Enzyme Replacement Therapy for PKU

1

A NOVEL APPROACH FOR ENZYME REPLACEMENT THERAPY: THE USE OFPHENYLALANINE HYDROXYLASE-BASED FUSION PROTEINS FOR THE TREATMENT

OF PHENYLKETONURIA *Ronen Eavri and Haya Lorberboum-Galski

From the Department of Cellular Biochemistry and Human Genetics, Faculty of Medicine, HebrewUniversity, Ein-Kerem, Jerusalem, Israel 91120

Running Title: Enzyme Replacement Therapy for PKUAddress correspondence to: Haya Lorberboum-Galski, Department of Cellular Biochemistry and HumanGenetics, Faculty of Medicine, Hebrew University, Jerusalem, Israel 91120. Tel. 972 2-6757465; Fax.972 2-6415848; E-Mail: hayalg@md.huji.ac.il

Metabolic diseases arise from mutationsin key enzymes of major metabolic pathways.One promising approach for the treatment ofsuch diseases is based on the administration ofa wild type enzyme to substitute the activity ofthe impaired enzyme by the use of enzymereplacement therapy, yet it is important todeliver this enzyme to the specific deficienttissue. We suggest a new concept for thetreatment of metabolic diseases using fusionproteins. We examined the feasibility of thisconcept in the well-characterized metabolicdisease, phenylketonuria (PKU), which resultsfrom a mutation in the liver enzymephenylalanine hydroxylase (PAH). PAH is akey enzyme in the metabolic pathway ofphenylalanine. Deficiency in PAH leads to highand persistent levels of this amino acid in theplasma of PKU patients, causing permanentneurological damage. Currently a low proteindiet is still considered the only effectivetreatment for most PKU patients. In order torestore PAH activity in the liver of PKUpatients, we constructed PAH-based fusionproteins with delivery moieties based on theHIV-TAT peptide, and fragments of the humanhepatocyte growth factor (HGF) aiming tospecifically target PAH to the liver. We showthat these new fusion proteins can be deliveredinto a variety of human liver cell lines andretain PAH activity after being internalized.We also show that plasma phenylalanine levelswere dramatically lowered in mice treated withPAH-based fusion proteins after IVadministration. We therefore suggest analternative concept for the treatment of PKUusing targeted fusion proteins, which may alsobe applied to the treatment of other metabolicdiseases.

Phenylketonuria (PKU) is a metabolic

disorder resulting from an abnormal form of theenzyme phenylalanine hydroxylase (PAH) (EC1.14.16.1). PAH catalyses the irreversiblehydroxylation of the amino acid phenylalanine(Phe) to tyrosine (Tyr) in the liver. This step isconsidered the rate-limiting step in the catabolicpathway of Phe (1). Up until now, more than 500mutations have been found in the PAH gene, mostof them leading to a dysfunctional enzyme(www.pahdb.mcgill.ca). The disease is inherited ina classic recessive way, with an incidence of 1 outof 10,000 newborns (OMIM 261600). PAH is aliver specific enzyme with a crucial role in themaintenance of a low and steady level of Phe inthe plasma. Untreated PKU patients suffer seriousneurological damage due to sustained high Phelevels in the blood. Children diagnosed with PKUare immediately transferred to a low protein dietfor life, in order to lower Phe levels in theirplasma and maintain this normal level (2).

New approaches for treatment have beensuggested in recent years, including gene therapy(3,4) and enzyme replacement therapy using theyeast-derived enzyme PAL which converts Pheinto a harmless metabolite (5). These new ideasare considered promising, but still remain outsideclinical practice, leaving reliance on the lowprotein diet as the only treatment used today forPKU.

In this paper we present a novel approachfor the treatment of PKU. This approach is basedon directing the PAH enzyme to the liver usingeither a common peptide or a selective homingligand, both aimed at delivering the enzyme to theliver and its fast clearance from the blood.

Recently new mechanisms for proteindelivery have been suggested. One new promisingapproach is fusion with small peptides calledprotein transduction domains (PTDs). These smallpeptides have been found to facilitate the

http://www.jbc.org/cgi/doi/10.1074/jbc.M703367200The latest version is at JBC Papers in Press. Published on June 12, 2007 as Manuscript M703367200

Copyright 2007 by The American Society for Biochemistry and Molecular Biology, Inc.

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

2

penetration of fused proteins through themembranes of a large variety of eukaryotic cells.Though the mechanism of action involved in thisinternalization process is not clearly understood,involving neither receptor and clatherin-mediatedendocytosis nor phagocytosis, these new PTDs areconsidered promising delivery tools. Out of manyPTDs discovered so far, the HIV transactivator oftranscription (TAT) peptide has been most widelystudied. TAT-fusion proteins were shown to bedelivered efficiently into cultured cells, intact tissueand live tissues when injected into mice (6,7).Proteins delivered by this peptide retained theiractivity after internalization. A recent studyconducted on the pharmacokinetics of TAT-β-galactosidase fusion protein, revealed that the liveris the main target for this fusion protein (8). Basedon these notions we first constructed a fusionprotein containing TAT fused to human PAH.

