Changes in immunologic properties of group A RBCs during treatment with an A-degrading exo-alpha-N-acetylgalactosaminidase.

BACKGROUND: Blood group A RBCs theoretically can be converted to universal-donor group O cells by incubation with alpha-N-acetylgalactosaminidase (A-zyme). The purpose of this study was to compare the extent of A-zyme treatment required to abolish several immunologic responses of group B or O recipients to group A RBCs. STUDY DESIGN AND METHODS: A(1) RBCs were incubated for 2 hours at 37 degrees C in buffer alone (control) or with 0.1 to 5.0 units (U) of A-zyme per mL of packed cells. They were then tested for 1) immune adherence (% rosetting) and activation of monocytic cells (% erythrophagocytosis, TNFalpha production), 2) activation of immune hemolysis, and 3) hemagglutination with Dolichos biflorus lectin and pooled human anti-A serum. RESULTS: A 2-hour incubation with < or =5.0 U of A-zyme per mL of packed cells abolished monocytic cell adherence and phagocytosis of group A(1) RBCs coated with IgG(3) A MoAb. Epitopes of A binding to IgG(3) anti-A A005 MoAb and BG-2 MoAb differed in A-zyme sensitivity. TNFalpha production in group B or O whole blood in response to the addition of A(1) cells varied, but it essentially was abolished when group A cells were treated with 0.1 to 1.0 U of A-zyme per mL of packed cells. A epitopes mediating immune hemolysis and hemagglutination with D. biflorus lectin were also cleaved by <5 U of A-zyme per mL of packed cells. In contrast, hemagglutination with polyclonal anti-A typing serum was diminished by only 1 to 2 serial titer dilutions. CONCLUSION: A epitopes mediating immune hemolysis and immune adherence to and activation of monocytic cells are highly sensitive to A-zyme cleavage, as compared to those mediating hemagglutination with monoclonal and polyclonal anti-A.

Purification and characterization of blood group A-degrading isoforms of alpha-N-acetylgalactosaminidase from Ruminococcus torques strain IX-70.

To cleave blood group A immunodeterminants from erythrocytes (Hoskins, L. C., Larson, G., and Naff, G. B. (1995) Transfusion 35, 813-821), we purified and characterized alpha-N-acetylgalactosaminidase (EC 3.2.1.49) activity from culture supernatants of the human fecal bacterium Ruminococcus torques strain IX-70. Three isoforms separated during hydrophobic interaction chromatography. Hydroxyapatite chromatography further resolved the most hydrophilic, isoform I, into isoforms IA and IB. The most hydrophobic, isoform III, differed from IA and IB by a more acidic pH optimum, greater heat resistance, greater sensitivity to alkylating agents, and anomalous retardation during gel filtration chromatography. Isoform IB differed from IA and III in N-terminal amino acid sequence and in sensitivity to EDTA inhibition. Each cleaved nonreducing alpha(1-->3)-N-acetylgalactosamine residues from human blood group A and AB mucin glycoproteins, Forssman hapten, and blood group A lacto series glycolipids. The apparent molecular mass of denatured isoform subunits of IA, IB, and III-PII (158, 173, and 201 kDa, respectively) bore no integer relationship to the apparent molecular mass of the native isoforms (265, 417, and 530 kDa), but the latter bore a ratio of 1.96:3.09:3.93 to the weight-average apparent molecular mass of native IA (135 kDa), suggesting that the isoforms are multimers of a 135-kDa sequence. Isoforms IA and III-PII had an identical N-terminal amino acid sequence which showed homologies to the N-terminal sequence of sialidases produced by Bacteroides fragilis SBT3182, another commensal enteric bacterium.

Blood group A immunodeterminants on human red cells differ in biologic activity and sensitivity to alpha-N-acetylgalactosaminidase.

