Defective Oligomerization of Arylsulfatase A as a Cause of Its Instability in Lysosomes and Metachromatic Leukodystrophy*

In one of the most common mutations causing metachromatic leukodystrophy, the P426L-allele of arylsulfatase A (ASA), the deficiency of ASA results from its instability in lysosomes. Inhibition of lysosomal cysteine proteinases protects the P426L-ASA and restores the sulfatide catabolism in fibroblasts of the patients. P426L-ASA, but not wild type ASA, was cleaved by purified cathepsin L at threonine 421 yielding 54- and 9-kDa fragments. X-ray crystallography at 2.5-Å resolution showed that cleavage is not due to a difference in the protein fold that would expose the peptide bond following threonine 421 to proteases. Octamerization, GCTCAT-GAGCCCCTGCTGCTCTATGAC GCTTCAGCTGCTCAAGC- GCCAGTTAGACGCAGC transformed the

version of a critical cysteine to C ␣ -formylglycine, and forms dimers. The dimers are transported to the Golgi, where they receive mannose 6-phosphate recognition markers and bind to mannose 6-phosphate receptors. The ASA-receptor complexes are transported to endosomes where the complexes dissociate. The receptors return to the Golgi, and the ASA is delivered to lysosomes. Within the lysosomes the dimers oligomerize in a pH-dependent manner to an octamer (1)(2)(3). The major natural substrate of ASA is cerebroside 3-sulfate, which is presented to the enzyme as a 1:1 complex with the sphingolipid activator protein B (4). Deficiency of ASA causes metachromatic leukodystrophy (MLD), a lysosomal storage disorder associated with progressive demyelination of the central and peripheral nervous system. In rare cases MLD is caused by the deficiency of the sphingolipid activator protein B (5).
One of the most frequent ASA alleles in MLD patients causes the substitution of Pro-426 by leucine. Homozygosity for the P426L-ASA allele is associated with the adult form of MLD, presenting first clinical symptoms at the age of 17 or later and characterized by a protracted clinical course. The combination of the P426L-ASA allele with one of the MLD alleles that causes the most severe form of the disease, the late infantile MLD, is usually associated with an intermediate form of MLD manifesting at ages between 6 and 16. This juvenile form of MLD is usually fatal in the second or third decade of life (6). The P426L-ASA allele is characterized by a decreased stability within lysosomes, whereas its trafficking to lysosomes and its catalytic activity are normal. In fibroblasts from MLD patients carrying a P426L-ASA allele, the mutant enzyme can be stabilized by the addition of inhibitors of cysteine proteinases. Under these conditions ASA activity increases and the catabolism of cerebroside 3-sulfate is normalized (7,8).
Protection of P426L-ASA from lysosomal degradation could therefore be a means to alleviate or even correct the metabolic defect in a substantial fraction of MLD patients. This prompted us to analyze the structural basis for the decreased lysosomal stability of the P426L-ASA and to characterize the lysosomal cysteine proteinase(s) responsible for its degradation.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis, Expression, and Purification of ASA-The cDNA encoding human ASA (9) was cloned as an EcoRI fragment into pMPSVEH (10) and subjected to site-directed mutagenesis using the QuikChange method (Stratagene) according to the instructions of the manufacturer. The following complementary primers were applied (only coding sequences given, mutated triplets underlined): GCTCAT-GAGCCCCTGCTGCTCTATGAC (P426L), GCTTCAGCTGCTCAAGC-GCCAGTTAGACGCAGC (A464R).
Escherichia coli DH5␣ was transformed with the mutagenized plasmids by electroporation (Electroporator 1000, Stratagene). The mutant cDNAs were checked by sequencing to preclude any PCR-derived er-* This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The different ASA forms were purified to homogeneity from secretions of the transfected cells by affinity chromatography (1) and dialyzed against 10 mM Tris-HCl, 150 mM NaCl, pH 7.4.
