Studies on the formation of N6-hydroxylysine in cell-free extracts of Aerobacter aerogenes 62-1.

We have investigated conditions optimal for the conversion of L-lysine to its N6-hydroxy derivative by partially purified cell-free extracts of Aerobacter aerogenes 62-1. The enzyme system was highly specific to L-lysine: the D-isomer and, the N2- or N6-derivatives of lysine, and alpha-amino acids were not hydroxylated. Most of the latter compounds had little effect onthe hydroxylation of L-lysine. However, -l-glutamic acid and L-glutamine enhanced the hydroxylation, with half-maximal activation achieved at 100 micrometers concentration of the effector. The Km values for pyruvate and L-(+)-lactate (compounds known to stimulate N-hydroxylysine formation) were found to be approx. 100 micrometers. The data show that N-hydroxylation of the amino acid precedes acylation in the biosynthesis of hydroxamic acid in A. aerogenes 62-1.

Comparison of the properties of a soluble form of glucocerebrosidase from human urine with those of the membrane-associated tissue enzyme.

Human urine contains a soluble form of glucocerebrosidase, an enzyme associated with the lysosomal membrane in cells and tissues. Urinary glucocerebrosidase is identical to the enzyme extracted from tissues with respect to the following parameters: Km for natural and artificial substrates, inhibition by conduritol B-epoxide, and stimulation by taurocholate. The enzyme is greater than 90% precipitable by polyclonal anti-(placental glucocerebrosidase) antiserum. Upon isoelectric focussing of urinary glucocerebrosidase multiple peaks of activity were observed. Partial deglycosylation (removal of sialic acid, N-acetylglucosamine and galactose) of the urinary enzyme increased the isoelectric point to a value identical to that of the main form found after partial deglycosylation of the placental enzyme. Upon polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate followed by immunoblotting, the immunopurified urinary enzyme shows the same molecular mass forms as the enzyme immunopurified from brain and kidney. In placenta the apparent molecular mass is somewhat higher but upon removal of sialic acid, N-acetylglucosamine and galactose the urinary and the placental enzyme show identical molecular masses of 57 kDa. We conclude that the enzymes extracted from urine and tissue are identical and that differences in apparent molecular mass and isoelectric point are probably due to heterogeneity in the oligosaccharide moieties of the molecules.

Glucocerebrosidase, a lysosomal enzyme that does not undergo oligosaccharide phosphorylation.

Labelling of cultured human skin fibroblasts from either control subjects or patients with mucolipidosis II (I-cell disease) with [32P]phosphate resulted in tight association of phosphate with immunoprecipitated glucocerebrosidase, a membrane-associated lysosomal enzyme. Endoglycosidase F digestion of the immunoprecipitated glucocerebrosidase did not release labelled phosphate, suggesting that the phosphate was not associated with the oligosaccharide moiety of this glycoprotein. Purification of the enzyme from cells labelled with [32P]phosphate and [35S]methionine by an immunoaffinity chromatography procedure, which included a washing step with detergent, resulted in complete separation of the phosphate label from the peak of glucocerebrosidase activity and methionine labelling. We conclude that oligosaccharide phosphorylation, which is essential for transport of soluble lysosomal enzymes to the lysosomes in fibroblasts, does not occur in glucocerebrosidase.

Retroviral coexpression of a multidrug resistance gene (MDR1) and human alpha-galactosidase A for gene therapy of Fabry disease.

Human alpha-galactosidase A (alpha-Gal A; EC.3.2.1.22) is a lysosomal exoglycosidase encoded by a gene on Xq22. Deficiencies of this enzyme result in Fabry disease, an X-chromosome-linked recessive disorder that leads to premature death in affected males. For treatment of genetic diseases, we have developed a retroviral vector system, pSXLC/pHa, that enables coexpression of drug-selectable markers with a second nonselectable gene as part of a bicistronic message using the promoter from the Harvey murine sarcoma virus and an internal ribosomal entry site (IRES) from encephalomyocarditis virus. Retroviral vectors based on this system that carry the human alpha-Gal A cDNA either upstream (pHa-alpha Gal-IRES-MDR) or downstream (pHa-MDR-IRES-alpha Gal) from the IRES relative to the drug-selectable MDR1 (P-glycoprotein) cDNA were constructed. Each of eight independent vincristine-resistant, pHa-alpha Gal-IRES-MDR-transfected clones and all four vincristine-resistant, pHa-alpha Gal-IRES-MDR retrovirus-transduced clones showed significantly higher activity of alpha-Gal A than the parental cells. More than 50% of the vincristine-resistant, pHa-MDR-IRES-alpha Gal-transfected clones and all four vincristine-resistant, pHa-MDR-IRES-alpha Gal retrovirus-transduced clones showed significantly higher activity of alpha-Gal A than the parental cells. In these bicistronic vectors, the cDNA whose translation was cap-dependent (upstream) was expressed at higher levels than when the same cDNA was translated in an IRES-dependent manner (downstream). These vectors may prove useful in the gene therapy of Fabry disease.