BACKGROUND: Epitopes of blood group A antigen can be enzymatically cleaved from red cells (RBCs), but the extent of cleavage required for normal survival in allogeneic blood transfusion recipients is unknown. Therefore, the cleavage rates were studied for A antigen epitope binding of 1) complement-activating anti-A, 2) Dolichos biflorus anti-A, lectin, and 3) hemagglutinating anti-A during incubation with a purified alpha-N-acetylgalactosaminidase, E.C. 3.2.1.49 (alpha-GalNAc'ase). STUDY DESIGN AND METHODS: Suspensions of group A RBCs were incubated with alpha-GalNAc'ase. Cells were removed at intervals, washed, and tested for loss of binding by monoclonal, polyclonal, and complement-activating anti-A, D. biflorus anti-A1 lectin, and Ulex europaeus anti-H lectin. RESULTS: A epitopes binding D. biflorus lectin were highly susceptible to alpha-GalNAc'ase; simultaneously with their loss, binding with U. europaeus lectin emerged. Loss of complement-mediated hemolysis was slower. A epitopes binding hemagglutinating anti-A were most resistant. Cleavage of A epitopes from membrane glycosphingolipids with short oligosaccharide chains was similarly resistant. Rates of cleavage from A1 and A2 RBCs were similar. CONCLUSION: RBC epitopes of blood group A differ in susceptibility to cleavage and biologic reactivity, which suggests that subsets mediating important biologic functions exist on functionally and topographically distinct membrane glycoconjugates.

Mucin degradation in the human colon: production of sialidase, sialate O-acetylesterase, N-acetylneuraminate lyase, arylesterase, and glycosulfatase activities by strains of fecal bacteria.

Oligosaccharide side chains of human colonic mucins contain O-acetylated sialic acids and glycosulfate esters. Although these substituents are considered to protect the chains against degradation by bacterial glycosidases, sialate O-acetylesterase, N-acetylneuraminate lyase, and glycosulfatase activities have been found in fecal extracts. To better define the source of these activities, we measured extracellular and cell-bound sialidase, sialate O-acetylesterase, N-acetylneuraminate lyase, arylesterase, and glycosulfatase activities produced by 23 isolates of human fecal bacteria grown anaerobically in a hog gastric mucin culture medium; these represented dominant populations of fecal anaerobes, facultative anaerobes, and the subset of mucin oligosaccharide-degrading bacteria. Every strain produced sialidase and high levels of arylesterase, and all but five facultative anaerobes produced sialate O-acetylesterase. Sialic acids containing 2 mol or more of O-acetyl ester per mol of sialic acid were cleaved from mucin glycoproteins more slowly by sialidases of mucin oligosaccharide-degrading stains than were sialic acids containing 1 or 0 mol, and only N-acetyl- and mono-O-acetylated sialic acids were recovered from enzyme digests of a mucin containing di-O-acetylated sialic acids. No detectable N-acetylneuraminate lyase activity was produced by any strain, but low activity was induced by increasing the glycoprotein-bound sialic acid concentration in the culture medium of six Escherichia coli strains. Using lactitol-6-sulfate as a substrate, we found weak glycosulfatase activity in the partially purified, concentrated enzyme mixture in the culture supernatants of four mucin oligosaccharide-degrading strains but in none of the unconcentrated culture fractions. We conclude that the presence of two or more O-acetyl groups on sialic acids inhibits enteric bacterial sialidases but that production of sialate O-acetylesterases by several populations of enteric bacteria lessens the likelihood that mucin oligosaccharide chains terminating in O-acetylated sialic acids are protected from degradation. Sialate O-acetylesterases have a role in bacterial degradation of mucin glycoproteins in the human colon.

Deantigenation of human erythrocytes by bacterial glycosidases--evidence for the noninvolvement of medium-sized glycosphingolipids in the Dolichos biflorus lectin hemagglutination.