Cathepsins-Cathepsin B was purified from rat liver (13), cathepsin H from Epstein-Barr virus transformed human B-lymphocytes (14), and cathepsin L from the culture medium of non-small cell lung cancer cell line EPLC 32 M1 (15). All cathepsins were purified to homogeneity, and the concentration of the cathepsins was determined by titration with E64 (16). The concentrations given for cathepsins refer to those of the active enzymes.
Digestions Purification, Edman Sequencing, and Mass Spectrometry of ASA Fragments-For a preparative digest 0.63 nmol of P426L-ASA (40 g) were incubated with 0.17 nmol of cathepsin L for 8 h at 37°C in 0.1 ml of 50 mM sodium acetate, pH 4.8, containing 90 mM NaCl, 2.5 mM EDTA, and 2.5 mM dithiothreitol. The peptides were separated by RP-HPLC using a C4 column (Aquapore butyl column, 2.1 ϫ 200 mm; BAI, Lautertal) with increasing concentration of acetonitrile (2% per minute) in 0.1% trifluoroacetic acid/H 2 O at a flow rate of 0.3 ml/min. The peptides were analyzed by MALDI-TOF mass spectrometry and N-terminal Edman sequencing (2).
Crystallization and Data Collection-Crystallization of P426L-ASA was done under similar conditions to those described for wt-ASA (3). Crystals were grown using the hanging-drop vapor diffusion method (17) at 18°C. The protein solution (ϳ5 mg/ml in 10 mM Tris-HCl, pH 7.4, and 150 mM NaCl) was mixed 1:1 with a reservoir solution prepared with 100 mM sodium acetate buffer, pH 5.3-5.4, containing 10 -13% polyethylene glycol 6000.
Crystals suitable for x-ray analysis grew within 2 weeks. They had a bipyramidal shape with a maximum size of 0.3 ϫ 0.3 ϫ 0.4 mm 3 and belonged to the tetragonal space group I422 (unit cell parameters: a ϭ b ϭ 131.4, c ϭ 192.0 Å). There is one molecule per asymmetric unit and a Matthews coefficient (18) of V m ϭ 3.3 Å 3 Da Ϫ1 corresponding to a solvent content of ϳ63%. X-ray diffraction data were collected at room temperature using MAR Research image plate detectors at EMBL c/o Deutsches Elektronen-Synchrotron, Hamburg, on the X11 synchrotron beamline with an x-ray wavelength of 0.91 Å. As crystal decay took place a complete data set with a reasonable resolution could only be obtained by merging data of three crystals. The data were processed with the programs DENZO and SCALEPACK (19) to a resolution of 2.5 Å. Statistics on data quality and refinement are summarized in Table I.
Structure Solution and Refinement-The structure was determined using the modified coordinates of the wild type (identification code: 1AUK, Protein Data Bank at Rutgers University) (3). Amino acids surrounding the mutation were omitted. These residues were fitted into the electron density of the omit maps. Solvent waters and the cation were included from the structure of the wild type after examination of the difference electron density. Additional waters were included if they had suitable stereochemistry and individual B-factors below 80 Å. The structure was refined with the program REFMAC (20) using maximum likelihood as the target. The final model was refined to an R-factor of 18.5% and an R free -factor of 25.0% calculated on 5% of the data chosen at random (see Table I). The model comprises 3552 protein atoms, 174 water molecules, a magnesium ion, and 28 atoms of a sugar residue. The program PROCHECK (21) was used to check the stereochemical and geometrical outliers in the final structure, and the results are comparable to those of the wild type ASA. Graphical modeling was performed using the program XtalView (22).
Rutgers Protein Data Bank Accession Numbers-The coordinates and structure factors have been deposited in the Rutgers Protein Data Bank with the entry codes 1e33 and r1e33sf, respectively.
