© 2001 Oxford University Press
Human Molecular Genetics, 2001, Vol. 10, No. 9
983–995
Molecular pathogenesis of a disease: structural consequences of aspartylglucosaminuria mutations Jani Saarela1, Minna Laine1, Carita Oinonen2, Carina von Schantz1, Anu Jalanko1, Juha Rouvinen2 and Leena Peltonen1,3,+ 1Department
of Medical Genetics, University of Helsinki and Department of Molecular Medicine, National Public Health Institute, Haartmaninkatu 8, P.O. Box 104, FIN-00251 Helsinki, Finland, 2Department of Chemistry, University of Joensuu, P.O. Box 111, FIN-87101 Joensuu, Finland and 3Department of Human Genetics, UCLA School of Medicine, 695 Charles E. Young Drive South, Box 708822, Los Angeles, CA 90095-7088, USA
Received 16 January 2001; Revised and Accepted 2 March 2001
A deficiency of functional aspartylglucosaminidase (AGA) causes a lysosomal storage disease, aspartylglucosaminuria (AGU). The recessively inherited disease is enriched in the Finnish population, where 98% of AGU alleles contain one founder mutation, AGUFin. Elsewhere in the world, we and others have described 18 different sporadic AGU mutations. Many of these are predicted to interfere with the complex intracellular maturation and processing of the AGA polypeptide. Proper initial folding of AGA in the endoplasmic reticulum (ER) is dependent on intramolecular disulfide bridge formation and dimerization of two precursor polypeptides. The subsequent activation of AGA occurs autocatalytically in the ER and the protein is transported via the Golgi to the lysosomal compartment using the mannose-6-phosphate receptor pathway. Here we use the three-dimensional structure of AGA to predict structural consequences of AGU mutations, including six novel mutations, and make an effort to characterize every known disease mutation by dissecting the effect of mutations on intracellular stability, maturation, transport and the activity of AGA. Most mutations are substitutions replacing the original amino acid with a bulkier residue. Mutations of the dimer interface prevent dimerization in the ER, whereas active site mutations not only destroy the activity but also affect maturation of the precursor. Depending on their effects on the AGA polypeptide the mutations can be categorized as mild, moderate or severe. These data contribute to the expanding body of knowledge pertaining to molecular pathogenesis of AGU. INTRODUCTION Mutations in genes encoding functional domains of human proteins result in human genetic diseases. Sickle-cell anemia was the first human genetic disorder providing evidence that a +To
single amino acid change can dramatically alter the properties of a protein (1). Since then, specific mutations of some 1100 human diseases have been identified; however, in only a fraction of these are consequences of mutations for the structure or cellular metabolism characterized in detail. The most common inherited disorder of glycoprotein catabolism is aspartylglucosaminuria (AGU; McKusick 208400). AGU results from the deficient activity of a lysosomal amidase, aspartylglucosaminidase [AGA, glycosylasparaginase, N4-(β-N-acetylglucosaminyl)-l-asparaginase, EC 3.5.1.26]. This recessively inherited condition is manifested by excessive accumulation of uncleaved glygoasparagines, mainly aspartylglucosamine [2-acetamido-1-(β-l-aspartamido)-1,2-dideoxyglucose], in lysosomes as well as elevated metabolite levels in urine (2). Patients exhibit a relatively uniform clinical phenotype characterized by progressive mental retardation from early childhood with minor connective tissue changes and premature death (3). AGU is enriched in Finland: one mutation, denoted AGUFin, represents 98% of the AGU alleles due to a founder effect (4). Previously, altogether 20 different AGU alleles have been characterized and 18 of them represent sporadic AGU alleles found outside Finland (5–16). The synthesis of the functional AGA molecule is relatively complex. It is translated as a single precursor polypeptide chain of 346 amino acids (5) and the signal peptide is cleaved co-translationally in the endoplasmic reticulum (ER), where initial folding of the polypeptide occurs. Also in the ER, two precursor polypeptides dimerize (17) and the dimeric precursor complex is activated by an autocatalytic cleavage of the peptide bond between amino acids D205 and T206, resulting in a tetrameric molecule consisting of two α- and two β-subunits (18,19). The cleavage of the precursor into α- and β-subunits exposes the α-amino group of T206 necessary for the catalytic mechanism of AGA. The α-amino group of T206 in the Nterminus of the β-subunit acts as a base, which increases the nucleophilicity of the side-chain OH group (20). In the Golgi, the glycosylation of the catalytically active molecule is completed and the lysosomal targeting signal, mannose-6phosphate, is specifically added to the oligosaccharides of AGA, directing the mature hydrolase to the lysosomal compartment. The phosphorylation of AGA is dependent on
whom correspondence should be addressed. Tel: +1 310 794 5631; Fax: +1 310 794 5446; Email:
[email protected]
984
Human Molecular Genetics, 2001, Vol. 10, No. 9
three lysine residues and a tyrosine in spatially defined positions on the surface of the mature AGA molecule (21). The three-dimensional structure of AGA is known based on X-ray crystallography. The details of the active site and the roles of individual amino acids for the catalytical activity are also known (20). Two heterodimers both contain one α- and one β-subunit arranged into four layers; two central β-sheet layers are surrounded by α-helical layers, demonstrating that AGA is a member of the N-terminal nucleophile protein family (22,23). The cell biological and biochemical characterization of mutated AGA polypeptides has included analyses of intracellular targeting and preliminary quantitative and qualitative analyses of polypeptides using immunoprecipitation and western blotting. However, detailed molecular characterization of the consequences of AGU mutations has not been performed in the majority of disease mutations. AGA represents a multimeric molecule which is functionally very sensitive to mutations. Firstly, the formation of the active molecule requires correct folding in the ER for the oligomerization and subsequent autocatalytic activation to take place (17). Secondly, in the Golgi, the three-dimensional structure of the molecule determines whether phosphorylation and transport of AGA to lysosomes occurs or not (21). Only if the AGA molecule is properly activated and transported to the lysosomes does the enzyme become functional in the correct subcellular location (24). Based on the three-dimensional structure of AGA we have here performed structural analyses of AGU mutations, including six novel mutations. We addressed the molecular pathogenesis of AGU by predicting the consequences of mutations for the three-dimensional structure of the AGA enzyme. Further, we expressed mutagenized polypeptides