Structural organization and expression of the mouse gene encoding alpha-galactosidase A.

alpha-Galactosidase A (alpha-D-galactoside galactohydrolase, EC 3.2.1.22; alpha GalA) is a lysosomal enzyme that hydrolyses the alpha-D-galactosyl residues from glycosphingolipids. Fabry disease, an inhibited X-linked recessive human metabolic disorder, results from a mutation in the alpha GalA gene at Xq22. As a prerequisite for generating a mouse model of Fabry disease by gene targeting, we have isolated and characterized the mouse alpha GalA gene and cDNA. A cloned mouse alpha GalA cDNA encoded a putative precursor protein of 419 amino acids (aa), including a 31-aa signal peptide (SP). The deduced aa sequence showed high homology (79%) with the human alpha GalA protein. Nucleotide sequence analysis of genomic clones revealed that the overall structure and organization of the gene was very similar to that of human alpha GalA. All exon-intron splice junctions conformed to the GT/AG consensus sequence. Comparison of genomic and cDNA sequences revealed the occurrence of two putative polyadenylation signals whose alternative use results in the two mouse alpha GalA transcripts of 1.4 and 3.6 kb. The 5'-flanking region of mouse alpha GalA had no typical TATA box. Several putative promoter-associated elements including Sp1, AP1 and a potential cAMP-responsive element (CRE) were identified. Northern blot analysis revealed the widespread tissue distribution of mouse alpha GalA transcripts. Lower expression levels, however, were observed in some tissues, implying tissue-specific differences in alpha GalA promoter function.

Gaucher's disease: lack of antibody response in 12 patients following repeated intravenous infusions of mannose terminal glucocerebrosidase.

An amplified ELISA has been employed for monitoring the safety of repeated intravenous infusions of modified human placental glucocerebrosidase. The enzyme infusions consisted of biweekly injections of macrophage targeted glucocerebrosidase over a 6 month duration. Serum samples collected throughout the study were assayed by use of an ELISA using alkaline phosphatase coupled to alcohol dehydrogenase for amplification. Using this protocol, 0.2-5 ng affinity purified immunoglobulin specific for glucocerebrosidase can be detected. Occasional false positives necessitate multiple repeat assays over time to accurately assess immunogenic response. Blinded ELISAs were performed on sera from both infused patients with Gaucher's disease and uninfused control patients and compared with apparent immunoglobulin concentration in 54 normal control sera. Although several samples showed apparently elevated immunoglobulin levels, repeat analyses failed to demonstrate high levels reproducibly. Furthermore, these sera were unable to neutralize enzyme or to precipitate radiolabelled enzyme, confirming the absence of antibody. Problems with high sensitivity ELISA formats are discussed.

Lectin-specific targeting of lysosomal enzymes to reticuloendothelial cells.

