Mild Feline Mucopolysaccharidosis Type VI

The missense mutation, L476P, in theN-acetylgalactosamine 4-sulfatase (4S) gene, has previously been shown to be associated with a severe feline mucopolysaccharidosis type VI (MPS VI) phenotype. The present study describes a second mutation, D520N, in the same MPS VI cat colony, which is inherited independently of L476P and is associated with a clinically mild MPS VI phenotype in D520N/L476P compound heterozygous cats. Biochemical and clinical assessment of L476P homozygous, D520N/L476P compound heterozygous, and D520N homozygous cats demonstrated that the entire range of clinical phenotypes, from severe MPS VI, to mild MPS VI, to normal are clustered within a narrow range of residual 4S activity from 0.5% to 4.6% of normal levels. When overexpressed in CHO-KI cells, the secreted form of D520N 4S was inactivated in neutral pH conditions. In addition, intracellular D520N 4S protein was rapidly degraded and corresponded to 37%, 14.5%, and 0.67% of normal 4S protein levels in the microsomal, endosomal, and lysosomal compartments, respectively. However, the specific activity of lysosomal D520N 4S was elevated 22.5-fold when compared with wild-type 4S. These results suggest that the D520N mutation causes a rapid degradation of 4S protein. The effect of this is partially ameliorated as a result of a significant elevation in the specific activity of mutant D520N 4S reaching the lysosomal compartment.

The mucopolysaccharidoses (MPS) 1 are a group of lysosomal storage disorders that involve the deficiency of specific enzymes required for the degradation of glycosaminoglycans (GAG). MPS VI is characterized by a deficiency of N-acetylgalactosamine 4-sulfatase (4S), which leads to the lysosomal accumulation and urinary excretion of the GAG dermatan sulfate (DS) (1). The severe MPS VI phenotype is characterized by growth retardation, coarse facial features, joint stiffness, cor-neal clouding, skeletal deformities, and organ and soft tissue involvement. As a result of DS storage in the heart valve and lung, the normal function of these organs is often compromised, leading to early death in affected individuals. The central nervous system does not appear to be affected, even in individuals with clinically severe MPS VI (2).
Patients with MPS VI are diagnosed by elevated levels of DS in their urine and a substantial reduction or lack of 4S activity in cells such as peripheral blood leukocytes and cultured skin fibroblasts. The clinical phenotype of MPS VI patients ranges from severe to relatively mild and has generally been shown to correlate with urinary DS and residual 4S activity levels (2). Generally, patients manifest symptomology associated with MPS VI when their 4S protein and activity levels are 5% or less of normal controls (3).
Currently, there is no safe and effective treatment available for MPS patients other than surgical intervention to alleviate symptoms whenever possible. MPS VI is a good candidate for both enzyme and gene replacement therapy protocols for a number of reasons. First, the complication of targeting recombinant enzyme across the blood-brain barrier is negated by the lack of central nervous system pathology in MPS VI patients. Second, circulating levels of recombinant enzyme can be efficiently targeted to the lysosomal compartment of MPS VI cells as a result of mannose 6-phosphate (M6P) receptor-mediated endocytosis of 4S (4).
The M6P moieties on lysosomal enzymes serve as high affinity ligands for binding to M6P receptors in the trans-Golgi network. Binding with M6P receptors permits lysosomal enzymes to be segregated from proteins destined for secretion or transport to other intracellular compartments. The ligand-receptor complex is then transported from the trans-Golgi network via a clathrin-coated vesicle to an acidified endosomal compartment, where the M6P receptor and lysosomal enzyme dissociate. Further processing of the enzyme occurs in the lysosome. The M6P receptor is either recycled from the endosome back to the Golgi for further targeting of endogenous de novo synthesized lysosomal enzymes or transported to the plasma membrane, where it functions to internalize exogenous lysosomal enzymes (4).
It is imperative that, where possible, enzyme and gene replacement therapy protocols for MPS disorders be extensively evaluated in large animal models prior to clinical application. A naturally occurring feline model for a severe form of MPS VI has previously been described (5), and a colony has been established in Adelaide to study the progression of the disease and evaluate the efficacy of therapy protocols. The intravenous administration of recombinant human 4S (rh4S) to severely affected MPS VI cats has been shown to dramatically reduce disease progression in a dose-dependent manner (6,7). However, three cats placed on enzyme replacement therapy dis-played significant antibody titers to the human enzyme, which may have reduced the availability and subsequent efficacy of 4S (8,9). When using animal models to evaluate therapy protocols, it may be advantageous to use species-specific material in an attempt to avoid the complication of immunological responses to foreign proteins. As an initial step toward speciesspecific enzyme and gene replacement therapy in the MPS VI cat, the feline 4S (f4S) cDNA has been isolated and recombinant enzyme expressed and characterized (10).
The isolation of the f4S cDNA has also permitted detailed molecular analysis of mutations within this MPS VI cat colony. We have previously identified a missense mutation, L476P, which causes the clinically severe form of feline MPS VI characteristic of the colony, and we have developed a rapid PCRbased screening method to detect this mutation. L476P homozygous cats contain very low levels of 4S activity in blood leukocytes and display symptoms and disease progression identical to the severe form of human MPS VI. However, a number of colony cats that had no clinical indications of MPS VI also had very low or, in some instances, undetectable 4S activity levels in their blood leukocytes despite being genotyped as normal with respect to the L476P mutation. These observations suggested the presence of a second mutation (10). This study describes the identification and molecular characterization of this second mutation as the missense mutation, D520N. A PCR-based screening method enabled the identification of L476P homozygous cats, D520N/L476P compound heterozygous cats, and D520N homozygous cats in the colony. Clinical assessments of these cats have previously demonstrated that the three genotypes are associated with a severe MPS VI phenotype, mild MPS VI phenotype, and normal phenotype, respectively (10,11). In this study, we determined the 35 S-labeled GAG storage and 4S activity levels in D520N/L476P compound heterozygous and D520N homozygous skin fibroblasts. In addition, the D520N mutation was engineered into the wild-type f4S cDNA and expressed in CHO-KI cells to determine its effect upon 4S synthesis, specific activity, intracellular processing, and trafficking.