and monitored polypeptide stability, maturation, transport and enzymatic activity. RESULTS Identification of novel AGU mutations To identify the molecular defect in AGU in altogether 10 uncharacterized patients originating from Italy, Pakistan, Finland or the United Kingdom, the patients’ genomic DNA was amplified by PCR with intron-specific primers (11). We determined the nucleotide sequences of each exon, exon– intron boundaries and untranslated regions, and identified six previously unknown mutations. Two Italian compound heterozygote patients had the double mutation 34G→T + 1000G→T in one allele, which results in V12L and E334X at the protein level. In this family, we were unable to find disease-causing mutations in the other allele. Some DNA variants were found in introns but no further analysis of these was carried out due to a lack of patient cells. In two AGU families originating from Pakistan, the three patients were homozygous for the mutation 372_375del resulting in a frameshift and premature stopcodon, T125fsX126. One Finnish patient had a 755G→A mutation resulting in G252E in one allele, and the other allele contained the AGUFin mutation. Interestingly, the 755G→A mutation is a de novo mutation, since the mutation was not found in the parents, whose parenthood was confirmed by genotyping with a 99.9% confidence level. The other Finnish patient had the AGUFin mutant allele and the other allele
contained the 770C→T mutation resulting in T257I. A British AGU patient had the 503G→A mutation in one allele resulting in W168X at the polypeptide level; on examining the other allele we were again unable to find disease-causing mutations in the exons, exon–intron boundaries, or in the 5′- or 3′untranslated regions. Some intronic DNA variants were found but no further studies were carried out due to a lack of patient cells. Finally, two affected siblings from Italy were found to be homozygous for the mutation 754G→C leading to a G252R substitution in the AGA polypeptide. All the mutations were confirmed in the non-coding strand by sequencing. The novel mutations are listed in addition to previously published mutations in Table 1 and schematically presented in Figure 1. Some AGU mutations hit highly conserved amino acids The nucleophile of AGA, T206, residing in the N-terminus of the β-subunit, is fully conserved across the species (Fig. 2). Other active site residues show high conservation levels as well, but the lysosomal targeting residues, glycosylation sites, and disulfide bridge-forming residues (C64–69, C163–179, C286–306, C317–345) are conserved in mammals only. The positions of missense point mutations resulting in AGU affect amino acids with variable levels of conservation across the species (Table 2). Three mutations hit residues in the active site, and three substitutions replace fully conserved residues whereas 12 substitutions change a less conserved residue. Figure 3 displays the location of AGU mutations in the threedimensional AGA structure. Mutations commonly disturb the packing of the AGA molecule In the light of three-dimensional information for the active AGA molecule, the mutations A42_A43insDA, G60D, S72P, G100E, A101V, C163S, G252E, G252R, T257I, G302R and C306R are located in a structural environment, which is insufficient to accommodate the replacing residue and thus disturb the correct packing of secondary structure elements. Figure 3 shows the locations of the point mutations at the polypeptide level and Figure 4 illustrates the detailed structural environment of the wild-type residues in the native AGA three-dimensional structure. Table 2 summarizes the mutation characterization data. In the case of the mutations R161Q+C163S and C306R the substitution of cysteine prevents the formation of a disulfide bridge in the initial folding of the AGA precursor polypeptide in the ER. G226 is located in the active site in a loop, which contains substrate-binding residues R234 and D237. Two mutations, A101V and F135S, reside on or near the interface between the two αβ-dimers of AGA and they were found to affect the dimerization of AGA. The replacement F135S introduces a hydrophilic residue in a hydrophobic environment in the dimer interface. Mutant AGA polypeptides have defects in intracellular processing We mutagenized AGA cDNA to create the new mutation constructs and expressed them as well as 16 previously published mutations. The effects of the mutations on the maturation of the AGA polypeptide were studied using the SVpoly expression vector and transient expression in COS-1
Human Molecular Genetics, 2001, Vol. 10, No. 9
985
Table 1. Summary of AGU mutations Type of variation
cDNA levela
Protein level
Remarks/References
Ethnic origin
Substitution
[34G →T + 1000G→T] + [?]
[V12L + E334X]
Compound heterozygote, mutation in one allele unknown
Italian
Deletion (7 bp) and substitution
[101_107del] + [302C→T]
[W34fsX45]b + [A101V]
Compound heterozygote 9
British
Insertion (6 bp)
[127_128insATGCGG] + [127_128insATGCGG]
[A42_A43insDA]
7
Tunisian
Substitution
[179G→A] + [179G→A]
[G60D]
7
German
Substitution
[192T→A] + [192T→A]
[C64X]
10
Puerto Rican
Deletion (2 bp) and substitution
[199_200del] + [482G→A + 488G→C]
[E65fsX70] + [R161Q + C163S]
Compound heterozygote 11
Finnish
Substitution
[214T→C] + [214T→C]
[S72P]b
14
Palestinian Arab
Substitution and substitution
[299G→A] + [404T→C]
[G100E] + [F135S]
Compound heterozygote 15
Canadian
Substitution
[302C→T] + [302C→T]
[A101V]
7
Italian
Deletion (1 bp)
[336del] + [336del]
[I112fsX127]
7
Dutch
Deletion (5 bp)
[367_371del] + [367_371del]
[T123fsX142]b
9
American
Deletion (4 bp)
[372_375del] + [372_375del]
[T125fsX126]
Substitution; in intron 3
[395–8A→G] + [395–8A→G] results in 394_395ins7
[A132fsX146]
Splicing defect; 8
Japanese
Substitution
[482G→A + 488G→C] + [482G→A + 488G→C]
[R161Q + C163S]
AGUFin mutation; 5
Finnish
Substitution and substitution
[482G→A + 488G→C] + [755G→A]
[R161Q + C163S]+[G252E]
Compound heterozygote
Finnish
Substitution and substitution
[482G→A + 488G→C] + [770C→T]
[R161Q + C163S]+[T257I]
Compound heterozygote
Finnish
Substitution
[503G→A] + [?]
[W168X]
Compound heterozygote, mutation in one allele unknown
British
Substitution or deletion (76 bp)
[677G→A] + [677G→A] or [EX6del]
[G226D] or [G208fsX210]
Two different mRNA species in a homozygous patient; 16
Canadian
Substitution
[754G→C] + [754G→C]
[G252R]
Deletion (1 bp)
[788del] + [788del]
[L263X]
13
Mauritian
Insertion (1 bp)
[800_801insT] + [800_801insT]
[P268fsX319]
7
Spanish-American
Pakistani
Italian
Substitution
[904G→A] + [904G→A]
[G302R]
7
Turkish
Substitution
[916T→C] + [916T→C]
[C306R]
7
American
Substitution; in intron 8
[941+1G→T] + [941+1G→T] results in EX8del
[S269fsX274]b
Splicing defect; 6
African-American
Deletion (2078 bp); begins in intron 8, ends in 3′-UTR
[940+1258_1041+1830del] + [940+1258_1041+1830del]
[G314fsX378]
12
American
aNumbering bNot
according to Ikonen et al. (5), GenBank accession no. X55330. expressed in this study.