In order to achieve even better selectivitytowards delivery of the enzyme to the liver, thenatural location and the site of biological activity ofPAH, we also constructed a different set of PAH-based fusion proteins using human hepatocytegrowth factor (HGF) as the specific deliveringsequence. HGF is a potent paracrine growth factorwith mitogenic and morphogenic activities that issynthesized by mesenchymal cells. Liver is themain target of HGF and biological responses to thisgrowth factor are mediated by the tyrosine kinasereceptor encoded by the Met oncogene (cMet)(9,10). It is the most potent mitogen forhepatocytes, and was considered to be the majorgrowth factor responsible for driving hepaticregeneration (10). Research conducted on thebinding properties of HGF revealed that the α-chain alone (including the N-terminus and 4 kringledomains) can bind to the cMet receptor. Thisfragment however, will not induce phosphorylationof the cMet receptor, hence mitogenic, motogenicand morphologenic responses will not be activated(11). Therefore we constructed PAH-fusionproteins with the HGF-α-chain as the deliveringsequence. We used various fragments of the α-chain to construct NK1-PAH, NK2-PAH, ΝΚ3-PAH and _HGF-PAH. These constructs contain theN-terminus fragment and kringle domains 1, 1-2, 1-3 and full length α-HGF, respectively.

All the PAH-based fusion proteins wereproduced in bacterial cells. We show here that

both TAT and HGF derived leadingpeptides/proteins can be used to deliver functionalPAH into liver cells. Most importantly, we showthat the Phe plasma level of C57BL normal micetreated with TAT-PAH was lowered as soon as 15minutes after the intravenous administration of thefusion protein. Phe plasma levels remained lowcompared to untreated mice for several hours. Thiswork demonstrates that fusion proteins could bepotential candidates for treatment of metabolicdiseases such as PKU.

Experimental Procedures

Construction of plasmids encoding TAT-PAH andPAH - In order to construct the pET28-PAHplasmid with a His-tag sequence at the 5’ side ofthe coding sequence, we first cut the plasmidcontaining the cDNA sequence of human PAH (agift from Dr. Shwartz, Sheba Hospital, TelHashomer, Israel) with SpeI. The 5’ sticky endwas removed with mung bean nuclease. Thisfragment was then cut with EcoRI generating a1358 bp insert that was ligated into the pET28avector (Novagen Madison, Wisconsin) cut withHindIII, followed by a mung bean nucleasereaction and a second cut with EcoRI. Escherichiacoli strain DH5 (Stratagene, La Jolla, California)was used for all plasmid transformations.Restriction and modifying enzymes were obtainedfrom Roche (Mannheim, Germany) or NewEngland BioLabs (Ipswich, Massachusetts).The TAT-PAH plasmid was constructed using the5’ primer: 5’ TTT GAA TTC GTC CAC TGCGGT CCT GGA A 3’; and the 3’ primer: 5’ TTTGTC GAC TTA CTT TAT TTT CTG GAG GGCA 3’. The PCR insert cut with EcoRI and SalI wasligated into the pTAT2.1 vector cut with the sameenzymes (pTAT2.1 was a gift from Dr. S. Dowdy(University of California at San Diego, La Jolla,CA)).Construction of plasmids encoding the NK1-PAH,NK2-PAH, NK3-PAH and _HGF-PAH fusionproteins - The PAH plasmid was cut with EcoRIand ClaI restriction enzymes, and ligated with anoligo linker containing a BstEII recognition sitedownstream to the PAH sequence. Linker sense:5’ AAT TCG TTA ACA TGT GGG TGA CCGAT 3’; antisense: 5’ CGA TCG GTC ACC CAATG TTA ACG 3’. This plasmid served as anintermediate plasmid.

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

3

Total RNA was isolated from fresh humanplacenta with the TriPure Isolation reagent(Boehringer Mannheim, GmbH, Frejburg,Germany) and then reverse transcribed into firststrand cDNA using a reverse transcription system(Promega, Madison, WI).The NK1, NK2, NK3 and _HGF sequences,flanked by ClaI restriction sites, were generated byPCR, using total human placenta cDNA, and thefollowing synthetic oligonucleotide primers: HGFCT: 5’ GGA TCG ATT CGC AAT TGT TTC GT3’; HGF NT: 5’ AGA TCG ATA TGT GGG TGACCA AA 3’; PK1L: 5’ CCA TCG ATT TCA ACTTCT GAA CAC TG 3’; PK2L: 5’ CCA TCG ATTTCA GTT GTT TCC AAA GG 3’; and PK3L: 5’CCA TCG ATA TCT TGT CCA TGT GAC AT3’. These primers covered different regions of theα chain coding region of the HGF protein. HGFCT and HGF NT primers were used to constructthe αHGF construct; HGF NT and PK1L primersfor the NK1 construct; HGF NT and PK2Lprimers for the NK2 construct; and HGF NT andPK3L primers for the NK3 construct.The intermediate plasmid and PCR products werethen cut by BstEII and ClaI and ligated to produceNK1-PAH, NK2-PAH, NK3-PAH and αHGF-PAH carrying plasmids (Ptrhp4, Ptrhp5, Ptrhp6,Ptrhp7 respectively).Protein expression and purification of PAH andTAT-PAH proteins - Escherichia coli strainBL21(DE3), which carries a T7 RNA polymerasegene in a lysogenic and inducible form, was usedfor the expression of all fusion proteins. Cellscarrying the hPAH-based plasmids were grown inLB medium containing either ampicillin (100µg/ml) or kanamycin (100 µg/ml) according to theexpression vector.After reaching an OD600 value of 0.8-1, thecultures were induced for 24 h at 22°C with 1 mMisopropyl-1-thio-D-galactopyranoside (IPTG,Sigma-Aldrich, St. Louis, USA) and supplementedwith 0.02 mM FeNH4(SO4)2 (Sigma-Aldrich). Thecells were collected by centrifugation, and thepellet was stored at -70°C for several hours. Thefrozen pellet was thawed and suspended in PBSwith 1mM phenylmethylsulphonylfluoride (PMSF,Sigma-Aldrich), followed by sonication andcentrifugation at 35,000 g for 30 min. Both thepellet (insoluble fraction) and the supernatant(soluble fraction) were collected. The solublefraction was then, filtered and loaded on HiTrap