Oligomeric State of wt-ASA and ASA Mutants-Purified wt-ASA or ASA mutants (2 g) was incubated for 3 h on ice in 50 l of 10 mM Tris-HCl, 150 mM NaCl (pH 7.0) or 100 mM NaAc/HAc, 150 mM NaCl (pH 4.8 -6.0) and then subjected to gel filtration chromatography on a Superdex 200-column (PC3.2/30, Amersham Biosciences, Inc., bed volume 2.4 ml), equilibrated in the respective buffers, and eluted at a flow rate of 50 l/min using the SMART system (Amersham Biosciences, Inc.). The fractions were examined for ASA activity (8).

Endocytosis of [ 35 S]ASA by Fibroblasts-[ 35 S
]ASA was affinity-purified from secretions of mpr Ϫ MEF cells expressing the wt-ASA, the P426L mutant, or the A464R mutant that had been metabolically labeled for 24 h in the presence of [ 35 S]methionine. Confluent dishes (60 mm) of human control or late infantile MLD skin fibroblasts, or mouse control or cathepsin L-deficient fibroblasts (29), were incubated for 6 -24 h in 2 ml of Dulbecco's modified Eagle's medium supplemented with 30 -50 ϫ 10 3 cpm of [ 35 S]ASA as described previously (23). Where indicated the cells were treated with 100 M leupeptin for 24 h prior to and during the endocytosis period. [ 35 S]ASA was immunoprecipitated from extracts of cells and medium (8). The immunoprecipitates were subjected to SDS-PAGE, and labeled polypeptides were visualized by phosphorimaging (BAS 1000 3 Bas 1000ϩ, Raytest).

Cathepsin L Is a Lysosomal Cysteine Proteinase Selectively
Degrading P426L-ASA-To identify the lysosomal cysteine proteinase that preferentially or selectively degrades the P426L-ASA, wt-ASA and P426L-ASA were purified to homogeneity and incubated at pH 4.8 with purified cathepsin B, H, and L, the three major lysosomal cysteine proteinases. wt-ASA and P426L-ASA were resistant to cathepsin H, whereas both were equally susceptible to degradation by cathepsin B (Fig. 1A). P426L-ASA but not wt-ASA was susceptible to cathepsin L (Fig. 1B). Varying the pH between 4.8 and 6.5 showed that degradation of the mutant was maximal between pH 4.8 and 5.5 (not shown).
The Cathepsin L Cleavage Site in P426L-ASA Is in Close Proximity to the Mutated Residue 426 -Digestion of the P426L mutant with cathepsin L produced three major fragments with apparent molecular masses of 54, 30, and 9 kDa (Fig. 1B). To separate the fragments, the digest was subjected to RP-HPLC. The 54-and 9-kDa fragments were obtained in amounts sufficient to allow N-terminal amino acid sequencing. The N terminus of the 54-kDa polypeptide corresponded to the N terminus of mature ASA (RPPNIVL), whereas the N terminus of the 9-kDa fragment was heterogenous. The major (72%) sequence started with AHEPLLLYDLSK, indicating that the N terminus of this peptide corresponds to position 422 of ASA. The minor (28%) sequence started with histidine 423 and the following sequence was identical to that of the major 9-kDa fragment. N-terminal sequencing of both 9-kDa peptides of the P426L mutant verified the presence of a leucine residue at position 426. If the 9-kDa peptides represent the C-terminal ASA fragments 422-507 and 423-507, their theoretical mass is 9031.4 and 8960.3 Da, respectively. By MALDI-TOF mass spectrometry the masses of the two 9-kDa peptides were determined as 9034.2 and 8960.2 Da. This shows that the 9-kDa peptides indeed represent a mixture of ASA 422-507 and ASA 423-507 (Fig. 2). The mass of 9-kDa peptides and of the 54-kDa fragment add up to the apparent size of the ASA precursor (62 kDa). This indicates that the two fragments represent most, if not all of the ASA sequence and that cathepsin L cleaves N-terminal from residue 422. Cathepsin L is known to prefer a hydrophobic residue in position Ϫ2 (24). ASA contains in position 420 a leucine residue supporting the notion that cathepsin L cleaves the P426L mutant between threonine 421 and alanine 422. Because cathepsin L has no aminopeptidase activity (16), the ASA 423-507 fragment may arise from a minor cathepsin L-cleavage site between alanine 422 and histidine 423. However, we cannot completely exclude the possibility that the initial cathepsin L cleavage site is located further upstream of threonine 421 and that the 9-kDa fragments are generated by cathepsin L-mediated processing of an originally larger C-terminal fragment.