cells. Figure 5 shows the representative data of immunoprecipitation analysis of wild-type and mutant AGA polypeptides after a 6 h chase. We have reported earlier that the activation of precursor polypeptide occurs in less than 5 min after synthesis in the ER (18). A 6 h chase allows the majority of the polypeptides to be processed if the molecules are folded sufficiently correctly (25). We monitored the polypeptides after chase times of 1, 3 and 6 h. The relative amount of precursor polypeptide remained virtually unchanged for the majority of the mutations when 1 and 6 h chase polypeptides are compared, based on the quantitation using scanning of the radioactive signals on films (data not shown). The only exceptions
were wild-type, V12L, F135S, G226D and T257I, for which the amount of precursor decreased from 1 to 6 h chase due to activation (wild-type, V12L and F135S) or abnormal processing (G226D and T257I). V12L, expressed separately from its concurrent mutation E334X, was the only mutant construct which showed normal processing of the precursor polypeptide, as well as active subunits, α and β, in amounts equal to wild-type AGA. In addition, the enzymatic activity was comparable to that in wild-type indicating that V12L is a polymorphism. F135S expressed minor amounts of normal Pro-α, α- and β-subunits and had some enzymatic activity. Normal sized precursor polypeptide was produced by the
986
Human Molecular Genetics, 2001, Vol. 10, No. 9
Figure 1. Map of the known AGU mutations. The mutations are indicated with labels as in Table 1. Nucleotide changes are presented above the exons. Corresponding changes in the polypeptide are presented below the bar representing the AGA precursor polypeptide. Substitutions are marked with a dot and deletions are marked with a triangle; the lines connecting the label to the exon illustrate the deleted region. Insertions are marked with two lines parting at the label end. The signal sequence (SS), in the N-terminus of the precursor polypeptide, is cleaved co-translationally in the ER from the AGA precursor polypeptide. Two precursors dimerize in the ER prior to activation. In the ER, the AGA precursor is autocatalytically cleaved between residues 205 and 206 revealing the amino group of T206, which is the nucleophile for the catalysis of glycoasparagine substrates. The α- and β-subunits of AGA comprise of amino acids 24–185 and 206–333, respectively. The areas represented by the gray boxes between subunits (residues 186–205) and at the C-terminus of β-subunit (residues 334–346) are cleaved off in lysosomes. The residue numbers of the cleaved peptides are deduced from the crystal structure. The products of normal AGA processing from precursor to active subunits are α and β, the abnormal processing resulting in α- and Pro-β-subunits, shown with amino acid numbers at the bottom. 1,4, concurrent mutations; 2,3, compound heterozygote mutations; 5, two different mRNA species in a patient.
following constructs: A42_A43insDA, G60D, G100E, A101V, F135S, R161Q+C163S, G226D, G252E and T257I. Expression of G252R, P268fsX319, S269fsX274 and E334X produced truncated precursor polypeptides, whereas G314fsX378 produced an elongated AGA-polypeptide. Abnormal processing to α- and Pro-β-subunits was observed for S72P, G226D and T257I constructs. These mutant precursors were not processed into subunits in the ER but were transported to the lysosomes, where the trimming of the C-terminus of the α-subunit occurred producing the abnormal Pro-βsubunit. S72P exhibits an enzymatic activity of about 30% of wild-type due to extracellular autocatalytic activation of its precursor and subsequent endocytosis into cells (14). The majority of mutations produced inactive polypeptides, which do not correspond to the sizes of normal AGA subunits. Table 3 lists the mutated AGA cDNA constructs expressed by us and by others, the corresponding AGA polypeptides produced in
transient expression, and the enzymatic activity of the polypeptides. Out of 27 AGU mutations studied, all but V12L revealed problems in intracellular processing and 25 prevented the activation cleavage in the ER based on the data from 1, 3 and 6 h chase experiments. In the case of these 25 mutations the low activity obviously results from the lack of activation and cleavage of the AGA precursor into the α- and β-subunits. Mutant AGA polypeptides are typically mistargeted The intracellular distribution of AGA polypeptides was studied using the SVpoly mammalian expression vector in BHK cells. BHK cells were chosen for immunofluorescence analysis to reduce overexpression and protein synthesis was stopped for 3 h with cycloheximide to allow the proteins to be transported from the ER. In 3 h most AGA polypeptides reach the lysosomes if they are folded sufficiently correctly and a noticeable
Human Molecular Genetics, 2001, Vol. 10, No. 9
987
Table 2. Summary of mutation characterization and amino acid conservation Mutant
Location of mutationa
Remarks
Conservation
V12Lb
Signal sequence
No effects
NDd
W34fsX45
α-subunit
Frameshift and truncation of polypeptide
A42_A43insDA
AH1, facing solvent
Changes the residue order of helix
Conserved in mammals Conserved in animals and bacteria
G60D
AH2, facing interior
Not enough space for D
C64X
AH2, facing interior
Truncation of polypeptide
E65fsX70
α-subunit
Frameshift and truncation of polypeptide
S72P
Loop structure between AH2/AS2 in the active site, stabilizes nucleophile
Not enough space for P, restricted backbone conformation, prevents activation
S or T in animals and bacteria
G100E
AS3, near the dimer interface
Not enough space for E
Fully conserved
A101V
AS3, near the dimer interface
Not enough space for V, prevents dimerization
Conserved in animals and plants
I112fsX127
α-subunit
Frameshift and truncation of polypeptide
T123fsX142
α-subunit
Frameshift and truncation of polypeptide
T125fsX126
α-subunit
Frameshift and truncation of polypeptide
A132fsX146
α-subunit
Frameshift and truncation of polypeptide
F135S
AH4, on to dimer interface
Structurally possible, located in a hydrophobic environment, decelerates dimerization
Fully conserved
R161Qc
AH5, facing solvent
Structurally possible
Conserved in mammals
C163Sc
Loop structure, facing solvent
Not enough space for S, prevents disulfide bridge formation and dimerization
Conserved in animals
W168X
α-subunit
Truncation of polypeptide
G226D
Loop structure, between BS2/BS3 in the active site
Enough space for D, extraordinary backbone conformation, prevents activation
Fully conserved
G252E
BS4, inside the core
Not enough space for E, extraordinary backbone conformation
Conserved in animals and bacteria
G252R
BS4, inside the core
Not enough space for R, extraordinary backbone conformation
Conserved in animals and bacteria
T257I
BS4 in the active site, oxyanion hole former
Not enough space for I, disrupts oxyanion hole, prevents activation
Fully conserved
L263X
BH1, near to dimer interface
Truncation of polypeptide
P268fsX319
β-subunit, at the dimer interface
Frameshift and truncation of polypeptide
S269fsX274
β-subunit, at the dimer interface
Frameshift and truncation of polypeptide
G302R
BS5, inside the core, near to T257
Not enough space for R
Conserved in animals
C306R
BS5, inside the core
Not enough space for R, prevents disulfide bond formation
Conserved in mammals
G314fsX378
β-subunit
Frameshift and elongation of polypeptide
E334Xb
β-subunit
Truncation of polypeptide
aAccording
to Figure 2. mutations. dNot determined. b,cConcurrent