Chelating HP columns (Amersham-PharmaciaBiotech, Uppsala, Sweden) pre-loaded withNiSO4. The proteins were eluted with 80 mMhistidine using the ÄKTA FPLC system(Amersham-Pharmacia Biotech).The purified TAT-PAH fusion protein and thePAH protein were eluted from the column, anddialyzed against phosphate-buffered saline (PBS),pH 7.4 containing 0.02mM FeNH4(SO4)2 (finalconcentration).Protein expression and purification of HGF-basedproteins - The NK1-PAH, NK2-PAH, NK3-PAHand αHGF-PAH fusion proteins were expressed asdescribed above for the TAT-PAH fusion proteins.Expressing cells were then sub-fractionized andboth the supernatant (soluble fraction), and thepellet (insoluble fraction) were collected. Thesoluble fraction was used for further analysis.Western blot analysis - Samples from the variousprotein fractions (5-20 µg protein/lane) wereloaded on 12% (w/v) SDS/PAGE gels. Theproteins were electrotransferred onto Immobilon-PTransfer membrane (Millipore, Bradford, USA),and blotted either with mouse anti-human PAHantibody (Chemicon Temecula, California USA1:10,000), or mouse anti-His-tag antibody(Amersham-Pharmacia Biotech, 1:5,000).Antibody binding was detected by enhancedchemiluminescence (ECL) as described by themanufacturer (Amersham-Pharmacia Biotech).Phenylalanine hydroxylation assay - L-[U-14C]Phe and L-[U-14C] Tyr were obtained fromAmersham-Pharmacia Biotech. The PAH-basedfusion proteins were tested for their enzymaticability to hydroxylate [C14] Phe to [C14] Tyr in anin vitro cell free assay (12) modified by the use of40 µM BH4, 400 µM FeNH4(SO4)2, 4 mg/mlcatalase, and 20 mM DTT (all obtained fromSigma) in a 100 mM Na-Hepes pH 7 buffer. Thereaction mixture was incubated for 2 h at 30°C andthen loaded onto 3 mm chromatography paper(Schleicher&Schuell, Dassel, Germany), andseparated with 73% buthanol and 27% acetic acidas the mobile phase. Overnight film exposurerevealed the relative presence of [C14] Phe and of[C14] Tyr in the reaction. Relative concentrationswere calculated by processing the film with theImageJ software (NIH, Bethesda, Maryland,USA).

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

4

Cell Lines - All cell lines were obtained from theAmerican type Culture Collection (ATCC,Manassas, VA, USA). Hepatocarcinoma HepG2and HuH7 cell lines were maintained inDulbecco's modified Eagle's medium (DMEM,Biological Industries, Beit-Haemek, Israel). Thecolon adenocarcinoma Colo 205 cell line wasmaintained in RPMI 1640 medium (BiologicalIndustries). All media were supplemented with10% fetal bovine serum (HyClone, Logan UT,USA), 2 mM L-glutamine, 100 units/ml penicillin,and 100 µg/ml streptomycin (BiologicalIndustries), cultured in 100 mm Petri dishes, andgrown in a humidified atmosphere of 5% CO2,95% air at 37°C. All cell cultures were tested formycoplasma contamination and found to benegative.Primary Hepatocytes - Mice primary hepatocyteswere obtained by perfusion of liver from C57BLfemale mice, age 7-8 weeks according to themethod described by Berry et al (13) with thefollowing modifications: liver perfusion rate was 5ml/min without recirculation of the perfusate.Cells were maintained under the same conditionsas HepG2 cells.Protein delivery into cultured cells - The humancell lines HepG2, HuH7 and Colo205 wereincubated with PAH-based fusion proteins forvarious time periods, the cells were collected bycentrifugation and washed twice with cold PBS.The cells were centrifuged at 500 g for 5 min at4°C and the packed cell volume determined. Thepacked cells were resuspended in two packed cellvolumes of buffer A (10 mM Hepes, pH7.5, 1.5mM MgCl2, 5 mM KCl) and allowed to swell onice for 15 min. The cells were then lysed byrapidly pushing them through a narrow pipette tip20 times. The cell homogenate was centrifuged at10,000 g for 5 min to obtain the cytoplasmicsupernatant (cytoplasmic fraction). Theinternalized fusion proteins were detected in thecytoplasmic fraction by assessing cytoplasmicPAH enzymatic activity.Analysis of cells treated with PAH-based fusionproteins by confocal microscopy - HepG2 cellsgrown on cover slips to 50-70% confluency weretreated with TAT-PAH (20 µ g/ml finalconcentration), NK1-PAH and NK2-PAH (2µg/ml final concentration) for various timeperiods. Cells were than washed with PBS, andwith 0.05% trypsin for 2 min at 37ºC, fixed in

