Trp-16 is essential for the activity of α-galactosidase and α-N-acetylgalactosaminidase

E L S E V I E R Biochimica et Biophysica Acta 1297 (1996) 99-104

Btl Biochi~ie~a e t B i o p h y s i c a / E t a

Trp-16 :is essential for the activity of a-galactosidase and a-N- acetyl g alactosamini dase

Alex Zhu *, Catherine Monahan, Zhong-Kun Wang The Lindsley F. Kimball Research Institute of The New York Blood Center 310 East 67 Street, New York, NY 10021, USA

Received 16 February 1996; revised 3 June 1996; accepted 5 June 1996

A b s t r a c t

By expressing site-directed mutants in the methylotrophic yeast strain Pichiapastoris, the role of a tryptophan residue at position 16 in the activity of a-galactosidase and a-N-acetylgalactosaminidase, two closely related exoglycosidases, was studied. A substitution of Trp-16 with an arginine residue in a-N-acetylgalactosaminidase abolished the enzyme activity, which was confirmed by replacing a 600 bp fragment containing the mutation with the corresponding wild-type sequence. The same tryptophen residue was then substituted with an alanine in both enzymes by site-directed mutagenesis to reveal a possible relationship between their active sites. The purified a-N-acetylgalactosaminidase mutant demonstrated a specific activity of 2.8 X 10 -2 U/mg and a Vmax/g m of 4.3 × 10 -2, which were both more than a thousandfold lower than corresponding values for the wild-type enzyme. Furthermore, the mutant failed to bind to an affinity resin, suggesting the involvement of Trp-16 in substrate-binding. In addition, the purified a-galactosidase mutant resulted in more than a 104- fo ld decrease in specific activity. Thus our data suggest that Trp-16 in both a-galactosidase and a-N-acetylgalactosaminidase is critical for enzymatic activity, which in tum supports the hypothesis that these two enzymes may share a catalytic mechanism involving similar residues in their active sites.

Keywords: a-Galactosidase; a-N-Acetylgalactosaminidase; Site-directed mutagenesis; Tryptophan; (Pichia pastoris)

1. I n t r o d u c t i o n

a-Galactosidase (a -GAL, EC 3.2.1.22) and a-N-

acetylgalactosaminidase ( a -NAGAL, EC 3.2.1.49) are exoglycosidases that cleave the terminal a-linked Gal and GalNAc, respectively, from glycoconjugates. Both enzymes have been identified and purified from a variety of prokaryotic and eukaryotic species [1-4]. They play an essential role in numerous metabolic pathways and their deficiency in humans may cause severe clinical disorders. For example, Fabry's disease, an X-linked inbom error of glycosphingolipid metabolism, results from the defective activity of lysosomal a -GAL [5] and Schindler's disease, an infantile neuroaxonal dystrophy characterized by the accumulation of glycoconiugates with a-N-acetylgalacto- saminyl moieties, results from deficient a -NAGAL activity [6]. In addition to their clinical relevance, a -GAL and a -NAGAL have been used in biomedical and biotech-

* Corresponding author at the Cell Biochemistry Laboratory. Fax: + 1 (212) 8790243.

nology settings, a -GAL purified from green coffee beans and a -NAGAL purified from chicken liver have been shown to effectively remove the terminal sugar residues that define antigens B and A, respectively, from glycoconjugates of the red cell surface, thus generating group O cells [7-9]. Since group O erythrocytes are considered 'universal donor cells', such an in vitro conversion procedure would have enormous impact on transfusion medicine [101.

cDNAs coding for a -GAL have been cloned from a variety of sources [11-15]; whereas, only two gene sequences for a -NAGAL (human and chicken) have been elucidated [16,17]. Amino-acid sequence alignment indicated a high homology (50-70%) between these two exoglycosidases [16]; furthermore, studies of genomic structures of human a -GAL and a -NAGAL genes revealed identical exons interrupted by the first six introns [18]. Although these two glycosidases have distinctive substrate specificities, the sequence alignment data and intron-exon structures suggest a -GAL and a -NAGAL are evolutionar- ily related and arose through duplication and divergence from a common ancestral gene. Many mutant a -GALs

0167-4838/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PII S0167-4838(96)00108-2