EXPERIMENTAL PROCEDURES
Mutation Analysis of the f4S cDNA-Prior to the identification of D520N, cats were only screened for the L476P mutation using a PCRbased screening method (10). Cat 237 was identified to be homozygous for the normal 4S allele at position 476 (Leu-476) via mutation analysis, but had no detectable leukocyte 4S activity. Cultured skin fibroblasts from cat 237 were used to isolate poly(A) ϩ RNA using oligo(dT) Dynabeads (Dynal). First strand cDNA synthesis (100-l reaction volume) was carried out at 37°C for 1 h in the presence of 1ϫ first strand buffer (Life Technologies, Inc.), 100 ng of poly(A) ϩ RNA, 2 g of oligo(dT) primers, 1 mM dNTPs, and 200 units of reverse transcriptase (Life Technologies, Inc.). PCR reactions (100-l reaction volume) contained 5 l of template DNA, 1ϫ PCR buffer (Boehringer Mannheim), 10% dimethyl sulfoxide, 25 mM MgCl 2 , 600 ng of each oligonucleotide, 400 M dNTPs, and 2 units of Taq DNA polymerase (Boehringer Mannheim). PCR conditions were 30 cycles of the following: denaturation at 94°C for 45 s, annealing at 60°C for 45 s, and then extension at 72°C for 1 min. PCR products encompassing the coding region of f4S were purified using a QIAquick PCR purification kit (QIAGEN) and directly sequenced using a fmol sequencing kit (Promega). The position and sequence of oligonucleotides used for PCR or PCR sequencing are listed in Table I of Ref. 10.
Screening the MPS VI Cat Colony for L476P and D520N-All animals used in these studies were established from heterozygous cats (MPS VI-3 family) obtained from Mark Haskins (University of Pennsylvania, Philadelphia, PA) (5). PCR products encoding f4S cDNA nucleotides 1353-1649 were amplified from cat venous blood using oligonucleotides f4S-7 and f4S-9 and Taq DNA polymerase (Boehringer Mannheim) as described previously (10). The L476P mutation was detected via restriction analysis of PCR products with HaeIII (Boehringer Mannheim). The presence of the mutant Pro-476 allele (CCC) results in the introduction of a HaeIII restriction site at nucleotide 1427 (10). This PCR product also spans the D520N mutation and can therefore be used to identify the second mutation. The D520N mutation was detected via restriction analysis of PCR products with BslI (New England Biolabs, Inc). The presence of the mutant Asn-520 allele (AAC) results in the destruction of a BslI site at nucleotide 1558.
Construction of D520N 4S-To determine whether the D520N mutation affects 4S activity, it was engineered into the normal f4S cDNA sequence. Nucleotides 1148 -1608 were PCR-amplified from cat 237 (homozygous for the D520N mutation) first strand cDNA using oligonucleotides 4SP-31 and f4S-11 (600 ng; see Table I of Ref. 10) and 2 units of Taq DNA polymerase. The PCR product was digested with SphI and XhoI (Boehringer Mannheim), gel-purified, and cloned into the normal f4S cDNA (10) digested with the same restriction endonucleases. Nucleotides 1148 -1608 of 24 individual clones were PCR-amplified using primers 4SP-31 and f4S-11. BslI restriction analysis of each PCR product was used to identify f4S clones containing the D520N mutation. Mutant f4S clones were sequenced with a fmol sequencing kit (Promega), and a clone that contained the D520N mutation and no other changes was then excised from pBluescript SK Ϫ with EcoRI and cloned into the mammalian expression vector pEFNeo (pEFNeoD520N) (12). Restriction analysis was used to identify recombinants containing the 4S construct in the correct orientation. The wild-type f4S cDNA, which has previously been cloned into pEFNeo (pEFNeof4S), was used as a positive control (10).
Percoll Granular Fractionation-G418-resistant mass cultures of CHO-KI cells, expressing wild-type 4S and D520N 4S (CHOf4S and CHOD520N, respectively) were grown to confluence in 75-cm 2 flasks. Two 75-cm 2 flasks of CHOf4S and CHOD520N cells were metabolically labeled with 100 Ci/ml EXPRE 35 S 35 S protein labeling mix (NEN Life Science Products) for 45 min. The labeling medium was removed and replaced with Ham's F-12 supplemented with 10% FCS. After 4 h, cells were harvested by treatment with trypsin-versene (Life Technologies, Inc), recovered by centrifugation, washed three times with phosphatebuffered saline (PBS) by centrifugation/resuspension, and finally resuspended in 3 ml of 0.25 M sucrose, 10 mM Hepes, pH 7. Intracellular organelles were released by six exposures to hypobaric shock (14). The lysate was centrifuged at 200 ϫ g for 5 min to remove nuclei. The supernatant was then layered on top of 17 ml of 0.25 M sucrose, 10 mM Hepes, 18% (v/v) Percoll (Amersham Pharmacia Biotech) and centrifuged at 50,000 ϫ g for 1 h (15). One-ml fractions were collected from the Percoll gradient and assayed for total protein (Bio-Rad protein assay), acid phosphatase (a marker for endosomes and lysosomes) (16), the lysosomal enzyme ␤-hexosaminidase (17), and total arylsulfatase using 4-methylumbelliferyl sulfate (4MUS) (18). Fractions corresponding to the microsomal, endosomal, and lysosomal compartments were pooled. Seventy-five microliters of each pooled fraction was used for the immunoquantification of 4S protein and Western blot analysis (see below). The remainder of each pooled fraction was combined with an equal volume of 2ϫ solubilization buffer (PBS containing 2% w/v sodium deoxycholate, 0.2% w/v SDS, 1% v/v Nonidet P-40) and used for immunoprecipitation (see below).