3.7% formaldehyde in PBS for 10 min at roomtemperature and then permiabilized with 0.2%Triton X100 for 30 min. Following a second wash,cells were incubated with mouse anti-human PAHantibody (PH8, 1:200, Chemicon Temecula) andthen with a FITC-conjugated goat anti-mouseantibody (1:250, Jackson Immunoresearch,Cambridgeshire, UK). Cells were then treated withRNAse (20 µg/ml) and then with propidium iodide(PI, 5 µg/ml) in order to visualize cell nucleus.Cells were analyzed with a confocal laser scanningmicroscope (NIKON C1, Nikon Corp., Tokyo,Japan).Cell proliferation assay - HuH7 and HepG2 cellswere used to investigate the proliferative effect ofPAH-based fusion proteins. 5 x 104 cells weregrown in 96 well plates for 24 h prior toincubation with the proteins. 20 µg/ml TAT-PAH,2 µg/ml NK1-PAH, and NK2-PAH protein wereadded to cells for a period of 72 h, after which Celltiter blue® reagent (Promega Madison Wisconsin)was used according to manufacturer's instructionsto analyze cell survival.HPLC analysis of plasma phenylalanine andtyrosine concentrations – 3 C57BL female mice,age 7-8 weeks were used for each time point. Micewere injected with 200µl of 100 µg/ml TAT-PAH,and sacrificed at various times after the injection.0.5 ml blood samples from the mice at each timepoint were collected in heparin tubes (BelliverIndustrial Estate, Plymouth UK). The tubes werecentrifuged at 7,000 g for 5 min at 4°C in order toseparate the plasma. 6% TCA was then added tothe supernatant and the tubes were centrifugedagain at 20,000 g for an additional 5 min, thesupernatant was then collected and diluted 1:10 inmobile phase for analysis. Phe levels weremeasured in mouse plasma by HPLC analysissimilar to that described by Atherton and Green(14) with the following modifications: the mobilephase was 7.5% acetonitrile (J.T.Baker, Deventer,Netherlands) 20 µl/l octylamine (Sigma-Aldrich),and 800 µ l/l 11.6M perchloric acid ( M e r k ,Darmstadt, Germany). Samples were separatedusing a C18 column (150 mm octadecylsilane; 4.6mm pore size; Knauer, Berlin, Germany). Phelevels were detected with the Spectroflow 773absorbance detector at 193 nm (Kratos, Kyoto,Japan) and the SP4400 integrator (Spectra Physics,

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

5

San Jose, California, USA). Analysis wasperformed on a Merck HPLC system.

RESULTS

Construction, expression, purification andcharacterization of the fusion proteins - W econstructed plasmids encoding the PAH-basedfusion proteins TAT-PAH, NK1-PAH, NK2-PAH,NK3-PAH, αHGF-PAH, and PAH alone. TAT-PAH contains the delivering moiety TAT, whichpossesses the ability to penetrate a wide variety ofeukaryotic membranes, but is mainly taken up bythe liver. NK1-PAH, NK2-PAH, NK3-PAH andαHGF-PAH consist of various fragments of thehuman αHGF sequence as the delivering moiety,in order to selectively target PAH to liver cells.PAH without a delivering domain, yet with aprotease cleavage sequence was used as a controlprotein. The coding sequence of the proteins wasconfirmed by restriction enzyme analysis andDNA sequence analysis (results not shown).Following transformation of E. coli BL21 (DE3)cells with one of the plasmids, expression of thefusion gene was controlled by the bacteriophageT7 late promoter. A schematic representation ofthe new fusion proteins is shown in Fig 1a.

Analysis of the soluble fraction of E. coliexpressing cells on SDS/PAGE gels revealed amajor band of the expected molecular mass of 53kDa corresponding to PAH and TAT-PAHproteins (Fig 1b). These bands reacted with antiPAH antibodies (Fig 1c) and with antibodiesagainst His-tag (data not shown). Immunoblottinganalysis of the soluble fractions of E. coli cellsexpressing NK1-PAH and NK2-PAH using antiPAH antibodies revealed bands of 76 kDa and 87kDa respectively (Fig 1d). Extracts of BL21 cellsexpressing the NK3-PAH (97 kDa) and theαHGF-PAH (109 kDa) fusion proteins revealedthat these proteins were found mainly in theinsoluble fraction (Fig 1d). The soluble fraction ofTAT-PAH and PAH was loaded on a Ni chelatingcolumn in order to affinity purify these PAH-based proteins. Both the TAT-PAH and controlPAH proteins were purified in a single step. Elutedproteins showed a major band of the expected sizeand were >95% pure, as determined by SDS-PAGE analysis (Fig 1b).

Enzymatic activity of the proteins - TAT-PAH and control PAH purified proteins weretested for their PAH activity by an in vitro cell freeassay. The proteins were incubated in a reactionmixture containing the co-factors and reagentsnecessary for activity, and the PAH activity wasmonitored by following the conversion of [C14]Phe substrate into [C14] Tyr. The soluble fractionof the PAH-based proteins was found to be activein vitro, exhibiting the enzymatic activity ofconverting Phe to Tyr (Fig 2). PAH and TAT-PAH proteins show a dose dependent activitywhen the protein concentration is >40 ng/ml.However, at the concentration range of 40-4000ng/ml, PAH exhibited a higher enzymaticactivity than that of TAT-PAH (Fig 2a). Furtheranalysis conducted on both proteins revealed ahigher Vmax value for the PAH enzyme (3.98µM·min-1·µg-1 ± 0.63) than for TAT-PAH (1.76µM·min-1·µg-1 ± 0.80). However, the Km values ascalculated from this plot were the same for bothproteins (9.55 µM ± 1.81 for PAH, and 9.47 µM ±0.94 for TAT-PAH) (Fig 2b). Thus, changes thatdecrease the specific activity are, most probably,induced in the PAH enzyme upon fusion with theTAT peptide. These changes however, do notcause a change in the affinity of the fusion proteinto its substrate. It should be pointed out that thePAH-based proteins showed PAH enzymaticactivity only in the soluble subcellular fraction.This is despite multiple efforts to denature andproperly refold the highly enriched insolublesubfraction under various conditions.

In vitro enzymatic activity test of the fourHGF targeted clones revealed that only the NK1-PAH and NK2-PAH fusion proteins retained theenzymatic activity of the PAH enzyme in thesoluble fraction (Fig 2c lines 1&2). The enzymaticactivity of these proteins was highly specific anddose dependent. The soluble fraction of E. colicells expressing NK3-PAH and _HGF-PAH didnot show any activity (Fig 2c lines 3&4). As forthe TAT-PAH protein, attempts made to denatureand properly refold the four clones from theinsoluble fraction failed and ended in adysfunctional enzyme. Consequently, allsubsequent experiments were performed withhighly purified PAH-based proteins from thesoluble subfraction of cells expressing TAT-PAHand PAH, or from the enriched soluble fraction of

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

6

cells expressing our chimeric proteins NK1-PAHand NK2-PAH.