100 A. Zhu et al. / Biochimica et Biophysica Acta 1297 (1996) 99-104

have been identified and characterized from patients with Fabry disease [19], however, the catalytic mechanism of the enzyme and its active site have yet to be elucidated at the molecular level. Chemical modification of c~-GAL isolated from coconut kernels suggested that a tryptophan, a tyrosine and two carboxyl groups were at or near the active site [20]; thus, an ac id -base catalytic mechanism of action for c~-GAL was proposed [21]. On the other hand, the biochemical and molecular aspects of the active site of c~-NAGAL are not as well understood. Availabil i ty of cDNA clones and ease of site-directed mutagenesis have made it possible to localize the critical residues involved in the active site of c~-GAL and c~-NAGAL, thereby shed- ding some light upon their catalytic mechanisms. For example we have substituted a tyrosine residue at position 108 (Tyr-108) of coffee bean ~-GAL with a phenylalanine by site-directed mutagenesis. The mutant was then expressed in a baculovirus system, which resulted in a 1000- fold reduction in Vmax/Krn value, indicating that Tyr-108 was a critical residue for enzyme activity [22]. Further- more, Tyr-108 is also conserved among all c~-NAGALs cloned to date, raising the possibili ty that the two families of enzymes may have similar residues involved in the active site and share a similar catalytic mechanism.

Evidence is presented here which indicates that the tryptophan residue at position 16 (Trp-16) is essential for the enzyme activities of coffee bean a - G A L and chicken liver c~-NAGAL. To the best of our knowledge, Trp-16 is the first example of residue determined to be critical for the activity of both c~-NAGAL and a -GAL, supporting the hypothesis that these two enzymes may share a catalytic mechanism involving similar residues in their active sites.

2. Mater ia l s a n d m e t h o d s

2.1. Fragment replacement

The cDNA encoding chicken liver a - N A G A L was cloned into the E c o R I site of the Pichia pastoris expression vector pHIL-S1, generating pHO-N(wt) (Fig. I A). During the subcloning and expression of the cDNA, a single point mutation at nucleotide 45 ( + 1 is the beginning of mature c~-NAGAL) was identified which in turn caused the substitution of tryptophan with arginine at position 16 of c~-NAGAL and the clone was designated pHO-N(W16R). In order to replace a fragment containing the mutation in pHO-N(W16R) with the corresponding

(A) Trp-16

Xho I [ Bgl II

Bg'II~o N(~ ) jc~ a

Arg-16

Bgl n , Fragment (ab) replacement ] ,

pHO-N(600R) pHO-N(600W)

~) 1 2 3 4 5 6 7 8 9

ccNAGAL 3L- 2~ ~./ .e.~Z t.~ L~" /:~ /.,7,, / . ' /

Fig. 1. Partial restriction map of pHO-N(wt) for construction of pHO-N(600W) and pHO-N(600R) and their expression in P. pastoris (A) The plasmid, pHO-N(wt), is composed of the vector pHIL-S1 and o~-NAGAL cDNA (Thick bar) cloned at EcoRI site. Trp-16 indicates the wild-type sequence for a-NAGAL. The plasmid pHO-N(W16R) is the same as pHO-N(wt) except for the substitution of Trp-16 with Arg. To construct pHO-N(600R), the XhoI/BgllI Fragment ab (shaded bar) in pHO-N(wt) was replaced by the corresponding mutant Fragment ab from pHO-N(W16R). Similarly. pHO-N(600W) was generated from pHO-N(W16R) with the mutant Fragment ab replaced by the wild-type XhoI/BgllI Fragment ab. (B) Both pHO-N(600W) and pHO-N(600R) were transformed into P. pastoris. The induced culture supernatants were examined by a Western blot (10 ILl each) and assayed for a-NAGAL activity (15 Ixl each). Lane 1 is 0.1 Ixg of a-NAGAL(nat) as a positive control, lane 2 is 10 ILl of the supernatant from pHO-N(600W) transformed cells. Lanes 3 through 9 are 10 Ixl of each supernatant from 7 randomly picked pHO-N(600R) transformants. The a-NAGAL activities (mU/ml) corresponding to the samples in each lane are shown in the panel below the Western blot.

