Supporting Information

Wang et al. 10.1073/pnas.1222742110 SI Materials and Methods mM; Sigma) or 5 μg/mL of RAP was added to inhibit receptor- Cell Line Maintenance. Cells were cultured in Dulbecco modified mediated uptake 30 min before and during the incubation, in EaglemediumorRPMImedium1640(LifeTechnologies)with10% parallel with untreated wells with -containing medium. (vol/vol) FBS (HyClone), 2 mM , and antibiotics. Cells Each experiment was performed in duplicate. For binding/in- LDLRneg were maintained at 37 °C in a humidified atmosphere containing ternalization assay, CHO -LRP1 cells were exposed at 4 °C for 20 min to IDUA fusion of the same IDUA ac- 5% CO2, and were routinely tested for Mycoplasma infection. tivity (500 U/mL) with or without M6P or RAP inhibitors. After Fluorimetric Enzyme Assay for α-L-Iduronidase. The catalytic activity multiple washes, cells were cultured in fresh medium at 37 °C for of α-L-iduronidase (IDUA) or β- was determined 1 h, followed by IDUA assay using cell lysate as described. with a fluorometric enzyme assay as previously described (1, 2), respectively, with modifications. For IDUA assay, cell pellets or tis- Analysis of Transcytosis with an in Vitro Blood–Brain Barrier Model. μ sues were homogenized in distilled water or lysis buffer (150 mM Upper chambers of six-well Transwell inserts (0.4 M; BD Bio- μ NaCl and 50 mM Tris·HCl with 1% Triton X-100) using an ul- sciences) were precoated with type I collagen (50 g/mL; Santa trasonic processor (GE). Aliquots (10 μL) of cleared lysate, plasma Cruz ). Bovine brain microvascular endothelial cells or culture medium were incubated at 37 °C for 15 to 120 min with (BMEC) (CRL-12414; American Type Culture Collection) were 2.5 mM fluorogenic substrate (25 μL) and 4-methylumbelliferyl cultured in them for 2 d, followed by the addition of medium ′ (4MU)-α-L-idopyranosiduronic acid sodium salt (Toronto Re- containing IDUA3 myc or IDUA fusion candidates at 4 °C for 30 μ search Chemicals) in 0.4 M NaFormate buffer (pH 3.2), followed min in the presence or absence of RAP (5 g/mL). After rinsing by addition of 1 mL of glycine carbonate buffer (0.1 M, pH 10.5). with PBS solution three times, the cells were transferred to a new β chamber and cultured at 37 °C with fresh culture medium for 3 h. For determination of -hexosaminidase activity, brain tissues were μ homogenized in 0.9% sodium chloride solution containing 1% Medium in the lower chambers (300 L) was collected at varying time points during incubation. Modified IDUA was cap- Triton X-100, followed by dilution (1:250 and 1:500) in distilled – – water. The diluted homogenates (10 μL) were incubated with tured by IP overnight at 4 °C using anti myc- coated μ β Protein A/G beads (30 μL) and quantified by IDUA enzyme assay 10 L of 1.2 mM 4MU- -N-acetylglucosaminide (Sigma) in 10 mM ′ citrate/20 mM phosphate buffer, pH 4.5, for 1 h at 37 °C. The re- as described earlier. Each experiment included IDUA3 myc in action was stopped by the addition of 500 μL of glycine carbonate parallel with all three selected IDUA-Rb proteins to control for buffer (pH 10.5). The fluorescent product released from each re- leakage of BMEC. Transcytosis assays involving RAP were per- formed in parallel with untreated assays. action was analyzed with an emission wavelength of 450 nm and an excitation wavelength of 365 nm by using a SpectraMax M5 fluo- Cross-Correction and Analysis of Lysosomal Morphology. To de- rometer (MDS Analytical Technologies). Samples were assayed in fi fi termine potential metabolic cross-correction by modi ed IDUA, duplicate and quanti ed in duplicate wells, together with buffer primary fibroblasts from a patient with controls in parallel. Protein was determined by Coomassie blue (MPS) I, which are suitable for evaluating treatment of the at- dye-binding assay (BioRad) or Pierce BCA Protein Assay (Thermo tenuated forms of the disease, were purchased from Coriell Cell Fisher Scientific). One unit of enzyme activity is defined by the Repositories. Cells that had been growth on poly-L-lysine (Sigma)– release of 1 nmol of 4MU in a 1-h reaction at 37 °C. Intracellular coated cover slides in the low chambers of Transwell inserts (0.4 IDUA activity was calculated as units per milligram of protein, and μM) were cocultured for 24 h with HEK293-based cell lines that extracellular IDUA activity as units per milliliter of medium or overexpressed IDUA or IDUA fusion candidates. Fibroblasts were plasma. Two to four brain specimens were assayed for each animal. washed with PBS solution and incubated with 75 nM LysoTracker fi μ red (Invitrogen) for 1 h at 37 °C. The slides were xed with 4% for Captured Activity. Medium (500 L) pre- (vol/vol) paraformaldehyde and mounted by using Vectashield conditioned by 24-h culture of HEK293-based cells that stably medium with DAPI (Vector Laboratories) for fluorescence mi- overexpressed different IDUA receptor-binding peptides (Rb) croscopy with a Leica DMI6000 B system. were incubated with protein A/G beads (50 μL; Invitrogen) pre- – μ coated with mouse anti c-Myc monoclonal antibody (5 L; Santa Gene Transfer by Hydrodynamic Injection. Plasmid DNA (50 μg) Cruz Biotechnology) overnight at 4 °C. For IP experiments to expressing a selected IDUA fusion protein from a liver-specific fi evaluate LRP1 binding, mouse embryonic broblast-based cells promoter were injected i.v. into the tail vein of 7- to 8 wk-old MPS I were exposed to IDUA fusion proteins with the same IDUA ac- mice in a volume of saline solution equivalent to 10% (vol/wt) of tivity (500 U/mL) at 4 °C for 20 min, followed by rinsing three body mass over a period of 5- to 8 s by using a 26-gauge insulin times with PBS solution. Immunoprecipitation (IP) was performed syringe. Mice were monitored for recovery to normal activity with cell lysates by using protein beads precoated with rabbit anti- within 5 min after the injection. Blood specimens were collected LRP1 polyclonal antibody (1:200; 0.4 mg/mL) (3). The binding periodically from the tail vein to monitor plasma IDUA activity. was conducted in parallel with untreated wells by the addition of 5 μg/mL of recombinant receptor-associated protein (RAP) (4) Animal Perfusion. To reduce the effect of blood IDUA on brain 30 min before and during the incubation of IDUA fusion proteins. enzyme levels, mice were anesthetized by i.p. injection with pen- Postprecipitation beads were washed three times, followed by tobarbital (Abbott Laboratories), followed by transcardial perfu- assay of IDUA activity in duplicate as described earlier. sion with cold saline solution via the aorta to remove blood from the vasculature. The success of this procedure was confirmed by a loss Inhibitory Uptake and Binding/Internalization Assays. To evaluate of color in the liver and the blood vessels that flank the midline of enzyme uptake mediated by specific receptor pathways, cells (2 × the rib cage. 105 per well) were seeded in 24-well plates. One day later, cells were incubated for 3 h at 37 °C with 5% CO2 with medium Depletion of Brain Capillaries. Freshly removed mouse brains were containing similar levels of various IDUA candidates. M6P (1 homogenized on ice with 10 strokes in capillary depletion buffer,

