Amino Acid Residues Forming the Active Site of Arylsulfatase A

Arylsulfatase A belongs to the sulfatase family whose members carry a Cα-formylglycine that is post-translationally generated by oxidation of a conserved cysteine or serine residue. The formylglycine acts as an aldehyde hydrate with two geminal hydroxyls being involved in catalysis of sulfate ester cleavage. In arylsulfatase A and N-acetylgalactosamine 4-sulfatase this formylglycine was found to form the active site together with a divalent cation and a number of polar residues, tightly interconnected by a net of hydrogen bonds. Most of these putative active site residues are highly conserved among the eukaryotic and prokaryotic members of the sulfatase family. To analyze their function in binding and cleaving sulfate esters, we substituted a total of nine putative active site residues of human ASA by alanine (Asp29, Asp30, Asp281, Asn282, His125, His229, Lys123, Lys302, and Ser150). In addition the Mg2+-complexing residues (Asp29, Asp30, Asp281, and Asn282) were substituted conservatively by either asparagine or aspartate. In all mutants V max was decreased to 1–26% of wild type activity. The K m was more than 10-fold increased in K123A and K302A and up to 5-fold in the other mutants. In all mutants the pH optimum was increased from 4.5 by 0.2–0.8 units. These results indicate that each of the nine residues examined is critical for catalytic activity, Lys123 and Lys302 by binding the substrate and the others by direct (His125 and Asp281) or indirect participation in catalysis. The shift in the pH optimum is explained by two deprotonation steps that have been proposed for sulfate ester cleavage.

Mammalian sulfatases are involved in the degradation of sulfated substrates like mucopolysaccharides, cerebroside sulfates, and sulfated steroids. The key residue for catalytic activity in sulfate ester cleavage has recently been shown to be a C␣-formylglycine (FGly; Refs. 1 and 2), 1 which is post-translationally generated from a cysteine in the endoplasmic reticulum (3). Failure of this amino acid modification leaves newly synthesized sulfatases inactive and is the cause of multiple sulfatase deficiency, a rare but fatal lysosomal storage disorder (Ref. 1; reviewed in Refs. 4 and 5). X-ray determination of the three-dimensional structures of two human sulfatases, arylsul-fatase A (ASA; Ref. 6) and N-acetylgalactosamine 4-sulfatase (arylsulfatase B, ASB; Ref. 7), revealed that the FGly residue is buried at the bottom of a cavity that is formed by positively and negatively charged amino acids. This cavity proved to be the active site by two lines of evidence. Cocrystallization of ASB with vanadate, a potent inhibitor of ASB, revealed that vanadate was covalently bound to FGly (7). In addition, an ASA mutant containing serine instead of FGly was found to be covalently sulfated at the position of FGly 69 after incubation with sulfate ester (see below and Ref. 8). The 2-fold disordered electron density of FGly was interpreted to be an aldehyde hydrate with two geminal hydroxyls protruding into the lumen of the cavity (6). In close vicinity to the aldehyde hydrate a metal ion (Mg 2ϩ in ASA and Ca 2ϩ in ASB) is complexed by three aspartates and one asparagine. Two lysines, two histidines, and a serine complete the arrangement of residues potentially involved in binding and cleavage of the sulfate group within the active site.
These data led to a proposal for the catalytic mechanism of sulfate ester cleavage (6) as follows (Fig. 1). In the first halfcycle, one of the two geminal oxygens of the aldehyde hydrate attacks the sulfur of the sulfate ester leading to a transesterification of the sulfate group onto the aldehyde hydrate. Simultaneously the substrate alcohol is released. In the second halfcycle, sulfate is eliminated from the enzyme-sulfate intermediate by an intramolecular rearrangement induced by the second oxygen. By this "intramolecular hydrolysis" the aldehyde group is regenerated. The essential role of FGly in forming the transient enzyme-sulfate ester has been shown indirectly (8). Mutants of ASA and ASB were expressed in which the FGly had been replaced by a serine. These mutants were able to cleave sulfate ester and to generate a sulfated enzyme intermediate with a serine sulfate ester in position 69 (ASA) or 91 (ASB). However, due to the lack of a second hydroxyl group, the release of the sulfate group regenerating the aldehyde did not occur.