signal should be visible outside the ER (14,18). After fixation the cells were stained with AGA-specific antibodies and co-stained with ER or lysosome-specific antibodies, and observed with confocal microscopy (Fig. 6). Table 3 lists the results of immunofluorescence analysis. The V12L mutation is located in the signal sequence, but had no effect on transport. S72P polypeptides were visible in the ER, Golgi and lysosomes (14). Similarly, in the case of F135S, G226D and T257I, part of the polypeptides remained in the ER while some lysosomal staining also occurred. The other missense mutations, excluding C306R, severely affected the initial folding since the AGA polypeptides remained in the ER. C306R and every
mutation presenting an early stop-codon had even more severe consequences since no AGA polypeptides could be detected in the cells. Out of 27 AGU mutations studied, 22 demonstrated significant problems in the lysosomal targeting of AGA polypeptides. DISCUSSION We have made an effort here to characterize molecular consequences of AGU mutations at the polypeptide level. Besides analyzing the previously identified mutations, we also monitored the effect of six novel AGU mutations identified
988
Human Molecular Genetics, 2001, Vol. 10, No. 9
Figure 2. Glycoasparaginase and asparaginase amino acid sequence alignment in 10 species. The aligned mammals are: Homo sapiens (GenBank accession no. X55330), Mus musculus (Q64191), Rattus norvegicus (Ref. 36), and Bos taurus (Ref. 36); the other eukaryotic animals are Spodoptera frugiperda (O02467), Drosophila melanogaster (AAF58918) and Caenorhabditis elegans (Q21697). Flavobacterium meningosepticum (Q47898) represents bacterial glycoasparaginases. Arabidopsis thaliana (Z34884) and Lupinus angustifolius (X60691) represent plant asparaginases. The sequence identity against the human AGA is shown in brackets. The residue numbers at the end of the lines include the signal sequences, which are not shown. Locations of α-helices and β-sheets are shown above the sequence, and the functional residue symbols are explained below the alignment. Conserved amino acids are shaded and the locations of mutations are indicated in bold in the human AGA sequence.
here. The 27 AGU mutations are spread rather evenly along the coding region of AGA and generally several mutations are found in each of the nine exons. The only exception is exon 5, which is devoid of mutations. The proteolytic cleavage site, essential for the activation of the precursor polypeptide is coded by exon 5. In addition, ∼50% of exon 5 encodes the
20 amino acid peptide between the α- and β-subunits, which gets proteolytically cleaved off in lysosomes. Among the different AGU alleles missense mutations (65%) are most prevalent and deletions (27%) are common as well. The percentage of substitutions involving glycine is quite high among the AGU alleles with point mutations, 6/15 (40%). It is
Human Molecular Genetics, 2001, Vol. 10, No. 9
989
frequency of glycines emphasizes their significance to backbone flexibility at the sites where enzymatic catalysis and activation take place. In addition, glycines provide much of the required flexibility in the initial folding, which involves making and breaking of disulfide bonds by the ER resident enzyme, protein disulfide isomerase. The mutations are grouped here according to the severity of the effect for the AGA polypeptide, based on intracellular maturation, lysosomal targeting or activity. They are considered to cause mild, moderate or severe effects. The clinical phenotype of AGU patients is reported to be relatively homogenous (24). However, we don’t have the clinical details of the progression or severity of symptoms for our patients, and thus we cannot at this point reliably analyze whether the severity of mutations at the polypeptide level is reflected in the clinical outcome. Some AGU mutations alter the dimer interface disturbing the dimerization of precursors P268fsX319, S269fsX274, G314fsX378 and E334X produced more stable polypeptides than C306R or L263X. This indicates that the amino acids in the α-helix BH1 and the following loop, which contains approximately D261–L267, might be important for polypeptide folding and stability. This hypothesis is supported by the AGA structure; these amino acids, similar to H124, are located between the αβ dimers of AGA and therefore might stabilize the interaction between two αβ dimers in the molecule. With H124 substitutions, the initial folding of precursors is defective, dimerization is prevented and the polypeptides face rapid degradation in the ER due to the important stabilizing role of H124 (19). This is most probably true for amino acids D261–L267 as well. In addition, Riikonen et al. (17) have shown that the large loop structure consisting of the C-terminal part of the α-subunit is crucial for dimerization. Mutations with mild effects Figure 3. The three-dimensional structure of the native human AGA enzyme (1APY, Protein Data Bank, http://www.rcsb.org/pdb/). (Top) One half of the AGA molecule, comprising one α- and one β-subunit, is light blue and the other is dark blue. The locations of missense mutations and the insertion A42_A43insDA are yellow in the polypeptide backbone and the corresponding amino acid labels are adjacent to the residues. (Bottom) The molecule has been rotated along the X-axis. Three regions in both αβ dimers provide an abundance of contacts between the two αβ dimers of AGA: Firstly, the loop containing residue H124 is indicated by an arrow and label I. Secondly, the amino acids in the α-helix BH1 and the following loops are indicated by an arrow and label II, respectively. Thirdly, the large loop structure at the C-terminal end of the α-subunit is indicated by an arrow and label III.
also noteworthy that no replacements of proline exist. Glycine and proline have specific roles in polypeptide structures: glycine enables tight turns of the polypeptide backbone and the small size of its side-chain fits in limited structural environments; on the other hand, proline reduces structural flexibility. The sequence of human AGA contains 31 glycine residues, half of which are close to active site amino acids. It has 17 proline residues, five of which are located very near the lysosomal targeting residues at the C-terminus of the α-subunit and provide rigidity to the polypeptide backbone. The mutation
AGU mutations that produce intracellularly stable precursor polypeptides and at least a fraction of normal subunits are defined here as mild. Intracellularly, these AGA polypeptides are seen in the Golgi compartment and lysosomes in addition to the ER. The polypeptides may also have detectable enzyme activity. The active site S72 stabilizes the α-amino group of the catalytic T206 with a hydrogen bond (20). The loop structure in the active site funnel, which contains S72, is unable to accommodate the rigid proline residue without disturbing packing, and S72P substitution results in abnormal maturation (14). A similar processing defect could be detected for the G226D mutation. G226 is located in the active site in a loop, which contains substrate binding residues R234 and D237. The tight backbone turn at G226 would be absent in the G226D substitution, since according to the Ramachandran plot such backbone torsion angles can only be allowed for glycine. The substitution might cause spatial problems and in any case the long side-chain of aspartic acid would occupy the active site preventing normal processing. Patients with the homozygous G226D mutation produce an additional mRNA species with exon 6 missing resulting in G208fsX210. This mRNA is prevalent and probably caused by a mis-splicing resulting from the alteration of an exonic splicing enhancer sequence (16). The predicted
990
Human Molecular Genetics, 2001, Vol. 10, No. 9
Table 3. Summary of immunoprecipitation, immunofluorescence and activity measurement data Construct
Immunoprecipitation Prec.