Wang et al. www.pnas.org/cgi/content/short/1222742110 1of6 followed by dextran-gradient centrifugation (16%, vol/vol) at (1 U/μL) for 30 min at 37 °C. Soluble GAG was quantified in du- 5,400 × g for 15 min at 4 °C (5). The supernatant suspensions of plicate by using 1,9-dimethylmethylene blue chloride dye as de- brain tissues and the pellets containing the capillaries were scribed (6) with modification. Absorbance of the color reaction was carefully separated and washed three times with PBS solution, measured at 656 nm by using a SpectraMax M2 microplate reader followed by analysis of IDUA activity. and compared with a standard curve generated with heparan sul- fate standard solutions (Sigma). Protein concentration was mea- fl Immuno uorescence Staining of Brain Tissues. A portion of the sured by using a BCA Protein Assay kit (Thermo Fisher Scientific). fi cortical region of the forebrain from perfused mice was post xed in All GAG values were normalized to the amounts of protein. 4% paraformaldehyde, followed by incubation overnight at 4 °C in 4% paraformaldehyde containing 30% (wt/vol) sucrose. Frozen Quantification of Transgene Frequency by Quantitative PCR. To sections (10 μm) were obtained and permeabilized with PBS so- quantify transgene frequency, genomic DNA isolated from the bone lution containing 0.1% Triton X-100 and treated with blocking marrow 5 mo after transplantation was analyzed by real-time quan- solution (PBS solution containing 5% horse serum and 0.1% Tri- titative PCR (qPCR) for IDUA cDNA and murine apolipoprotein B ton X-100). Sections were incubated overnight at 4 °C with the (apoB) DNA withTaqMan primer/probes.The primers and probe for primary antibody sheep anti-hIDUA (1:100; R&D Systems), IDUA cDNA were designed by Primer Expression 3.0 software and mouse c-Myc antibody (9E10; 1:500; Santa Cruz Biotechnology), the sequences were as follows: 5′-CTGGTCTGGTCGGATGAA- rabbit anti-NeuN (1:200; EnCor Biotechnology), rat anti-mouse CA -3′ (sense), 5′-CCGTCCTGAGAGAACTGGATCT -3′ (an- CD31 (1:200; BD Pharmingen), and rabbit anti-GFAP (1:250; tisense), and 5′-FAM-TCCAAGTGCCTGTGGAC-3′ (probe). Abcam). Slices were washed three times with PBS solution and Amplification of endogenous apoB with VIC-labeled probe serves incubated with a secondary antibody of the appropriate species, as an internal control for normalization of DNA as described pre- including Alexa 488 donkey anti-sheep (1:500; Invitrogen), Alexa viously (7). The standard curve for IDUA cDNA frequency (ranging 488 goat anti-mouse (1:500; Invitrogen), Alexa 568 goat anti-rabbit (1:500), and Alexa 568 goat anti-rat (1:500). Slices were mounted from 0.1% to 100%) was established from (Ct) values using a set of with Vectashield medium containing DAPI (Vector Laboratories), standard samples generated from a mixture of 3T3-IDUA cells and analyzed by using a DMI6000 B microscope system. (containing approximately one copy IDUA determined by ) and 3T3 cells. Each real-time reaction contained 100 ng of Quantification of . Two to three aliquots of genomic DNA, 300 nM of IDUA or ApoB primers, 200 nM frozen brain specimens were homogenized in 10% vol/weight of TaqMan IDUA or ApoB probe, and 10 μL of TaqMan 2× Uni- water. Equivalent amounts of brain samples (∼1 mg of protein) versal Master Mix (Applied Biosystems). Amplification con- were defatted by treatment with chloroform:methanol (1:2) solu- ditions were 2 min at 50 °C and 10 min at 95 °C for the initial cycle, tion, washed with 100% ethanol, followed by digestion with papain followed by 40 cycles at 95 °C for 15 s and 60 °C for 1 min. Each in 100 mM sodium acetate buffer, pH 5.5, containing 5 mM cys- sample was analyzed in two separate qPCR assays with duplicate teine and 5 mM EDTA (Sigma). Samples were treated with DNase amplifications.

