Proteolytic processing of the alpha-chain of the lysosomal enzyme, beta-hexosaminidase, in normal human fibroblasts

The two subunits of beta-hexosaminidase undergo many post-translational modifications characteristic of lysosomal proteins, including limited proteolysis. To identify proteolytic cleavage sites in the alpha-chain, we have biosynthetically radiolabeled the transient forms, isolated these by immunoprecipitation, gel electrophoresis, and electroelution, and subjected them to automated Edman degradation. The position of the NH2-terminal amino acid was inferred from the elution cycle of the radioactive amino acid and the primary sequence encoded in the alpha-chain cDNA. The amino terminus of the precursor obtained by in vitro translation of SP6 alpha-chain mRNA in the presence of microsomes was leucine 23. The same amino terminus was found in precursor alpha-chain synthesized by normal human fibroblasts (IMR90) in a 1- or 3-h pulse or secreted by these cells in the presence of NH4Cl. The alpha-chain isolated after a 3-h pulse followed by a 5-h chase (intermediate form) included a mixture of molecular species of which the amino terminus was arginine 87 (most abundant), histidine 88, or leucine 90. After a 20-h chase (mature form) the latter species predominated. This mature form of the alpha-chain remained fully reactive with antibody raised against the carboxyl-terminal 15 amino acids, indicating little if any proteolysis at the carboxyl terminus. Thus synthesis and maturation of the alpha-chain of beta-hexosaminidase includes two major proteolytic cleavages: the first, between alanine 22 and leucine 23, removes the signal peptide to generate the precursor form, whereas the second occurs between the dibasic amino acids, lysine 86 and arginine 87. The second cleavage is followed by trimming of 3 additional amino acids to give the mature form of the alpha-chain.     

Proteolytic processing of human lysosomal arylsulfatase A.

Arylsulfatase A purified from human placenta contained an unreported component with an apparent molecular mass of 7 kDa in addition to the two known components with apparent molecular masses of 58 and 50 kDa. The detailed relationship between the 58 kDa component and the 50 kDa component is as yet unknown. The present study was undertaken to define the structure of the subunits of the sulfatase. The N-terminal sequence of the 50 kDa component was identical to that of the 58 kDa component. Furthermore, the peptide maps of the 50 kDa component, which was separately digested with trypsin and Achromobacter proteinase I, were quite similar to those of the 58 kDa one. Through sequence analysis of the incompatible peaks in the peptide maps, the 50 kDa component was found to lack a sequence from Val-445 to the C-terminus. On the other hand, the N-terminal sequence of the 7 kDa component began with Ala-448, though there was a minor sequence commencing with Thr-449. These observations suggest that the 50 and 7 kDa components were produced by limited proteolysis near the C-terminus of the 58 kDa component. Through analysis using unreducing SDS-PAGE, the 58 and the 7 kDa components were found to be linked by disulphide bonds. Arylsulfatase A purified from human liver was also composed of the same subunits as the placental one. This finding suggests that human arylsulfatase A undergoes similar proteolytic processing regardless of the tissue involved.     

Synthesis of a human lysosomal enzyme, beta-hexosaminidase B, using the baculovirus expression system

Human lysosomal beta-hexosaminidase exists in two major forms: the A isoform is composed of both alpha and beta chains, while the B form is a homopolymer of beta chains. Deficiency of beta-hexosaminidase underlies the GM2 gangliosidoses. We have produced active beta-hexosaminidase B in cultured insect (Sf9) cells by isolation of a recombinant insect virus (baculovirus) containing the cDNA for the beta chain within the viral polyhedron gene and infection of Sf9 cells with this construct. That portion of the enzyme secreted into the medium, 50%, was purified with concanavalin A Sepharose and subsequent affinity chromatography to yield beta-hexosaminidase B that is 75% pure. The product has an N-terminal amino acid sequence, specific activity, and size (M(r) 62,000) similar to that of the enzyme present in cultured human fibroblasts. However, endo H sensitivity studies revealed that the oligosaccharide structures present on recombinant beta-hexosaminidase B differ from those found on the enzyme synthesized in the human system. In addition, these structures lack the mannose 6-phosphate recognition marker that targets degradative hydrolases to lysosomes. Despite these differences, recombinant beta-hexosaminidase B does serve as a specific substrate for the mannose phosphorylating enzyme, N-acetylglucosaminyl phosphotransferase. Furthermore, the oligosaccharide moieties phosphorylated in vitro match those phosphorylated in vivo, pointing to the conformational integrity of the recombinant enzyme. Generous amounts of easily obtained, easily purified, and properly folded beta-hexosaminidase B will facilitate physical structural analysis of the enzyme.     

