www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 337 (2005) 701–707

BBRC

Improved intracellular delivery of glucocerebrosidase mediatedby the HIV-1 TAT protein transduction domainq

Kyun Oh Lee, Nga Luu, Christine R. Kaneski, Raphael Schiffmann,Roscoe O. Brady, Gary J. Murray *

Developmental and Metabolic Neurology Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health,

Building 10, Room 3D04, 10 Center Drive, Bethesda, MD 20892, USA

Received 3 May 2005Available online 26 September 2005

Abstract

Enzyme replacement therapy (ERT) for Gaucher disease designed to target glucocerebrosidase (GC) to macrophages via mannose-specific endocytosis is very effective in reversing hepatosplenomegaly, and normalizing hematologic parameters but is less effective inimproving bone and lung involvement and ineffective in brain. Recombinant GCs containing an in-frame fusion to the HIV-1 trans-ac-tivator protein transduction domain (TAT) were expressed in eukaryotic cells in order to obtain active, normally glycosylated GC fusionproteins for enzyme uptake studies. Despite the absence of mannose-specific endocytic receptors on the plasma membranes of variousfibroblasts, the recombinant GCs with C-terminal TAT fusions were readily internalized by these cells. Immunofluorescent confocalmicroscopy demonstrated the recombinant TAT-fusion proteins with a mixed endosomal and lysosomal localization. Thus, TAT-mod-ified GCs represent a novel strategy for a new generation of therapeutic enzymes for ERT for Gaucher disease.Published by Elsevier Inc.

Keywords: Glucocerebrosidase; Gaucher disease; Enzyme replacement therapy; TAT; Protein transduction

Lysosomal storage diseases (LSD) are a group of hered-itary metabolic disorders, caused by defects in catabolic en-zymes that result in pathological accumulation ofunmetabolized substrates within the lysosomes of affectedindividuals and clinical diseases with a broad range ofseverity and organ system involvement [1,2]. Gaucher dis-ease (MIM 230800) is the most common LSD with a prev-alence of 1 in 50,000 [3,4]. This autosomal recessive LSD iscaused by defects in the enzyme glucocerebrosidase (GC;EC 3.2.1.45) that is essential for the catabolism of glucocer-

0006-291X/$ - see front matter. Published by Elsevier Inc.

doi:10.1016/j.bbrc.2005.05.207

q Abbreviations: LSD, lysosomal storage diseases; ERT, enzyme replace-ment therapy; GC, glucocerebrosidase; PTD, protein transductiondomain; TAT, trans-activator; HIV-1, human immunodeficiency virustype 1; HSV, herpes simplex virus; PVDF, polyvinylidene difluoride; MOI,multiplicity of infection; 4-MUG, 4-methylumbelliferylglucopyranoside;HRP, horseradish peroxidase; FITC, fluorescein isothyocyanate; BBB,blood–brain barrier.* Corresponding author. Fax: +1 301 496 9480.E-mail address: murrayg@ninds.nih.gov (G.J. Murray).

ebroside [5]. Deficiency of this enzyme leads to abnormallyhigh concentrations of glucocerebroside in the lysosomes ofmononuclear phagocytes. Over 200 different mutationshave been identified in the gene encoding GC [4,6]. Thebroad spectrum of manifestations in Gaucher disease in-cludes hepatosplenomegaly, anemia, thrombocytopenia,bone lesions, and sometimes involvement of brain, lungsor kidneys. Patients with Gaucher disease are categorizedas type 1: non-neuronopathic, type 2: acute neuronopathicor type 3: chronic neuronopathic [3,7].

Elucidation of the lectin-specific uptake of circulatingglycoproteins [8,9] led to a strategy of partial enzymaticdeglycosylation of native GC to direct the enzyme to mac-rophages via mannose-specific endocytosis [10]. In patientswith type 1 Gaucher disease ERT reverses visceral storageof glucocerebroside resulting in dramatic reduction inhepatosplenomegaly and normalization of hematologicalparameters [11]. Despite this success, in one studyonly two-thirds of patients on high dose ERT show bone

702 K.O. Lee et al. / Biochemical and Biophysical Research Communications 337 (2005) 701–707

marrow improvement [12] and in another reversal of boneloss in splenectomized patients was not seen [13]. Further-more, ERT has not been shown to be useful for arrestingthe neurological progression of the disease or reducingaccumulation of glucocerebroside in the brain [7,14].