1. Wang D, Worsham DN, Pan D (2008) Co-expression of MGMT(P140K) and alpha-L- receptor-related protein. Multiple regulatory sites for 39-kDa receptor-associated iduronidase in primary hepatocytes from mucopolysaccharidosis type I mice enables protein. J Biol Chem 267(13):9035–9040. efficient selection with metabolic correction. J Gene Med 10(3):249–259. 5. El-Amouri SS, Cao P, Miao C, Pan D (2013) Secreted luciferase for in vivo evaluation of 2. Zheng Y, et al. (2003) Treatment of the mouse model of mucopolysaccharidosis I with systemic protein delivery in mice. Mol Biotechnol 53(1):63–73. retrovirally transduced bone marrow. Mol Genet Metab 79(4):233–244. 6. Barbosa I, et al. (2003) Improved and simple micro assay for sulfated glycosaminoglycans 3. Basford JE, et al. (2011) Hepatic deficiency of low density lipoprotein receptor-related quantification in biological extracts and its use in skin and muscle tissue studies. protein-1 reduces high density lipoprotein secretion and plasma levels in mice. J Biol Glycobiology 13(9):647–653. Chem 286(15):13079–13087. 7. Worsham DN, Schuesler T, von Kalle C, Pan D (2006) In vivo gene transfer into adult 4. Williams SE, Ashcom JD, Argraves WS, Strickland DK (1992) A novel mechanism for stem cells in unconditioned mice by in situ delivery of a lentiviral vector. Mol Ther controlling the activity of alpha 2-macroglobulin receptor/low density lipoprotein 14(4):514–524. IDUA activityin Medium 100000 100000 **

