling et al.

to migrate like wild-type TPH. Interestingly, all of the mu-tations maintained robust enzyme activity.

Figure 4C illustrates the elution profiles for wild-type TPH and the leucine zipper mutations. Disruption ofthe hydrophobic interactions within the tetramerizationdomain of TPH with the mutant L435A resulted in aheterogeneous mixture of tetramers, dimers, and mono-mers. In the case of L435A, approximately 50% of the totalactivity displayed an elution profile characteristic of eithera dimer or a monomer of TPH. Specifically, this value wasdetermined by summing the activity for L435A in thefractions representing tetramer, dimer, and monomer(Ve/Vo 1.00–1.63) and dividing that value into the sumof TPH activity for dimers and monomers (Ve/Vo 1.20–1.63). This Ve/Vo range correlates to molecular weightsof 47–172 kDa according to the standard curve for theS-200-HR column. The activity in the final two fractions

was not incorporated into this calculation because theMW values fall outside the range of a monomer. A sum-mary of all site-directed mutagenesis and the correspond-ing effects on macromolecular assembly for both TPH andTH are presented in Table II.

DISCUSSIONThe recent crystallizations of TH and PAH provide

valuable information regarding the subunit/subunit assem-bly of these two aromatic amino acid hydroxylases (TH:Goodwill et al., 1997, 1998; PAH: Erlandsen et al., 1997;Kobe et al., 1997, 1999; Fusetti et al., 1998). However,the instability and thermooxidative qualities of TPH haveforced research on this important enzyme to lag behindthat on TH and PAH. For this reason, information re-garding the structure and regulation of TH and PAH wasutilized to infer structural information concerning TPH.

Fig. 4. Size-exclusion chromatography elution profiles. A: Compositeelution profiles for wt TPH, K111E, and E223K. TPH activity isrepresented in the units nmol hr21 ml21 on the abscissa. Each fractioncollected and assayed is represented as a Ve/Vo ratio on the ordinate.The average activity applied to the column for each of the mutants wasas follows (in nmol hr21 mg total protein21): wt TPH 13.7 6 4.2;K111E 10 6 1.0; E223K 7.8 6 1.5. B: Composite elution profiles forwtTPH, K111A, K111R, E223A, and E223L. The average activityloaded onto the column for the mutants is as follows (in nmol hr21 mgtotal protein21): wtTPH 13.7 6 4.2; K111A 7.4 6 0.7; K111R 10.6 61.6; E223A 6.7 6 0.5; E223L 6.6 6 1.0. C: Composite elution profilesfor wt TPH and the mutants of the tetramerization domain, L435A,L442A, and L435A/L442A. The average activity loaded onto thecolumn for the mutants is as follows (in nmol hr21 mg total protein21):L435A 12.5 6 2.7; L442A 17.0 6 1.1; L435A/L442A 11.4 6 3.2. Peakfractions of standard molecular weight markers are indicated [bA,b-amylase (200,000 Da); ADH, alcohol dehydrogenase (150,000 Da);Hb, hemoglobin (66,500 Da)].

Mutagenesis of TH and TPH Binding Domains 317

The recent crystallization of the catalytic domain ofTH confirmed the theory first postulated by Liu and Vrana(1991), that a leucine zipper in the carboxyl terminal endof TH is required for the tetramerization of the protein(Goodwill et al., 1997). The TH crystal structure alsorevealed what was termed a dimerization interface. Thisinterface is composed of an ionic salt bridge interactionbetween lysine 170 and glutamate 282 on separate subunitsof TH and neighboring hydrogen bonding interactions.This salt bridge is conserved in TPH, but not PAH (Fig.1A,B). Mutagenesis primers were designed to disrupt thispotential dimerization interface in both TPH and TH. Inaddition, site-directed mutagenesis was utilized to assessthe contribution of individual hydrophobic residues(leucines 435 and 442) within the leucine zipper (tet-ramerization domain) of TPH (Fig. 1C).