The principles and methods used for enzymatic modification of the carbohydrate portion of glucocerebrosidase are similar to those performed by Ashwell and Morell, Stahl, and others. It is difficult to explain the lack of uptake of native enzyme through binding of the high-mannose type glycopeptide to Man/GlcNAc receptors since approximately 20% of the total oligosaccharides on the native enzyme are high mannose type. Possibly a requirement for multiple sites of attachment to the receptor is not met by a single high-mannose type oligosaccharide per molecule. Alternatively, the presence of complex type oligosaccharides on this enzyme, demonstrated by structural studies, may mask the mannose site and thus account for the poor uptake of native enzyme. The ability to successfully deglycosylate any protein or enzyme in order to specifically target a selected cell type requires that there be (1) an available source of pure enzyme; (2) specific exoglycosidases of high specific activity available either commercially or relatively easily purified; (3) chemical or biochemical means available for the testing of the product, preferably at each step; and (4) a means of separating the glycosidases used from the desired enzyme product. The characteristic and unique accumulation of glucocerebroside only in cells of the monocyte- histiocyte series, makes Gaucher's disease an excellent prototype for the study of enzyme replacement therapy. The principles demonstrated for the enzymatic deglycosylation of glucocerebrosidase may be applied to the cell-specific delivery of other glycoproteins as well. Other lysosomal diseases in which storage occurs in multiple cell types may be ameliorated by administration of macrophage-directed enzymes if, by so doing, storage material can be digested during the normal phagocytic turnover of senescent cells. Consideration of the kinetics of degradation and the structural features affecting the stability of enzymes in vivo are prerequisites to improving the bioengineering of targeted lysosomal enzymes. Animal and culture models have been developed for the study of glucocerebrosidase delivery to specific cell types and substrate degradation. Other studies have progressed toward a definition not only of the receptors at the plasma membrane involved in the internalization of exogenous enzyme, but also of internal receptors or properties of the lysosome responsible for intracellular protein trafficking. A complete understanding of the forces acting to direct endogenous or exogenously supplied enzyme to a given subcellular organelle will require a synthesis of experimental results from all areas of glycoprotein research.

Immunoelectron microscopic localization of mannose-terminal glucocerebrosidase in lysosomes of rat liver Kupffer cells.

Knowledge of the cellular distribution and subcellular localization of mannose-terminal glucocerebrosidase after intravenous infusion is necessary for understanding the efficacy of targeted enzyme replacement therapy for Gaucher's disease. Selective uptake of mannose-terminal glucocerebrosidase by Kupffer cells in rat liver has been previously demonstrated biochemically. In this study we used immunohistochemical and immunogold labeling techniques to provide direct visual proof for the localization of the delivered enzyme. Light microscopy confirmed biochemical data identifying non-parenchymal cells as the primary target of the modified glucocerebrosidase. Using a primary antibody specific for glucocerebrosidase and a secondary gold-conjugated antibody, we used immunoelectron microscopy to quantify the extent and distribution of exogenous enzyme in various cell types in rat liver and its localization within their respective subcellular organelles. Thirty minutes after intravenous administration of mannose-terminal glucocerebrosidase, enzyme was localized primarily in lysosomes of Kupffer cells. Of eight intact Kupffer cells counted, 16 of 21 lysosomes (78%) contained immunogold conjugates (average concentration 293 gold particles/micron 2). Of 589 particles counted in these lysosomes, 485 (82%) were localized within the lumen of the lysosome; only 104 (18%) were membrane-associated. Five of the 21 lysosomes counted were negative for gold. No gold particles were found in the mitochondria of Kupffer cells and very few particles (8.2/microns 2) were found over the nucleus. The density of gold particles was also low over the nucleus (7.2/microns 2), mitochondria (8.8/microns 2), and lysosomes (7.9/microns 2) of hepatocytes. No specific labeling was observed in erythrocytes, platelets, lymphocytes, pit cells, fat-storing cells, or bile duct. Background labeling of control liver sections from rats receiving saline injection was 8.2 +/- 1.4 gold particles/microns 2. We conclude that mannose-terminal glucocerebrosidase is delivered to the lysosomes of Kupffer cells in liver and that it is distributed both within the lumen (82%) and over the membrane (18%) of the lysosome, with a slight preferential association with the membrane. These findings may provide insights into the design of more effective therapeutic enzyme preparations for the treatment of Gaucher's disease.

Peptide: N-glycosidase F: studies on the glycoprotein aminoglycan amidase from Flavobacterium meningosepticum.

Peptide: N-glycosidase from Flavobacterium meningosepticum was isolated in a homogeneous state and its physico-chemical characterization was accomplished. The reliability of the previously recorded assay procedures was assessed. Using an octaglycopeptide derived from ovomucoid a rapid and sensitive FPLC method was developed for the assay of enzymatic activity. Peptide: N-glycosidase was found to effect deglycosylation of glycoproteins bearing complex and/or multiantennary glycans even in their native state. In contrast, glycoproteins with high mannose and/or hybrid carbohydrates required denaturation to become susceptible to deglycosylation by the enzyme.