Three-dimensional Structure of P426L-ASA-The close proximity of the cathepsin L cleavage site to the site of the P426L mutation suggested that the P426 L substitution caused a local structural change uncovering a cathepsin L cleavage site following residue 421.
Therefore, the crystal structure of the P426L mutant was determined. X-ray data of P426L-ASA were collected, and the structure was solved using the coordinates of the wt-ASA omitting the amino acids surrounding the mutation. The structure was refined to a resolution of 2.5 Å, and overall it shows no difference to the structure of the wt-ASA (data to 2.1 Å), the root mean square deviation of a fit between the C␣ atoms in both structures being 0.2 Å. The crystals were obtained at pH 5.3 and contained octamers. The asymmetric unit is occupied by a monomer, the rest of the oligomeric moiety being generated by the symmetry operators. The ASA-monomer has a hat-like form made up of one large and one small ␤-sheet surrounded by ␣-helices (Fig. 3B).
The cathepsin L cleavage site and the mutation are found in ␤18, which is made up by the residues 421 to 430 and lies in the small ␤-sheet (Fig. 3, A and B). Residues 424 to 426 form a ␤-bulge, which seems to be caused by the conformation of the neighboring cis-Pro-425. The leucine residue in position 426 could well be fitted into the electron density found in the omit maps (not shown). It is surrounded by hydrophobic residues of the same dimer. The shortest interatomic distances to neighboring atoms of the leucine lie in the range of 3.7 Å. Thus the interactions of the side chain of Leu-426 with the surrounding residues are sterically not hindered.
A more detailed structural comparison of the region around the cleavage site following threonine 421 and the residue 426 in both wt and P426L-ASA shows that all atoms but those of the cis-Pro-425 and Leu-426 lie at the same positions within the experimental uncertainty. All atoms of the cis-Pro-425 are shifted by 0.8 -1.2 Å, whereas the side chain of Leu-426 is different due to the mutation (Fig. 3C).
Thus, the structural alterations seen in the crystal structure of P426L-ASA as compared with the wild type ASA are minor and cannot explain why the former but not the latter is cleaved by cathepsin L.
The P426L Substitution Impairs the Octamerization of ASA-At acidic pH the ASA dimers oligomerize to octamers, which are mainly stabilized by hydrophobic interactions. Adjacent to the mutated residue 426 are cis-Pro-425 and Glu-424, which both contribute to the hydrophobic interactions that stabilize the octamers. Glu-424 is of critical importance for the pH dependence of the ASA oligomerization. The side chain of Glu-424 is disordered among two discrete sites (Fig. 4). In one conformation the protonated form of Glu-424 establishes intramolecular hydrogen bonds to the carbonyl oxygen of His-423 (distance 2.7 Å) and to N ⑀2 of Gln-460 (distance 3.0 Å). In the alternative conformation it binds to the carbonyl oxygen of Phe-398 of the neighboring dimer (distance 2.9 Å). Deprotonation of the Glu-424 side chain at neutral pH will prevent binding to the carbonyl oxygen of His-423 and favor the conformation toward the surface of the ASA molecule. In this conformation, however, the negative charge of the Glu-424 side chain has a repulsive effect on the hydrophobic residues of the other dimer thus explaining the failure of ASA to octamerize at neutral pH.