In the present studies, the hypothesis was tested thata conserved, ionic salt bridge in TPH (K111 and E223) isnecessary for proper macromolecular assembly. This saltbridge in TH (K170 and E282) has been hypothesized toserve as a dimerization interface between two monomersof the TH protein (Goodwill et al., 1997). As was pre-dicted by Goodwill and collegues, disruption of this in-teraction in TH significantly alters the macromolecularassembly of TH. A shift from tetramer to dimer formationis observed in TH following mutagenesis of K170. Theobserved molecular weight for TH K170E (172 kDa) isslightly larger than would be theoretically expected for adimer of TH. It is believed that this increase can beaccounted for in the fact that a K170E dimer may have anextended axial ratio. This could result in an increasedvolume of distribution caused by two subunits of THassuming an ellipsoid shape. A similar phenomenon wasobserved for dimers of TPH and TH, using SEC (TPH:Yohrling et al., 1999; TH: Vrana et al., 1994b). Themutants (TPH V28R-L31R and TH CD19) both exhib-ited a diffuse band peak suggesting molecular heterogeni-ety. The dimeric TH CD19 had an observed molecular

weight of 152 kDa, which is similar to that reported forTH K170E in this report (172 kDa).

To address whether this dimerization domain was func-tional in TPH, site-directed mutagenesis was utilized to dis-rupt the hypothesized salt bridge domain (K111–E223). Eachresidue was converted to amino acids with a variety ofdifferent properties. Regardless of the type of mutation, tet-ramer formation was maintained. A small, unexplainable shiftin the elution profiles was observed in two TPH salt bridgemutants tested (K111A and E223A). This shift is translatedinto a smaller molecular weight tetramer (197 kDa) than thatreported for wild-type TPH (206 kDa). Whereas we believethis change to be nonsignificant, the effect may be due inlarge part to a partial collapse of the native TPH structurefollowing the addition of the various point mutations. It isour conclusion that the proposed salt bridge interaction,though essential for TH macromolecular assembly, does notplay a significant role in maintaining the quaternary structureof TPH. It may be true that the mutations in TPH did disruptthe dimerization interface, but stronger hydrophobic inter-actions (e.g., C-terminal tetramerization domain) are work-ing to maintain tetrameric assembly in TPH. However, if thisis true, it would argue that TPH is a tetramer of monomers.This is in opposition to what has been reported for TH(Goodwill et al., 1997), and the experiments presented in thisreport cannot distinguish between these possibilities.

All aromatic amino acid hydroxylases possess a 4,3-hydrophobic alpha helix on the carboxy terminal end ofthe protein. This helix assumes a coiled-coil formationwith helices from other subunits to allow for tetrameriza-tion of TH and TPH (Liu and Vrana, 1991; Lohse andFitzpatrick, 1993; Vrana et al., 1994; Mockus et al., 1997;Goodwill et al., 1997). In TPH and TH, a leucine heptadrepeat (leucine zipper) is found within the 4,3-hydrophobic repeat. Although the leucine zipper region isnot 100% conserved in PAH, it has been demonstratedthat the hydrophobic motif is required for the formation oftetramers in PAH (Knappskog et al., 1996; Hufton et al.,1998). Hufton and collegues showed that the conversionof L448 in PAH to an alanine (L448A) created a proteinwith a tetramer to dimer ratio of 19:81. Likewise, for THit has been reported that the mutation L480A was suffi-cient in converting TH to a dimer (Vrana et al., 1994b).Recently, mutations of hydrophobic residues within aproposed 4,3-hydrophobic alpha helix in the regulatorydomain of TPH were demonstrated to convert TPH to adimeric species of 127 kDa (Yohrling et al., 1999). Acomparison of the hydrophobic mutation results aboveand the salt bridge mutational analyses presented in thisreport are in agreement with previous observations sug-gesting that individual ionic interactions may be less sta-bilizing to protein structure than their equivalent hydro-phobic interactions (for review see Sindelar et al., 1998).In addition, Kobe et al. (1999) have recently shown thatthe dimerization interface for PAH consists of 26 polarcontacts between the regulatory and catalytic domains oftwo PAH subunits. If we assume a similar interaction to betrue for TPH, it is logical that disruption of one polar

TABLE II Summary of Nondenaturing Size-ExclusionChromatography

Sample Observed MWa (kDa) Molecular Structure

wt TPH 206 TetramerTPH V28R-L31Rb 127 DimerTPH L435A 47–197 Tetramer, dimer, monomerTPH K111A 197 TetramerTPH E223A 197 TetramerTPH E223L 197 TetramerTPH E223K 214 Tetramerwt TH 206 TetramerTH K170E 172 DimerTH L480Ac 150 DimerTH CD19c 152 Dimera MW determined from peak of average elution profiles (n 5 3) for eachprotein.b Construct from Yohrling et al. (1999).c Mutants created in Vrana et al. (1994b). In a sequence alignment of rat THwith rabbit TPH (Fig. 1C), L480 of TH corresponds to L428 of TPH.