Stability of Intracellular and Secreted 35 S-Labeled D520N 4S-G418-resistant mass cultures of CHOf4S and CHOD520N were grown to confluence in 25-cm 2 tissue culture flasks and then metabolically labeled with 100 Ci/ml EXPRE 35 S 35 S protein labeling mix for 1 h. Labeling medium was removed, and cells were rinsed in serum-free medium and then chased with 6 ml of Ham's F-12 supplemented with 10% FCS and 5 mM M6P (to prevent re-uptake of secreted precursor 4S). After 4, 10, 24, and 48 h, the chased medium in one flask was collected and clarified by centrifugation (1000 ϫ g, 5 min, 4°C). At each time point, cells from one flask in each group were also harvested with trypsin-versene, washed, and resuspended in 5 ml of 1ϫ solubilization buffer (PBS containing 1% w/v sodium deoxycholate, 0.1% w/v SDS, 0.5% v/v Nonidet P-40) and used for immunoprecipitation (see below).
Western Blot Analysis-Thirty microliters of each pooled endosomal and lysosomal CHOD520N 4S and CHOf4S fraction (see "Percoll Granular Fractionation") was microcentrifuged for 10 min at 4°C to remove Percoll and subjected to Western blot analysis as described previously (19) using a rabbit polyclonal antibody raised against rf4S (titer 1/131,000) as the primary antibody.
Immunoquantification of Wild-type 4S and D520N 4S-The human 4S monoclonal antibodies F58. 3 and F22.1 were tested for their ability to cross-react with wild-type 4S and D520N 4S secreted from CHO-KI cells using a sandwich ELISA as described in Ref. 20. Rabbit and murine polyclonal ␣rf4S sera were raised against immunopurified rf4S (titer 1/131,000 and 1/800,000, respectively). The specificity of both rf4S polyclonal antibodies was demonstrated by immunoprecipitation of 4S present in CHOf4S and CHOD520N samples. A sandwich ELISA was used to quantitate 4S protein in CHOf4S and CHOD520N microsomal, endosomal, and lysosomal fractions isolated from Percoll gradients. Approximately 11 mg of total IgG was purified from 1.5 ml of rabbit ␣rf4S serum using a 1-ml protein G column (Amersham Pharmacia Biotech). Total IgG isolated from rabbit ␣rf4S serum was diluted in 0.1 M NaHCO 3 , pH 10, to a final concentration of 10 g/ml and added to individual wells of a polyvinyl chloride plate (Costar, Cambridge, MA) and incubated at 4°C overnight (100 l/well). Each well was then aspirated, and the remaining reactive sites were blocked by adding 200 l of buffer A (0.02 M Tris-HCl, pH 7.0, 0.25 M NaCl, 1% (w/v) ovalbumin) and incubating for 2 h at 22°C. The wells were then aspirated, washed three times with 0.02 M Tris-HCl, pH 7.0, 0.25 M NaCl (buffer B), and then incubated with 100 l of samples diluted in buffer A at 4°C overnight. Each well was aspirated, washed three times with 200 l of buffer B, and then incubated with 100 l of the murine ␣rf4S polyclonal serum diluted 1/10,000 in buffer A. After a 4-h incubation at 22°C, each well was aspirated, washed, and incubated with an affinity-purified, horseradish peroxidase-conjugated sheep anti-mouse immunoglobulin second antibody (1:1000 dilution (v/v) in buffer A) (Silenus Laboratories) for 1 h at 22°C. The antibody was removed and the wells washed three times with 200 l of buffer B. The wash buffer was aspirated, and 100 l of peroxidase substrate (ABTS substrate kit, Bio-Rad) was added to each well. After 10 -20 min, color development was quantified by measuring absorbance at 414 nm on an automated ELISA reader (Titertek Multiscan; Flow Laboratories). All results were extrapolated through a standard curve and were expressed as nanograms of 4S protein/mg of total cell protein. Analysis of conditioned media from CHO-KI cells expressing D520N 4S, before and after inactivation at pH 7, showed that the immunoquantification assay can detect both active and inactive D520N 4S with equal efficiency.