Characterization of human placental beta-hexosaminidase I2. Proteolytic processing intermediates of hexosaminidase A

The lysosomal hydrolase beta-hexosaminidase (beta-N-acetylhexosaminidase, EC 3.2.1.52) exists as two major isozymes in normal human tissue: an acidic A-form and a basic B-form. There are also minor forms of intermediate pI known as I-forms. Increases in one or more of these intermediates have been associated with various disease states. Although the two major isozymes have been extensively studied, the structure and biosynthetic origins of the I-forms are unknown. Characterization of a placental hexosaminidase I-form, presented in this report, demonstrates that it is composed of two forms of partially processed hexosaminidase A. The major form contains an intact pro-alpha chain and a pro-beta chain lacking 2 residues from its amino terminus (Ala and Arg). The minor form also contains an alpha and a beta subunit, but each has undergone further proteolytic processing. The amino terminus of each of these partially processed polypeptide chains matches one of those previously found on stable processing intermediates in a single normal human fibroblast cell line. These data confirm that similar processing intermediates exist in human placenta, suggesting that this I-form lacks a unique enzymatic function in vivo. A sequence of normal proteolytic processing events is postulated.     

Translation initiation in the HEXB gene encoding the beta-subunit of human beta-hexosaminidase

The human lysosomal enzyme beta-hexosaminidase (EC 3.2.1.52) is a glycoprotein composed of dimers of alpha- and/or beta-subunits. The subunits of the enzymes are synthesized in the rough endoplasmic reticulum and transported through the Golgi apparatus to the lysosome. As such, each subunit contains an amino-terminal signal peptide that directs the nascent polypeptide into the lumen of the endoplasmic reticulum. The signal peptide cleavage site of the beta-polypeptide is known, but its NH2 terminus has not been determined due to the presence of three candidate initiation codons upstream of the cleavage site. In this study, we identified the mRNA cap site, confirming the presence of all three AUGs in the majority of HEXB mRNA. To identify the site of translation initiation, we mutated the three ATGs by deletion and site-directed mutagenesis and showed that all three AUG codons can be used for translation initiation after expression in COS cells. Furthermore, in each case, a fully processed, i.e. mature lysosomal, and enzymatically active beta-hexosaminidase was produced indicating that a functional signal peptide was synthesized. However, expression of a frameshift mutation in the normal construct, created by insertion of a single nucleotide between the first and second ATG, resulted in no significant enzyme activity or beta-subunit protein. We conclude, therefore, that the first in-frame ATG is used exclusively in vivo, in keeping with the scanning model of eukaryotic translation initiation. Interestingly, substitution of all three ATGs with CTG resulted in a significant amount of mature beta-hexosaminidase, showing that under these conditions, initiation could occur from non-AUG codons. Translation initiation from the first AUG gives the prepro-beta-polypeptide a signal peptide of 42 amino acids that has an unusually long hydrophobic core more typical of membrane spanning domains. Such a large hydrophobic core has not been found in other cleavable signal peptides.     

Characterization of the human HEXB gene encoding lysosomal beta-hexosaminidase

The lysosomal enzyme beta-hexosaminidase A contains alpha- and beta-subunits that are encoded by the HEXA and HEXB genes, respectively. The human HEXB gene has been isolated and characterized. It is 45 kb long and is split into 14 exons. Of the 13 introns, 12 interrupt the coding sequences at homologous positions in the HEXA and HEXB genes. The 5' flanking region contains the functional HEXB gene promoter. While a fine-structure analysis has yet to be done, we note that the sequence is GC rich and has several GC boxes and one CAAT box. There are also sequences related or identical to a progesterone response element and an AP-1 binding motif.     