The P426L substitution induces slight structural changes in the region of residues 425 to 426 (see Fig. 3C). These changes could affect the octamerization of ASA by altering the pK of the ␥-carboxyl group of Glu-424 and/or the hydrophobic interactions stabilizing the octamer. The cathepsin L cleavage site in the P426L mutant is located at the periphery of the contact site between the dimeric subunits of the octamer. In the octameric form of ASA the peptide bond between residues 421 and 422 is likely to be inaccessible to the protease (Fig. 5). We therefore determined the quaternary structure of ASA by gel filtration chromatography using a Superdex 200 column.
At pH 5.4 wt-ASA eluted as an octamer, whereas two-thirds of the P426L mutant eluted as a dimer (Fig. 6A). When the octamer/dimer ratio was determined at pH 4.8 -7.0, it became apparent that below pH 5.6 wt-ASA exists essentially as an octamer and above pH 6.0 as a dimer. For full octamerization of the P426L mutant the pH must be below pH 4.8 (Fig. 6B). The pH, at which half of the ASA exists as octamer, is shifted by the P426L substitution from pH 5.8 to 5.2.
The P426L substitution could affect the hydrophobic interactions that stabilize the octamer. ASA was therefore subjected to gel chromatography in the presence of 20% acetonitrile. The pH was chosen to achieve complete octamerization, i.e. pH 5.6 for wt-ASA and pH 4.8 or the P426L mutant (see Fig. 6). The presence of 20% acetonitrile decreased the octamer fraction from 100 to 60% for wild type ASA and from 100 to 80% for the P426L (not shown). This indicates that a shift of the pK of the ␥-carboxyl group of Glu-424 rather than a weakening of hydrophobic interactions impairs the octamerization of the P426L mutant.
Lysosomal Instability of A464R-ASA, a Mutant Forming Dimers at Acid pH-The analysis of the P426L mutant showed that the decreased stability within lysosomes correlates with an impaired octamerization between pH 5 and 6. To determine whether the inability of ASA to octamerize causes susceptibility toward lysosomal proteinases, we attempted to create an ASA mutant that does not octamerize. For this purpose, alanine 464, which is exposed at the hydrophobic interface participating in octamerization, was substituted by arginine. This substitution indeed prevented octamerization. The A464R mutant eluted as a dimer both at pH 4.8 and 7.0 (Fig. 7).
For determination of the intralysosomal stability, wt-ASA, the P426L mutant, and the A464R mutant were purified from the secretions of stably expressing mpr Ϫ MEF cells that had been metabolically labeled with [ 35 S]methionine. Fibroblasts from patients with the late infantile form of metachromatic leukodystrophy, which lack ASA polypeptides, were incubated for 6 -24 h in the presence of 35 S-labeled sulfatase, and the internalized [ 35 S]ASA polypeptides were immunoisolated from cell extracts. In parallel the internalized ASA was determined by measuring the ASA activity in the cell extracts.
When [ 35 S]wt-ASA was offered to the fibroblasts, 18.7% of the labeled ASA were recovered in the cells after an endocytosis period of 24 h as a 62-kDa polypeptide (Fig. 8, lane 2). When analyzed by catalytic activity measurement, 19.5% of wt-ASA was internalized within 24 h. Treatment of the recipient cells with 100 mM leupeptin prior to and during the endocytosis to inhibit degradation by lysosomal cysteine proteinases had no effect on the amount of internalized wt-[ 35 S]ASA (Fig. 8,  lane 3).
When [ 35 S]P426L-ASA was offered for internalization, 1.3% of the added ASA was recovered after 24 h in the cells, as determined both as [ 35 S]polypeptides (Fig. 8, lane 2) and as catalytic activity. Treatment with leupeptin increased the recovery of [ 35 S]P426L-ASA from 1.3 to 7.4% (Fig. 8, lane 3). This agrees with earlier observations (7) that the P426L substitution causes susceptibility of the mutant ASA to degradation by lysosomal cysteine proteinases. Interestingly, when control fibroblasts were used as recipient cells, the recovery of internalized [ 35 S]P426L-ASA increased slightly (2.8% instead of 1.3%) and a 47/49-kDa doublet was observed in addition to the 62-kDa form (Fig. 8, lane 4; note the 4-fold longer exposure time of this lane).