A. Zhu et al. / Biochimica et Biophysica Acta 1297 (1996) 99-104 101

wild-type sequence from pHO-N(wt), two restriction enzyme digestions were necessary for both plasmids. Refer- ring to Fig. 1A, fragments ab (0.6 kb) and bc (1.5 kb) were isolated from XhoI /Bgl I I /XbaI triple digestion and fragment ca (7.4 kb) was isolated from XhoI/XbaI double digestion. Gel-purified fragments bc and ca from pHO-N(wt) were ligated with fragment ab from pHO- N(W16R), generating pHO-N(600R). Similarly, gel-purified fragments bc and ca from pHO-N(W16R) were ligated with Fragment ab frcm pHO-N(wt), generating pHO- N(600W). Automated DNA sequencing determined that fragment ab of pHO-N(W16R) differed from fragment ab of pHO-N(wt) by a single nucleotide change (T ~ C), which resulted in the W16R mutation.

culture supematant was collected by removing cells fol- lowing centrifugation at 3500 × g for 30 min. The samples were concentrated with a 10000 molecular weight cutoff ultra-filtration membrane (Millipore) and then applied to a cation exchange column, Macro prep S-50 (Bio-Rad) according to the procedure previously reported [25]. The expressed protein was eluted from the column with a NaC1 gradient ranging from 50 mM to 500 mM. By analyzing different fractions on a SDS-PAGE stained with Coomassie brilliant blue, the samples showing a single band on the gel were combined and stored at -20°C for further characterization.

2.5. Protein characterization

2.2. Site-directed mutagenesis

The tryptophan at position 16 of both a-NAGAL and a-GAL, was replaced by an alanine as a result of site-directed PCR mutagenesis according to published procedures [23,24]. Thus, the oligodeoxynucleotide, 5'-CTGGTTG- GCCGCGGAGCGGTTC-3', was used to introduce the mutation W16A for a-NAGAL, generating the plasmid pHO-N(W16A) and the oligodeoxynucleotide, 5'-GTG- GAACAGCCGCGAATCATTTC-3' was used to introduce the mutation WI6A for a-GAL, generating the plasmid pHO-G(W16A). The wild-type (wt) plasmid pHO-G(wt) was generated from subchming a-GAL cDNA into EcoRI site of pHIL-S 1.

2.3. Pichia pastoris transJbrmation and selection of clones

P. pastoris strain GS-115 (His4-) from Invitrogen was transformed by both the spheroplast method recommended by the manufacturer and by electroporation with the Cell Porator (Gibco) at 440 V, low [l and 10 /zF. The yeast transformants (His + Mut +) were tested directly for enzyme activity by a screening process we reported before [25]. 5-bromo-4-chloro-3-indolyl-a-D-galactopyranoside, (X-a-gal, 1 mg/ml in phosphate buffer, pH 6.5) was used for a-GAL and 5-bromo-4-chloro-3-indolyl-a-D-2- acetylamido-2-deoxygalactopyranoside (X-a-galNAc, 1 mg/ml in phosphate buffer, pH 3.64) for a-NAGAL. The expression of recombinant proteins was examined by a Western blot.

2.4. Sample preparation and protein purification

The enzyme assay consisted of incubating the a-GAL samples with 1.25 mM of p-nitrophenol-a-galactopyranoside (PNP-a-Gal), at pH 6.5 and 37°C and incubating the a-NAGAL samples with 2.5 mM of p-nitrophenol-a-N- acetylgalactosaminoside (PNP-a-galNAc) at pH 3.65 and 37°C. The assay reaction was terminated by adding 0.2 M borate buffer (pH 9.8), and then the absorbance of the reaction at 405 nm was measured. One unit (U) of activity is defined as the amount of enzyme that hydrolyzes 1 Ixmol of substrate per minute under the assay conditions. Protein concentration was measured by using BCA Protein Assay (Pierce) with bovine serum albumin (BSA) as the standard. Protein samples were analyzed on 12% acryl- amide gels (precast gel from Bio-Rad). Western blots were visualized using the ProtoBlot system (Promega) with polyclonal antisera raised against either purified chicken liver a-NAGAL or coffee bean a-GAL, according to the recommended protocol. Substrate concentrations used in the kinetic studies, range from 0.1-2.0 mM PNP-a-gal for a-GAL, 0.2-5 mM PNP-a-galNAc for a-NAGAL (W16A). The enzyme activity was measured at various time points to calculate the initial velocity (V i) at a certain substrate concentration ([S]). For each enzyme, at least five [S]-V i values were needed to calculate Vma x and K m using the computer program Enzpack 3 ® (Biosoft, Cambridge, UK) based on the method of Lineweaver-Burk.