Pro-α
Immunofluorescence
α
Pro-β
β
β′
ER
Golgi
Activity in the cells Lysosomes
WT (%)
WT AGA
P
P
P
P
P
P
100
V12La
P
P
P
P
P
P
100
W34fsX45b
BGg
A42_A43insDA
P
P
BGg
G60D
P
P
BGg
C64X
BGg
E65fsX70
BGg
S72Pc
P
G100E
P
P
BGg
A101V
P
P
BGg
P
P
P
P
P
30
I112fsX127
BGg
T123fsX142b
BGg
T125fsX126
BGg
A132fsX146
BGg
F135S
P
R161Q+C163S
P
P
P
P
P
P
P
12 BGg
P
BGg
W168X P
P
G252E
P
P
BGg
G252R
Pe
P
BGg
T257I
P
P
P
P
P
P
P
P
P
BGg
G226D
P
BGg BGg
L263X P268fsX319
Pe
P
BGg
S269fsX274d
Pe
P
BGg
G302R
P
P
BGg
G314fsX378
Pf
P
BGg
E334Xa
Pe
P
BGg
BGg
C306R
COS-1
BGg
P, present. aConcurrent mutations. bPark et al. (9). cPeltola et al. (14). dFisher and Aronson (6). eTruncated precursor. fElongated precursor. gBackground ≤ 7% wild-type.
effect of the G208fsX210 mutation would be severe because practically the whole β-subunit is deleted from the molecule (see later). Naturally occurring active site mutations of AGA are rare. So far, the only mutations of residues directly involved in catalysis are the homozygous S72P and the novel heterozygous mutation T257I. T257 and G258 form the oxyanion hole that stabilizes a covalent enzyme–substrate transition state (20). Isoleucine is unable to replace T257 in the catalytic reaction and an abnormal processing resembling that observed for S72P and G226D could be detected. Polypeptides with active site mutations R234A, R234Q, R234K and T257A (20) and several
others substituting H204, D205 or T206 (19) were also processed abnormally. These mutations disrupt the autocatalytic activation of the AGA precursor in the ER by altering the threedimensional structure of the active site. The consequences of the heterozygous F135S mutation are different from the other mild mutations. The structural environment of F135 is hydrophobic and the substitution of phenylalanine for hydrophilic serine most probably eliminates local stabilizing Van der Waals interactions. F135 makes hydrophobic contacts with T122 and L127, thus stabilizing the loop containing H124. H124 has been shown to be important for dimerization (19). In addition, F135 has direct Van der
Human Molecular Genetics, 2001, Vol. 10, No. 9
991
Figure 4. The spatial environment of mutations in native AGA crystal structure (1APY, Protein Data Bank). The native AGA amino acid residues that are substituted for by AGU mutations are marked with red labels next to the α-carbon atom. The small dots around these residues represent atomic contacts between neighboring residues; their Van der Waals surfaces are partially represented by the dots. The brighter the color of the dots, the smaller is the inter-atomic distance. Should the affected amino acid have side-chain–side-chain contacts with other amino acids, the labels of those residues are marked in white next to the α-carbon. A42 and A43 are located in an α-helix (A). The spatial environment around A43 is restricted; the side-chain of A43 has contacts to the side-chains of M209 and A56, and the main-chain atoms of A43 have contacts to other main-chain atoms. In the same manner, the spatial environment of native residues G60 (B), S72 (C), G100 (D), A101 (E), C163 (G), G252 (I), T257 (J),G302 (K) and C306 (L) is limited. F135 is located in a hydrophobic pocket and its side-chain has many contacts to neighboring side-chains (F), some of which are located in the other half of the native enzyme (indicated with an asterisk). The disulfide bridge between C163 and C179 is disrupted in the C163S patient mutation (G). G226 is located in a hydrophilic environment in the active site funnel of AGA (H); the nucleophile T206 is shown, although G226 and T206 have no direct contacts. The C306R mutation prevents the formation of the disulfide bridge between C306 and C286 (L).
992
Human Molecular Genetics, 2001, Vol. 10, No. 9
Figure 5. Representative data of the immunoprecipitation analysis of the wild-type and mutant AGA polypeptides after a 6 h chase. AGA polypeptides from COS-1 cells transfected with wild-type or mutant AGA-SVpoly were labelled and subsequently immunoprecipitated either in native form (WT, V12L, G226D, T257I) with antibody against the native enzyme, or in the denatured form (other mutations) with antibody against the denatured subunits of AGA. The polypeptides in the latter group were not visible if immunoprecipitated as in the former group. The polypeptides were resolved on 14% SDS–PAGE and observed with autoradiography. On both sides of the gel picture, the locations of the previously characterized AGA polypeptides are indicated. Normal AGA subunits, Pro-α, α, β, and β′, are marked with a dot. The abnormal Pro-β-subunit is pointed out with an arrow. +, mutations expressing active AGA. A42_A43insDA provides an example with two precursor polypeptides of slightly divergent size representing heterogeneous glycosylation of polypeptides. BG denotes the null-transfected COS-1 cell background expression as a control. 1, concurrent mutations; 2, compound heterozygote mutations.