When [ 35 S]A464R-ASA was offered for internalization 2.1% of the labeled ASA was recovered intracellularly after endocytosis for 24 h as the 62-kDa form. In addition, a major 54-kDa form and a minor 47/49-kDa doublet were recovered, representing 8.1 and 1.3% of the offered ASA (Fig. 8, lane 2). The 54-kDa form appears to be catalytically active, because 9.5% of the added A464R-ASA activity was recovered in the cell after a 24-h endocytosis period. Internalization into leupeptin-treated cells increased the recovery of the 62-kDa form from 2.1 to 14%, whereas the fraction of smaller forms decreased from 9.4% to less than 1% (Fig. 8, lane 3). This indicates that the internalized A464R-ASA is rapidly degraded by lysosomal cysteine proteinases to a 54-kDa form, which is only slowly further degraded to the 47/49-kDa doublet. In contrast to P426L-ASA internalization of A464R-ASA by control fibroblasts neither affected the amount nor the composition of the internalized [ 35 S]polypeptides (Fig. 8, lane 4).
Stability of P426L-ASA in Cathepsin L-deficient Cells-The in vitro and in vivo data had suggested that defective oligomerization of the P426L-ASA is a cause for its rapid degradation by cathepsin L. To assess the role of cathepsin L for the instability of P426L-ASA in lysosomes, we offered [ 35 S]P426L-ASA for internalization to control and cathepsin L-deficient mouse fibroblasts (29). The P426L-ASA internalized by control mouse fibroblasts was degraded to products indistinguishable from those seen in human control fibroblasts (compare Fig. 9, lane 2, with Fig. 8, lane 4). In addition, trichloroacetic acid-soluble radiolabeled degradation products were released into the medium (not shown). Treatment with leupeptin inhibited the degradation of internalized P426L-ASA, increased its intracellular recovery 3-fold (Fig. 9, lane 3), and decreased the appearance of trichloroacetic acid-soluble degradation products in the medium. This indicates that P426L-ASA is degraded by thiolproteinases in mouse fibroblasts as rapidly as in human fibroblasts.
In cathepsin L-deficient mouse fibroblast P426L-ASA was degraded to a 54-kDa form, which was stable (Fig. 9, lane 4). The latter was indicated by the observation that treatment with leupeptin prevented cleavage of the 62-kDa precursor to the 54-kDa form, and increased the recovery of internalized P426L-ASA less than 1.2-fold (Fig. 9, lane 5).
After internalization [ 35 S]wt-ASA was as stable in mouse fibroblasts as in human fibroblasts (not shown). [ 35 S]A464R-ASA was rapidly degraded in control mouse fibroblasts to a 54-kDa polypeptide (Fig. 9, lane 2). The 54-kDa polypeptide was further degraded as indicated by the release of trichloroacetic acid-soluble material into the medium. Treatment with leupeptin inhibited degradation to the 54-kDa form and in- creased the recovery of internalized [ 35 S]A464R-ASA 3-fold (Fig. 9, lane 3). In cathepsin L-deficient mouse fibroblasts, internalized A464R-ASA was degraded to the 54-kDa form. As observed for the 54-kDa form of P426L-ASA, the 54-kDa form of A464R-ASA was stable in cathepsin L-deficient fibroblasts. Treatment with leupeptin inhibited formation of the 54-kDa form and increased the recovery of internalized [ 35 S]A464R-ASA only 1.2-fold (Fig. 9, lanes 4 and 5). In summary, these data show that deficiency of cathepsin L stabilizes the two octamerization-impaired ASA mutants at the level of the 62and 54-kDa forms (see below).