Fresh human A1 erythrocytes, washed and pretreated in phosphate buffer with or without papain, were incubated at 37 degrees C with blood group-degrading enzymes from the human fecal Ruminococcus torques strain IX-70. The effects were assayed as changes in hemagglutination patterns, and blood group activities of alkali stable glycolipid extracts from the enzyme-treated cells using Dolichos biflorus anti-A1 lectin, Ulex europaeus type 1 anti-H lectin, and various monoclonal anti-A antibodies. Hemolysis was negligible (less than or equal to 1% after 6 h), and the osmotic fragility increased slightly only after papain treatment. The papain-untreated A1 erythrocytes lost D. biflorus agglutinability within minutes at room temperature with the unfractionated bacterial enzyme mixture IX-70 (42 mU 1,3-alpha-N-acetylgalactosaminidase (alpha-GalNAc'ase)/ml), but remained A active by strong agglutination with BioClone anti-A antibody even after 6 h of incubation. Thin layer chromatographic (TLC) immunostaining of extracted lipids showed hydrolysis of D. biflorus binding glycosphingolipids with more than six monosaccharides after 1 h, i.e., at a slower rate than the loss of D. biflorus agglutinability. Disappearance of these glycosphingolipids after 1 h paralleled the appearance of U. europaeus agglutinability and the strong binding of this lectin to glycolipid extracts in TLC immunoassays. A partly purified 1,3-alpha-GalNAc'ase (XI-117) (100 mU/ml) and a 1,2-alpha-fucosidase fraction (XI-50) containing alpha-GalNAc'ase (10 mU/ml) did not degrade blood group A active glycosphingolipids but completely abolished the D. biflorus agglutinability within 6 h. Papain pretreatment exposed U. europaeus receptors on the cell surface without changing the A1 hemagglutination pattern. It also facilitated a complete degradation of D. biflorus and U. europaeus reactive glycolipids with the IX-70 enzyme mixture within 6 h. The D. biflorus lectin was a good discriminator of A1/A2 subjects using erythrocyte lipid extracts but had a low affinity for the blood group A type 3 and type 4 glycosphingolipids in the TLC-overlay technique. In conclusion this study shows that (i) loss of D. biflorus A1 hemagglutination does not correlate with a loss of D. biflorus binding glycosphingolipids and (ii) loss of D. biflorus binding glycosphingolipids does not correlate with a loss of D. biflorus agglutinability. The results indicate that the serological D. biflorus agglutinability of A1 erythrocytes is not dependent on medium-sized glycosphingolipids (hexa- to dodecaglycosylceramides).

Enhancing effects of bile salts on the degradation of glycosphingolipids by glycosidases from bacteria of the human fecal flora.

Different concentrations of ionic and non-ionic detergents were examined for optimization of the in vitro degradations of intestinal glycosphingolipids by alpha- and beta-glycosidases from human fecal bacteria. In 5 mM Triton X-100 the enzymes hydrolyzed glycosphingolipids with lactoseries type 1 and 2 chains essentially to lactosylceramide (LacCer). In 5 mM sodium di- and trihydroxy bile salts lactosylceramide was degraded to glycosylceramide (GlcCer) in varying extent by enzymes from all five strains. The minimal bile salt concentrations for optimal 1,4-beta-galactosidase activities varied between 1 and 20 mM, i.e., close to or above the critical micellar concentrations (cmc). Dihydroxy bile salts were the most efficient in promoting conversion of LacCer to GlcCer at concentrations below 10 mM and conjugation with a taurine residue did not markedly lower the GlcCer yield. The optimal detergent concentrations for hydrolyses of the p-nitrophenyl (pnp) glycosides Gal beta 1-pnp and GalNAc alpha 1-pnp were approximately 0.05 mM for Triton X-100 and 0.5 mM for sodium taurodeoxycholate, i.e., clearly below their reported cmc values. Galabiosylceramide, globotria- and globotetraosylceramides, not degraded in the Triton X-100 micelles, were also resistant to hydrolysis using the sodium bile salts as detergents. In contrast, lactotetraosylceramide and isoglobotriaosylceramide were significantly more degraded by enzymes from a Ruminococcus gnavus strain and gangliotetraosylceramide by enzymes from a Bifidobacterium bifidum and a Bifidobacterium infantis strain using bile salt detergents. All strains but R. gnavus released terminal GalNAc from para-Forssman but not from the globotetraosylceramide or Forssman structures using 5 mM sodium deoxycholate as detergent. GM1 desialylation by two Ruminococcus torques strains and the R. gnavus and B. bifidum strains were enhanced under identical conditions. We conclude that the observed effects on glycosphingolipid hydrolyses reflects variations in the micellar presentation of the substrates. In addition, detergents seem to have a direct stimulating effect on the glycosidases, however at concentrations 10-100-times below the ones optimal for glycolipid degradations. These results with optimized bile salt concentrations, further support our previous observations that these five fecal bacterial strains produce enzymes with selected specificities towards glycosphingolipid core chains of the lactoseries type 1 and 2.

Bacteria of the human intestinal microbiota produce glycosidases specific for lacto-series glycosphingolipids.