Waals interactions with residues from the other half of the native enzyme. The delayed activation seen in biochemical experiments probably results from impaired dimerization in the ER, caused by the lack of stabilizing interactions between two precursor polypeptides. This results in low levels of the AGA subunits and decreased activity compared with wild-type AGA. The specific activity of F135S AGA was not studied. Mutations with moderate effects AGU mutations are classified here as moderate when they produce inactive precursor polypeptides, which remain unprocessed. These precursor polypeptides are retained in the ER and the majority of polypeptides face rapid degradation. The A42_A32insDA mutation adds two amino acids to the middle of an α-helix. The insertion most probably disturbs local folding in the ER, since the structural environment of A43 is restricted and the mutation changes the residue order of the helix. The consequences of the homozygous G60D mutation are very similar since the side-chain of aspartic acid is excessively bulky for local structure and results in misfolding of the polypeptide. Likewise, the substitution of valine for alanine at 101 disturbs the three-dimensional structure of AGA due to space limitations and probably also because the side-chain of A101 is in hydrophobic contact with the side-chain of L126. L126 resides in a loop containing H124, which provides crucial contacts between the two αβ dimers of AGA (19). The most common AGU mutation worldwide is R161Q+C163S, designated the AGUFin mutation. Ikonen et al. (26) have demonstrated that the causative C163S substitution is responsible for defective folding and premature degradation of the mutant AGA. Detailed molecular characterization shows that replacement of C163 with serine causes structural problems in addition to preventing the formation of the disulfide bridge between C163 and C179. This causes destabilization of the loop structure that provides contacts between the two αβ dimers of AGA. The main reason for the the mutated polypeptides to
remain in the ER as monomeric precursor polypeptides is evidently that the initial folding is defective and dimerization is prevented (17). C163 is also located very near to K177 and Y178, which form one of the three phosphotransferase recognition sites, and by interfering with phosphorylation the C163 substitution could block the lysosomal targeting (21). The unstable precursor molecules produced by the novel mutations G252E and G252R can be explained by the fact that according to the Ramachandran plot the backbone torsion angles at 252 are only allowed for glycine. G302R, a neighbor of the oxyanion hole residue T257, also made the precursor unstable. Mutations with severe effects AGU mutations resulting in remarkably unstable AGA polypeptide are defined here as severe. Trace amounts of mutant polypeptides in this category are seen after a 6 h chase, but they do not represent normal AGA precursor or subunits. AGA polypeptide is undetectable in immunofluorescence analysis of transfected cells even with AGA antibodies against the denatured subunits. These mutants are inactive and the polypeptides are severely shortened due to the premature stopcodons except in the case of C306R. Peltola et al. (10) indicated that the early stop-codon in C64X faces partial readthrough during translation in the patients’ fibroblasts. The patients’ AGA mRNA was unstable, but traces of normal sized inactive precursor polypeptide were detected due to substitution of C64 with an unidentified amino acid. In transient expression the polypeptide was unstable. C64 forms a disulfide bridge with C69 that stabilizes the AGA molecule (17). The lack of the stabilizing covalent bond leads to the formation of a dysfunctional gene product. W34fsX45, I112fsX127, T123fsX142, T125fsX126 and A132fsX146 produce unstable polypeptides, as the novel mutation T125fsX126 demonstrates. Yet another mutation with a similar effect is a novel mutation, W168X, which was found in a British patient. The early stop-codon makes the
Human Molecular Genetics, 2001, Vol. 10, No. 9
993
Figure 6. Distribution of the wild-type (WT) (A–C), G100E (D–F), R161Q+C163S (G–I), G226D (J–L) and G252R (M–O) AGA polypeptides in transiently transfected BHK cells observed with confocal immunofluorescence microscopy. Double immunofluorescence staining of the AGA polypeptides (red, A, D, G, J, M) with lysosomes and late endosomes (green, B and K) or ER (green, E, H, N) using the following antibodies: native AGA antibody (A, C, J, L), antibody against the denatured subunits of AGA (D, F, G, I, M, O), anti-LGP 120 for lysosomes-late endosomes (B, C, K, L), and anti-PDI for ER (E, F, H, I, N, O). Yellow indicates an overlap of the AGA polypeptides (red) and subcellular markers (green) (C, F, I, L, O). An untransfected cell is visible above a transfected cell (D–F). Scale bar, 20 µm.
AGA polypeptides completely unstable. Riikonen et al. (25) have demonstrated that even co-expression of both of the subunits from separate constructs is insufficient for the correct folding of the AGA molecule, and the resulting polypeptides
are rapidly degraded. In these mutations the observation that the transfected cells contained no AGA polypeptides is logical since the mutated gene products lack the whole β-subunit and are therefore severely misfolded and rapidly degraded.
994
Human Molecular Genetics, 2001, Vol. 10, No. 9
Similarly, mutant polypeptides with L263X substitution were unstable although the polypeptides contain the correct active site amino acids. The lack of more than 80 β-subunit residues resulted in severe misfolding of the polypeptides and rapid degradation. Replacement of the disulfide bond forming C306 with arginine results in total misfolding and rapid degradation of the synthesized polypeptides. The substitution disturbs the packing of the molecule. This is the determining factor in the misfolding, since C286S and C306S polypeptides were normally processed and active, demonstrating that the disulfide bridge between C306 and C286 is not crucial for the early folding and stability of AGA (17). Conclusions A rare disorder, AGU, provides a model for the characterization of the molecular details of a human disease. The results demonstrate that structural changes caused by missense mutations are the most common cause of the AGU disease. They cause folding problems and destabilization of AGA polypeptides by introducing bulky residues to limited spaces in threedimensional structure, affect active site residues, create electrostatic disturbances, or prevent disulfide bond formation. The predicted and observed consequences of AGA mutations are in good agreement. These data also indicate that three regions, a helix and a loop at the dimer interface and a large loop structure on the surface, provide crucial stabilizing interactions between two precursors in the ER. This molecular characterization of AGU mutations contributes to the expanding body of knowledge pertaining to human disease pathogenesis. After molecular characterization of the consequences of these mutations, details of AGA dimerization and activation are now better defined and a more comprehensive structure–function understanding of AGA is possible. MATERIALS AND METHODS Patients and sequencing analyses DNA sample of one AGU patient was received from the United Kingdom and samples of two affected siblings were received from Italy. In addition, samples of two AGU families originating from Pakistan, samples of two families from Finland and samples of one family from Italy were obtained. Altogether the DNA of 10 new AGU patients was sequenced. The diagnosis was ascertained by the referring center on the basis of clinical findings, demonstration of urinary glycoasparagines and assay of AGA activity in cultured fibroblasts or peripheral blood leukocytes. In one Finnish family, with the patient carrying a de novo mutation, parenthood was confirmed with genotyping. The genotyping was performed with an AmpFLSTR Profiler PCR amplification kit according to instructions of the manufacturer (Applied Biosystems). PCR amplification of exons was performed with intron-specific primers (11) and sequencing of both strands of the AGA gene was performed with an ABI-Prism 377 DNA sequencer (Perkin Elmer). Sequencher software was used to analyze the data (Gene Codes).