DISCUSSION
Cleavage of P426L-ASA by Cathepsin L-One of the most frequent mutations causing metachromatic leukodystrophy is the substitution of Pro-426 by leucine. The ASA deficiency is caused by the intralysosomal instability of the P426L-ASA. Addition of cysteine proteinase inhibitors to fibroblasts of patients carrying the P426L allele restores the ASA activity partially and normalizes the breakdown of cerebroside 3-sulfate, the major physiological substrate of ASA (6,8).
Here we show that cathepsin L is a candidate for the lysosomal cysteine proteinase degrading P426L-ASA. Although degradation of wt-ASA by purified cathepsin L was negligible, P426L-ASA was degraded. The initial cathepsin L-cleavage site during degradation of P426L-ASA could be localized to the peptide bond between residues 421 and 422. The cleavage site is in the near neighborhood of the P426L substitution and fulfills the known structural requirements for cathepsin L cleavage sites. In vivo the 54-kDa product is further degraded in a cathepsin L-dependent manner, indicating that the 54-kDa form represents an intermediate of P426L degradation by cathepsin L.
Defective Octamerization of P426L-ASA at Acidic pH-The x-ray crystallography did not show a major structural change of the P426L-ASA octamer compared with wt-ASA. Exposure of peptide bonds that normally are inaccessible therefore cannot explain the increased susceptibility to cathepsin L.
We noted that both the P426L substitution and the cathepsin L cleavage site following residue 421 are close to Glu-424, which controls the oligomeric state of ASA. At neutral pH ASA exists as a dimer. Protonation of Glu-424 at acidic pH permits the oligomerization of ASA to octamers, which are mainly stabilized by hydrophobic interactions (3). Substitution of Glu-424 by alanine, glutamine, or arginine allows octamer formation at neutral pH, indicating that it is the negative charge of Glu-424 that prevents octamerization of wt-ASA at neutral pH. 2 Examination of the oligomeric state of P426L-ASA revealed that the P426L substitution shifts the pH required for 50% octamerization by 0.6 pH units from pH 5.8 to pH 5.2. Thus, within lysosomes a substantial fraction of the P426L-ASA exists as a dimer, whereas for wt-ASA the octameric state is favored in lysosomes. This suggested that the unusually high intralysosomal stability of ASA (25) is linked to its octamerization and that mutations, impairing octamerization of ASA in lysosomes, decrease its lysosomal stability.
ASA Dimers Have a Reduced Lysosomal Stability-To corroborate a causal relationship between the existence of ASA as a dimer within lysosomes and a decreased lysosomal stability,  (lanes 2 and 3) and cathepsin L-deficient mice (lanes 4 and 5) was done as described in Fig. 8 4). The presence of leupeptin prevented the formation of the 54-kDa fragment but did not increase the intracellular recovery of ASA. Lower part, in control mouse fibroblasts 19% of the added [ 35 S]A464R-ASA were recovered, 54% as 62-kDa precursor and 46% as 54-kDa fragment. The presence of leupeptin inhibited the formation of the 54-kDa fragment and increased the amount of intracellularly recovered ASA to 55% (lanes 2 and 3). In cathepsin L-deficient fibroblasts 46% of the added [ 35 S]A464R-ASA were recovered intracellularly, 54% as 62-kDa precursor and 46% as 54-kDa from. The presence of leupeptin inhibited the formation of the 54-kDa form but did not affect the intracellular recovery of ASA (lanes 4 and 5).
Ala-464 in ASA was substituted by arginine. The expectation that this substitution would impair ASA octamerization was met. The A464R-ASA dimers do not octamerize at pH 4.8. A464R-ASA was internalized by fibroblasts but rapidly degraded to smaller molecular weight forms similar to P426L-ASA. The major degradation product of A464R-ASA, a 54-kDa polypeptide, is rather stable in human fibroblasts and retains catalytic activity. This is in contrast to the P426L-ASA, of which only trace amounts of 47/49-kDa degradation intermediates were recovered intracellularly. We conclude from these data that octamerization of ASA impairs the accessibility of peptide bonds cleavable by lysosomal cysteine proteinases. In the case of the P426L-ASA, the degradation results in a rapid loss of catalytic activity and immunoreactive peptides, whereas in the case of the 464R-ASA a 54-kDa intermediate is generated, which is catalytically active. It is degraded only slowly to lower molecular mass products, including the 47/49-kDa degradation intermediate seen for P426L-ASA. In control fibroblasts the P426L-ASA is slightly more stable than in ASAdeficient fibroblasts (see Fig. 8). This suggests that, in the endosomal/lysosomal compartment, mixed P426L/wt-ASA octamers are formed, which stabilize the P426L-ASA.