Analysis of GAG Storage in Cultured Fibroblasts-Previous experiments in our laboratory have demonstrated a direct correlation between the amount of intracellular lysosomal 4S and 35 S-labeled storage material in Na 2 35 SO 4 -labeled human MPS VI fibroblasts (21). Analysis of GAG storage in cultured feline skin fibroblasts was determined using the method described by Harper et al. (21), with the following modifications. Triplicate cultures of normal, D520N homozygous, D520N/ L476P compound heterozygous, and L476P homozygous skin fibroblasts were grown to confluence in 25-cm 2 flasks. Growth medium was Dulbecco's modified Eagle's medium supplemented with 10% (v/v) FCS. Fibroblasts were labeled with 15 Ci/ml Na 2 35 SO 4 (543 mCi/mmol; NEN Life Science Products) for 24 h in Ham's F-12 medium supplemented with 10% (v/v) FCS. Labeled cells were then rinsed with PBS (5 ml/flask), harvested by treatment with trypsin-versene, and transferred to new 25-cm 2 flasks containing 7 ml of Dulbecco's modified Eagle's medium supplemented with 10% (v/v) FCS. After 48 h, cells were harvested, lysates were prepared as described below and assayed for total protein, 35 S radioactivity, ␤-hexosaminidase, and 4S activity using the trisaccharide substrate.
Determination of 4S Expression-Cultured skin fibroblasts were harvested by treatment with trypsin-versene and washed with PBS. Fibroblast lysates were prepared by seven cycles of freeze-thaw in 0.5 M NaCl, 20 mM Tris-HCl, pH 7. Fibroblast lysates were dialyzed for 16 h against 50 mM sodium formate, pH 3.5, and then specifically assayed for 4S by incubation with the radiolabeled trisaccharide substrate GalNAc4S-GlcA-GalitolNAc4S for 18 h (22). Each fraction isolated from CHOf4S and CHOD520N Percoll gradients was assayed for total arylsulfatase activity using the fluorogenic substrate, 4MUS (18). CHO-KI cells express very low levels of total arylsulfatase activity (approximately 0.12 nmol/min/mg). We have therefore assumed that the majority of arylsulfatase activity detected in transfected CHO-KI cells corresponds to arylsulfatase B (4S).
Endocytosis of Wild-type 4S and D520N 4S-Conditioned medium from G418-resistant mass cultures of CHOf4S and CHOD520N cells were used as a source of wild-type 4S and D520N 4S, respectively. CHOf4S and CHOD520N cells were cultured in Ham's F-12 medium supplemented with 10% (v/v) FCS. Twenty-four-hour conditioned CHOf4S and CHOD520N medium was clarified by centrifugation, assayed directly with 4MUS, and diluted in growth medium to a final 4S activity of 7 nmol/min/ml. Triplicate cultures of confluent L476P homozygous fibroblasts in 25-cm 2 flasks were metabolically labeled with 15 Ci/ml Na 2 35 SO 4 as described above. Cells were then harvested by treatment with trypsin-versene and transferred to new 25-cm 2 flasks containing growth medium supplemented with 7 nmol/min/ml wildtype 4S activity or D520N 4S activity (derived from CHOf4S or CHOD520N conditioned medium, respectively) in the presence or absence of 5 mM M6P. After 48 h, cell lysates were prepared and assayed for total protein, 35 S radioactivity, ␤-hexosaminidase, and 4S activity toward 4MUS.
Determination of D520N 4S Activity in Neutral and Acidic pH Conditions-Twenty-four-hour conditioned medium (Ham's F-12 supplemented with 10% (v/v) heat-inactivated FCS) was harvested from confluent, G418-resistant mass cultures of CHOf4S and CHOD520N cells, clarified by centrifugation (2000 ϫ g for 5 min at 4°C) and used immediately. The pH of CHOf4S and CHOD520N medium at 4°C was sequentially reduced from pH 7.2 down to pH 4.4 with 1 M glacial acetic acid. At various pH levels, a 300-l aliquot was removed and kept at 4°C. Samples were then incubated at 37°C for 360 min. After 30, 70, 195, and 360 min at 37°C, a 10-l aliquot was removed and assayed for arylsulfatase activity using 4MUS (18). After 360 min, the pH of CHOf4S pH 7.2 and CHOD520N pH 7.2 samples was reduced to pH 5.5 with 0.1 M glacial acetic acid. Samples were then incubated for another 16 h at 37°C to determine whether the reduction in D520N 4S activity was reversible. After 2, 4, and 16 h, a 10-l aliquot was removed and assayed for arylsulfatase activity using 4MUS.
Densitometric Analysis-Autoradiographs were scanned with a Molecular Dynamics personal densitometer and images analyzed with Molecular Dynamics ImageQuant software.

RESULTS
Identification of D520N-A colony cat (cat 237), with no detectable 4S activity and a normal L476 genotype, was used in the present study. PCR products encoding the f4S cDNA were amplified from cat 237 fibroblast first strand cDNA, sequenced directly and compared with the normal f4S cDNA (10). At nucleotide 1558 of the published sequence (23), a G 3 A base substitution was identified in the 4S sequence from cat 237, which resulted in a change at codon 520 from GAC (aspartic acid) to AAC (asparagine) (D520N, Fig. 1). No other changes were identified, suggesting that D520N is likely to be the cause of the reduced 4S activity in cat 237.
Detection of L476P and D520N-Cats were screened for the L476P and D520N alleles by restriction enzyme analysis of PCR products as described under "Experimental Procedures." Fig. 2 depicts a representative family tree of the genotypes of five cats determined by restriction enzyme analysis with HaeIII and BslI. Identification of L476P and D520N with this PCR-based screening method has enabled cats within the colony to be divided into six groups based on their genotypes: D520N heterozygous, D520N homozygous, L476P heterozy- gous, L476P homozygous, and D520N/L476P compound heterozygous. In addition, neither L476P or D520N were detected in nine normal cats (genetically unrelated to this colony) and nine colony-bred cats with 4S activity levels in the normal range. None of the 19 MPS VI cats genotyped to be homozygous for the L476P mutation contained the D520N mutation. Similarly, none of the 39 D520N homozygotes carried the L476P mutation, suggesting that the L476P and D520N alleles are inherited independently of each other.