Synthesis and processing of lysosomal alpha-fucosidase in cultured human fibroblasts

The lysosomal enzyme alpha-L-fucosidase from human skin fibroblasts is synthesized as a 53 kDa glycosylated precursor which is then proteolytically processed to a 50 kDa mature form. This was confirmed by pulse-chase labeling studies with chase times up to 72 h. In fibroblasts treated with 1-deoxymannojirimycin to prevent trimming of high mannose oligosaccharides, endoglycosidase H (endo H) treatment completely deglycosylated and reduced the size of immunoprecipitated alpha-fucosidase by 4-5 kDa, suggesting the presence of two oligosaccharide units. Endoglycosidase H and endo F studies on untreated alpha-fucosidase suggested the presence of one complex-type and one high mannose-type unit, and that the final processing from 53 to 50 kDa did not involve the removal of carbohydrate. Processing was inhibited by the thiol proteinase inhibitor Ep-459, but not by Ep-475 or leupeptin. Since Ep-459 treatment increased both alpha-fucosidase activity (3-fold) and the amount of immunoprecipitable alpha-fucosidase protein in normal human skin fibroblasts, this suggests a role for cysteine-like proteinases either directly or indirectly in lysosomal hydrolase processing and turnover. Subcellular fractionation studies revealed that the proteolytic processing of the 53 kDa precursor to the 50 kDa mature form occurred in the lysosome, or some other dense organelle.     

Presence of a lysosomal enzyme, arylsulfatase-A, in the prelysosome-endosome compartments of human cultured fibroblasts

Although endosomes and lysosomes are associated with different subcellular functions, we present evidence that a lysosomal enzyme, arylsulfatase-A, is present in prelysosomal vesicles which constitute part of the endosomal compartment. When human cultured fibroblasts were subfractionated with Percoll gradients, arylsulfatase-A activity was enriched in three subcellular fractions: dense lysosomes, light lysosomes, and light membranous vesicles. Pulsing the cells for 1 to 10 min with the fluid-phase endocytic marker, horseradish peroxidase, showed that endosomes enriched with the marker were distributed partly in the light lysosome fraction but mainly in the light membranous fraction. By pulsing the fibroblasts for 10 min with horseradish peroxidase conjugated to colloidal gold and then staining the light membranous and light lysosomal fractions for arylsulfatase-A activity with a specific cytochemical technique, the endocytic marker was detected under the electron microscope in the same vesicles as the lysosomal enzyme. The origin of the lysosomal enzyme in this endosomal compartment was shown not to be acquired through mannose 6-phosphate receptor-mediated endocytosis of enzymes previously secreted from the cell. Together with our recent finding that the light membranous fraction contains prelysosomes distinct from bona fide lysosomes and was highly enriched with newly synthesized arylsulfatase-A molecules, these results demonstrate that prelysosomes also constitute part of the endosomal compartment to which intracellular lysosomal enzymes are targeted.     

Analysis of the glycosylation and phosphorylation of the lysosomal enzyme, beta-hexosaminidase B, by site-directed mutagenesis

Lysosomal enzymes require a mannose 6-phosphate recognition marker, constructed on asparagine-linked oligosaccharide chains, for targeting to lysosomes. We have identified the glycosylation sites of human beta-hexosaminidase B and have determined the influence of individual oligosaccharides on the phosphorylation, lysosomal targeting, and catalytic activity of the enzyme. The five potential glycosylation sites of the hexosaminidase beta-chain were modified individually by site-directed mutagenesis, and the constructs were expressed in COS 1 cells. By this analysis, we determined that four of the five potential sites were glycosylated. Two of the four oligosaccharides were preferentially phosphorylated. The absence of these two preferentially phosphorylated oligosaccharides resulted in greatly reduced amounts of the lysosomal form of the enzyme with increased secretion into the medium. The catalytic activity of beta-hexosaminidase B was not significantly altered by the absence of individual oligosaccharides suggesting the folding and assembly of the enzyme was not disrupted.     