In Vivo Cleavage of P426L-ASA by Cathepsin L-The in vivo data had shown that ASA mutants with impaired octamerization are susceptible to partial (A464R-ASA) or complete (P426L-ASA) proteolysis by lysosomal cysteine proteinases. The in vitro data pointed to a role of cathepsin L in the degradation of P426L-ASA.
The comparison of the degradation of the P426L-and A464R-ASA mutants in normal and cathepsin L-deficient fibroblasts clearly demonstrated the essential role that cathepsin L plays in the degradation of these two ASA mutants. In the absence of cathepsin L the ASA mutants are degraded to a 54-kDa form, which accumulates. This indicates that cathepsin L is not essential for the initial cleavage of the 62-kDa precursor yielding the 54-and 9-kDa fragments. Although not being essential, cathepsin L is the major cysteine proteinase catalyzing the cleavage of the 62-kDa precursor in control fibroblasts. This is based on the observation that the yield of the 54-kDa form in cathepsin L-deficient fibroblasts was 2-to 3-fold lower than expected from the amount of 62-kDa precursor that is internalized and degraded in control mouse fibroblasts in the absence of leupeptin. The degradation of the 54-kDa fragment, however, requires cathepsin L.
We noted one difference for the degradation of P426L-and A464R-ASA. Transient accumulation of the 54-kDa fragment in human or mouse fibroblasts was seen only for A464R-ASA. The 54-kDa fragment of P426L-ASA was rapidly degraded via 47/49-kDa intermediates to lower molecular weight products that did not react with ASA antiserum and were, at least partly, released into the medium as trichloroacetic acid-soluble material. In fact the 54-kDa fragment was only seen when its degradation was blocked due to the absence of cathepsin L. Because the two mutations at residues 426 and 464 are located in the C-terminal 9-kDa fragment, the reason for the differential stability of the 54-kDa fragment is not obvious. A possible explanation could be that the dissociation of the 9-kDa fragments from the 54-kDa fragments is affected by the mutations and that the dissociation controls the accessibility of the 54-kDa fragment for cathepsin L. It should also be noted that the stability of the 54-kDa fragment derived from A464R-ASA depended on the cell type. Although it was efficiently degraded in mouse fibroblasts, it was rather stable in human fibroblasts.
Medical Aspects-The normal phenotype of individuals with about 5% residual ASA activity due to compound heterozygosity for an ASA null allele and an ASA pseudodeficiency allele (26) demonstrates that low residual ASA activity is sufficient to sustain normal sulfatide metabolism. The slow progression of the disease in patients that carry one or two copies of the P426L-ASA allele suggests that a repeated transient elevation of ASA activity may be sufficient to clear the lysosomal compartment from nondigested cerebroside 3-sulfate and to prevent progression of the disease. The present data suggest that lowering the lysosomal pH to induce a shift from dimeric to octameric P426L-ASA may provide a means to impair rapid degradation of P426L-ASA by lysosomal cysteine proteinase and to restore catabolism of cerebroside 3-sulfate. Furthermore, the knowledge of the cathepsin L cleavage site makes possible the development of peptide inhibitors that protect the P426L-ASA in the lysosomal environment. Because the P426L-ASA allele accounts for 16 -25% of all ASA alleles in MLD patients (6,27,28), about one-fifth of all MLD patients could benefit from such a therapeutic approach.