Analysis of 4S Activity and GAG Storage in Feline Skin Fibroblasts-Previously, very low or non-detectable levels of 4S activity toward the trisaccharide substrate were demonstrated in D520N homozygous and D520N/L476P compound heterozygous peripheral leukocyte samples (10). By extending the incubation time with the trisaccharide substrate to 18 h and using skin fibroblasts, it was possible to detect 4S activity in these cats. Skin fibroblasts were cultured from L476P homozygous, D520N/L476P compound heterozygous, and D520N homozygous cats and used to determine residual 4S activity levels and 35 S-GAG storage levels (see "Experimental Procedures"). A direct correlation was observed between 4S activity levels and 35 S-labeled GAG storage. L476P homozygous fibroblasts contained very low levels of 4S activity and a 11-fold elevated level of 35 S-labeled GAG storage when compared with normal fibroblasts (Table I). Skin fibroblasts isolated from D520N homozygous and D520N/L476P compound heterozygous cats contained approximately 4.6% and 3.1% of normal 4S activity levels, respectively, and demonstrated a reproducible 2-3-fold elevated level of 35 S-GAG storage when compared with normal skin fibroblasts (Table I). These results suggest that the D520N mutation, in combination with the L476P mutation or on its own, results in a significant defect in GAG turnover in skin fibroblasts. Despite a reduced rate of GAG metabolism in skin fibroblasts, the residual 4S activity present in these cats appears to prevent lysosomal vacuolation in skin tissue (11).
Endocytosis of Wild-type 4S and D520N 4S by MPS VI Fibroblasts-An uptake experiment was done in order to characterize the endocytosis and lysosomal targeting of the mutant D520N 4S enzyme secreted from CHO-KI cells. L476P homozygous MPS VI fibroblasts were used in this study because of their low levels of 4S activity. D520N 4S and wild-type 4S enzyme used in the uptake experiment were derived from CHOD520N and CHOf4S conditioned media and contained 32.5 and 44.6 nmol/min/ml arylsulfatase activity, respectively. The CHOD520N and CHOf4S conditioned media were diluted in growth medium to a final arylsulfatase activity of 7 nmol/ min/ml. MPS VI (L476P homozygous) feline fibroblasts were metabolically labeled with Na 2 35 SO 4 for 24 h and then incubated with equal amounts of D520N 4S or wild-type 4S activity (7 nmol/min/ml) over 48 h in the presence or absence of 5 mM M6P.
The results show that MPS VI (L476P homozygous) fibroblasts contained very low levels of arylsulfatase activity and in this experiment showed an 8-fold elevated level of 35 S-labeled GAG storage when compared with normal fibroblasts (Table  II). The level of intracellular 35 S-labeled GAG storage material in MPS VI fibroblasts was reduced to below normal levels following endocytosis of either wild-type 4S or D520N 4S (Table  II). These results suggest that the mutant D520N 4S enzyme can correct the 35 S-GAG storage phenotype in MPS VI fibroblasts. Although equal amounts of D520N 4S activity and wildtype 4S activity were placed on the MPS VI fibroblasts, the amount of intracellular D520N 4S activity resulting from endocytosis appeared to be significantly lower for the D520N enzyme in comparison to wild-type 4S. The activity of the lysosomal enzyme ␤-hexosaminidase was unaffected by the endocytosis of D520N 4S or wild-type 4S.
The presence of 5 mM M6P significantly reduced the level of endocytosed D520N 4S and wild-type 4S enzyme activity, suggesting that both enzymes are predominantly endocytosed via the M6P receptor pathway (Table II). However, even in the presence of 5 mM M6P, there is a small amount of enzyme uptake that results in arylsulfatase activity of approximately 15% of normal, which is sufficient to efficiently clear the 35 Slabeled storage product in these cells (Table II). In all of the MPS VI cells exposed to enzyme, storage is reduced to levels below that seen in the normal control. We believe this represents intrinsic differences in the cell lines used rather than a significant phenomenon.
The mannose 6-phosphorylation of mutant D520N 4S was confirmed in a 32 P-labeling experiment. G418-resistant mass cultures of CHOf4S and CHOD520N were metabolically labeled with Na 2 32 PO 4 , immunoprecipitated, and analyzed via SDS-PAGE and autoradiography. The results suggested that the D520N mutation does not measurably affect the specific phosphorylation of intracellular or secreted 4S protein (results not shown).