3. Results and discussion

3.1. Trp-16--* Arg mutation is responsible for the loss of a-NAGAL activity

P. pastoris transformants chosen for characterization were grown in either a fermentor or shake-flask culture [25]. While the conditions for the fermentor culture have been standardized [25], modifications of the shake-flask culture protocol (Invitrogen) included a medium containing 1% casamino acids, (I.8 mM biotin, 1% glycerol, and 1.34% yeast nitrogen base in 0.1 M potassium phosphate buffer (pH 5.0). After 4 days of enzyme induction, the

In the course of studying recombinant a-NAGAL expressed in P. pastoris, a clone was identified which produced a-NAGAL at a level comparable to that of a- NAGAL(wt); however, this clone was found to be enzymatically inactive. DNA sequencing analysis revealed a single nucleotide mutation ( -TGG-~ -CGG-) that resulted in the substitution of tryptophan at position 16 with arginine. Based on amino-acid sequence alignment, the Trp-16

102 A. Zhu et aL / Biochimica et Biophysica Acta 1297 (1996) 99-104

is conserved not only in human and chicken c~-NAGALs but also in c~-GALs cloned from sources including human, yeast and plants [28-30], suggesting that the Trp-16 residue might play a functionally important role. Data from our previous work [16] and others [18] suggested that although ce-NAGAL and c~-GAL have distinctive substrate specificities they may have evolved from a common ancestral gene. In order to confirm that the inactivation of this a- NAGAL-producing clone is the result of Trp-16--* Arg mutation and not due to other changes in the protein or host cell line, the X h o I / B g l I Fragment ab containing the (W16R) mutation was replaced with the corresponding Fragment ab from the wild-type plasmid. Thus, the plasmid pHO-N(600R) is identical to pHO-N(wt) except for the T ~ C mutation and the plasmid pHO-N(600W) is identical to pHO-N(W16R) except for the wild-type codon for Trp-16. The screening for enzyme activity demonstrated that P. pastoris transformed with pHO-N(600W) produced active c~-NAGAL while pHO-N(600R) transformants were enzymatically inactive, suggesting that the Trp-16 ~ Arg mutation is responsible for the loss of c~- NAGAL activity. In order to ensure that the lack of c~-NAGAL activity in pHO-N(600R) transformants was not simply the failure to produce any enzyme, it is necessary to examine the expression of the recombinant enzyme by a Western blot. Not all P. pastoris transformants are capable of expressing recombinant proteins [25]; thus, seven pHO-N(600R) transformants were randomly chosen for growth in flasks. As a control, one pHO-N(600W) transformant, which tested positive was also chosen for Western blot analysis. After methanol induction, the culture supernatants were collected both for the c~-NAGAL assay with PNP-c~-galNAc and for the Western blot with a polyclonal antibody raised against a-NAGAL purified from chicken liver (Fig. 1B). The supernatant from pHO- N(600W) transformed cells (lane 2) expressed a-NAGAL, which demonstrated the enzymatic activity comparable to that of ce-NAGAL (nat) (lane 1), indicating that a-NAGAL activity was restored by replacing Arg-16 with Trp. On the other hand, five of the seven pHO-N(600W) transformants expressed the enzyme as shown in lanes 3 -9 (Fig. 1B), but remained enzymatically inactive, which confirmed that the Trp-16--* Arg mutation is responsible for the loss of c~- NAGAL activity.

3.2. Substitution o f Trp-16 with Ala abolished ce-GAL activity as well as c~-NAGAL activity

In order to rule out the possibility that the activity of o~-NAGAL(W16R) was abolished simply because of the introduction of a strongly positive charged group from the arginine residue, Trp-16 was substituted with an alanine residue by site-directed mutagenesis. The resulting plasmid, pHO-N(W16A), was generated and transformed into P. pastoris for the expression of the mutant enzyme, o~-NAGAL(W 16A). Similar to the studies of pHO-N(600R)

described above, c~-NAGAL(W16A) was expressed nor- mally in P. pastoris; however, the enzymatic activity remained undetectable under standard conditions, further suggesting that Trp-16 is essential for a-NAGAL activity.

Once Trp- 16 was confirmed to be critical for c~-NAGAL activity, the importance of the corresponding residue in c~-GAL was investigated. Based on sequence alignment between these two enzymes, the corresponding Trp in coffee bean c~-GAL is also at position 16 ( + 1 is the beginning of mature c~-GAL). Site-directed mutagenesis was used to alter the codon TGG for Trp-16 in the plasmid pHO-G(wt) to GCG coding for Ala-16, generating pHO- G(W16A). Examined by the screening procedure and Western blot, the mutant protein a-GAL(WI6A) was produced in P. pastoris but enzymatically inactive.