The limited ability of biological macromolecules to crosscell membranes, especially the blood–brain barrier (BBB),has led to the design of fusion proteins containing proteintransduction domains (PTD) [15,16]. The HIV-1 encodedtrans-activator (TAT) protein, essential for viral geneexpression and replication, was shown to freely enter cellswhen added exogenously to the culturemedium [17,18]. Sub-sequent studies have demonstrated the capacity of TAT tocarry heterologous protein ‘‘cargo’’ [19] and that the basic11-amino acid TAT(47–57) peptide (47YGRKKRRQRR57R) is sufficient for protein transduction across theBBB [15]. TAThas been shown to be capable of transportingsome proteins across the BBB but not others [20–22]

In the present study, we investigated the potential of the11 amino acid TAT domain fused to GC as a means to im-prove delivery of the therapeutic protein into cells. Wehave expressed active TAT-GC fusion proteins in variousanimal and human cell lines including type 1 and type 2Gaucher fibroblasts in order to test the ability of theTAT domain to facilitate transduction of active enzymeinto various cells lacking functional mannose-specific endo-cytic receptors.

Materials and methods

Plasmid construction and lentivirus production. Plasmids pCR 2.1

TOPO, pcDNA3.1/V5-His B, pLenti6/V5-DEST, and pDONR obtained

Table 1Oligonucleotide primers used in this study

Primer Sequence (5 0 to 30)

PF1 aptatacggtcgtaaaaaacgtcgtcagcgtcgtcgtgcccgccPR1 cacctgatgcccacgacactgcctgaagtagaagcPF2 aptacggtcgtaaaaaacgtcgtcagcgtcgtcgttgatggagcPR2 ctggcgatgccacaggtaggtgtgaatggPF3 gggtaagcatcatggctggcagc

PR3 ctcgagtggcgatgccacaggtaPR4 ctcgagcgacgacgctgacgaPR5 accggtacgacgacgctgacgaPR6 gcctccacctccaccctggcgatgccaPR7 ctcgagcgtcgtcgctgtctccgcttcttcctgcctccacctccacPF4 ggggacaagtttgtacaaaaaagcaggctcaccgtgacaatgc

PF5 caacaagtttgtacaaaaaagcaggctccgcggccgccccttcaPR8 ggggaccactttgtacaagaaagctgggtaggctgatcagcgg

a p represents end nucleotide phosphorylation.

from Invitrogen were used according to the supplied protocols. Full-lengthb-galactosidase and GC were subcloned into the pcDNA3.1/V5-His B toobtain pcDNA3.1/V5-His/lacZ and pcDNA3.1/V5-His/GC. ChimericcDNA fragments that encode the recombinant GCs containing the 11-amino acid TAT PTD (47YGRKKRRQRR57R) on the N-terminus orthe C-terminus of the mature form were made by using PCR-based site-directed mutagenesis in a kit from Stratagene [23]. NTAT-GC-V5-Hiscoding for GC with N-terminal TAT was constructed by inserting TATbetween the 19-amino acid signal peptide sequence and the mature proteinsequence. Three different chimeric GCs containing C-terminal fusions withTAT were also designed. GC-5G-CTAT-V5-His was designed with fiveglycine residues inserted as a flexible spacer between GC and TAT. Inorder to test for any synergistic effect on the transduction properties ofjuxtaposing the 6·His-tag with the TAT fusion on the C-terminus of GCas suggested [19], we have removed the intervening V5 epitope from GC-CTAT-V5-His to generate GC-CTAT-His.