Subcellular Localization of D520N 4S-Microsomal, endosomal, and lysosomal fractions were isolated from EXPRE 35 S 35 Slabeled CHO-KI cells overexpressing D520N 4S and wild-type 4S by Percoll granular fractionation. The amounts of total protein across each gradient were similar (data not shown). The data in Fig. 3 (A and B) demonstrates the enrichment of endosomes and lysosomes at the middle and bottom of each Percoll gradient, respectively. The top of each Percoll gradient (fractions 1-3) contained a small amount of ␤-hexosaminidase and a significant amount of acid phosphatase activity consistent with some lysis of organelles prior to their separation (Fig.  3, A and B). The intracellular trafficking of newly synthesized (45-min pulse label and 4-h chase) 35 S-labeled D520N 4S and 35 Slabeled wild-type 4S through the microsomal, endosomal, and lysosomal compartments appeared to be similar. However, approximately 60% of the newly synthesized lysosomal D520N 4S protein was detected as a 64-kDa precursor polypeptide, whereas the majority (90%) of lysosomal wild-type 4S protein was detected as a 44-kDa mature form (Fig. 4). These observations suggest that the rate of maturation of D520N 4S in the lysosome was significantly reduced in comparison to wild-type 4S. The small amount of organelle lysis observed in this experiment may explain the presence of mature 4S in the microsomal fractions. Both 35 S-labeled D520N 4S and 35 S-labeled wildtype 4S precursor protein were secreted into the culture medium in approximately equal amounts (Fig. 4). The 46-kDa polypeptide in the endosomal fraction is occasionally observed in analyses of 4S processing; however, its significance is unclear. 2 Fig . 3C and Table III illustrate the enrichment of arylsulfatase activity in the lysosomal compartment of CHO-KI cells expressing wild-type 4S. As would be expected, the majority of the activity (203 nmol/min/mg, approximately three quarters) is found in the lysosomal compartment. In comparison, expression of D520N 4S resulted in a substantially reduced level of arylsulfatase activity in the lysosomal compartment that was equivalent to 17% (35 nmol/min/mg) of the level in cells expressing wild-type 4S. Lesser reductions in the relative levels of arylsulfatase activity were seen in the microsomal (33% of wild-type) and endosomal (52% of wild-type) compartments of CHO-KI cells expressing D520N 4S (Table III).
The human 4S monoclonal antibodies F58.3 and F22.1, which have previously been shown to cross-react with f4S (20), did not cross-react with D520N 4S in a sandwich ELISA. A sandwich ELISA with a murine and rabbit polyclonal antibody raised against rf4S was therefore used to determine 4S protein levels in CHOD520N and CHOf4S microsomal, endosomal, and lysosomal fractions. Immunoquantification of 4S protein levels demonstrated a rapid degradation of D520N 4S mutant protein during synthesis in the endoplasmic reticulum and transport to the endosome and lysosome. 4S protein levels in CHOD520N microsomal, endosomal, and lysosomal enriched fractions were 38%, 15%, and 0.76% of wild-type 4S protein levels, respectively (Table III). Therefore, while there is almost complete degradation of D520N 4S in the lysosome, there is a corresponding 22.5-fold increase in specific activity that results in a much less severe reduction in the level of D520N 4S activity in the lysosomal compartment. (Table III). A smaller increase in D520N 4S specific activity was also evident in the endosomal compartment (3.6-fold increase when compared with endosomal wild-type 4S specific activity, Table III). The specific activity of D520N 4S in the microsomal compartment (58 nmol/ min/g) was partially reduced when compared with the specific activity of microsomal wild-type 4S (66 nmol/min/g) ( Table  III).
Destabilization of Mature D520N 4S-To test the effect of the D520N missense mutation on 4S stability, G418-resistant mass cultures of CHOf4S and CHOD520N were metabolically pulse-labeled with EXPRE 35 S 35 S protein labeling mix. Cells were then chased for varying lengths of time in the presence of 5 mM M6P (to prevent re-uptake of 35 S-labeled secreted precursor 4S). Immunoprecipitation showed that the intracellular and secreted forms of 35 S-labeled wild-type 4S are stable over a 2-day chase period (Fig. 5, A and B). The stability of 35 S-labeled precursor and mature D520N 4S secreted into the medium was not affected over a 2-day chase period (Fig. 5B). In contrast, the mature form of intracellular 35 S-labeled D520N 4S protein was degraded within 4 -10 h (Fig. 5A).
Western Blot Analysis-Precursor 4S protein was detected in the endosomal enriched fractions isolated from CHOf4S and CHOD520N cells via Western blot analysis (Fig. 6). As ex-2 D. Brooks, personal communication.  (11). L476P homozygous cats have a clinically severe form of MPS VI, which has previously been characterized (5). One skin biopsy was analyzed from a normal cat. For each MPS VI genotype, skin biopsies were obtained from three individual cats. Triplicate cultures of each primary fibroblast line were metabolically labeled with Na 2 35 SO 4 for 24 h. Cells were then washed, incubated with growth medium for another 48 h, harvested, and assayed for total cell protein, 35 S radioactivity (cpm/mg total cell protein), and 4S activity using the trisaccharide substrate. Results are expressed as the mean Ϯ standard error. WT ϭ wild type. Allele (Fig. 6).

Inactivation of D520N 4S in Neutral pH
Conditions-Conditioned medium from CHOD520N cells was incubated at 37°C in various pH conditions. D520N 4S was rapidly inactivated with time upon incubation at 37°C, pH 7.2. The rate of D520N 4S inactivation was reduced with increasingly acidic conditions (Fig. 7). ELISA demonstrated no difference in D520N 4S pro-tein levels in CHOD520N conditioned medium samples following a 24-h incubation at pH 7.2 or pH 5.9. These observations suggest that the D520N mutation rapidly inactivates 4S secreted from CHO-KI cells in neutral pH conditions but does not result in the degradation of the secreted enzyme. The inactivation of D520N 4S in neutral pH conditions was not reversible. Following complete D520N 4S inactivation in neutral pH con-  (14) and separated on Percoll gradients. Twenty 1-ml fractions were collected from the top of each gradient down to the bottom. Fractions were assayed for acid phosphatase (A, a marker for endosomes and lysosomes), the lysosomal enzyme ␤-hexosaminidase (B), and total arylsulfatase activity (C). No arylsulfatase activity was detected in untransfected, fractionated CHO-KI cells when incubated with 4MUS under the same conditions (data not shown). The arylsulfatase activity detected in fractionated CHOD520N and CHOf4S cells (C) therefore corresponds to arylsulfatase B (4S) activity. Fractions corresponding to the microsomal (fractions 1-3), endosomal (fractions 7-10), and lysosomal (fractions [15][16][17][18][19] compartments were pooled and used for immunoprecipitation (Fig. 4) and Western blot analysis (Fig.  6).