3.3. Purification and characterization o f mutant enzymes

Further characterization and quantitative comparisons of the catalytic activity of the mutated enzymes with their wild-type counterparts became possible after the enzyme expression and purification. Coffee bean c~-GAL (wt) has been expressed in P. pastoris and purified to apparent homogenicity by a single chromatography on the Macro Prep S-50 column [25]. Recently the expression and purification of chicken c~-NAGAL (wt) have been accomplished in our laboratory by applying a similar procedure (unpublished data). Small aliquots (between 1-3.8 Ixg) of purified recombinant proteins were loaded onto a 0.1% SDS-12% PAGE and stained with Coomassie brilliant blue. As shown in Fig. 2, all c~-NAGAL preparations: wt, W16R, and W16A (lanes 2, 3 and 4, respectively), contained double bands of an approximate molecular mass of 50 kDa. Both bands were immunologically cross-reactive with the antibody against ce-NAGAL purified from chicken liver (data not shown). Although the recombinant c~-NAGAL pro-

1 2 3 4 5 6 7 8

97.4 66.2

45

31

21.5

14.4

Fig. 2. SDS-PAGE of purified a-NAGALs and a-GALs. Protein samples were loaded onto a 0.1% SDS-12%PAGE followed by Coomassie brilliant blue staining. Lanes 1-4 are a-NAGAL samples: nat (1.3 txg), wt (1.2 /,zg), W16R (3.8 l, tg) and W16A (1.0 l, zg), respectively. Lanes 5-7 are o~-GAL samples: W16A (2.0 I~g), wt (1.3 I~g), and nat (0.7 txg), respectively. Lane 8 shows a low molecular weight marker (Bio-Rad) with sizes given in kDa.

A. Zhu et al. / Biochimica et Biophysica Acta 1297 (1996) 99-104 103

duced in P. pastoris migrated more slowly than its native counterpart, c~-NAGAL(nat), as shown in Fig. 1B and Fig. 2, this discrepancy was eliminated by treating the recombinant proteins with N-glycosidase F (unpublished data). Thus the double bands observed on the Coomassie stained gels may represent recombinant ce-NAGAL with different degrees of hyperglycosylation generated by P. pastoris host cells. On the other hand, purified c~-GALs: wt and W16A (lanes 5 and 6 in Fig. 2), contained only a single band, which migrated to the same position as native enzyme purified from coffee bean (lane 7). Our previous study indicated that both native and recombinant cr-GAL were not glycosylated [26].

The specific activities and kinetic parameters of the purified recombinant enzymes were compared with those of both the wild-type and native enzymes. As shown in Table 1, c~-NAGAL(W16A) has a specific activity of 2.8 × I0 -2 U / m g and Vmax/Km of 4.3 × 10 - 2 which are more than a thousand fold lower than corresponding values for ce-NAGAL(wt) or a-NAGAL(nat). Furthermore, the specific activity of c~-NAGAL(WI6R) is 4.0X 10 4 U/mg, which is five orders of magnitude lower than c~-NAGAL(wt). Note that ce-NAGAL(W16R) demonstrated only 1% of the activity of c~-NAGAL(W16A), suggesting that Arg at position 16 has an additional ad- verse effect on enzyme activity. Nevertheless, the data confirmed the importance of Trp-16 for ce-NAGAL activity. In addition, both c~-NAGAL mutants (W16A and W16R) lost the ability to specifically bind to E-amino- caproyl- c~-o-galactopyrano sylamine-Sepharose conjugate [27], which has been used as an affinity resin for purification of native c~-NAGAL [4]. Thus the data suggest that Trp-16 may be directly involved in the substrate-binding site of ~-NAGAL.

The impact of the Trp--16-~ Ala mutation on enzyme activity was more profound in c~-GAL than in c~-NAGAL. As shown in Table l, the substitution of Trp-16 with Ala in c~-GAL resulted in more than 10 - 4 fold decrease in its specific activity, suggesting that Trp-16 is essential for ce-GAL activity.