PCR products were first subcloned into a TA cloning vector, pCR 2.1-TOPO, and then the inserts cut and inserted into the pcDNA3.1/V5-His Bexpression vector by digestion with EcoRI/XhoI for pcDNA3.1/V5-His/NTAT-GC, pcDNA3.1/V5-His/GC-CTAT, and pcDNA3.1/V5-His/GC-5G-CTAT and EcoRI/AgeI for pcDNA3.1/His/GC-CTAT. The oligonu-cleotide primers used are shown in Table 1. The Gateway system(Invitrogen) was used to produce lentiviral constructs. The attB-flankedb-galactosidase and GC fragments were amplified using specific attBprimers (PF4, PF5, and PR8 in Table 1) and pcDNA3.1/V5-His/lacZ,pcDNA3.1/V5-His/GC, pcDNA3.1/V5-His/NTAT-GC, pcDNA3.1/V5-His/GC-CTAT, pcDNA3.1/V5-His/GC-5G-CTAT, and pcDNA3.1/His/GC-CTAT plasmids as the PCR templates. The PCR products were usedfor BP recombination using the pDONR vector. Then the resultingpDONR plasmids were used for LR recombination with pLenti6/V5-DEST vector, finally producing pLenti6/V5-6His/lacZ, pLenti6/V5-6His/GC, pLenti6/V5-6His/NTAT-GC, pLenti6/V5-6His/GC-CTAT, pLenti6/V5-6His/GC-5G-CTAT, and pLenti6/6His/GC-CTAT. Production of len-tiviral stocks was carried out in 293FT according to the recommendedprocedure.

Cell culture. Type 1 (DMN 89. 19; N370S/RecNciI) and Type 2 (DMN89. 57; D409H/?) Gaucher fibroblast cells obtained from patients under anapproved National Institutes of Health protocol were immortalized by

Use

cctgcatccctaaaagc pCR 2.1-NTAT-GCpCR 2.1-NTAT-GC

aagggcgaattctgc pCR 2.1-GC-CTATpCR 2.1-GC-CTATpcDNA3.1/V5-His/GCpcDNA3.1/V5-His/NTAT-GCpcDNA3.1/V5-His/GC-CTATpcDNA3.1/V5-His/GCpcDNA3.1/V5-His/GC-CTATpcDNA3.1/His/GC-CTATpLenti6/V5-His/GC-5G-CTAT

c pLenti6/V5-His/GC-5G-CTATagctgag pLenti6/V5-His/GC

pLenti6/V5-His/NTAT-GCpLenti6/V5-His/GC-CTATpLenti6/V5-His/GC-5G-CTATpLenti6/His/GC-CTAT

ccatg pLenti6/V5-His/lacZgtttaaactcaatg pLenti6/V5-His/lacZ

pLenti6/V5-His/GCpLenti6/V5-His/NTAT-GCpLenti6/V5-His/GC-CTATpLenti6/V5-His/GC-5G-CTATpLenti6/His/GC-CTAT

Fig. 1. Schematic representations of designed chimeric proteins in thestudy: b-gal-V5-His (1), GC-V5-His (2), NTAT-GC-V5-His (3),GC-CTAT-V5-His (4), GC-5G-CTAT-V5-His (5), and GC-CTAT-His(6). b-Galactosidase (b-gal), glucocerebrosidase (GC), leader sequence ofthe glucocerebrosidase (Sig), the 11 amino acids of the HIV-1 Tat PTD(TAT), histidine tag (His), V5 epitope (V5), and 5-glycine space (5G) areindicated in the boxes, respectively.

A B

DC

Fig. 2. Western blot analysis showing expression of the chimeric proteinsin 293FT and NIH/3T3 cells: b-gal-V5-His (1), GC-V5-His (2), NTAT-GC-V5-His (3), GC-CTAT-V5-His (4), GC-5G-CTAT-V5-His (5), andGC-CTAT-His (6). The blots were probed with anti-V5-HRP antibody(A,B) or anti-GCR antibody (C,D).

K.O. Lee et al. / Biochemical and Biophysical Research Communications 337 (2005) 701–707 703

SV40 transformation [24]. The GC specific activity of type 1 (DMN 89. 19)and type 2 (DMN 89. 57) cells based on activity units per milligram totalprotein homogenate was equivalent to 6.1% and 2.3% of normal fibroblastactivity, respectively (data not shown). 293FT (Invitrogen), HT1080(ATCC), NIH/3T3 (ATCC), and fibroblasts were grown in Dulbecco�smodified Eagle�s medium (DMEM; Gibco-BRL) containing 10% heat-inactivated fetal bovine serum at 37 �C in a 5% CO2 incubator.

Western blot analysis. The total protein concentration of extractedprotein was determined by using a BCA assay kit (Pierce). Samples weresubjected to SDS–polyacrylamide gel electrophoresis (20 lg protein/lane)and electrophoretically transferred to a polyvinylidene difluoride (PVDF)membrane (Invitrogen) and then probed with monoclonal anti-V5-HRP(Invitrogen) or polyclonal rabbit anti-GCR antisera. For the latter anti-body HRP-conjugated goat anti-rabbit IgG antibody (Cappel) was used asthe secondary antibody. Blots were visualized using the ECL chemilumi-nescence system (Amersham).