The function of the other residues forming the active site remained speculative. In ASA one oxygen of the aldehyde hydrate is in close vicinity to the Mg 2ϩ ion, which is complexed by residues Asp 29 , Asp 30 , Asp 281 , and Asn 282 , and the other oxygen is close to the imidazole ring of His 125 . Whereas Mg 2ϩ is thought to coordinate the oxygen that attacks the sulfur in the first step of sulfate cleavage, His 125 is a candidate to deprotonate the second hydroxyl of the aldehyde hydrate, which is involved in the release of the sulfate and the regeneration of the aldehyde. By modeling a sulfate anion into the active site, the positively charged residues Lys 123 , Lys 302 , His 229 , and residue Ser 150 were proposed to fix the sulfate group (6). To obtain experimental evidence for the function of the putative active site residues, we replaced residues Asp 29 , Asp 30 , Asp 281 , Asn 282 , His 125 , His 229 , Lys 123 , Lys 302 , and Ser 150 by alanine in ASA and analyzed V max , K m , and the pH optimum of the * 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.
Determination of Distances and Potential Hydrogen Bonds between the Residues of the Active Site of ASA-The atomic coordinates of human ASA (identification code 1AUK, Protein Data Bank of the Brookhaven National Laboratory) were analyzed for the distances between the functional groups of the active site residues using the computer program XtalView (9). Interatomic distances in the range of 2.5-3.2 Å between potential proton donors and acceptors were considered to be bridged by hydrogen bonds (10). All these potential hydrogen bonds are indicated in Fig. 3. The coordinates for the aldehyde hydrate residue of FGly 69 were obtained from G. Lukatela (Berlin, Germany).
Site-directed Mutagenesis-Single amino acid substitutions were performed replacing the codons for the putative active site residues Asp 29 , Asp 30 , Asp 281 , Asn 282 , His 125 , His 229 , Lys 123 , Lys 302 , and Ser 150 of human ASA by codons for alanine. In a second series of mutations the codons for Asp 29 , Asp 30 , and Asp 281 were replaced by codons for asparagine, and the codon for Asn 282 was replaced by that of aspartate. Mutagenesis was performed using the pAlter cloning kit (Promega Corp.), the cDNA for human ASA (11) cloned into the pAlter mutagenesis vector (12), and appropriate primers. Following mutagenesis mutant cDNAs were cloned as EcoRI fragments into the pMPSVEH vector downstream of the myeloproliferative sarcoma virus promoter.
Expression and Analysis of ASA Mutants-The plasmids carrying the mutated cDNA were stably transfected into mouse embryonic fibroblasts deficient for both mannose 6-phosphate receptors (mpr Ϫ MEF cells; Ref. 13). After selection and expansion of the clones stably expressing the mutants, sulfatases were purified from cell secretions by affinity chromatography and dialyzed against 10 mM Tris/HCl, 150 mM NaCl, pH 7.5. V max , K m , and the pH optimum of sulfatase activity of the purified sulfatases were determined using 2, 5, 10, and 20 mM pnitrocatechol sulfate as substrate as described earlier (12). All kinetic data were determined under initial rate conditions. Purity and amount of ASA proteins were calculated from the integral of the ASA peak at 280 nm after reversed phase HPLC as described (12). To determine the percentage of Cys 69 converted to FGly 69 , purified mutants were digested with trypsin. The peptides containing Cys 69 and FGly 69 were purified by reversed phase HPLC and quantified by Edman sequencing (12).

RESULTS AND DISCUSSION
Putative Active Site Residues of Arylsulfatase A-The amino acid residues supposed to form the active site of ASA are shown in Fig. 3 in a three-dimensional view. FGly 69 is shown as an aldehyde hydrate with two geminal hydroxyl groups. The Mg 2ϩ ion is octahedrically coordinated by one hydroxyl oxygen of Thus, a hydrogen bond may be formed between the neighboring oxygens of FGly 69 and Asp 281 leading to an at least partial deprotonation of this hydroxyl of FGly 69 . The free oxygen would then be able to start a nucleophilic attack on the sulfur of the sulfated substrate. Coordination to Mg 2ϩ may help to stabilize this catalytically active form of FGly 69 . On the other side His 125 has an atomic distance of 2.7 Å to the second oxygen of the FGly hydrate. Therefore, His 125 is in an appropriate position to de-protonate and thereby activate this second oxygen of the FGly hydrate leading to the regeneration of the oxogroup of the aldehyde and the cleavage of the enzymesulfate intermediate. Lys 123 , Lys 302 , His 229 , and Ser 150 have been proposed to contribute to the binding of the sulfate group and to the activation of the sulfur atom by increasing its electrophilia (6). In addition, these residues may form hydrogen bonds with other residues of the active site: Lys 123 with Asp 29 and Ser 150 as well as Lys 302 with Asp 30 . All residues supposed to be involved in sulfate binding or in catalysis are able to form hydrogen bonds to neighboring residues, resulting in a net-like connection of the residues to a functional catalytic system (Fig. 3).