Synthesis and assembly of a catalytically active lysosomal enzyme, beta-hexosaminidase B, in a cell-free system

The synthesis and dimerization of beta-chains during the formation of catalytically active beta-hexosaminidase B were studied in a cell-free system. beta-chain mRNA, transcribed from the cloned cDNA with SP6 polymerase, was translated in a rabbit reticulocyte protein-synthesizing system in the presence of dog pancreas microsomal membranes and oxidized glutathione. Under these conditions, the primary beta-chain translation product was translocated into the microsomal vesicles and modified by the addition of N-linked oligosaccharide chains. After transfer into the microsomal vesicles, the beta-polypeptide assumed a conformation resembling the native state as determined by antibody reactivity. Like the authentic precursor enzyme, the microsomally located chains were assembled into dimers and were catalytically active. In intact human fibroblasts, dimerization of beta-chains occurred within 15 min after their synthesis, consistent with a site of assembly in the rough endoplasmic reticulum. The cell-free expression system was also useful in establishing the functionality of beta-chain initiator methionine codons. By expression of beta-chain mRNAs with altered methionine codons, we demonstrated that polypeptides initiating from any of the first three methionine codons in the beta-chain sequence contain a functional signal sequence and form catalytically active enzymes.     

Two mutations produce intron insertion in mRNA and elongated beta-subunit of human beta-hexosaminidase

An elongated beta-subunit of the lysosomal enzyme beta-hexosaminidase was found in fibroblast strains derived from two patients with juvenile Sandhoff disease and two asymptomatic individuals sharing an unusual isoenzyme pattern: a low level of residual A (alpha beta) isoenzyme activity (3-6% of normal for the juvenile Sandhoff and 9-10% for the asymptomatic strains) without B (beta beta) isoenzyme activity. The elongated beta-subunit was abnormal in other ways: It reacted with antiserum against the unfolded polypeptide, it was not phosphorylated on mannose residues, it was not processed to the mature form, and it was degraded rapidly. The increased length of the beta-subunit was caused by two different mutations. Cells from two juvenile Sandhoff and one asymptomatic individuals had the previously described G----A transition in intron 12 that creates a splice site, causing an in-frame insertion of 24 intronic nucleotides into mRNA (Nakano, T., and Suzuki, K. (1989) J. Biol. Chem. 264, 5155-5158). The second mutation was found in cells from the asymptomatic girl whose A+B- isoenzyme pattern had been designated "Hexosaminidase Paris" (Dreyfus, J. C., Poenaru, L., Vibert, M., Ravise, N., and Boue, J. (1977) Am. J. Hum. Genet. 29, 287-293); duplication of a region straddling the junction of intron 13 and exon 14 generates an alternate splice site, causing an in-frame insertion of 18 nucleotides into mRNA. Although the two new splice sites are used preferentially, the normal sites may be used to some extent, accounting for the residual A isoenzyme activity.     

Dynamics of lysosomal cholesterol in Niemann-Pick type C and normal human fibroblasts

The dynamics of endolysosomal cholesterol were investigated in Niemann-Pick type C (NPC) cells and in human fibroblasts treated with class 2 amphiphiles to mimic NPC cells. We showed through new approaches that the massive pools of endolysosomal cholesterol in these cells are not trapped but, rather, circulate to the cell surface at about the normal rate. This flux spared NPC and amphiphile-treated cells from disruption by the extraction of their plasma membrane cholesterol with cyclodextrin. Nocodazole, a microtubule-depolymerizing agent, reversed the resistance of NPC and U18666A-treated cells to cholesterol depletion, apparently by reducing the flux of endolysosomal cholesterol to the plasma membrane. Neither nocodazole nor bafilomycin A1 (an inhibitor of the vacuolar proton pump) acted in the same way as the NPC mutation or class 2 amphiphiles: both agents decreased plasma membrane cholesterol at the expense of the endolysosomal pool and both blocked the actions of the amphiphile, U18666A. Finally, the resistance of NPC cells to lysis by amphotericin B was shown not to reflect a reduction in plasma membrane cholesterol arising from a block in lysosomal cholesterol export but rather the diversion of the amphotericin B to cholesterol-rich endolysosomes. We conclude that the large pool of endolysosomal cholesterol in NPC and amphiphile-treated fibroblasts is dynamic and that its turnover, as in normal cells, is dependent on microtubules.     

Transport and processing of beta-hexosaminidase in normal and mucolipidosis-II cultured fibroblasts. Effect of monensin and nigericin.