Enzyme assay. Cells were harvested 48–72 h post-infection, washedwith phosphate-buffered saline, and extracted in a buffer containing50 mM citric acid/potassium phosphate, pH 6.0, 0.2% (w/v) Triton X-100,and 1% (w/v) sodium taurocholate (Sigma). The cells were disrupted bysonication, and cellular debris was pelleted by centrifugation for 20 min at4 �C and 20,000g. Cell extract (15–20 ll) was added to GC assay buffer(final volume 200 ll) containing 100 mM potassium phosphate, pH 6.0,15 mM 4-methylumbelliferylglucopyranoside (4-MUG; Sigma), 0.15% (w/v) Triton X-100, 0.125% (w/v) sodium taurocholate (Sigma), and 0.1% (w/v) bovine serum albumin (BSA). The reaction was incubated 15 min to 1 hat 37 �C and stopped with 800 ll of stop solution containing 100 mMglycine and 100 mM NaOH. Fluorescence was measured using a fluo-rometer (Photometer 1101M; Eppendorf) or a microtiter plate fluorometer(Victor 1420 multilabel counter; Perkin-Elmer). One GC activity unit (U)releases 1 nmol of 4-methylumbelliferone per hour.

Protein transduction. 293 FT cells were infected with pLenti6/V5-His/lacZ, pLenti6/V5-His/GC, pLenti6/V5-His/GC-CTAT, pLenti6/V5-His/GC-5G-CTAT, and pLenti6/His/GC-CTAT at an MOI of 2–5, and pro-teins were extracted from the cells using a buffer containing 50 mMpotassium citrate/phosphate, pH 6.0, 1% (w/v) sodium taurocholate(Sigma) and passed through PD-10 desalting columns, (Amersham).Protein concentrations and GC activity were determined and 500 U ofeach of the recombinant enzymes to be tested was added to the media ofrecipient cells. For the protein transduction assay type 1 Gaucher (89. 19),type 2 Gaucher (89. 57), and NIH/3T3 cells (each 3 · 105 cells per a well of6-well plate) were plated the day before transduction and incubated withthe recombinant GCs for 2 h at 37 �C. After incubation, cells were har-vested and lysates were prepared for fluorimetric enzyme assay.

Immunocytochemistry. Cells grown in chamber slides were fixed in 3%paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, for 30 min, washed3 times (5 min each) with 0.1 M phosphate buffer, blocked, and perme-abilized with PBS containing 5% goat serum and 0.1% saponin for 1 h.Slides were next incubated for 1 h with anti-V5-FITC antibody (QEDbioscience) diluted 1:200 in PBS containing 2% normal goat serum and0.1% saponin, and then washed 3 times (5 min each) with PBS containing1% goat serum. Finally, the chamber slides (Lab-Tek) were mounted withFluor-Save antifade mounting reagent (Calbiochem) and photographedusing a Zeiss LSM 510 META confocal microscope.

Results

Design of chimeric GCs with TAT

We designed constructs expressing various chimericGCs with or without TAT to investigate whether theTAT PTD facilitates transduction of GC (Fig. 1). Tran-sient expression studies using plasmid vectors did not resultin high enough levels of enzyme to be used for later trans-duction experiments and were discarded in favor of a com-mercially obtained lentiviral expression system. Expression

of GC with an intact termination codon and hence lackingthe V5 and 6· His fusion (not shown) showed no differencein expression of enzyme protein or activity compared withthe V5 and 6· His fusion construct, GC-V5-His. This latterconstruct without TAT was used as a positive control forcomparison of enzyme activity and protein transductioncapability of various N- and C-terminal TAT-containingGC fusion proteins.

Expression of the chimeric GCs in 293FT and NIH/3T3 cells

To produce the recombinant proteins and determinetheir activities, 293FT and NIH/3T3 cells were infectedindividually with the various designed lentiviral constructs(Table 1; pLenti6/V5-His/lacZ, pLenti6/V5-His/GC, pLen-ti6/V5-His/NTAT-GC, pLenti6/V5-His/GC-CTAT, pLen-

ti6/V5-His/GC-5G-CTAT, and pLenti6/His/GC-CTAT).Western blot analysis using anti-V5-HRP confirmed thatthe V5 epitope-tagged proteins, with the exception ofNTAT-GC-V5-His, are full-length (Figs. 2A and B). Theabsence of the V5 epitope on GC-CTAT-His explains thelack of reactivity of this product with the anti-V5-HRP.