Internalization of PAH based proteinsassessed by confocal microscopy - HepG2 andHuH7 hepatomcarcinoma cells were incubatedwith 20 µg/ml (final concentration) TAT-PAH,and 2 µg/ml (final concentration) of NK1-PAHand NK2-PAH proteins for various times. A seriesof experiments using FITC-conjugated anti-PAHantibodies and confocal microscopy wereperformed in order to visualize the internalizationof the TAT-PAH protein into cells.

The images of treated cells revealed TAT-PAH within the cytoplasm of both HepG2 (Fig 3)and HuH7 cells (data not shown) after 30 minincubation. TAT-PAH was distributed evenlythroughout the cytoplasm of treated cells, andshowed a stronger signal with time, as comparedto control cells (Fig 3). Images of HepG2 cellstreated with TAT-PAH showed a stronger signalafter 2 h incubation then that of HuH7 cells treatedfor the same period of time (data not shown), mostprobably indicating a more efficient delivery ofthe fusion proteins into HepG2 cells.

We used two hepatocyte cell lines in orderto test the ability of HGF targeting to deliver thePAH based proteins into real liver target cellsexpressing cMet. HepG2 and HuH7 cell lines werefirst analyzed by RT-PCR for expression of cMetand both of these lines were found to express thisreceptor in high levels (data not shown). Analysisof confocal images of HepG2 cells treated withNK1-PAH and NK2-PAH showed that thesefusion proteins internalize into the cells within 10minutes, and that this internalization increaseswith time (Fig 3). No increase in PAH wasobserved in the cells treated with the control PAHprotein lacking a delivering domain (Fig 3).

Activity of PAH based proteins withintreated cells - In order to assess the PAH activityof the fusion proteins after they were deliveredinto the cells, we lysed the treated cells with ahypotonic buffer. The reaction mixture containingthe co-factors and [C14] Phe were then added to thecell lysates. This assay was based on theassumption that cell lysates of treated cells wouldshow stronger PAH activity than untreated cells inan in vitro cell free system. We found that lysatesof cells treated with PAH based proteins containedactive PAH (Fig. 4). We found that both HuH7and HepG2 cells had background PAH activity in

the cell free system assay. However, this activitywas amplified in cells treated with TAT-PAH. Theamplification of Tyr formation from Phe wasobserved in the homogenates of cells treated withTAT-PAH as after only 30 min incubation (Fig4a). The strongest effect of TAT-PAH on bothHuH7 and HepG2 cells was seen after anincubation period of 3 h with an almost tenfolddecrease in Phe concentration. This decrease wasobserved for as long as 6 h. Longer incubationperiods seem to decrease this effect. (Fig 4a).Similar results were observed in Colo 205 cellstreated with TAT-PAH (data not shown).

HepG2 and HuH7 are both hepatocarcinomacell lines, strongly expressing the HGF receptorcMet. We next compared the decrease of [C14] Phein the extracts of cells treated with the variousproteins. A significant decrease in the [C14] Phesignal was observed in the lysets of cells treatedwith TAT-PAH (20 µg/ml final concentration),NK1-PAH and NK2-PAH (2 µg/ml finalconcentration). Treatment with NK1-PAH had agreater effect (20 fold) on the decrease of C14Phethen did NK2-PAH (10 fold) (Fig 4b, P<0.01).We found no significant decrease in the [C14] Phesignal of the lysates of cells that were treated withthe PAH protein lacking any delivering moiety ascompared to PBS treated control cells (Fig 4b).These results confirmed the specific activity of thetargeted fusion proteins.

In order to test the internalization andactivity of our fusion proteins in primary cells, weisolated mouse primary hepatocytes and incubatedthem with TAT-PAH. Treated cells showed analmost two fold increase in radioactive Tyrformation from radioactive Phe as compared tocontrol cells (Fig 5a). Thus, TAT-PAH isinternalized into primary mouse hepatocytes aswell as into cells of established lines.

Toxicity and proliferation – The delivery offusion proteins into cells can cause undesiredeffects such as non-specific toxicity or cellproliferation. Moreover, since HGF is known toinduce proliferation of cells, it was important torule out this possibility in cells treated with HGF-based fusion proteins. Therefore, PAH-basedproteins were tested for their effect on the viabilityof HuH7 and HepG2 cells exposed to theseproteins for 72 h. We used a proven toxic fusionmolecule GnRH-PE as a positive control (15). Wefound no significant difference in the survival rate

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

7

of HuH7 cells treated with 2 µg/ml NK1-PAH orNK2-PAH as compared to PBS treated controlcells (data not shown). Thus, HGF-PAH fusionproteins did not cause any toxic effect to the cells,and did not contribute to the proliferation of thetreated cells.

HPLC analysis of Phe and Tyrconcentrations in plasma of mice treated withTAT-PAH - PAH-based proteins should have theability to reduce blood Phe concentration. In orderto test the effect of TAT-PAH administration onblood Phe levels, 3 mice at each time point wereadministered 20 µg of highly purified TAT-PAHintravenously. Plasma was collected and analyzedby HPLC to determine Phe and Tyr levels. Phewas dramatically reduced with a decline startingafter 15 minutes. The lowest Phe concentrationwas measured 1 h following TAT-PAH injectionand stayed less than 20% of control for as long as6 h (Fig. 5b). No significant rise in Tyrconcentration was observed in these plasmasamples.