Flavobacterium meningosepticum peptide:N-glycosidase: influence of ionic strength on enzymatic activity.

Flavobacterium meningosepticum peptide:N-glycosidase-mediated deglycosylation of N-linked glycan strands of glycoproteins has been found to be strongly influenced by the ionic strength of the assay medium. By use of a modification of a previously published assay procedure for quantitative analysis of glycan release we have been able to improve reproducibility and thus to compare the extent of deglycosylation achieved under a variety of conditions of ionic strength. We have observed that enzyme activity is adversely affected by high ionic strength buffers such as those recommended for deglycosylation of various glycoproteins and recommend the use of low ionic strength buffers for routine use.

Efficient correction of Fabry mice and patient cells mediated by lentiviral transduction of hematopoietic stem/progenitor cells.

A deficiency in alpha-galactosidase A (alpha-gal A) activity causes Fabry disease. Virus-based delivery of genes can correct cells and establish a sustained supply of therapeutic proteins. Recombinant lentiviral vectors (LVs) show promise in this context. We first demonstrate LV-mediated marking of peripheral blood (PB) cells by transduction/transplantation of hematopoietic stem/progenitor cells. Stable enGFP expression was observed in PB for 37 weeks. Next, we transplanted Fabry mice with bone marrow mononuclear cells (BMMNCs) transduced a single time with a LV encoding the human alpha-gal A cDNA. Sustained expression of functional alpha-gal A in Fabry mice was observed over 24 weeks. Plasma alpha-gal A activity from treated Fabry mice was two-fold higher than wild-type controls. Increased alpha-gal A activity, often to supra-normal levels, and reduction of globotriaosylceramide, a glycolipid that accumulates in Fabry disease, was observed in all organs assessed. In secondary bone marrow transplantations, Fabry mice showed multilineage marking of PB, splenocytes and BMMNCs, along with therapeutic levels of alpha-gal A activity in plasma and organs over 20 weeks. Lastly, we transduced mobilized PB CD34(+) cells from a Fabry patient and observed corresponding enzymatic increases. Thus a single LV-mediated transduction of primitive hematopoietic cells can result in sustained correction for Fabry disease.

Image-guided, direct convective delivery of glucocerebrosidase for neuronopathic Gaucher disease.

OBJECTIVE: To determine if convection-enhanced delivery (CED) of glucocerebrosidase could be used to treat targeted sites of disease progression in the brain and brainstem of a patient with neuronopathic Gaucher disease while monitoring enzyme distribution using MRI. METHODS: A CED paradigm in rodents (n = 8) and primates (n = 5) that employs co-infusion of a surrogate MRI tracer (gadolinium diethylenetriamine penta-acetic acid [Gd-DTPA]) with glucocerebrosidase to permit real-time monitoring of distribution was developed. The safety and feasibility of this delivery and monitoring paradigm were evaluated in a patient with type 2 Gaucher disease. RESULTS: Animal studies revealed that real-time, T1-weighted, MRI of Gd-DTPA accurately tracked enzyme distribution during CED. Targeted perfusion of clinically affected anatomic sites in a patient with neuronopathic Gaucher disease (frontal lobe and brainstem) with glucocerebrosidase was successfully performed. Real-time MRI revealed progressive and complete filling of the targeted region with enzyme and Gd-DTPA infusate. The patient tolerated the infusions without evidence of toxicity. CONCLUSIONS: Convection-enhanced delivery can be used to safely perfuse large regions of the brain and brainstem with therapeutic levels of glucocerebrosidase. Co-infused imaging surrogate tracers can be used to monitor and control the distribution of therapeutic agents in vivo. Patients with neuronopathic Gaucher disease and other intrinsic CNS disorders may benefit from a similar treatment paradigm.

Cardiovascular malformations in Oryzias latipes embryos treated with 2,4,5-trichlorophenoxyacetic acid (2,4,5-T).

Exposure of fertilized Oryzias latipes eggs to concentrations of 2,4,5-trichlorophenoxyacetic acid up to 14 ppm had no gross developmental effects. Exposure to concentrations above 14 ppm resulted in multiple malformations of the heart and blood vessels. Embryos failed to hatch when exposed to concentrations above 25 ppm.