Fig. 3. GC activity in 293FT (A) and NIH/3T3 cells (B) after infectionwith the designed lentiviral constructs: pLenti6/V5-His/lacZ (1), pLenti6/V5-His/GC (2), pLenti6/V5-His/NTAT-GC (3), pLenti6/V5-His/GC-CTAT (4), pLenti6/V5-His/GC-5G-CTAT (5), and pLenti6/His/GC-CTAT(6). Specific activity of the GC was expressed as nmol/h/mg protein(mean ± SD, n = 3).

704 K.O. Lee et al. / Biochemical and Biophysical Research Communications 337 (2005) 701–707

Western blot analysis using anti-GCR, a rabbit polyclonalanti-human GC antibody (Figs. 2C and D), showed bandsrelated to the expected recombinant products as well asendogenous human GC produced by the 293FT (humanembryonic kidney) cells. GC-V5-His, GC-CTAT-V5-His,and GC-5G-CTAT-V5-His showed slightly increasedmolecular weights compared with the endogenous GC.The GC-CTAT-His did not show a distinctive size increasecompared with the endogenous GC in the SDS–PAGE.Since the antibody did not show cross-reactivity with themouse GC in NIH/3T3 cells (derived from mouse embry-onic fibroblasts), no signal is seen from endogenous mouseGC. The pattern of expression of human GC in NIH/3T3cells is similar to that obtained from the 293FT cells. Wedid not observe evidence for expression of the NTAT-GC-V5-His on Western blot. TAT inserted between the19 amino acid signal peptide and the mature protein mayprevent recognition by the signal peptidase in the endoplas-mic reticulum or may hamper assumption of the correctconformation and proper processing for the mature pro-tein, possibly resulting in premature proteolysis.

Enzymatic analysis of the expressed chimeric GCs

Expression of the chimeric GCs in 293FT and NIH/3T3cells may also be assessed by means of GC activitymeasurement using 4-MUG as a substrate (Fig. 3). Cellstransduced with pLenti6/V5-His/GC, pLenti6/V5-His/GC-CTAT, pLenti6/V5-His/GC-5G-CTAT, and pLenti6/His/GC-CTAT constructs showed approximately 2- to 3-foldhigher GC activity than untreated cells. The pLenti6/V5-His/GC transduced 293FT and NIH/3T3 cells showedhighest GC activity. The pattern of GC expression assessedby enzyme activity measurement is consistent with the re-sult of Western blot analysis indicating both the highest le-vel of protein expression and enzyme activity in thepLenti6/V5-His/GC transduced cells. The enzyme activityobtained from pLenti6/V5-His/GC transfected cells wassimilar to that obtained from cells transfected with a vectorcontaining the wild-type GC coding sequence (data notshown). Taken together, these results also suggest thatthe fusion protein containing TAT has similar specificactivity to the native protein. The GC activity in pLenti6/His/GC-CTAT infected 293FT and NIH/3T3 cells wasslightly higher than those of pLenti6/V5-His/GC-CTATand pLenti6/V5-His/GC-5G-CTAT infected cells. This indi-cates that all the newly translated chimeric GCs, with theexception of NTAT-GC-V5-His, are enzymatically activeproteins. No product of the N-terminal fusion constructcould be detected. The results of the Western blot analysisand the activity assay indicate that TAT modification tothe C-terminus of GC does not inhibit enzyme activity.Qualitatively similar expression levels, albeit at lower lev-els, were observed when the above constructs were usedto transfect SV40 transformed fibroblasts from patientswith Gaucher disease. Low expression of other unrelatedproteins such as b-galactosidase in Gaucher cells transfec-

ted with the control lentivirus construct pLenti6/V5-His/lacZ suggests that the best explanation for the low expres-sion level in these cells is the low competence of these cellsfor lentiviral infection.