318 Yohrling et al.

contact, within the larger context of 26 contacts, wouldresult in little structural disorganization of TPH. However,when one residue in this region was mutated in TH(K170E), it disrupted TH structure. Therefore, in spite ofthe significant conservation of tertiary structure betweenthe three enzymes, subtle nuances in the interface domains(e.g., the disparate lengths of the regulatory domains forthe three AAAH) may account for these observed differ-ences.

Using site-directed mutagenesis, the present studiesdemonstrated that an individual leucine residue (L435) inthe tetramerization domain of TPH plays an importantrole in proper subunit assembly. The L435A mutant wassufficient to disrupt normal tetrameric assembly. Hetero-geneous elution patterns consistent with a mixture oftetramers, dimers, and monomers were observed. Similarto that observed for K111A and E223A, rightward shift inthe elution profile for L435A was observed compared towild-type TPH (Fig. 4C). In quantifying the structuraldisruption of L435A, it was calculated that .50% of theenzymatically active protein for L435A resided in thedimer-monomer range for TPH (see Results). To disruptmacromolecular assembly further, the mutation L435A/L442A was created. However, this mutation failed toexacerbate the disruption observed in the single-pointmutation L435A (Fig. 4C).

In conclusion, these data and previous studies dem-onstrate that hydroxylase tetramer formation relies heavilyon hydrophobic interactions within/between amino andcarboxyl termini (Vrana et al., 1994b; Mockus et al., 1997;Hufton et al., 1998; Yohrling et al., 1999). The crystalstructure of TH has identified K170–E282 as an intersub-unit binding domain between two monomers of TH. Aswas predicted from the crystal structure for the TH cata-lytic domain, mutagenesis of this site was sufficient todisrupt tetrameric assembly of TH. The cognate intersub-unit salt bridge (K111–E223) is not required for tetramer-ization of TPH. It can also be concluded from this studythat an individual residue within the carboxyl terminalleucine zipper of TPH (L435A), and its correspondinghydrophobic interactions with other TPH subunits, isessential for the formation of homogenous tetramers. Thislatter finding is in total agreement with previous studies ofPAH and TH and provides further support for the uni-versal nature of the carboxy terminal hydrophobic repeatin aromatic amino acid hydroxylase structure.

ACKNOWLEDGMENTSThe authors thank Dr. S. John Mihic, Wake Forest

University, for his technical assistance with the site-directed mutagenesis studies. This work was supported byPHS grants RO1 GM-38931 (to K.E.V.) and T32-AA07565 (to S.M.M.).

REFERENCESBeevers SJ, Knowles RG, Pogson CI. 1983. A sensitive radiometric assay

for tryptophan hydroxylase applicable to crude extracts. J Neurochem40:894–897.

Cornwell-Jones CA, Decker MW, Chang JW, Cole B, Goltz KM, Tran T,McGaugh JL. 1989. Neonatal 6-hydroxydopa, but not DSP-4, elevatesbrainstem monoamines and impairs inhibitory avoidance learning in de-veloping rats. Brain Res 493:258–268.

Daubner SC, Lohse DL, Fitzpatrick PF. 1993. Expression and character-ization of catalytic and regulatory domains of rat tyrosine hydroxylase.Protein Sci 2:1452–1460.

Delgado PL, Moreno FA. 1998. Hallucinogens, serotonin and obsessivecompulsive disorder. J Psychoact Drugs 30:359–366.

D’Sa CM, Arthur RE Jr, Jennings I, Cotton RGH, Kuhn DM. 1996a.Tryptophan hydroxylase: Purification by affinity chromatography oncalmodulin-Sepharose. J Neurosci Methods 69:149–153.