Mannose 6-phosphate receptor-mediated uptake of modified low density lipoprotein results in down regulation of hydroxymethylglutaryl-CoA reductase in normal and familial hypercholesterolemic fibroblasts.

Monophosphotetramannosyl-1-deoxymannitol-1-yl-low density lipoprotein (Man-6-P-LDL) was prepared by covalent attachment of the pentasaccharide omega-(6-phospho-tetra(alpha 1-3)mannosyl(alpha 1-2)mannose to amino groups on low density lipoprotein. Normal human fibroblasts were shown to specifically bind, internalize, and degrade 125I-labeled Man-6-P-LDL. Specificity for the mannose 6-phosphate (Man-6-P) receptor was demonstrated by competitive displacement with cold Man-6-P-LDL, Man-6-P, or mannose. No displacement was seen with cold LDL. Kd is estimated to be less than or equal to 2 X 10(-9) M. Degradation of 125I-labeled Man-6-P-LDL in familial hypercholesterolemic fibroblasts showed the same time course and specificity as observed in normal fibroblasts. Man-y-P-LDL was also able to deliver cholesterol to the cytosol where down regulation of the enzyme 3-hydroxy-3-methylglutaryl CoA reductase was observed in both normal and familial hypercholesterolemic fibroblasts. Down regulation could be blocked by Man-6-P in both cell lines. The possible uses of agents such as Man-6-P-LDL as research probes and therapeutic tools directed to specific cell types are discussed.

Purification and characterization of bovine brain glucocerebrosidase.

Glucocerebrosidase was isolated from bovine brain by cholate extraction, ammonium sulfate fractionation, acid precipitation at pH 5.35, and hydrophobic chromatography. The purification is about 2400-fold with a specific activity of about 286,000 nmole/hr/mg protein. Molecular weight as determined by chromatography on Bio-Gel P-200 was 138,000. On SDS-polyacrylamide gel electrophoresis the enzyme protein resolved into two bands with apparent molecular weights of 63,000 and 56,000. These bands are cross-reactive to monospecific polyclonal antibody to homogeneous human placental glucocerebrosidase. The enzyme was found to be a complex glycoprotein based on its lectin binding specificity. Brain enzyme was found to be similar to placental glucocerebrosidase in its pH optima, heat stability at 52 degrees C, and substrate affinity. Enzyme kinetics were measured in the presence of conduritol-beta-epoxide, an irreversible inhibitor, and gluconolactone, a competitive inhibitor.

Ricin linked to monophosphopentamannose binds to fibroblast lysosomal hydrolase receptors, resulting in a cell-type-specific toxin.

The receptor specificity of the plant seed toxin ricin, which ordinarily binds to galactose-containing receptors, has been altered by coupling monophosphopentamannose residues to ricin by reductive amination and by reversibly binding lactose to the modified ricin. The added monophosphopentamannose residues provide ricin with the recognition factor common to fibroblast lysosomal hydrolases and enable the modified ricin (Man6P-ricin) to bind to the fibroblast Man6P receptor and inhibit protein synthesis in the cells via this receptor. The addition of lactose to Man6P-ricin saturates the galactose site on Man6P-ricin and prevents the binding of Man6P-ricin to cells via galactose-containing ricin receptors. The Man6P receptor-mediated toxicity of Man6P-ricin, identified in human fibroblasts by competition by Man6P and blockade by alkaline phosphatase treatment, was not detected in HeLa cells or human amnion cells. Consequently, in the presence of lactose, the fibroblasts were 8 and 13 times more sensitive than amnion and HeLa cells, respectively. These results show that highly toxic cell-type-specific reagents can be made by the proper alteration of toxin receptor specificities. An attempt to construct a highly toxic altered toxin by adding Man6P residues to diphtheria toxin fragment A was unsuccessful. A possible explanation is that in Man6P-ricin the ricin B chain performs some entry function, even though the initial binding step occurs at the Man6P receptor.

Modifying exogenous glucocerebrosidase for effective replacement therapy in Gaucher disease.