Based on a chemical modification study of cr-GAL isolated from coconut kernels, Mathew and Balasubrama- niam [21] proposed a general acid-base mechanism for c~-GAL, involving a deprotonated carboxyl group for stabi- lizing a carbonium ion intermediate and a second proto- nated carboxyl group for donating a proton. The second carboxyl group (Asp/Glu residue), acting as a general acid, should have a higher p K a value, presumably due to the presence of a tryptophan and a tyrosine residue in the vicinity. Data generated from the study of coffee bean c~-GAL by applying inhibitors [31] and sterochemistry [32] are consistent with the catalytic mechanism proposed. Although the Glu/Asp residues functioning as general acid and base in the catalysis have yet to be identified, Trp-16, as well as Tyr-108 which was reported previously [22], may be candidates for influencing the carboxyl group acting as general acid as proposed by Mathew and Bala- subramaniam [21].

The reason for choosing Ala in place of Trp-16 in the mutagenesis study of ce-GAL is to eliminate the indole ring of Trp, which is most likely the functional group. A substitution of Trp with non-aromatic residues such as Ala, Gly or Leu has been successfully used in studies of catalytic mechanisms of various enzymes [33-35]. Al- though the Trp residue at position 16 has been shown to be critical for the cr-GAL activity, it remains to be addressed whether the indole ring of Trp-16 is absolutely necessary or its function in the catalysis may be substituted by other aromatic amino acids in the same position. This question can be answered by marking c~-GAL mutants with Trp-16

Tyr or Trp-16 --* Phe substitution. Comparison of deduced amino acid sequences of

chicken liver c~-NAGAL and coffee bean a-GAL demon- strates 60% homology if conserved amino-acid substitu- tions are included, or 43% homology if only considering invariant residues (PROSIS program from Hitachi Software Engineering). Our data support the hypothesis that these two exoglycosidases may have similar residues at the active site and may share a similar catalytic mechanism. The data presented here will prove important in the future

Table 1 Comparison of wild-type and row:ant tx-NAGALs and oL-GALs produced in P. pastoris

Enzyme Specific activity ( U / m g ) K m (raM) Vma x ( /~mol/min per mg) Vmax/K m

ot NAGAL(nat) 42.2 0.8 75.7 94.6 ozNAGAL(wt) 38.9 0.8 60.9 76.1 o~NAGAL(W16A) 2.8 X 10 2 4.7 0.2 4.3 x 10 -2 otNAGAL(W16R) 4.0 X 10 -4 N / A N / A N / A a GAL(nat) 31.9 0.4 44.8 112.0 a GAL(wt) 31.2 0,4 46.9 117.3 aGAL(W16A) 7.7 × 10 -5 N / A N / A N / A

The specific activity and Vmax//K'm values were measured for purified recombinant proteins according to the procedures described in Section 2. The enzyme activity levels of a-NAGAL(W16R) and ot-GAL(W16A) were so low that measuring the kinetic parameters was both difficult and inaccurate, therefore, these numbers were nol available (N/A).

104 A. Zhu et a l . / Biochimica et Biophysica Acta 1297 (1996) 99-104

deduct ion of the structures of c~-GAL and c~-NAGAL by

X-ray crystal lography which in turn will aid in the elucida-

tion o f the catalytic mechan i sms of these enzymes .

Acknowledgements

The authors wish to thank Dr. Jack Golds te in for his

support and critical reading of the manuscript . This work

was supported in part by grant no. N00014-93-1-0466

from the Naval Medica l Research and D e v e l o p m e n t C o m -

mand (to J. Goldstein).

References

[1] Dey, P.M., Del Campillo, E.M. and Lezica, R.P. (1983) J. Biol. Chem. 258, 923-929.

[2] Dean, K.J. and Sweeley, C.C. (1979) J. Biol. Chem. 254, 9994- 10000.

[3] Uda, Y., Li, S.-C. and Li, Y.-T. (1977) J. Biol. Chem. 252, 5194-5200.

[4] Hata, J., Dhar, M., Mitra, M., Harmata, M., Haibach, F, Sun, P. and Smith, D. (1992) Biochem. Int. 28, 77-86.

[5] Desnick, R.J. and Eng, C.M. (1993) in Dermatology in General Medicine (Fitzpatrick, T.B,, Eisen, A.Z., Wolff, K., Freedberg, I.M. and Autin, K.F., eds.), 4th Ed., pp. 1916-1937, McGraw-Hill, New York.

[6] Schindler, D., Bishop, D.F., Wolfe, D.E., Wang, A.M., Egge, H., Lemieux, R.U. and Desnick, R.J. (1989) N. Engl. J. Med. 320, 1735-1740.