The univalent-cation ionophores monensin (4.0 microM) and nigericin (0.5 microM) inhibited the abnormal excretion of beta-hexosaminidase from mucolipidosis-II cultured fibroblasts by 62 and 76% respectively, with a corresponding intracellular accumulation of the enzyme. As shown by lectin binding, the enzyme which accumulated in monensin-treated cells did not contain galactose residues, whereas the corresponding enzyme from nigericin-treated cells was galactosylated. The results suggest that monensin acts at an early point in the process of hydrolase glycosylation, and nigericin acts later, both presumably within the Golgi region, allowing the accumulation of different glycosylated forms of the enzyme. The intra- and extra-cellular distribution of beta-hexosaminidase in ionophore-treated normal cells was essentially unchanged, whereas concanavalin A precipitability of excreted enzyme was increased and its ability to be taken up by deficient fibroblasts was decreased. The bivalent-cation ionophore A23187 (1 microM) reduced beta-hexosaminidase excretion from mucolipidosis-II cells by 82% and by 96% when used with EGTA (1 mM). However, there was no intracellular accumulation of enzyme, suggesting that the effect of this ionophore was restricted to the inhibition of synthesis. It therefore appears that the actual transport of beta-hexosaminidase in mucolipidosis-II cells is affected by univalent-cation ionophores in a selective manner. These findings suggest that individual ionophores could be used to identify the sites of hydrolase oligosaccharide processing in the Golgi region by causing intermediate glycosylated forms of the transported hydrolase to accumulate in a specific Golgi compartment preceding the blocking site of the ionophore.     

Correction of I-cell defect by hybridization with lysosomal enzyme deficient human fibroblasts

I-cell fibroblasts with a multiple intracellular lysosomal enzyme deficiency were hybridized with cells from patients with different types of single lysosomal enzyme defects. Fusion with G_M2 gangliosidosis, type 2, (Sandhoff disease) fibroblasts resulted in a restoration of the hexosaminidase activity, in a normalization of the electrophoretic mobility of the isoenzymes, and in a decreased activity in the medium. Fusion of I-cells with fibroblasts from G_M1 gangliosidosis, type 1, led to enhancement of β-galactosidase (β-gal) activity. This complementation must be the result of the presence of normal polypeptide chains in I-cells, whereas the other cell types provide a factor that causes the intracellular retention of the enzymes. Restoration of β-gal was also observed in heterokaryons after fusion of I-cells with β-galactosidase/neuraminidase-deficient (β-gal⁻/neur⁻) variants, indicating that the neuraminidase(s) and the posttranslational modification of β-gal are affected in a different way in I-cell disease and in β-gal⁻/neur⁻ variants. Fusion of I-cells with mannosidosis fibroblasts resulted in a restoration of the acidic form of α-mannosidase and in a decrease of the extracellular activity of both this enzyme and the hexosaminidase enzyme, indicating that fusion of I-cells with different types of fibroblasts with a single lysosomal enzyme deficiency not only leads to complementation for one particular enzyme but also to a correction of the basic defect in I-cells.     

Analysis of the glycosylation and phosphorylation of the alpha-subunit of the lysosomal enzyme, beta-hexosaminidase A, by site-directed mutagenesis.

The glycosylation and subsequent phosphorylation of mannose residues is a pivotal modification during the biosynthesis of lysosomal enzymes. We have identified the sites of N-linked glycosylation and oligosaccharide phosphorylation on the alpha-subunit of beta-hexosaminidase and have determined the influence of the oligosaccharides on the folding and transport of the enzyme. The potential glycosylation sequences, either singly or in combination, were eliminated through site-directed mutagenesis of the cDNA. By expression of the mutant cDNAs in COS-1 cells, each of the three glycosylation sites on the alpha-subunit was found to be modified by an oligosaccharide. One of the three oligosaccharides was the preferred site of phosphorylation. The absence of any individual oligosaccharide did not diminish the expression of the catalytic activity associated with the alpha-chain, implying proper folding and assembly of subunits. A profound effect was observed, however, when all three oligosaccharides were absent. The unglycosylated alpha-subunit, resulting from genetic alteration of all three glycosylation sites or synthesis of the wild-type protein in the presence of tunicamycin, was catalytically inactive. It was found to be improperly folded into an insoluble aggregate, linked through inappropriate disulfide bonds. The unglycosylated protein was trapped in the lumen of the endoplasmic reticulum and was found in a complex with the Ig heavy chain-binding protein, BiP. The properties of the nonglycosylated, misfolded alpha-subunit were similar to some mutant alpha-subunits in Tay-Sachs disease patients. The results indicate that the oligosaccharides are essential, although not in a site-specific manner, for proper folding and cellular transport of the alpha-subunit.     