TABLE III
Specific activity of D520N 4S and wild-type 4S in fractionated CHO-KI cells Intracellular organelles were isolated from CHOD520N and CHOf4S cells by Percoll granular fractionation. Fractions corresponding to the microsomal (M), endosomal (E), and lysosomal (L) compartments were pooled and assayed for arylsulfatase activity (nmol/min/mg total cell protein) using the fluorogenic substrate 4MUS or 4S protein (ng/mg total cell protein) using a sandwich ELISA. The amounts of arylsulfatase activity and 4S protein in each CHOD520N fraction are also expressed as a percentage of normal levels in CHOf4S cells and are indicated in parentheses. As untransfected, fractionated CHO-KI cells contain no detectable arylsulfatase activity, we have made the assumption that the arylsulfatase activity detected in fractionated CHOD520N and CHOf4S cells corresponds to arylsulfatase B (4S) activity. The specific activity of the 4S in each subcellular compartment (nmol/ min/g 4S) was calculated by dividing the arylsulfatase activity (nmol/ min/mg total cell protein) by the amount of corresponding 4S protein (g/mg total cell protein). The -fold increase in specific activity (relative to wild type 4S) for CHOD52ON in the endosomal and lysosomal compartments is indicated in parentheses.  Table I), also contain marginally elevated levels of urinary DS and lysosomal vacuolation in some chondrocytes. The small increase in residual 4S activity (from 3.1% to 4.6% of normal levels) appears to be sufficient to prevent the occurrence of degenerative joint disease in D520N homozygous cats (11). Although D520N homozygous and D520N/L476P compound heterozygous cats contain low 4S activity, their normal growth and outward appearance (11) suggest that this residual 4S activity is sufficient to metabolize the majority of natural substrate in vivo. This is encouraging for gene and enzyme replacement therapy protocols for severely affected MPS patients, for whom only a small amount of recombinant enzyme may be required to prevent the majority of somatic symptoms.
Characterization of D520N/L476P compound heterozygous cats was complicated by the presence of two mutant 4S alleles, which together cause a mild biochemical and clinical MPS VI phenotype. To determine the effect of each mutation on 4S synthesis, subcellular distribution, and specific activity, both mutant alleles were engineered into the wild-type 4S cDNA and expressed. Expression of L476P 4S in CHO-KI cells has previously shown that this mutant of 4S is synthesized as an inactive precursor polypeptide at reduced levels and is not cleaved to the mature form (10). In this study, the D520N 4S mutant was overexpressed in CHO-KI cells to permit detailed molecular characterization.
Recently, the crystal structure of human 4S has been determined (24). As the human and feline forms of 4S are 91%  7. Inactivation of D520N 4S activity is pH-dependent. Conditioned medium was harvested from CHOD520N cells and adjusted to pH 5.9 (E), 6.2 (Ⅺ), 6.3 (q), 6.4 (OE), 6.6 (f), and 7.2 (ࡗ) using 1 M acetic acid. Aliquots (300 l) were incubated at 37°C. At the indicated time points, samples were assayed for arylsulfatase activity with 4MUS. Wild-type 4S activity was stable over these time periods at all pH values tested (data not shown). identical at the amino acid level, the tertiary structure of f4S is likely to be similar to that of human 4S. Asp-518 (Asp-520 of the f4S sequence) is in close proximity to the postulated phosphotransferase binding domain of 4S (24). From this observation, we initially speculated that the D520N mutation may cause a functional disruption of the phosphotransferase binding domain, leading to reduced mannose 6-phosphorylation of 4S in the cis-Golgi, and thereby disrupt lysosomal targeting. However, uptake studies of D520N 4S and wild-type 4S (secreted from CHO-KI cells) into L476P homozygous fibroblasts demonstrated that both enzymes were efficiently endocytosed via the M6P receptor pathway (Table II). Metabolic labeling with Na 2 32 PO 4 confirmed that the specific phosphorylation of precursor D520N 4S is indistinguishable from wild-type 4S. Furthermore, immunoprecipitation of 35 S-labeled CHOD520N organelles fractionated on Percoll gradients suggested that the D520N mutation has no significant effect on the lysosomal targeting of 4S (Fig. 4). Collectively, these results suggest that the mannose 6-phosphorylation and lysosomal targeting of 4S protein is not altered by the D520N mutation.
Two monoclonal antibodies (F58.3 and F22.1) raised against human 4S, and shown previously to cross-react with f4S (20), did not cross-react with D520N 4S in a sandwich ELISA. This suggests that the epitopes on f4S that are recognized by these antibodies are altered by the D520N amino acid substitution. Therefore, a sandwich ELISA was developed using a murine polyclonal rf4S antibody and a rabbit polyclonal rf4S antibody and used to immunoquantitate 4S protein levels in CHOD520N and CHOf4S cells. The lowest amount of mutant protein (0.76% of normal levels) was detected in the lysosomal compartment of CHOD520N cells (Table III). When compared with wild-type 4S, the specific activity of D520N 4S was elevated by 22.5-fold in the lysosomal compartment of CHO-KI cells (Table III). These observations suggest that the D520N 4S undergoes rapid degradation in the lysosome, but the consequences of this are partially ameliorated by the increased specific activity of the small amount 4S mutant protein in the lysosome. The instability of D520N 4S and the low level of expression prevented accurate immunoquantification of 4S protein levels in skin fibroblasts from D520N homozygous and D520N/L476P compound heterozygous cats.