TAT-mediated delivery of GC into cultured cells

To examine the protein transduction of the various fu-sion proteins, we prepared enzyme extracts using sodiumtaurocholate containing buffers, removed excess detergentby use of a desalting column, concentrated the enzyme,and then added equivalent units of each enzyme to themedia of type 1 (DMN 89. 19) and type 2 (DMN 89. 57)Gaucher fibroblasts at 37 �C. Comparing the uptake 2 hafter treatment (Fig. 4) all cells treated with C-terminalTAT-containing enzymes showed increased GC activitycompared with GC-V5-His suggesting that the proteinshave been taken up by the cells. A slight though non-signif-icant increase in GC activity was seen after protein trans-duction with the GC-5G-CTAT-V5-His, when comparedwith the other C-terminal TAT-containing constructs.

Fig. 4. Protein transduction into type 1 and type 2 Gaucher fibroblasts:Imiglucerase (1) GC-V5-His (2), GC-CTAT-V5-His (3), GC-5G-CTAT-V5-His (4), and GC-CTAT-His (5). Specific activity of the GC wasexpressed as nmol/h/mg protein (mean ± SD, n = 3).

K.O. Lee et al. / Biochemical and Biophysical Research Communications 337 (2005) 701–707 705

The possible structural flexibility of TAT-PTD caused by5-glycine space between the protein and TAT-PTD may al-low the domain to interact more easily with unknown com-ponents required for protein internalization.

Additional evidence for HIV-1 TAT-PTD mediated GCimport into cells and their cellular distribution was ob-tained by immunocytochemical analysis using confocalfluorescent microscopy after protein transduction in

Fig. 5. Localization of the recombinant proteins in HT1080 cells after proteinCTAT-V5-His (G–I). Cells were probed with LysoTracker Red DND-99 (A,D,in (C,F,I).

NIH/3T3 (data not shown) and HT1080 (Fig. 5) cells.GC-CTAT-V5-His and GC-5G-CTAT-V5-His showedstrong fluorescence inside the cells suggesting that the pro-teins were effectively delivered into the cells whereas GC-V5-His transduced cells (Fig. 5A) did not show any fluores-cence in the cells. To determine the localization within thetransduced cells, lysosomes were stained with LysoTrackerRed DND-99 (Molecular Probes) and the red fluorescencewas merged with the green fluorescence of the anti-V5-FITC antibody (Fig. 5). We could observe partial co-local-ization of GC-CTAT-V5-His and GC-5G-CTAT-V5-Hiswith lysosomes labeled with a low-pH indicator; however,the transduced GC-CTAT-V5-His and GC-5G-CTAT-V5-His proteins may have also been more generally distrib-uted within the cytosol. This result is consistent with theresult of GC assay after protein transduction which dem-onstrated that the TAT-modified GCs are transduced intothe recipient cells.

Discussion

ERT for Gaucher disease presents some unique andunusual challenges. The distribution of the stored glycolip-id primarily within macrophages and related cell-types, therelatively hydrophobic nature of the enzyme GC, and theabsence of mannose-6-phosphate residues on the native en-zyme led investigators to exploit the mannose-specific

transduction: GC-V5-His (A–C), GC-CTAT-V5-His (D–F), and GC-5G-G) and anti-V5-FITC antibody (B,E,H), and merged fluorescence is shown

706 K.O. Lee et al. / Biochemical and Biophysical Research Communications 337 (2005) 701–707

endocytic route to improve the efficacy of this drug. How-ever, the very efficiency of the mannose-specific targetingmay limit the ability to eliminate stored glycolipids in othertissues. The present work was undertaken to assess the po-tential of using a PTD for improving the general distribu-tion of GC as therapy for Gaucher disease irrespective ofwhether the enzyme is administered exogenously or a prod-uct of gene therapy. Despite its apparent success, standardERT for Gaucher disease based on the internalization ofthe systemically administered mannose-terminated GCmay be limited in effectiveness to those organs containingcell-types expressing high levels of mannose-specificendocytic receptors [8,9]. Even in organs in which thisreceptor exists, the efficiency of the liver and spleen in filter-ing blood, combined with the rapidity with which man-nose-specific endocytosis occurs, limits the availability ofenzyme to other organs especially the lungs and bone mar-row. This limitation in ERT may be overcome by develop-ment of novel strategies for recognition and internalizationof therapeutic enzyme independently of the mannose-spe-cific recognition route. Such a strategy may have broadimplications in many other LSDs in which the aberrantstorage cells to be targeted are not of reticuloendothelialorigin.