Important therapeutic principles were established in developing effective enzyme replacement therapy for patients with Gaucher disease. The background and sequence of the investigations that led to effective delivery of exogenous glucocerebrosidase to the lipid-storing macrophages in patients with Gaucher disease are described. The principle of targeting the intravenously injected enzyme to the mannose lectin on the surface of these cells by engineering the glycoform of the enzyme is a useful model of an essential requirement for effective enzyme therapy. Similar strategies are expected to be effective for the treatment of a number of hereditary metabolic disorders of humans.

Studies on the galactose-binding site of ricin and the hybrid toxin Man6P-ricin.

N-acetylimidazole (NAI) was used to O-acetylate the plant seed toxin ricin. O-acetylation of one to two tyrosine residues per molecule of ricin inhibited ricin binding to Sepharose 4B and decreased toxicity by 90% in a protein synthesis inhibition assay in HeLa cells. Lactose, known to block the binding site on the ricin B subunit, protected ricin from NAI modification of binding or toxicity. Thus NAI, under these conditions, can be a lactose site-specific inhibitor. The lactose site-specific modification of the hybrid toxin, Man6P-ricin, performed under the same conditions, exhibited the same 90% inhibition of Man6P receptor-mediated toxicity as the galactose-containing receptor-mediated toxicity of either Man6P-ricin or ricin. Thus the ricin B chain lactose-binding site appears to be essential for the high potency of Man6P-ricin via the new cell type-specific Man6P receptor. Treatment of fibroblasts with neuraminidase exposes galactose residues, thus increasing the sensitivity to ricin eight fold. The Man6P receptor-mediated toxicity of Man6P-ricin is not affected by this treatment, although the galactose-inhibited route is potentiated eight fold. The Man6P-ricin hybrid appears to require the ricin B chain galactose-binding site to enter the cytosol after initially binding to the Man6P receptor. These data provide some insights into the proper design of hybrid toxins. We discuss a number of possible models for hybrid toxin entry.

Studies on the turnover of glucocerebrosidase in cultured rat peritoneal macrophages and normal human fibroblasts.

The kinetics of glucocerebrosidase synthesis and degradation in rat peritoneal macrophages and in human fibroblasts have been studied using conduritol B epoxide (CBE), an irreversible and specific inhibitor of mammalian glucocerebrosidase. In cultured fibroblasts, higher concentrations of CBE and/or longer times were required for inhibition of glucocerebrosidase than were necessary for inhibition of the macrophage enzyme. However, inhibition of activity in cell extracts from both cell types showed identical time and concentration dependence. After the removal of CBE from cultures, enzyme activity returned to normal with a half-time of 48 h for macrophages and 40 h for fibroblasts. The reappearance of enzyme activity was prevented by an inhibitor of protein synthesis. Both the rate of synthesis and degradation of glucocerebrosidase enzyme protein were independent of the presence of CBE. The calculated rate of degradation of glucocerebrosidase was confirmed using metabolically labelled enzyme in cell cultures. The rate of synthesis for macrophages is 1.8 ng enzyme h-1 mg cell protein-1 and the rate of degradation is 1.4% h-1 (0.014 h-1). These values were 2.0 ng h-1 mg-1 and 0.018 h-1 for fibroblasts.

Glucocerebrosidase deficiency and lysosomal storage of glucocerebroside induced in cultured macrophages.

A cell culture model stimulating the genetic deficiency of glucocerebrosidase has been developed, utilizing macrophages and conduritol B epoxide (CBE), the specific irreversible inhibitor of the enzyme. Rat peritoneal macrophage glucocerebrosidase was completely inhibited when cells were treated with 10 microM CBE for 16 h or 100 microM CBE for 2 h. The t1/2 of inactivation was 30 min at 10 microM concentration. When cells were washed free of CBE, the enzyme activity reappeared linearly with time, reaching 50% of control activity 48 h after removal of the inhibitor. CBE-treated macrophages have normal phagocytic activity toward [3H]glycine-coupled latex beads and a normal number of mannose receptors. CBE was found to have no effect on other lysosomal enzymes. When [14C]glucocerebroside, encapsulated in multilamellar liposomes with alpha-D-mannopyranoside covalently coupled to the surface, was fed to glucocerebrosidase-depleted macrophages, the radiolabelled glycolipid accumulated and was undegraded. Subcellular fractionation on a Percoll density gradient demonstrated that the stored glucocerebroside in the CBE-treated macrophages was localized in lysosomes.