Five strains of human fecal bacteria, of the Ruminococcus and Bifidobacterium genera, produce extracellular alpha- and beta-glycosidases that degrade intestinal mucin oligosaccharides and glycosphingolipids of the lacto-series type 1 chain. We have tested the activities and substrate specificities of these enzymes using para-nitrophenyl glycosides and glycosphingolipids of different core chains (lacto, neolacto, globo, isoglobo, galabio, and ganglio), carrying different blood group determinants (A, H, X, Y, Forssman, and para-Forssman), and with different degrees of sialylation (mono- to tetra-sialo). Lactotetraosylceramide and neolactotetraosylceramide were the only core glycosphingolipids degraded by enzymes from these strains, resulting in lactosylceramide and glucosylceramide as the major end products. R. gnavus strain VI-268 did not degrade lactotetraosylceramide but only neolactotetraosylceramide yielding lactotriaosylceramide and lactosylceramide as the major end products. All strains but R. gnavus VI-268 also produced lactosylceramide from a bi-antennary 10-sugar glycosphingolipid with two blood group H determinants based on a lactotetraosylceramide core. Apart from strain specific blood group A-degrading (R. torques strain VIII-239 and IX-70, R. gnavus strain VI-268 and B. infantis VIII-240) and Forssman-degrading (R. torques VIII-239 and IX-70) activities, all strains also degraded the H-5, X-5, and Y-6 glycosphingolipids. All strains released N-acetylneuraminic acid from the gangliosides sialosyl-neolactotetraosylceramide, GD3, GD1a, GD1b, GT1b, and GQ1b corresponding to 2,3-alpha- and 2,8-alpha-N-acetylneuraminidase activities. The R. torques strains VIII-239 and IX-70 also partially desialylated GM1 to lactotetraosylceramide. The para-nitrophenyl glycoside degradations were often incompatible with the data from the glycosphingolipids degradations.(ABSTRACT TRUNCATED AT 250 WORDS)

Degradation of human intestinal glycosphingolipids by extracellular glycosidases from mucin-degrading bacteria of the human fecal flora.

Certain normal strains of human fecal bacteria are unique in producing extracellular glycosidases that degrade the oligosaccharide chains of gut mucin glycoproteins. We have studied the action of such glycosidases partially purified from the cell-free supernates of five of these strains on intestinal glycosphingolipids isolated from human meconium. The glycolipids were sialosyl-lactosylceramide, lactosylceramide, and fucolipids with A, B, H, Lea, or Leb blood group determinants. In addition to the strain-specific high blood group A-degrading activities (Ruminococcus torques strains VIII-239 and IX-70), B-degrading activity (Ruminococcus AB strain VI-268), and H-degrading activities (all strains) corresponding to alpha 1-3-N-acetylgalactosaminidase, alpha 1-3-galactosidase and alpha 1-2-fucosidase, respectively, all strains also degraded sialosyl-lactosylceramide and Lea and Leb antigenic glycolipids, indicating the presence of alpha 2-3-neuraminidases and alpha 1-4-fucosidases. Enzyme preparations from Bifidobacterium infantis strain VIII-240 and R. torques strain VIII-239 hydrolyzed the Lea active glycolipid directly to lactosylceramide, suggesting the presence of endo-beta 1-3-N-acetylglucosaminidase activities. Similar endo-beta-N-acetylglucosaminidase activities were identified in four of the five enzyme preparations. The enzymes produced by R. AB strain VI-268 lacked this activity as well as beta 1-3-galactosidase, and thus degradation stopped at lactotetraosylceramide. With enzyme preparations from the other strains lactosylceramide was the single major degradation product from complex glycosphingolipids with less than 30% further degradation to glucosylceramide within 48 h. We conclude that glycosidases from mucin-degrading strains of human enteric bacteria degrade oligosaccharide chains of lactoseries fucolipids and gangliosides of intestinal origin primarily to lactosylceramide. Since several genera of enteric bacteria bind preferentially to lactosylceramide in vitro, mucin-degrading strains may have an important ecological role in host-microbial associations in the human gut.

Mucin degradation in human colon ecosystems. Isolation and properties of fecal strains that degrade ABH blood group antigens and oligosaccharides from mucin glycoproteins.