In vitro mutagenesis Mutagenesis was performed with a QuikChange or Chameleon site-directed mutagenesis kit (Stratagene) on an AGA cDNA template (26). The gene was cloned into the BamHI site of the mammalian expression vector SVpoly containing the SV40 early promoter (27). After this, the insert was sequenced to exclude the possibility of unwanted mutations. Transfection, metabolic labelling and immunoprecipitation COS-1 cells (CRL-1650; ATCC, Manassas) were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum and antibiotics. For transfection, the cells were seeded on 3 cm plates at a density of 3 × 105 cells per well. Transfection was performed with the FuGENE 6 transfection reagent (Roche) using 1.5 µg of the wild-type or mutant AGA cDNA construct per well. Following a 48 h incubation the cells were metabolically labelled for 1 h with [35S]Cys (Amersham Pharmacia Biotech) followed by a 1, 3 or 6 h chase (28). Immunoprecipitation was carried out with fixed Staphylococcus aureus cells (Calbiochem) (29) using AGA-specific polyclonal antibody against either the denatured AGA-subunits or the native enzyme (30). The polypeptides were denatured by boiling them 5 min prior to the addition of antibodies against the denatured AGA subunits. The labelled polypeptides were separated by 14% SDS–PAGE under reducing conditions (31) and visualized by autoradiography. Assay for AGA activity The AGA activity measurement was performed subsequent to metabolic labelling. The modified assay (14) is based on colorimetric measurement of liberated N-acetylglucosamine (32). Immunofluorescence analysis and confocal microscopy BHK-21 cells (CCL-10; ATCC, Manassas) were seeded on coverslips on 3 cm plates at a density of 5 × 104 cells per well, grown to 50–60% confluency and transfected like COS-1 cells. Following a 48 h incubation the cells were treated with 50 µg/ml cycloheximide (Sigma) for 3 h to stop protein synthesis. Thereafter, the cells were fixed with 4% paraformaldehyde, pH 7.5 for 30 min. The cells were permeabilized for 15 min with 0.2% saponin (Sigma) in phosphate-buffered saline containing 0.5% bovine serum albumin (Sigma), and thereafter incubated 45 min either with a 1:200 dilution of antibody against the denatured AGA-subunits or with a 1:600 dilution of rabbit polyclonal antibody against the native enzyme. ER was detected with a mouse antibody against protein disulfide isomerase (1:800; StressGen Biotechnologies), and lysosomes-late endosomes with a mouse antibody against LGP 120 (1:200; a gift from Jean Gruenberg, Department of Biochemistry, Geneva, Switzerland). The cells were co-stained with a 1:200 dilution of tetramethylrhodamine isothiocyanate (TRITC)-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories) to localize AGA and with 1:200 dilution of fluorescein isothiocyanate (FITC)-conjugated goat antimouse IgG (Jackson ImmunoResearch Laboratories) to localize ER or lysosomes. Specimens were viewed with a 63× objective on a Leica DMR confocal immunofluorescence microscope (Leica) with TCS NT software.
Human Molecular Genetics, 2001, Vol. 10, No. 9
Computer-assisted analysis For all computer-assisted analysis we used the crystal structure of the native AGA molecule (22) (1APY, Protein Data Bank, http://www.rcsb.org/pdb/). The figures were calculated and created with Grasp (33), XtalView (34) and Probe (35). ACKNOWLEDGEMENTS We thank Paula Hakala, Tuula Manninen and Ritva Timonen for excellent technical assistance. We are grateful to Dr Rita Barone (Catania, Italy), Dr Alan Cooper (Manchester, UK), Professor Paola Rossi and Dr Margherita Santucci (Bologna, Italy), and especially Professor Pertti Aula (Helsinki, Finland) for the DNA and blood samples of the AGU families. This work was supported by grants from The Sigrid Juselius Foundation, The Ulla Hjelt Fund of the Pediatric Research Foundation, Center of Excellence in Disease Genetics of The Academy of Finland Grant 40803, The Academy of Finland Grant 64240 (C.O. and J.R.) and The Rinnekoti Research Foundation, Finland. REFERENCES 1. Ingram, V.M. (1956) A specific chemical difference between the globins of normal human and sickle-cell anemia hemoglobin. Nature, 178, 792–794. 2. Jenner, F.A. and Pollitt, R.J. (1967) Large quantities of 2-acetamido-1(beta-L-aspartamido)-1, 2-dideoxyglucose in the urine of mentally retarded siblings. Biochem. J., 103, 48–49. 3. Autio, S. (1972) Aspartylglucosaminuria: analysis of thirty-four patients. J. Ment. Defic. Res. Monogr. Ser., 1, 1–39. 4. Syvänen, A.-C., Ikonen, E., Manninen, T., Bengtström, M., Söderlund, H., Aula, P. and Peltonen, L. (1992) Convenient and quantitative determination of the frequency of a mutant allele using solid-phase minisequencing: application to aspartylglucosaminuria in Finland. Genomics, 12, 590–595. 5. Ikonen, E., Baumann, M., Grön, K., Syvänen, A.-C., Enomaa, N., Halila, R., Aula, P. and Peltonen, L. (1991) Aspartylglucosaminuria: cDNA encoding human aspartylglucosaminidase and the missense mutation causing the disease. EMBO J., 10, 51–58. 6. Fisher, K.J. and Aronson, N.N., Jr. (1991) Deletion of exon 8 causes glycosylasparaginase deficiency in an African American aspartylglucosaminuria (AGU) patient. FEBS Lett., 288, 173–178. 7. Ikonen, E., Aula, P., Grön, K., Tollersrud, O., Halila, R., Manninen, T., Syvänen, A.-C. and Peltonen, L. (1991) Spectrum of mutations in aspartylglucosaminuria. Proc. Natl Acad. Sci. USA, 88, 11222–11226. 8. Yoshida, K., Yanagisawa, N., Oshima, A., Sakuraba, H., Iida, Y. and Suzuki, Y. (1992) Splicing defect of the glycoasparaginase gene in two Japanese siblings with aspartylglucosaminuria. Hum. Genet., 90, 179–180. 9. Park, H., Vettese, M.B., Fensom, A.H., Fisher, K.J. and Aronson, N.N., Jr. (1993) Characterization of three alleles causing aspartylglucosaminuria: two from a British family and one from an American patient. Biochem. J., 290, 735–741. 10. Peltola, M., Chitayat, D., Peltonen, L. and Jalanko, A. (1994) Characterization of a point mutation in aspartylglucosaminidase gene: evidence for a readthrough of a translational stop codon. Hum. Mol. Genet., 3, 2237–2242. 11. Isoniemi, A., Hietala, M., Aula, P., Jalanko, A. and Peltonen, L. (1995) Identification of a novel mutation causing aspartylglucosaminuria reveals a mutation hotspot region in the aspartylglucosaminidase gene. Hum. Mutat., 5, 318–326. 12. Jalanko, A., Manninen, T. and Peltonen, L. (1995) Deletion of the C-terminal end of aspartylglucosaminidase resulting in a lysosomal accumulation disease: evidence for a unique genomic rearragement. Hum. Mol. Genet., 4, 435–441. 13. Park, H., Rossiter, M., Fensom, A.H., Winchester, B. and Aronson, N.N., Jr. (1996) Single base deletion in exon 7 of the glycoasparaginase gene causes a mild form of aspartylglucosaminuria in a patient of Mauritian origin. J. Inherit. Metab. Dis., 19, 76–83. 14. Peltola, M., Tikkanen, R., Peltonen, L. and Jalanko, A. (1996) Ser72Pro active-site disease mutation in human lysosomal aspartylglucosaminidase: abnormal intracellular processing and evidence for extracellular activation. Hum. Mol. Genet., 5, 737–743.