Synthesis of lysosomal alpha-mannosidase in normal and mannosidosis fibroblasts

The biosynthesis and secretion of lysosomal alpha-mannosidase was studied in metabolically labelled fibroblasts from controls and two patients with mannosidosis. Normal fibroblasts secrete alpha-mannosidase as a 110kDa polypeptide. Intracellularly alpha-mannosidase is represented by several polypeptides with apparent Mrs ranging from 40 to 67kDa. In two mannosidosis cell lines none of intra- and extracellular polypeptides of alpha-mannosidase were detectable. The mannosidosis fibroblasts secreted acid alpha-mannosidase activity at one third of the normal rate. In contrast to normal cells the secretion was not enhanced by NH4C1 and the secreted activity was not immunoprecipitable, indicating that the acid alpha-mannosidase activity secreted by mannosidosis fibroblasts is not related to the lysosomal alpha-mannosidase.     

Evidence for lysosomal enzyme recognition by human fibroblasts via a phosphorylated carbohydrate moiety

Adsorptive endocytosis of five different lysosomal enzymes from various human and non-human sources was susceptible to inhibition by mannose and l-fucose, methyl α-d-mannoside, α-anomeric p-nitrophenyl glycosides of mannose and l-fucose, mannose 6-phosphate and fructose 1-phosphate. A few exceptions from this general scheme were observed for particular enzymes, particularly for β-glucuronidase from human urine. The inhibition of α-N-acetylglucosaminidase endocytosis by mannose, p-nitrophenyl α-d-mannoside and mannose 6-phosphate was shown to be competitive. The loss of endocytosis after alkaline phosphatase treatment of lysosomal enzymes supports the hypothesis that the phosphorylated sugars compete with a phosphorylated carbohydrate on the enzymes for binding to the cell-surface receptors [Kaplan, Achord & Sly (1977) Proc. Natl. Acad. Sci. U.S.A. 74, 2026–2030]. Endocytosis of `low-uptake' forms of α-N-acetylglucosaminidase and α-mannosidase was likewise susceptible to inhibition by sugar phosphates and by alkaline phosphatase treatment, suggesting that `low-uptake' forms are either contaminated with `high-uptake' forms or are internalized via the same route as `high-uptake' forms. The existence of an alternative route for adsorptive endocytosis of lysosomal enzymes is indicated by the unaffected adsorptive endocytosis of rat liver β-glucuronidase in the presence of phosphorylated sugars and after treatment with alkaline phosphatase.     

Proteolytic activation and glycosylation of N-acylethanolamine-hydrolyzing acid amidase, a lysosomal enzyme involved in the endocannabinoid metabolism

N-acylethanolamine-hydrolyzing acid amidase (NAAA) is a lysosomal enzyme hydrolyzing bioactive N-acylethanolamines, including anandamide and N-palmitoylethanolamine. Previously, we suggested that NAAA is glycosylated and proteolytically cleaved. Here, we investigated the mechanism and significance of the cleavage of human NAAA overexpressed in human embryonic kidney 293 cells. Western blotting with anti-NAAA antibody revealed that most of NAAA in the cell homogenate was the cleaved 30-kDa form. However, some of NAAA were released outside the cells and the extracellular enzyme was mostly the uncleaved 48-kDa form. When incubated at pH 4.5, the 48-kDa form was time-dependently converted to the 30-kDa form with concomitant increase in the N-palmitoylethanolamine-hydrolyzing activity. The purified 48-kDa form was also cleaved and activated. However, the cleavage did not proceed at pH 7.4 or in the presence of p-chloromercuribenzoic acid. The mutant C126S was resistant to the cleavage and remained inactive. These results suggested that this specific proteolysis is a self-catalyzed activation step. We next determined N-glycosylation sites of human NAAA by site-directed mutagenesis addressed to asparagine residues in six potential N-glycosylation sites. The results exhibited that Asn-37, Asn-107, Asn-309, and Asn-333 are actual N-glycosylation sites. The glycosylation appeared to play an important role in stabilizing the enzyme protein.