10000 ** 10000 (unit/ml/TE)

1000 1000

100 100

10 10 (unit/mg/TE) 1 1 IDUA Acitivity in Lysate Lysate in Acitivity IDUA

Intracellular Extracellular

Fig. S1. Catalytic function of IDUA-Myc fusion proteins. Human HEK293 cells were harvested 3 d after cotransfection of the IDUA-containing and GFP- containing plasmids, after a 24-h medium collection. Protein concentration was determined by Coomassie blue dye-binding assay. All activities were nor- malized by transfection efficiencies (TE) determined by using FACS analysis for GFP+ cells (62–70%). Data were derived from two experiments in duplicate wells (**P < 0.01, Student t test).

Wang et al. www.pnas.org/cgi/content/short/1222742110 2of6 A B 1000 200 Bead 100 160 80-85 IDUA Activity in -

kD depleted Medium 10 120 (unit/ml)

1 80

(unit/ml bead) 0.1 40

Captured IDUA Activity Activity IDUA Captured 0.01 0 Anti-Myc ++ + Anti-Myc + + + IgG ++ Captured IgG + + + Medium IDUA IDUA3’Myc IDUA5’Myc IDUA IDUA3’Myc IDUA5’Myc

Fig. S2. Binding ability of released form myc-tagged IDUA. Medium preconditioned by 24-h culture of HEK293 cells overexpressing natural or myc-tagged IDUA were normalized to 200 U/mL, followed by IP with monoclonal antibody against myc tag or IgG control. (A) Western blot analysis of immunoprecipitated proteins. (Upper) Rabbit anti-Myc polyclonal . (Lower) Coomassie blue staining of SDS gel showing coated antibodies as loading controls. (B) IDUA activity in beads and medium after IP. The same initial amounts of functional IDUA were applied to all IP reactions.

A B LysoTracker Myc Merge

3T3 + IDUA

3T3 + IDUA3’Myc

Fig. S3. M6P receptor-specific uptake of IDUA-myc fusion proteins. (A) Inhibition of uptake in NIH 3T3 cells by M6P competition. NIH 3T3 cells were exposed for 2 h to the same volume of medium preconditioned by 24-h culture of enzyme-overexpressing HEK293 cells in the presence of various concentration of M6P inhibitor. The IDUA enzyme activity in medium were 2,267 U/mL for IDUA and 2,230 U/mL for IDUA3′Myc. Data were derived from two experiments, each performed in duplicate wells. (B) Colocalization of endocytosed IDUA3′Myc with marker. The 3T3 cells grown on cover slides (in lower chambers) were cocultured with cells stably overexpressing IDUA or IDUA3′Myc on Transwell inserts.