D’Sa CM, Arthur RE Jr, Kuhn DM. 1996b. Expression and deletionmutagenesis of tryptophan hydroxylase fusion proteins: delineation of theenzyme catalytic core. J Neurochem 67:917–926.

Erlandsen H, Martinez A, Knappskog PM, Haavik J, Hough E, Flatmark T.1997. Crystallization and preliminary diffraction analysis of a truncatedhomodimer of human phenylalanine hydroxylase. FEBS Lett 406:171–174.

Fusetti F, Erlandsen H, Flatmark T, Stevens RC 1998. Structure of tet-rameric phenylalanine hydroxylase and its implications for phenylketo-nuria. J Biol Chem 273:16962–16967.

Goodwill KE, Sabatier C, Marks C, Raag R, Fitzpatrick PF, Stevens RC.1997. Crystal structure of tyrosine hydroxylase at 2.3 Å and its implicationfor inherited neurodegenerative diseases. Nature Struct Biol 4:578–585.

Goodwill KE, Sabatier C, Stevens RC. 1998. Crystal structure of tyrosinehydroxylase with bound cofactor analogue and iron at 2.3 Å resolution:self-hydroxylation of Phe300 and the pterin- binding site. Biochem37:13437–13445.

Grenett HE, Ledley FD, Reed LL, Woo SLC. 1987. Full-length cDNA forrabbit tryptophan hydroxylase: functional domains and evolution of aro-matic amino acid hydroxylases. Proc Natl Acad Sci USA 84:5503–5534.

Grima B, Lamouroux A, Blanot F, Biguet NF, Mallet J. 1985. Completecoding sequence of rat tyrosine hydroxylase cDNA. Proc Natl Acad SciUSA 82:617–621.

Hufton SE, Jennings IG, Cotton RGH. 1998. Structure/function analysisof the domains required for the multimerisation of phenylalanine hydrox-ylase. Biochim Biophys Acta 1382:295–304.

Kandasamy SB, Williams BA. 1984. Hypothermic and anti-pyretic effectsof ACTH (1–24) and alpha-melanotropin in guinea pigs. Neuropharma-cology 23:49–53.

Knappskog PM, Flatmark T, Mallet J, Ludecke B, Bartholome K. 1995.Recessively inherited L-dopa-responsive dystonia caused by a point mu-tation (Q381K) in the tyrosine hydroxylase gene. Hum Mol Genet4:1209–1212.

Knappskog PM, Flatmark T, Aarden JM, Haavik J, Martinez A. 1996.Structure/function relationships in human phenylalanine hydroxylase.Effect of terminal deletions on the oligomerization, activation and coop-erativity of substrate binding to the enzyme. Eur J Biochem 242:813–821.

Kobe B, Jennings IG, House CM, Feil SC, Michell BJ, Tiganis T, ParkerMW, Cotton RGH, Kemp BE. 1997. Regulation and crystallization ofphosphorylated and dephosphorylated forms of truncated dimeric phe-nylalanine hydroxylase. Protein Sci 6:1352–1357.

Kobe B, Jennings IG, House CM, Michell BJ, Goodwill KE, SantarsieroBD, Stevens RC, Cotton RGH, Kemp BE. 1999. Structural basis of autoregulation of phenylalanine hydroxylase. Nature Struct Biol 6:442–448.

Kuhn DM, Ruskin B, Lovenberg W. 1980. Tryptophan hydroxylase: therole of oxygen, iron, and sulfhydryl groups as determinants of stability andcatalytic activity. J Biol Chem 255:4137–4143.

Kumer SC, Mockus SM, Rucker PJ, Vrana KE. 1997. Amino terminaldeletion analysis of tryptophan hydroxylase: PKA phosphorylation occursat serine-58. J Neurochem 69:1738–1745.

Mutagenesis of TH and TPH Binding Domains 319

Kwok SC, Ledley FD, DiLella AG, Robson KJ, Woo SL 1985. Nucleotidesequence of a full-length complimentary DNA clone and amino acidsequence of human phenylalanine hydroxylase. Biochemistry 24:556–561.

Laemmli UK. 1970. Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227:680–685.

Liu X, Vrana KE. 1991. Leucine zippers and coiled-coils in the aromaticamino acid hydroxylases. Neurochem Int 18:27–31.