We demonstrated the effectiveness of TAT-modified chi-meric constructs pLenti6/V5-His/GC, pLenti6/V5-His/GC-CTAT, pLenti6/V5-His/GC-5G-CTAT, and pLenti6/His/GC-CTAT to express functional GC in 293FT, NIH/3T3cells, and in fibroblasts derived from patients with Gaucherdisease. We have shown that treatment of Gaucher fibro-blasts and NIH/3T3 cells with the TAT-modified GCsresulted in enzyme accumulation in all cells tested. Previousinvestigations have used TAT fusions produced in bacteriaand included a denaturation and renaturation step prior totesting [16]. Using a much less stringent test for successthan the ability to cross the blood–brain barrier as usedby many previous investigators, we have succeeded in dem-onstrating that a normally glycosylated protein producedin a eukaryotic expression system and without prior dena-turation can be internalized into cultured cells in a recep-tor-independent manner.

In the present work, we examined one possible strategyfor improving enzyme therapy by decreasing cell-type spec-ificity but increasing the ability of enzyme to freely crosscell membranes. After systemic administration of proteinor viral vectors containing the cDNA for GC, the thera-peutic effect of the TAT-containing enzyme may be greaterthan that of the native enzyme or one subjected to sequen-tial enzymatic deglycosylation as has been used to preparemannose-terminated GC. The short residence time of glu-cocerebrosidase after internalization [25] may be improvedif an endosomal compartment or the cytosol acts as a depotfor transduced GC-TAT where it resides in an active formand is transported in a sustained manner to the affectedlysosomes. Additional experiments in whole animals arenecessary to clarify and extend these results. These findingsmay help us to develop more effective ERT and gene ther-

apy for patients with Gaucher disease as well as for otherlysosomal storage disorders. Nonetheless, potential im-mune responses and toxicity associated with the PTDs in vi-vo remain to be examined.

Acknowledgments

The authors acknowledge the excellent assistance provid-ed by Dr. Carolyn L. Smith of the NINDS Light ImagingFacility in the immunocytochemical analyses and Mr. J.W.Nagle of the NINDS DNA Sequencing Facility in sequenc-ing the various constructs used in this work. This work wassupported by the Post-doctoral Fellowship Program ofKor-ea Science & Engineering Foundation (KOSEF).

References

[1] E.F. Neufeld, The biochemical basis for mucopolysaccharidoses andmucolipidoses, Prog. Med. Genet. 10 (1974) 81–101.

[2] R.O. Brady, Sphingolipidoses, Annu. Rev. Biochem. 47 (1978) 687–713.

[3] E. Beutler, G.A. Grabowski, The metabolic and molecular bases ofinherited disease, in: C.R. Scriver, W.S. Sly, B. Childs, A.L. Beaudet,D. Valle, K.W. Kinzler, B. Vogelstein (Eds.), Vol. II, eighth ed.,McGraw-Hill, New York, 2001, pp. 3635–3668.

[4] E. Beutler, T. Gelbart, Glucocerebrosidase (Gaucher disease), Hum.Mutat. 8 (1996) 207–213.

[5] R.O. Brady, J.N. Kanfer, D. Shapiro, Metabolism of glucocerebro-sides, II: evidence of enzymatic deficiency in Gaucher�s disease,Biochem. Biophys. Res. Commun. 18 (1965) 221–225.

[6] G.A. Grabowski, Genetic diseases and testing in Ashkenazi Jews: partI, Genet. Test. 1 (1997) 5–12.

[7] R. Schiffmann, R.O. Brady, New prospects for the treatment oflysosomal storage diseases, Drugs 62 (2002) 733–742.

[8] A.G. Morell, G. Gregoriadis, I.H. Scheinberg, J. Hickman, G.Ashwell, The role of sialic acid in determining the survival ofglycoproteins in the circulation, J. Biol. Chem. 246 (1971) 1461–1467.

[9] D. Achord, F. Brot, A. Gonzalez-Noriega, W. Sly, P. Stahl, Humanb-glucuronidase. II. Fate of infused human placental b-glucuronidasein the rat, Pediatr. Res. 11 (1977) 816–822.