Wang et al. www.pnas.org/cgi/content/short/1222742110 3of6 A 10000000 Intracellular Extracellular 1000000

100000

10000

1000

100

IDUA Activity Activity IDUA 10

1 (unit/ml or mg protein) (unit/ml or mg

B

80-85

Fig. S4. Expression and release profile of LDL receptor superfamily-targeted fusion IDUA. (A) IDUA catalytic activities in cell lysate or 24-h precondition medium (106 cells/3 mL) from different 3T3 cell-based fusion IDUA-overexpressing cell lines. Stable, IDUA-expressing cell lines were obtained by cotransfecting cells with two plasmids expressing fusion IDUA or EGFP from CMV promoter, and subsequent selection of transfectants by G418. All assays were performed in triplicate. Error bars indicate SDs. (B) Western blot analysis of myc-tagged fusion IDUA in medium. The rabbit anti-myc polyclonal antibody was used to each lane loaded with 50 μL 24-h precondition medium collected as described in B. Culture medium contained 10% FBS.

1 1 P P R R L L g g - g - e e g e n n e n R n 1 R L 1 P L D P D L R R L L L O O H EF EF C CH M M

LRP-1 -actin

Fig. S5. Western blot analysis for LRP1 protein expression in genetically engineered cell lines. Whole-cell lysate (20 μg per lane) from indicated cell lines were loaded, and blots were developed with antibodies specific for human LRP-1 or murine β-actin (as loading controls).

Wang et al. www.pnas.org/cgi/content/short/1222742110 4of6 120% HEK293 100% CHO-LDLRneg 80% MEF-LRP1neg

60%

40%

IDUA3'Myc Uptake IDUA3'Myc 20%

0% 0 500 1000 1500 2000 2500 M6P Inhibitor ( M)

Fig. S6. Dose-dependent inhibition of IDUA3′Myc uptake in different cell lines by M6P competitor. All cells were cultured for 2 h with similar amounts of IDUA3′Myc (∼500 U/mL) in the absence or presence of increasing amounts of M6P inhibitor. The data were derived from two individual experiments with all cell lines tested in parallel in duplicated cultures. Error bars indicate SDs.

10000

MPS/IDUA3'Myc 1000 MPS/IDUAe1

100 MPS/IDUAe2 Carrier 10

1 IDUA in Plasma (unit/ml) Plasma in IDUA 0.1 ND Day 0Day 1Day 2Day 3

Duration after Injection

Fig. S7. Time-dependent overexpression of IDUA fusion proteins in blood circulation of MPS mice after hydrodynamic tail-vein injection. Blood samples were collected and analyzed at indicated time points from MPS I mice hydrodynamically injected with plasmids expressing fusion IDUA from liver-specific promoter (n = 4–8). The data were derived from two or three individual injection experiments, with each sample assayed in duplicate reactions. Plasma IDUA levels were undetectable (ND) in uninjected MPS control mice. Error bars indicate SDs.

1

0.1

(copy/cell) 0.01 y = 5E+07e-0.771x Transgene Frequency Frequency Transgene R² = 0.999 0.001 15 25 35 Threshold Cycle (Ct)

Fig. S8. Standard curve for real-time qPCR using TaqMan primer/probes. The standard curve for IDUA cDNA frequency was established by using genomic DNA samples isolated from a set of mixture of 3T3-IDUA cells (containing approximately one copy IDUA cDNA determined by Western blot) and 3T3 cells (zero copies). Data were derived from two sets of standard samples, each amplified twice in duplicate reactions.

Wang et al. www.pnas.org/cgi/content/short/1222742110 5of6 Table S1. Peptide sequence design as receptor-binding domain candidates for fusion IDUA construction Rb ID Size (monomer size × n) Location Accession no.

myc 10 + 5* Residual 410–419 P01106 apoB 39 Residual 3371–3409 P04114 apoE1 18 (9 × 2) Residual 159–167 P02649.1 apoE2 30 (15 × 2) Residual 159–173 P02649.1 apoE3 17 Residual 151–167 P02649.1 apoE4 20 Residual 148–167 P02649.1 apoE5 34 (17 × 2) Residual 151–167 P02649.1 apoE6 40 (20 × 2) Residual 148–167 P02649.1

apoE, apolipoprotein E. *A 5-aa space-linker (IDILE) was included following myc-tag because of the addition of a polylinker upstream to the stop codon.

Wang et al. www.pnas.org/cgi/content/short/1222742110 6of6