Lohse DL, Fitzpatrick PF. 1993. Identification of the intersubunit bindingregion in rat tyrosine hydroxylase. Biochem Biophys Res Commun197:1543–1548.

Ludecke B, Dworniczak B, Bartholome K. 1995. A point mutation in thetyrosine hydroxylase gene associated with Segawa’s syndrome. HumGenet 95:123–125.

Ludecke B, Knappskog PM, Clayton PT, Surtees RAH, Clelland JD,Heales,SJR, Brand MP, Bartholome K, Flatmark T. 1996. Hum MolGenet 5:1023–1028.

Mann JJ, Malone KM, Nielsen DA, Goldman D, Erdos J, Gelernter J. 1997.Possible association of the tryptophan hydroxylase gene and suicidalbehavior in depressed patients. Am J Psychol 154:1451–1453.

Mockus SM, Kumer SC, Vrana KE. 1997a. Carboxyl terminal deletionanalysis of tryptophan hydroxylase. Biochim Biophys Acta 1342:132–140.

Mockus SM, Kumer SC, Vrana KE. 1997b. A chimeric tyrosine/tryptophan hydroxylase: the tyrosine hydroxylase regulatory domainserves to stabilize enzyme activity. J Mol Neurosci 9:35–48.

Mockus SM, Vrana KE. 1998. Advances in the molecular characterizationof tryptophan hydroxylase. J Mol Neurosci 10:163–179.

Moran GR, Daubner SC, Fitzpatrick PL. 1998. Expression and character-ization of the catalytic core of tryptophan hydroxylase. J Biol Chem273:12259–12266.

Nielsen DA, Dean M, Goldman D. 1992. Genetic mapping of the humantryptophan hydroxylase gene on chromosome 11, using and intronicconformational polymorphism. Am J Hum Genet 51:1366–1371.

Nielsen DA, Goldman D, Virkkunen M, Tokola R, Rawlings R, LinnoilaM. 1994. Suicidality and 5-hydroxyindoleacetic acid concentration asso-ciated with a tryptophan hydroxylase polymorphism. Arch Gen Psychiatr51:34–38.

Reinhard JF, Smith GK, Nichol CA. 1986. A rapid and sensitive assay fortyrosine-3- monooxygenase based upon the release of 3H2O and adsorp-tion of [3H]-tyrosine by charcoal. Life Sci 39:2185–2189.

Schauenburg H and Dressler D. 1992. Gilles de la Tourette Syndrome.Nervearzt 63:453–461.

Sindelar CV, Hendsch ZS, Tidor B. 1998. Effects of salt bridges on proteinstructure and design. Protein Sci 7:1898–1914.

Tabaddor K, Wolfson LI, Sharpless NS. 1978. Ventricular fluid ho-movanillic acid and 5- hydroxyindolacetic acid concentrations in patientswith movement disorders. Neurology 28:1249–1253.

Volicer L, Direnfeld LK, Freedman M, Albert M, Langlias PJ, Bird ED.1985. Serotonin and 5- hydroxyindolacetic acid in CSF. Difference inParkinson’s disease and dementia of the Alzheimer’s type. Arch Neurol42:127–29.

Vrana KE, Rucker P, Kumer SC. 1994a. Recombinant rabbit tryptophanhydroxylase is a substrate cAMP-dependent protein kinase. Life Sci 55:1045–1052.

Vrana KE, Walker SJ, Rucker P, Liu X. 1994b. A carboxyl terminal leucinezipper is required for tyrosine hydroxylase tetramer formation. J Neuro-chem 63:2014–2020.

Vrana SL, Dworkin SI, Vrana KE. 1993. Radioenzymatic assay for tryp-tophan hydroxylase: [3H2O] release assessed by charcoal adsorption.J Neuorsci Methods 48:123–129.

Walker SJ, Liu X, Roskoski R Jr, Vrana KE. 1994. Catalytic core of rattyrosine hydroxylase: Terminal deletion analysis of bacterially-expressedenzyme. Biochim Biophys Acta 1206:113–119.

Yohrling IV, GJ, Mockus SM, Vrana KE. 1999. Identification of amino-terminal sequences contributing to tryptophan hydroxylase tetramer for-mation. J Mol Neurosci 12:23–34.

320 Yohrling et al.