995
15. Laitinen, A., Hietala, M., Haworth, J.C., Schroeder, M.L., Seargeant, L.E., Greenberg, C.R. and Aula, P. (1997) Two novel mutations in a Canadian family with aspartylglucosaminuria and early outcome post bone marrow transplantation. Clin. Genet., 51, 174–178. 16. Coulter-Mackie, M.B. (1999) A novel exonic mutation in the aspartylglucosaminidase gene causes exon skipping. J. Inherit. Metab. Dis., 22, 682–683. 17. Riikonen, A., Rouvinen, J., Tikkanen, R., Julkunen, I., Peltonen, L. and Jalanko, A. (1996) Primary folding of aspartylglucosaminidase: significance of disulfide bridges and evidence of early multimerization. J. Biol. Chem., 271, 21340–21344. 18. Ikonen, E., Julkunen, I., Tollersrud, O.-K., Kalkkinen, N. and Peltonen, L. (1993) Lysosomal aspartylglucosaminidase is processed to the active subunit complex in the endoplasmic reticulum. EMBO J., 12, 295–302. 19. Saarela, J., Laine, M., Tikkanen, R., Oinonen, C., Jalanko, A., Rouvinen, R. and Peltonen, L. (1998) Activation and oligomerization of aspartylglucosaminidase. J. Biol. Chem., 273, 25320–25328. 20. Tikkanen, R., Riikonen, A., Oinonen, C., Rouvinen, J. and Peltonen, L. (1996) Functional analysis of active site residues of human lysosomal aspartylglucosaminidase: implications for catalytic mechanism and autocatalytic activation. EMBO J., 15, 2954–2960. 21. Tikkanen, R., Peltola, M., Oinonen, C., Rouvinen, J. and Peltonen, L. (1997) Several cooperating binding sites mediate the interaction of a lysosomal enzyme with phosphotransferase. EMBO J., 16, 6684–6693. 22. Oinonen, C., Tikkanen, R., Rouvinen, J. and Peltonen, L. (1995) Threedimensional structure of human lysosomal aspartylglucosaminidase. Nat. Struct. Biol., 2, 1102–1108. 23. Brannigan, J.A., Dodson, G., Duggleby, H.J., Moody, P.C.E., Smith, J.L., Tomchick, D.R. and Murzin, A.G. (1995) A protein catalytic framework with an N-terminal nucleophile is capable of self-activation. Nature, 378, 416–419. 24. Aula, P., Jalanko, A. and Peltonen, L. (2001) Aspartylglucosaminuria. In Scriver, C.R., Beaudet, A.L., Sly, W.S., Valle, D., Childs, B., Vogelstein, B. and Kinzler, K.W. (eds), The Metabolic and Molecular Bases of Inherited Disease. McGraw-Hill, New York, USA, 8th edition, pp. 3535–3550. 25. Riikonen, A., Tikkanen, R., Jalanko, A. and Peltonen, L. (1995) Immediate interaction between the nascent subunits and two conserved amino acids Trp34 and Thr206 are needed for the catalytic activity of aspartylglucosaminidase. J. Biol. Chem., 270, 4903–4907. 26. Ikonen, E., Enomaa, N., Ulmanen, I. and Peltonen, L. (1991) In vitro mutagenesis helps to unravel the biological consequences of aspartylglucosaminuria mutation. Genomics, 11, 206–211. 27. Stacey, A. and Schnieke, A. (1990) SVpoly: a versatile mammalian expression vector. Nucleic Acids Res., 18, 2829. 28. Tikkanen, R., Enomaa., N., Riikonen, A., Ikonen, E. and Peltonen, L. (1995) Intracellular sorting of aspartylglucosaminidase: the role N-linked oligosaccharides and evidence of a Man-6-P-independent lysosomal targeting. DNA Cell. Biol., 14, 305–312. 29. Proia, R.L., d’Azzo, A. and Neufeld, E.F. (1984) Association of α- and βsubunits during the biosynthesis of β-hexosaminidase in cultured human fibroblasts. J. Biol. Chem., 259, 3350–3354. 30. Halila, R., Baumann, M., Ikonen, E., Enomaa, E. and Peltonen, L. (1991) Human leucocyte aspartylglucosaminidase: evidence for two different subunits in a more complex native structure. Biochem. J., 276, 251–256. 31. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680–685. 32. Reissig, J.L., Strominger, J.L. and Leloir, K.M. (1955) A modified colorimetric method for the estimation of N-acetylamino sugars. J. Biol. Chem., 217, 959–966. 33. Nicholls, A., Sharp, K.A. and Honig, B. (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins, 11, 281–296. 34. McRee, D.E. (1999) XtalView/Xfit-A versatile program for manipulating atomic coordinates and electron density. J. Struct. Biol., 125, 156–165. 35. Word, J.M., Lovell, S.C., LaBean T.H., Taylor, H.C., Zalis, M.E., Presley, B.K., Richardson, J.S. and Richardson, J.S. (1999) Visualizing and quantifying molecular goodness-of-fit: small-probe contact dots with explicit hydrogen atoms. J. Mol. Biol., 285, 1711–1733. 36. Liu, Y., Dunn, G.S. and Aronson, N.A., Jr. (1996) Purification, biochemistry and molecular cloning of an insect glycosylasparaginase from Spodoptera frugiperda. Glycobiology, 6, 527–536. 37. den Dunnen, J.T. and Antonarakis, S.E. (2000) Mutation nomenclature extensions and suggestions to describe complex mutations: a discussion. Hum. Mutat., 15, 7–12.
996
Human Molecular Genetics, 2001, Vol. 10, No. 9