DISCUSSION

We present here a novel approach for thetreatment of PKU, by targeting a functional humanPAH enzyme into liver cells. This enzyme willreplace the activity of the mutant PAH found inthe liver of PKU patients. In order to test thehypothesis that an active human PAH enzyme canbe delivered into liver cells, we constructed theTAT-PAH fusion protein using the HIV proteintransduction domain TAT as a delivering moiety.Moreover, in order to achieve better selectivitytowards the liver, the primary target for the PAHenzyme, we constructed a series of liver-targetedfusion proteins including NK1-PAH, NK2-PAH,NK3-PAH and αHGF-PAH using the human HGFas the targeting moiety.

By fusing human PAH to TAT and tofragments of HGF we were able to produce andpurify enzymatically active fusion PAH-basedproteins in bacterial cells. We demonstrated thatTAT-PAH protein can be delivered to a variety ofhuman cell lines and retain PAH activity for at least6 h after internalization, and longer in some cases(see below). We were also able to demonstrate thatby fusing PAH to NK1 and NK2 of human HGF,these proteins can enter human hepatocytes. Onceinside the cells, these proteins sustained PAH

activity. Finally, we treated C57BL mice withTAT-PAH proteins and showed that plasma Phelevels were lowered in the treated mice 15 minafter i.v. administration, and remained low forseveral hours.

Our PAH-based fusion proteins weredesigned and constructed as independent PAHmonomers. The PAH enzyme is naturally producedas a protease resistant tetramer consisting of fourindependent monomers, each with the capability ofhydroxylating Phe into Tyr. The PAH tetramer isformed by combining four monomers through atetramerization domain at the C-terminus of theenzyme (16). Thus, internalized PAH-based fusionproteins, designed as monomers, could besubjected to degradation by cellular proteases.However, we specifically constructed our PAH-based fusion proteins as monomers with the C-terminal free. This may enable these proteins toform protease resistant tetramers once they enterthe cell. Indeed, analysis of semi-native westernblots of TAT-PAH revealed bands correspondingto high molecular weights (data not shown). It istempting to speculate that fusion of TAT togetherwith a 6x His tag at the N-terminus of human PAHmight still enable the formation of tetramersthrough the C-terminus domain of this enzyme.Bands corresponding to higher molecular weightswere also observed -in western blots of NK1-PAHand NK2-PAH (results not shown).

We found that TAT-PAH, NK1-PAH andNK2-PAH fusion proteins were rapidlyinternalized into cells, and once inside the cells,these proteins were distributed throughout thecytoplasm (Fig 3). Though the final concentrationsof the NK1-PAH and NK2-PAH proteins werelow and about 1/10 of the final concentration ofTAT-PAH, the intensity of florescent signal wasstronger in cells treated with NK1-PAH and NK2-PAH. This suggests that delivery of PAH is moreefficient using fragments of the αHGF as thedelivering moiety, though farther analysis needs tobe performed in order to confirm this idea.

Analysis of lysates from cells treated withTAT-PAH, NK1-PAH and NK2-PAH retrievedfrom the soluble fraction showed a decrease inC14Phe and an increase in C14Tyr formation. Thelysates of cells treated with the insoluble fractionsdid not show any additional PAH activity.Therefore, the increase in radioactive Tyrformation can only be explained by the

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

8

contribution of PAH from the active fusionproteins that entered the cells. We were able toassess the stability of these proteins inside the cellsby conducting the PAH enzymatic assay in a cellfree system after these proteins were internalizedinto the cells. TAT-PAH activity was observed inHuH7 cells up to 24 h. This activity lasted evenlonger in HepG2 cells, as long as 48 h (Fig 4a).This difference in stability of TAT-PAH can beexplained by a possible variation in the expressionpattern of cellular proteases between these celllines.

Since the NK1-PAH and NK2-PAHconstructs are based on the HGF α subunit it wasimportant to assess if addition of these sequencesdid not induce proliferation of the liver cellstested. We found no change in proliferation ofliver cells treated with HGF-based fusion proteins,as assessed by a cell viability assay (data notshown). This is consistent with previous studiesconducted on the binding properties of HGF whichrevealed that αHGF and other smaller fragmentsof this factor can enter the cell without causingother downstream responses related to HGF(17,18). This enables the use of the HGF-αsubunit as a targeting moiety, while avoidingundesired biological responses that are induced bythe growth factor itself.

The ultimate test in the treatment of PKU isto maintain stable low plasma Phe levels. Normalplasma Phe levels are 60 µmol/l and are constantaround this level from childhood until maturity.Children or adults displaying persistent values ofPhe above 120 µmol/l are defined as havinghyperphenylalaninemia. Subsequent to the cellculture assays, we conducted an in vivoexperiment in order to test whether the injection ofPAH based fusion proteins can affect plasma Phelevels of mice. Surprisingly plasma Phe levelsstarted to decrease 15 min after intravenous TAT-PAH injection and had decreased dramatically by30 min. Phe levels remained less than 20% of thelevels in control untreated mice for as long as 6 h.It is important to point out that this decrease ofPhe levels was seen in healthy C57BL mice. Thiseffect would probably be stronger in the PKUmouse model since plasma Phe levels of thismouse are much higher, and the main objective intreatment is lowering and maintaining a normalPhe concentration.

The possibility of targeting a protein to aspecific tissue or cell has existed for some time.However, until now this approach has beenspecifically aimed at destroying a specificpathogenic cell population. In this paper wepresent for the first time the possibility of targetingan active enzyme to specific cells that lack a keymetabolic function. We have utilized the idea oftargeting a protein, not to eliminate a specific cellas demonstrated previously by utilizing fusionproteins as targeted cytotoxic molecules (19), butrather to bestow a functional activity to thedeficient cell.