We previously reported that the oligosaccharide chains of hog gastric mucin were degraded by unidentified subpopulations numbering approximately 1% of normal human fecal bacteria. Here we report on the enzyme-producing properties of five strains of mucin oligosaccharide chain-degrading bacteria isolated from feces of four healthy subjects. Four were isolated from the greatest fecal dilutions yielding mucin side chain-degrading activity in culture, and thus were the numerically dominant side chain-degrading bacteria in their respective hosts. Three were Ruminococcus strains and two were Bifidobacterium strains. Two Ruminococcus torques strains, IX-70 and VIII-239, produced blood group A- and H-degrading alpha-glycosidase activities, sialidase, and the requisite beta-glycosidases; these strains released greater than 90% of the anthrone-reacting hexoses from hog gastric mucin during growth in culture. The Bifidobacterium strains lacked A-degrading activity but were otherwise similar; these released 60-80% of the anthrone-reacting hexoses but not the A antigenic structures from hog gastric mucin. Only Ruminococcus AB strain VI-268 produced blood group B-degrading alpha-galactosidase activity, but this strain lacked beta-N-acetylhexosaminidases to complete degradation of B antigenic chains. When this strain was co-cultured with a strain that produced beta-N-acetylhexosaminidases, release of hexoses from blood group B salivary glycoprotein increased from 50 to greater than 90%, and bacterial growth was enhanced. The glycosidases required for side chain degradation were produced by these strains in the absence of mucin substrate, and a substantial fraction of each activity in stationary phase cultures was extracellular. In contrast, none of 16 other fecal Bacteroides, Escherichia coli, Streptococcus faecalis, and Bifidobacterium strains produced ABH blood group-degrading enzymes; other glycosidases produced by these strains were predominantly cell bound except for extracellular beta-N-acetylhexosaminidases produced by the five S. faecalis strains. We conclude that certain Bifidobacterium and Ruminococcus strains are numerically dominant populations degrading mucin oligosaccharides in the human colon due to their constitutive production of the requisite extracellular glycosidases including blood group antigen-specific alpha-glycosidases. These properties characterize them as a functionally distinct subpopulation of normal human enteric microflora comprised of specialized subsets that produce blood group H antigen-degrading glycosidases alone or together with either blood group A- or B-degrading glycosidases.

In vitro degradation of gastric mucin. Carbohydrate side chains protect polypeptide core from pancreatic proteases.

The polypeptide core of mucin glycoprotein subunits resists cleavage by proteases. To determine if the carbohydrate side chains of these subunits protect the underlying polypeptide core from proteolytic cleavage, we compared the effect of pancreatic proteases on hog gastric mucin before and after cleavage of its carbohydrate moieties by bacterial glycosidases from an anaerobic human fecal culture supernate. Hog gastric mucin was resistant to pancreatic proteases: less than 10% of the mucin protein became soluble in 80% vol/vol ethanol after 24-h incubation with trypsin, alpha-chymotrypsin, and elastase, and its elution pattern from Sephadex G-200 remained unchanged after elastase treatment, with 90% eluting at the void volume. By contrast, after removal of 50% of its carbohydrates, mucin was susceptible to pancreatic proteases: 50% of mucin protein became soluble in 80% vol/vol ethanol after 24-h incubation with alpha-chymotrypsin and elastase, and 24% of mucin protein became soluble in 80% vol/vol ethanol after 24-h incubation with trypsin; after elastase treatment, its elution from Sephadex G-200 was markedly retarded. We conclude that the carbohydrate side chains of hog gastric mucin glycoprotein protect the underlying polypeptide core from proteolysis and that degradation of the carbohydrate side chains by glycosidases from fecal bacteria renders the polypeptide core susceptible to pancreatic proteases.

Pancreatitis with arthropathy and subcutaneous fat necrosis. Evidence for the pathogenicity of lipolytic enzymes.

The occurrence of peripheral fat necrosis in exceptional cases of pancreatic disease is not well understood. We report studies on such a patient with arthropathy and subcutaneous nodules. Examination of serial serum samples demonstrated striking elevations of the pancreatic enzymes phospholipase A, 3-3.4 units/ml (normal 0.17-0.41); lipase, 7-39 Sigma-Tietz units/ml (normal less than 1); immunoreactive trypsin, 912-3,207 ng/ml (normal 12-41). The distinguishing characteristic of the patient's synovial fluid was a marked elevation of hydrolized fatty acids (680 mg/dl versus 19 +/- 19 in control inflammatory joint fluids). Synovial fluid fatty acid distribution was identical to values for tissue fat. In contrast, serum fatty acid levels and distribution were normal. No associated proteinase inhibitor or significant immunologic abnormality was detected. Certain properties of adipose cells and lipolytic enzymes may help explain the characteristically selective necrosis of fat cells observed in this syndrome.

Mucin degradation in human colon ecosystems. Degradation of hog gastric mucin by fecal extracts and fecal cultures.