When compared with wild-type 4S, the specific activity of D520N 4S was elevated by 3.6-fold and 22.5-fold, respectively, in the increasingly acidified endosomal and lysosomal compartments of CHO-KI cells (Table III). These observations suggested that the specific activity of D520N 4S may be pHsensitive. D520N 4S secreted from CHO-KI cells was rapidly inactivated at pH 7.2. With increasingly acidic conditions, secreted D520N 4S became less susceptible to inactivation (Fig.  7). These results provide further evidence for the sensitivity of D520N 4S to pH. This may result in a loss of specific activity in the pH neutral endoplasmic reticulum and Golgi compartments. However, only a small reduction in the specific activity of microsomal D520N 4S (58 nmol/min/g) was observed, when compared with the specific activity of microsomal wild-type 4S (66 nmol/min/g) ( Table III). The extent of decreased specific activity would depend on the time for which D520N 4S resides in these pH neutral compartments.
The increased specific activity of lysosomal D520N 4S may be a result of the effect of acidic pH on the enzyme. In addition, the increased specific activity of lysosomal D520N 4S may be associated with the mature form of the mutant molecule. Metabolic labeling with EXPRE 35 S 35 S protein labeling mix demonstrated that D520N 4S is targeted to the lysosomal compartment of CHO-KI cells and cleaved to a mature enzyme, which is rapidly degraded (Figs. 4 and 5). The 22.5-fold in-crease in D520N 4S specific activity may be associated with the highly unstable mature form of the mutant enzyme in the hydrolytic environment of the lysosomal compartment. This would be consistent with the observation that secreted D520N 4S enzyme, which is predominantly precursor form, has a normal specific activity at acidic pH (Fig. 7).
The studies of D520N 4S subcellular distribution in CHO-KI cells (Table III, Fig. 3) clearly demonstrate why D520N homozygous and D520N/L476P compound heterozygous cats develop a mild MPS VI phenotype despite the relatively high levels of 4S activity measured in fibroblast lysates (4.6% and 3.1% of normal levels, respectively; Table I). It is evident that only a small proportion of this activity, approximately 10%, will be in the correct cellular compartment; i.e. the lysosome. Therefore, the effective levels of 4S activity in D520N homozygous and D520N/L476P compound heterozygous animals will be approximately 0.5% and 0.3%, respectively.
A C-terminal extension mutation in the human 4S gene which is associated with an increased catalytic efficiency has previously been reported in a juvenile MPS VI patient (25). The mutation of a stop codon to a glutamine codon extended the open reading frame of the human 4S gene by 50 codons (*534Q). When overexpressed in LTK Ϫ cells, the majority of mutant *534Q enzyme was rapidly degraded in the trans-Golgi network. The small amount of *534Q mutant enzyme that was sorted to lysosomes was associated with an 8.6-fold increase in catalytic efficiency (25). Characterization of the *534Q mutation (25) and the D520N mutation in this study suggest that alterations of the C-terminal region of 4S can result in a higher susceptibility to proteolytic degradation but can also cause a significant increase in the specific activity of the molecule. These observations raise the possibility of engineering a derivative of 4S with an elevated specific activity, which could be used in gene and enzyme replacement therapies. However, to be effective, such a derivative would need to be resistant to proteolytic degradation.
In summary, we have identified a novel mutation, D520N, that causes a mild MPS VI phenotype in D520N/L476P compound heterozygous cats. Biochemical analysis of cultured skin fibroblasts from D520N/L476P compound heterozygous and D520N homozygous cats suggest that only small amounts of residual lysosomal 4S activity (probably less than 1% of normal lysosomal levels) are required to prevent the majority of somatic symptoms associated with severe feline MPS VI. However, significant chondrocyte vacuolation and degenerative joint disease were observed in D520N/L476P compound heterozygous cats (11). These observations suggest that the residual 4S enzyme activity present in these cats, while significant, is not sufficient to completely turn over the high levels of GAG present in the extracellular matrix of cartilage. Successful correction of chondrocytes by therapy strategies faces two problems: their higher requirement for 4S and the inaccessibility of these cells to circulating recombinant enzyme, as evident from enzyme replacement therapy studies (6). The mild feline MPS VI phenotype identified in this study will be useful for the evaluation of therapies directed to correction of lysosomal storage in chondrocytes.
The D520N mutation was engineered into the wild-type f4S cDNA. When expressed in CHO-KI cells, the secreted form of D520N 4S was susceptible to inactivation in neutral pH conditions. With increasingly acidic conditions, the mutant D520N 4S enzyme became more resistant to inactivation. The intracellular D520N 4S protein was rapidly degraded, with less than 1% of normal 4S protein levels reaching the acidified lysosomal compartment of CHO-KI cells. The small amount of D520N 4S that was sorted to the lysosomal compartment was associated with a 22.5-fold increase in specific activity. The lack of severe MPS VI clinical symptoms in D520N homozygous and D520N/ L476P compound heterozygous cats is therefore a result of a significant increase in the specific activity of the D520N 4S mutant enzyme reaching the lysosomal compartment. The subsequent residual levels of 4S activity in the lysosomes of these affected cats appear to be sufficient to metabolize the majority of GAG in vivo and thereby prevent severe MPS VI disease progression.