The putative active site residues are highly conserved among 19 eukaryotic and prokaryotic sulfatases, which are known to exhibit sulfatase activity (Fig. 2). Most of these residues are part of short sequences of conserved amino acids. Cys 69 in ASA and its equivalents in the other sulfatases are converted to FGly. The sulfatase of Klebsiella pneumoniae is the only biochemically characterized sulfatase where a serine rather than a cysteine is encoded by the DNA. However, this serine is also converted to FGly (14). Asp 29 , Asp 30 , Lys 123 , His 125 , His 229 , Asp 281 , Asn 282 , and Lys 302 are conserved or conservatively replaced in 14 -19 sulfatases. Ser 150 is only found in ASA but replaced by threonine in 6 sulfatases. The high conservation of these residues indicates that the structure of the active site and the catalytic mechanism of sulfate cleavage have been conserved in the sulfatase family.
Expression of Active Site Mutants-To obtain experimental evidence for their role in sulfate ester cleavage and sulfate binding, nine of the putative active site residues of ASA were individually substituted by alanine: Asp 29 , Asp 30 , Asp 281 , and Asn 282 as the metal coordinating residues, His 125 as one of the two FGly 69 activating residues (the other one being Asp 281 ), and Lys 123 , Lys 302 , His 229 , and Ser 150 as possible sulfate binding residues. Cys 69 has been substituted by alanine earlier (12), and the mutant C69A proved to be completely inactive. Arg 73 that is interacting with Asp 281 (Fig. 3) is another putative FIG. 1. Proposed scheme of sulfate ester cleavage by sulfatases. In the first half-cycle (I) one of the hydroxyls of the aldehyde hydrate attacks the sulfur of the sulfate ester forming an enzyme-sulfate ester, whereas the residual alcohol is released. Sulfate is eliminated in the second half-cycle (II) by an intramolecular rearrangement cleaving the ester bond and forming an aldehyde group at the enzyme. Finally, the aldehyde hydrate is regenerated by the entry of water (III). This scheme is based on the reaction mechanism described by Lukatela et al. (6). active site residue. However, the R73A mutant has been expressed earlier (12) and was found to be unstable, thus preventing its further characterization. The cDNAs encoding the mutant forms of ASA were stably expressed in mpr Ϫ MEF cells.
Due to deficiency in mannose 6-phosphate receptors, these cells secrete the newly synthesized lysosomal enzymes such as ASA (13). Clones stably expressing the mutant sulfatases were selected and expanded. The sulfatases were purified from the medium by affinity chromatography and analyzed for purity by analytical reversed phase chromatography and SDS-polyacrylamide gel electrophoresis. In all preparations ASA polypeptides represented 80 -90% of total protein. The major contamination (5-10%) was bovine serum albumin originating from the culture medium (data not shown).
Kinetic Properties of Active Site Mutants-We analyzed V max , K m , and pH optimum for the cleavage of p-nitrocatechol sulfate by the different substitution mutants (Table I). To control whether a loss of activity could be accounted for by a diminished conversion of Cys 69 to FGly, all mutant proteins were analyzed for the presence of either FGly or Cys in position 69. In all mutants FGly was found to account for 49 -95% of residue 69. In wild type control ASA 94% of residue 69 were converted to FGly (Table I). Earlier experiments have shown that this modification efficiency depends on the expression level, and values of 60 -90% have been observed (1). Thus, the analysis of the modification efficiency of the ASA mutants indicates that a decrease of FGly formation can account for a not more than 50% decrease of V max in the mutants.
Substitution of three of the Mg 2ϩ complexing residues (Asp 29 , Asp 30 , and Asp 281 ) by alanine led to a strong decrease of V max to 1-5% compared with wild type ASA, whereas substitution of Asn 282 by alanine decreased V max to 24%. This reflects the importance of the negatively charged aspartates and, to a lower degree, of the polar asparagine for enzyme activity by ensuring the exact positioning of a catalytically active metal ion or the general three-dimensional structure of the active site. Because of their close vicinity to the first oxygen of FGly 69 , a direct involvement of the Mg 2ϩ and the ␦ oxygen of Asp 281 in catalysis by stabilizing the deprotonated form of this FGly oxygen is likely. The importance of the metal complexing residues (Asp 29 , Asp 30 , Asp 281 , and Asn 282 ) for the general structure of the active site is supported by the fact that in all of these mutants the K m was increased 3-8-fold. It should be noted, however, that modeling of the substrate into the cavity did not predict a direct contact to one of the substituted residues. Asp 29 and Asp 30 may influence substrate binding by their coordination of Lys 123 and Lys 302 , respectively, which are strong candidates for binding the sulfate group (see below).
Whereas the Mg 2ϩ and Asp 281 are thought to be involved in the activation of FGly 69 during the cleavage of the substratesulfate ester, His 125 is proposed to catalyze the desulfation of the enzyme-sulfate ester leading to the elimination of sulfate and to the regeneration of the aldehyde. Substitution of His 125 by alanine led to a decrease of V max to 11% and to a 7-fold increase of the K m , indicating a participation in the catalytic process but also pointing to an influence on the binding of the substrate (see below).