[7] Goldstein, J., Siviglia, G., Hurst, R. and Lenny, L. (1982) Science 215, 168-170.

[8] Goldstein, J. (1989) Trans. Med. Rev. 111,206-212. [9] Lenny, L.L., Hurst, R., Zhu, A., Goldstein, J. and Galbraith, R.A.

(1995) Transfusion 35, 899-902. [10] Wilson, R.B. and Spitalnik, S.L. (1994) Transfusion, 34, 189-191. [11] Den Herder, I.F., Rosell, A.M.M., Van Zuilen, C.M., Punt, P.J. and

Van den Hondel, C.A.M.J.J. (1992) Mol. Gen. Genet. 233, 404-410.

[12] Bishop, D.F., Calhoun, D.H., Bernstein, H.S., Hantzopoulos, P., Quinn, M. and Desnick, R.J. (1986) Proc. Natl. Acad. Sci. USA 83, 4859-4863.

[13] Zhu, A. and Goldstein, J. (1994) Gene 140, 227-231 [14] Overbeeke, N., Fellinger, A.J., Toonen, M.Y., Van Wassenaar, D.

and Verrips, C.T. (1989) Plant Mol. Biol. 13, 541-550. [15] LiljestriSm, P.L. (1985) Nucleic Acids Res. 13, 7257-7268. [16] Zhu, A. and Goldstein, J. (1993) Gene 137, 309-314. [17] Wang, A.M., Bishop, D.F. and Desnick, R.J. (1991) J. Biol. Chem.

265, 21859-21866. [18] Wang, A.M. and Desnick, R.J. (1991) Genomics 10, 133-142. [19] Okumiya, T., Ishii, S., Kase, R. Kamei, S., Sakuraba, H. and Suzuki,

Y. (1995) Hum. Genet. 95, 557-561. [20] Mathew, C.D. and Balasubramaniam, K. (1986) Phytochemistry 25,

2439-2443. [21] Mathew, C.D. and Balasubramaniam, K. (1987) Phytochemistry 26,

1299-1300. [22] Zhu, A., Wang, Z.-K. and Goldstein, J. (1995) Biochim. Biophys.

Acta 1247, 260-264. [23] Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. and Pease, L.R.

(1989) Gene 77, 51-59. [24] Mikaelian, I. and Sergeant, A. (1992) Nucleic Acid Res. 20, 376. [25] Zhu, A., Monahan, C., Zhang, Z., Hurst, R. Leng, L. and Goldstein,

J. (1995) Arch. Biochem. Biophys. 324, 65-70. [26] Zhu, A., Leng, L., Monahan, C., Zhang, Z., Hurst, R., Lenny, L. and

Goldstein, J. (1996) Arch. Biochem. Biophys. 327, 324-329. [27] Harpaz, N., Flowers, H.M. and Sharon, N. (1974) Biochim. Bio-

phys. Acta 31,213-221. [28] Bishop, D.F., Calhoun, D.H., Bernstein, H.S., Hantzopoulos, P.,

Quinn, M. and Desnick, R.J. (1986) Proc. Natl. Acad. Sci. USA 83, 4859-4863.

[29] Liljestr~Sm, P.L. (1985) Nucleic Acids Res. 13, 7257-7268. [30] Overbeeke, N., Fellinger, A.J., Toonen, M.Y., Van Wassenaar, D.

and Verrips, C.T. (1989) Plant Mol. Biol. 13, 541-550. [31] Wang, Y.-F., Takaoka, Y. and Wong, C.-H. (1994) Angew. Chem.

Int. Ed. Engl. 33, 1242-1244. [32] Weiser, W., Lehmann, J., Matsui, H., Brewer, C.F. and Hehre, E.J.

(1992) Arch. Biochem. Biophys. 292, 493-498. [33] Lentz, M.R., Webster, R.G. and Air, G.M. (1987) Biochemistry 26,

5351-5358. [34] SCgaard, M., Kadziola, A., Haser, R. and Svenson, B. (1993) J. Biol.

Chem. 278, 22480-22484. [35] Aoki, D., Appert, H.E., Johnson, D., Wong, S.S. and Fukuda, M.N.

(1990) EMBO J. 9, 3171-3178.

Documents

Trp-16 is essential for the activity of α-galactosidase and α-N-acetylgalactosaminidase