In recent years new approaches for thetreatment of PKU have been proposed. The first isbased on gene therapy, suggesting the replacementof the mutant gene by the wild type sequenceencoding the normal PAH gene. This method,successfully tested on mice, can undoubtedly beused as the ultimate therapy (3,4). Unfortunatelygene therapy is still considered not applicable tohumans due to technical problems that have yet tobe resolved. Liver replacement in PKU patients bythe transplantation of a healthy organ will cure thedisease but subject the patient to a wide variety ofrejection complications (20).

Another promising potential cure has beensuggested for PKU recently based on the PALenzyme derived from the yeast Rhodosporidiumtoruloides. This enzyme is capable of degradingPhe by converting it to a harmless metabolitederivative that is cleared from the body throughthe kidneys. Oral administration of PAL has beenshown to lower Phe levels in the blood of treatedanimals but is subject to proteolysis. Intravenousadministration of PAL has been proven effectivebut will be recognized by the immune system asnon-self (21). Current work on PAL involvespegylation in order to reduce this effect (22).

Enzyme replacement therapy involves theadministration of the wild type PAH enzyme (23).Yet for the administration of this enzyme to thebloodstream to have an effect, it must be deliveredto the liver – the site of action of PAH. Althoughall of these new approaches for treatment areconsidered promising, the low protein diet stillremains the only clinical tool in PKU. Treatmentof metabolic deficiencies by enzyme replacementtherapy has already been proven effective in theGaucher Disease type 1 (24). We like othersbelieve that this could also be the most promising

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

9

approach for the treatment of PKU. We believethat targeting the PAH enzyme to the liver by theuse of a specific targeting moiety has the potentialfor the greatest benefits.

Using the approach we have presented here,human PAH enzyme can be delivered quickly andeffectively into the preliminary site of action – theliver. After the PAH enzyme has internalized intothe liver, it can act as a part of the normalmetabolic pathway in order to lower Phe levels inthe blood of PKU patients. This will also decreasethe time of circulation of the enzyme in the blood,

and thus can decrease the possible harmfulimmune effects. The idea of using PAH-basedfusion proteins can be regarded as a potentialalternative in the treatment of PKU.

Moreover, this approach, of targeting thewild type enzyme to a specific tissue, could alsobe applied to treatment of any other metabolicdeficiency for which the enzyme involved isknown and cloned, and the target tissue identified.Thus, this new targeted enzyme replacementapproach may open new paths in the treatment ofmetabolic diseases.

REFERENCES

1. Kaufman, S. (1999) Proc Natl Acad Sci U S A 96(6), 3160-31642. Scriver, C. R. (1995) Clin Biochem 28(2), 137-1443. Ding, Z., Georgiev, P., and Thony, B. (2006) Gene Ther 13(7), 587-5934. Chen, L., and Woo, S. L. (2005) Proc Natl Acad Sci U S A 102(43), 15581-155865. Sarkissian, C. N., Shao, Z., Blain, F., Peevers, R., Su, H., Heft, R., Chang, T. M., and Scriver, C.

R. (1999) Proc Natl Acad Sci U S A 96(5), 2339-2344.6. Schwarze, S. R., Ho, A., Vocero-Akbani, A., and Dowdy, S. F. (1999) Science 285(5433), 1569-

15727. Wadia, J. S., and Dowdy, S. F. (2005) Adv Drug Deliv Rev 57(4), 579-5968. Cai, S. R., Xu, G., Becker-Hapak, M., Ma, M., Dowdy, S. F., and McLeod, H. L. (2006) Eur J

Pharm Sci 27(4), 311-3199. Prat, M., Narsimhan, R. P., Crepaldi, T., Nicotra, M. R., Natali, P. G., and Comoglio, P. M.

(1991) Int J Cancer 49(3), 323-32810. Birchmeier, C., Birchmeier, W., Gherardi, E., and Vande Woude, G. F. (2003) Nat Rev Mol Cell

Biol 4(12), 915-92511. Parr, C., Hiscox, S., Nakamura, T., Matsumoto, K., and Jiang, W. G. (2000) Int J Cancer 85(4),

563-57012. Shiman, R., and Gray, D. W. (1980) J Biol Chem 255(10), 4793-480013. Barry, M. E., AM; Barritt, GJ. (1991) Isolated Hepatocytes Preparation, Properties and

Applications, Elsevier, Amsterdam14. Atherton, N. D., and Green, A. (1988) Clin Chem 34(11), 2241-224415. Nechushtan, A., Yarkoni, S., Marianovsky, I., and Lorberboum-Galski, H. (1997) J Biol Chem

272(17), 11597-1160316. Hufton, S. E., Jennings, I. G., and Cotton, R. G. (1998) Biochim Biophys Acta 1382(2), 295-30417. Lokker, N. A., and Godowski, P. J. (1993) J Biol Chem 268(23), 17145-1715018. Chan, A. M., Rubin, J. S., Bottaro, D. P., Hirschfield, D. W., Chedid, M., and Aaronson, S. A.

(1991) Science 254(5036), 1382-138519. Foss, F. (2006) Semin Oncol 33(1 Suppl 3), S11-1620. Levy, H. L. (1999) Proc Natl Acad Sci U S A 96(5), 1811-181321. Sarkissian, C. N., and Gamez, A. (2005) Mol Genet Metab 86 Suppl 1, S22-2622. Wang, L., Gamez, A., Sarkissian, C. N., Straub, M., Patch, M. G., Han, G. W., Striepeke, S.,

Fitzpatrick, P., Scriver, C. R., and Stevens, R. C. (2005) Mol Genet Metab 86(1-2), 134-14023. Gamez, A., Wang, L., Straub, M., Patch, M. G., and Stevens, R. C. (2004) Mol Ther 9(1), 124-

12924. Grabowski, G. A. (2005) Curr Opin Pediatr 17(4), 519-524

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

10

FOOTNOTES

We would like to thank Ruth Belostotsky and Matan Rapoport for their valuable contribution to thiswork. This work was supported by a grant from the Israel Science Foundation (ISF) grant number035358.