[10] F.S. Furbish, C.J. Steer, N.L. Krett, J.A. Barranger, Uptake anddistribution of placental glucocerebrosidase in rat hepatic cells andeffects of sequential deglycosylation, Biochim. Biophys. Acta 673(1981) 425–434.

[11] N.W. Barton, R.O. Brady, J.M. Dambrosia, A.M. DiBisceglie, S.H.Doppelt, S.C. Hill, H.J. Mankin, G.J. Murray, R.I. Parker, C.E.Argoff, R.P. Grewal, K.T. Yu, Replacement therapy for inheritedenzyme deficiency-macrophage targeted glucocerebrosidase forGaucher�s disease, N. Engl. J. Med. 324 (1991) 1464–1470.

[12] L.W. Poll, J.A. Koch, R. Willers, H. Aerts, A. Scherer, D.Haussinger, U. Modder, S. vom Dahl, Correlation of bone marrowresponse with hematological, biochemical, and visceral responses toenzyme replacement therapy of nonneuronopathic (type 1) Gaucherdisease in 30 adult patients, Blood Cells Mol. Dis. 28 (2002) 209–220.

[13] R. Schiffmann, H. Mankin, J.M. Dambrosia, R.J. Xavier, C. Kreps,S.C. Hill, N.W. Barton, D.I. Rosenthal, Decreased bone density insplenectomized Gaucher patients receiving enzyme replacementtherapy, Blood Cells Mol. Dis. 28 (2002) 288–296.

[14] K.E. Bove, C. Daugherty, G.A. Grabowski, Pathological findings inGaucher disease type 2 patients following enzyme therapy, Hum.Pathol. 26 (1995) 1040–1045.

[15] S.R. Schwarze, A. Ho, A. Vocero-Akbani, S.F. Dowdy, In vivoprotein transduction: delivery of a biologically active protein into themouse, Science 285 (1999) 1569–1572.

K.O. Lee et al. / Biochemical and Biophysical Research Communications 337 (2005) 701–707 707

[16] B. Gupta, T.S. Levchenko, V.P. Torchilin, Intracellular delivery oflarge molecules and small particles by cell-penetrating proteins andpeptides, Adv. Drug Deliv. Rev. 57 (2005) 637–651.

[17] A.D. Frankel, C.O. Pabo, Cellular uptake of the Tat protein fromhuman immunodeficiency virus, Cell 55 (1988) 1189–1193.

[18] M. Green, P.M. Loewenstein, R. Pusztai, J.S. Symington, Anadenovirus E1A protein domain activates transcription in vivo andin vitro in the absence of protein synthesis, Cell 53 (1988) 921–926.

[19] S. Fawell, J. Seery, Y. Daikh, C. Moore, L.L. Chen, B. Pepinsky, J.Barsoum, Tat-mediated delivery of heterologous proteins into cells,Proc. Natl. Acad. Sci. USA 91 (1994) 664–668.

[20] E. Kilic, G.P. Dietz, D.M. Hermann, M. Bahr, Intravenous TAT-Bcl-Xl is protective after middle cerebral artery occlusion in mice, Ann.Neurol. 52 (2002) 617–622.

[21] G. Cao, W. Pei, H. Ge, Q. Liang, Y. Luo, F.R. Sharp, A. Lu, R. Ran,S.H. Chen, J. Graham, In vivo delivery of a Bcl-xL fusion protein

containing the TAT protein transduction domain protects againstischemic brain injury and neuronal apoptosis, J. Neurosci. 22 (2002)5423–5431.

[22] H. Xia, Q. Mao, B.L. Davidson, The HIV Tat protein transduc-tion domain improves the biodistribution of b-glucuronidaseexpressed from recombinant viral vectors, Nat. Biotechnol. 19(2001) 640–644.

[23] T.A. Kunkel, Rapid and efficient site-specific mutagenesis withoutphenotypic selection, Proc. Natl. Acad. Sci. USA 82 (1985) 488–492.

[24] W.W. Colby, T. Shenk, Fragments of the simian virus 40 transform-ing gene facilitate transformation of rat embryo cells, Proc. Natl.Acad. Sci. USA 79 (1982) 5189–5193.

[25] G.J. Murray, K.L. Oliver, F.S. Jin, R.O. Brady, Studies on theturnover of exogenous mannose-terminal glucocerebrosidase in ratliver lysosomes, J. Cell Biochem. 57 (1995) 208–217.