The enteric flora of rats plays a major role in the degradation of their gut mucin glycoproteins, but the role of human enteric flora in degrading gut mucin glycoprotein has not been as well defined. We have studied the degradation of hog gastric mucin, which is structurally similar to human gastric mucin, in anaerobic human fecal cultures and by partially purified enzyme fractions from human fecal extracts and anaerobic culture supernates. Extensive degradation of the mucin carbohydrate moieties occurred in all systems, but the degradation of mucin protein was less. During a 48-h incubation, the average percent degradation of mucin carbohydrates and mucin protein were respectively 96% and 57% in eight fecal cultures, 66% and 15% by four fecal culture supernates, and 78% and 43% by three fecal extracts. Degradation of mucin protein was greater during bacterial growth in fecal cultures than during incubation with fecal culture supernates, suggesting that bacterial protease activity was predominantly cell bound. We conclude that, as in the rat, the human enteric microflora degrades mucin carbohydrate moieties extensively and the mucin protein to a lesser extent.

Mucin degradation in human colon ecosystems. Evidence for the existence and role of bacterial subpopulations producing glycosidases as extracellular enzymes.

Recent work indicates that subpopulations of human fecal bacteria, averaging approximately 1% of the total viable fecal flora, degrade the oligosaccharide side chains of hog gastric mucin, which structurally resembles human epithelial mucins. Here we report studies to determine whether degradation of mucin oligosaccharides is related to glycosidase production by bacteria growing in anaerobic fecal cultures. Triplicate cultures containing hog gastric mucin were inoculated with serially diluted feces from each of seven healthy subjects. When the stationary growth phase was attained, mucin oligosaccharide degradation and both cell-bound and extracellular activities of four glycosidases were measured in each culture. Cell-bound beta-d-galactosidase, beta-N-acetylglucosaminidase, and sialidase were present in bacteria growing at all levels of fecal inocula, including 10(-11) g. In contrast, extracellular activities were present in every culture inoculated with 10(-4)-10(-7) g feces, but were diminished or absent in cultures inoculated with 10(-8)-10(-11) g feces. Bacterial autolysis was an unlikely cause of extracellular glycosidase activity, since p-nitrophenyl-alpha-l-fucosidase remained cell bound in cultures at every level of fecal inoculum. Degradation of mucin oligosaccharides was associated with extracellular, but not with cell-bound beta-d-galactosidase, beta-N-acetylglucosaminidase, and sialidase. Among the seven subjects, the estimated most probable numbers (MPN) of fecal bacteria producing extracellular beta-d-galactosidase, beta-N-acetylglucosaminidase, and sialidase ranged from 10(6)-10(10)/g dry fecal wt, were comparable to the MPN of mucin-degrading bacteria, and were significantly smaller than the MPN of total fecal bacteria. We interpret these findings as evidence for the existence of bacterial subpopulations in the normal fecal flora that produce extracellular glycosidases, and that these subpopulations have a major role in degrading the complex oligosaccharides of mucin in the gut lumen.

Degradation of blood group antigens in human colon ecosystems. I. In vitro production of ABH blood group-degrading enzymes by enteric bacteria.

Human feces contain enzymes produced by enteric bacteria that degrade the A, B, and H blood group antigens of gut mucin glycoproteins. We have studied their production in fecal cultures to determine if such cultures can be a source for enzyme purification and to explore how blood group antigen-degrading enzymes are adapted in individual human colon ecosystems. They were present in fecal cultures from each of 27 healthy subjects, including ABH nonsecretors. Heat-sensitive obligate anaerobes are their major source. From 39 to 85% of the total enzyme activity produced by growing cultures was extracellular. Commercial hog gastric mucin and salivary glycoproteins, including Lea saliva which lacks A, B, and H antigens, enhance production of A-, B-, and H-degrading activity in anaerobic fecal cultures irrespective of the glycoprotein's blood group specificity. There is evidence that the host's ABO blood type and secretor status affects the specificity of blood group-degrading enzymes produced by his fecal bacteria in vitro. Thus, fecal inocula from B secretors incubated with hog gastric mucin (A and H specificity) or with Lea saliva produced greater levels of B-degrading than A- or H-degrading activity, and inocula from A secretors in similar media produced greater levels of A-degrading than B- or H-degrading activity. Blood group-degrading enzymes produced in fecal cultures are glycosidases and not proteases. The B-degrading enzyme cleaves the B antigenic determinant alpha-D-galactose from the oligosaccharide side chains of mucin glycoproteins with B specificity. Anaerobic fecal cultures containing blood group substances are a feasible source for purifying blood group antigen-degrading enzymes. Prior adaptation to blood group antigens in the gut mucins of type A and type B secretors affects the specificity of the enzymes produced in vitro.