Sulfated substrates are expected to be bound by positively charged residues, thereby orientating the sulfate group next to the catalytically active oxygen of FGly 69 and increasing the electrophilic property of the sulfur. When the potentially positively charged residues in the vicinity of FGly 69 (His 125 , His 229 , Lys 123 , and Lys 302 ) were substituted by alanine, 5-fold (His 125 and His 229 ) and more than 10-fold (Lys 123 and Lys 302 ) increases of the K m were observed. Because a 3-5-fold increase of K m was also observed in other mutants, which are unlikely to interact directly with the substrate molecule, only the two lysines (Lys 123 and Lys 302 ) remain as strong candidates for binding the sulfate group. Additionally, the V max of all of these four mutants was decreased to 6 -11% of wild type ASA. In the H125A mutant this is obviously due to the interference with FGly coordination (see above). For K123A and K302A it may be ascribed to their direct vicinity to Asp 29 and Asp 30 , which are essential for the coordination of the Mg 2ϩ ion (see above).
Substitution of Ser 150 , a residue close to Lys 123 , led to a strong decrease of V max (2% of wild type) without affecting the K m . A direct participation of Ser 150 in catalysis cannot be expected, because of the large distance to FGly 69 (4.9 Å). At least part of the loss of activity may be explained by the close vicinity to Lys 123 , which coordinates Asp 29 , one of the metal complexing residues. The K m of the S150A mutant was in the range of the K m of wild type ASA, indicating that Ser 150 is not essential for binding of p-nitrocatechol sulfate as substrate.
The strong decrease of activity in mutants containing alanine instead of the metal complexing aspartates may be explained by the loss of negative charges coordinating the Mg 2ϩ ion and for Asp 281 additionally by the loss of direct FGly 69 activation. Substitution of aspartate by asparagine replaces one oxygen with a nitrogen. By its free pair of electrons the latter may partly take over the function of the negative charge of the oxygen. To analyze the effect of a reciprocal exchange of aspartate and asparagine in the metal binding site, we substituted Asp 29 , Asp 30 , and Asp 281 by asparagine and Asn 282 by aspartate. Surprisingly the kinetic parameters of all of these mutants were similar to those of the respective alanine substitution mutants (Table I). Only the D30N mutant showed a higher V max as compared with D30A, indicating that one oxygen of Asp 30 could partly be replaced with a nitrogen. Taken together, these results demonstrate that even relatively small changes in the metal binding complex, induced by aspartate/ asparagine exchanges, affect catalytic activity and substrate binding. In line with this, the metal complexing residues are highly conserved in the sulfatase family (Fig. 2).
The pH optimum of all mutants was shifted to a more alkaline pH by 0.2-0.8 units as compared with wild type ASA (Table I and Fig. 4). This change indicates that proton abstraction is a crucial aspect in the catalytic mechanism of sulfate ester cleavage. In wild type ASA both oxygens of the aldehyde hydrate have to be activated by abstraction of a proton to facilitate the cleavage of the sulfated substrate and the release of the sulfate group from the enzyme (see above and Fig. 1). Proton abstraction may be impaired by substituting either His 125 and Asp 281 , which are directly involved in these steps,  with alanine or by point mutations that alter the overall structure of the active site. An increase in pH promotes the deprotonation and could partially compensate for the loss of catalytic deprotonation. The latter appears to depend on the exact positioning of Asp 281 , Mg 2ϩ , and His 125 next to FGly 69 . In line with this, the decrease in V max observed in all of these mutants can at least partly be ascribed to an impaired deprotonation of the aldehyde hydrate.
Conclusions-The amino acid residues neighboring FGly 69 in the active site of ASA are highly conserved among eukaryotic and prokaryotic members of the sulfatase family, indicating a similar structure of the reactive center and a common reaction mechanism. Crystal structure analysis revealed that these residues are probably interconnected by a net of hydrogen bonds, thus forming the active site. The structural data are consistent with the biochemical finding that all ASA mutants, in which putative residues of the active site were replaced by alanine, have a markedly decreased V max . The increased pH optimum observed for all mutants is in good agreement with the decreasing V max , because in both of the two half-reactions of sulfate ester cleavage a deprotonation step is involved. Among the residues that have been proposed to bind the sulfate group, Lys 123 and Lys 302 were shown to be most critical for substrate binding. Their substitution by alanine led to a marked increase of K m . The moderate 3-5-fold increase of K m in the other mutants, except for S150A, supports the view of the active site of ASA (and other sulfatases) as a highly interconnected network of polar residues, easily disturbed by the replacement of one of its elements.