The abbreviations used are: PKU, phenylketonuria; PAH, phenylalanine hydroxylase; ; PTDs, proteintransduction domains; TAT, HIV transactivator of transcription; HGF, hepatocyte growth factor; NK1,N’-terminus and first kringle domain of HGF; NK2, N’-terminus and first and second kringle domains ofHGF; BH4, tetrahydrobiopterin.

FIGURE LEGENDS

Fig 1. Schematic representation of PAH-based proteins (a), their expression and purification (b-d).(a) Schematic representation of PAH-based proteins (PAH – control protein; TAT-PAH; NK1-PAH;NK2-PAH; NK3-PAH; αHGF-PAH. *aa sequence of PAH before proteolytic cleavage of the thrombinsite on the N-terminus).(b) SDS-PAGE analysis of the purified fractions of (1) TAT-PAH and (2) PAH.These proteins were purified using affinity chromatography as described in Methods.(c) Western blot analysis of purified: (1) TAT-PAH and (2) PAH using antibodies against PAH.(d) Western blot analysis of (1) NK1-PAH, (2) NK2-PAH, (3) NK3-PAH and (4) HGF-PAH, usingantibodies against PAH. W: whole cell fraction of E. coli; S: soluble fraction; I: insoluble fraction.PAH, human phenylalanine hydroxylase; TAT, transactivator of transcription peptide; NK1, NK2, NK3,N’ terminal amino acid sequences of the α subunit of the human hepatocyte growth factor includingkringles 1; 1 and 2; 1, 2 and 3 respectively; HGF, α subunit of the human hepatocyte growth factor.

Fig 2. In vitro enzymatic activity of PAH-based proteins.(a) PAH and TAT-PAH dose dependent activity, based on intensity of the 14C Tyr signals of an in-vitroenzymatic activity assay of 2 h. (as described in methods).(b) Lineweaver-Burk plot of PAH and TAT-PAH (Determination of the Km and Vmax of PAH and TAT-PAH: 1µg of either PAH or TAT-PAH proteins were incubated in the presence of increasingconcentrations of 14C Phe: 2µM, 4µM, and 10µM for time points: 5min, 10min, and 20 min. Formation ofC14Tyr was quantified for each reaction, and used in order to calculate the V0 for each substrateconcentration. Once V0 values were determined, they were used in order to calculate the Vmax and Kmvalues. (c) Chromatogram showing conversion of 14C Phe to 14C Tyr by (W) whole cell fraction and (S) solublefraction of E. coli expressing (1) NK1-PAH, (2) NK2-PAH, (3) NK3-PAH, (4) αHGF-PAH and (5)unrelated protein. (M) Equal amounts of 14C Phe and 14C Tyr.

Fig 3. Internalization of PAH-based fusion proteins visualized by confocal microscopy.HepG2 cells were treated with TAT-PAH (20 µg/ml final concentration), NK1-PAH and NK2-PAH (2µg/ml final concentration) for 10 min, 30 min, 1 h and 2 h. Cells were permeabilized and incubated withanti-PAH antibody followed by FITC-conjugated secondary antibody, and analyzed by a confocal laserscanning microscope (NIKON C1, Nikon corporation, Tokyo, Japan) All images are at X1000magnification.

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

11

Fig 4. PAH enzymatic activity within extracts of cells treated with PAH-based proteins -“within-cell”activity.Cells were treated with TAT-PAH (20 µg/ml final concentration) for various times. Treated cells werethen washed and homogenized. Cell homogenates were assayed for PAH activity by conversion of 14CPhe to 14C Tyr (as described in Methods).(a) Line graphs of relative amounts of Phe (as compaired to PBS-treated controls) in HuH7 and HepG2cell extracts after 10 min to 72 h incubation with TAT-PAH (Results are expressed as means ± SEM of 3independent repeats for each time point).(b) PAH activity within extracts of HuH7 cells treated with PBS, PAH and TAT-PAH (20 µg/ml finalconcentration), NK1-PAH and NK2-PAH (2 µg/ml final concentration) for 3 h. These graphs are basedon the analysis of the amount of 14C Phe that remained in the cell extracts following a 2 h in-vitro activityassay (as described in Methods). Results are expressed as means ± SEM; overall significance wasdetermined by one-way ANOVA with LSD test used for post hoc comparisons. Groups that share lettersare not significantly different.

Fig 5. Effect of TAT-PAH on mouse primary cell lines and in vivo.(a) PAH activity in extracts of mouse primary hepatocytes treated with PBS and TAT-PAH (20 µg/mlfinal concentration) for 3 h. Results are expressed as means ± SEM of 3 independent repeats.(b) Plasma Phe levels relative to PBS-treated control mice. Plasma Phe concentrations were measured byHPLC after i.v. injection of TAT-PAH (20 µg) (as described in Methods). 3 C57Bl mice were used foreach time point. Results are expressed as means ± SEM.

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

12

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

13

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

14

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

15

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Enzyme Replacement Therapy for PKU

16

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from

Ronen Eavri and Haya Lorberboum-Galskihydroxylase-based fusion proteins for the treatment of phenylketonuria

A novel approach for enzyme replacement therapy: The use of phenyalanine

published online June 12, 2007J. Biol. Chem. 

  10.1074/jbc.M703367200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on January 28, 2020http://w

ww

.jbc.org/D

ownloaded from