Mixed Glycosidases Research Articles

Composition and Conformation of Cryptosporidium parvum Oocyst Wall Surface Macromolecules and Their Effect on Adhesion Kinetics of Oocysts on Quartz Surface

We characterized the composition and conformation of Cryptosporidium parvum ( C. parvum ) oocyst wall surface macromolecules and studied their effect on interactions between C. parvum oocyst and quartz surface. Proteinase K and mixed glycosidases were used to modify C. parvum oocyst surface macromolecules. The peptides released by proteinase K and carbohydrates hydrolyzed by mixed glycosidases were respectively analyzed with liquid chromatography/nanoelectrospray ionization tandem mass spectrometry (LC-MS/MS) and phenol-sulfuric acid assay to determine the composition of C. parvum oocyst wall surface macromolecules. Surface potential and polarity of the untreated and proteinase treated C. parvum oocysts revealed information about the conformation of oocyst wall surface macromolecules. The results illustrated that C. parvum oocyst wall is covered by a fluffy layer of glycoproteins. Adhesion kinetics of untreated and proteinase K treated C. parvum oocysts on quartz surface were studied in a radial stagnation point flow cell over a wide range of ionic strength to investigate the effect of C. parvum oocyst wall surface macromolecules on oocysts-quartz interactions. The adhesion rate coefficient of proteinase K treated C. parvum oocysts significantly decreased compared to that of untreated oocysts. This observation indicated that the fluffy layer on C. parvum oocysts wall leads to weaker van der Waals interaction and stronger steric repulsion.

Partial characterization of a non-proteinaceous, low molecular weight antigen of Eimeria tenella.

A low molecular weight (LMW) antigen of Eimeria tenella, initially identified using a murine monoclonal antibody (mAb C(3)4F(1)) raised against E. tenella sporozoites, was partially characterized using enzymatic degradation. solvent extraction, and immunization into various inbred lines of mice. The LMW antigen could be isolated using Folch extraction (methanol/chloroform/ water) and the epitope recognized by mAb C(3)4F(1) was resistant to degradation by alpha-amylase, pronase, and proteinase K, but was sensitive to sodium m-periodate treatment or digestion using mixed glycosidases (from Turbo cornutus). These observations suggest that the antigenic epitope recognized by mAb C(3)4F(1) is carbohydrate-dependent and, based on our ability to isolate the LMW antigen by Folch extraction, the epitope probably resides on a polar glycolipid. The inability of sporozoite-immunized nude mice to elicit a serum antibody response to this molecule indicates that it acts as a T-dependent antigen. Furthermore, sporozoite-immunized male CBA/N mice (with an X-linked immunodeficiency) also failed to elicit a serum antibody response to this molecule, which is consistent with a carbohydrate antigenic epitope. We propose that this antigenic molecule be designated ET-GL1 to reflect its origin and probable structure (E. tenella glycolipid 1).

Some Biochemical Properties of the Components of Staphylococcus aureus Binding to Human Platelets

The binding properties of Staphylococcus aureus in relation to human platelets were investigated. Protease digestion (pronase E, proteinase K, trypsin), heat treatment (80 degrees C, 30 min), and sonication for 5 min significantly reduced the binding abilities of the staphylococcal cells to 0% (p < .01), 50 +/- 5% (p < .05), and 38 +/- 9% (p < .05), respectively, while mixed glycosidases did not. Inhibition experiments indicated that protein A and various sugars were ineffective. A binding study using biotinylated cell surface fractions extracted from the whole cells of S. aureus indicated that the proteins having apparent molecular weights of 14400 and 16500 estimated by sodium dodecylsulfate-polyacrylamide gel electrophoresis were involved in the binding between S. aureus and human platelets.

Sulfhydryl-dependent attachment of Treponema denticola to laminin and other proteins

Attachment of Treponema denticola ATCC 35405 to laminin, a major basement membrane protein, and to other proteins was studied. Microdilution plates were coated with the proteins, and the attachment of T. denticola was measured by the enzyme-linked immunosorbent assay technique. Compared with bovine serum albumin (BSA), T. denticola had a high affinity to laminin, fibronectin, fibrinogen, and gelatin, as well as to type I and type IV collagens. Attachment to RGD peptide (Gly-Arg-Gly-Asp-Ser, the integrin recognition sequence) was only about 30% of that to laminin and was comparable to attachment to BSA. Tests with laminin fragments obtained through elastase digestion showed that the spirochetes attached well to an A-chain 140-kDa fragment involved in eukaryote cell attachment but did not attach to a 50-kDa fragment that includes the heparin binding site. Pretreatment of T. denticola with soluble laminin, fibronectin, gelatin, BSA, or fibrinogen had no effect on the attachment of the bacteria to laminin or fibronectin. A wide variety of compounds were tested for their possible inhibitory actions on the attachment. While most treatments of T. denticola ATCC 35405 had little or no effect on the attachment to proteins, sulfhydryl reagents p-chloromercuribenzoic acid (pCMBA) and oxidized glutathione inhibited the attachment by 70 to 99%, depending on the protein. When T. denticola was first allowed to attach to proteins, addition of pCMBA or oxidized glutathione could no longer reverse the attachment. Heat treatment of the spirochetes also markedly reduced the attachment to laminin, gelatin, and fibrinogen but not to BSA. Mixed glycosidase treatment of the spirochetes inhibited the attachment by 20 to 80%. None of the above treatments of the substrate proteins had any marked effect on the spirochete attachment. The results indicate that T. denticola has the capacity to bind to many different kinds of proteins by utilizing specific attachment mechanisms. The binding appears to involve protein SH groups and/or carbohydrate residues on the surface of T. denticola.

Characterization of bovine rotavirus VP6 and VP7 as glycoproteins using monoclonal antibodies

Bovine rotavirus proteins were analysed by a panel of monoclonal antibodies. Glycosylated epitopes were identified on both inner and outer capsid proteins (VP6 and VP7 respectively). VP7 possessed a periodate insensitive epitope which was, however, sensitive to endoglycosidase H, mixed glycosidases and to protease treatment. This epitope was not detected on viruses grown in the presence of 2-deoxy-D-glucose or tunicamycin. An epitope was detected on VP6 which was sensitive to periodate oxidation. The blotted protein reacted with a glycan assay kit; yet the epitope was not affected by endoglycosidase H and was found on viruses grown in the presence of 2-deoxy-D-glucose or tunicamycin. These results suggest that VP7 and VP6 epitopes are carbohydrate dependent. The VP7 epitope contains an N-linked carbohydrate moiety in contrast to the VP6 epitope which appears to contain O-linked glycosyl units.

Characterization of bovine rotavirus VP6 and VP7 as glycoproteins using monoclonal antibodies

Bovine rotavirus proteins were analysed by a panel of monoclonal antibodies. Glycosylated epitopes were identified on both inner and outer capsid proteins (VP6 and VP7 respectively). VP7 possessed a periodate insensitive epitope which was, however, sensitive to endoglycosidase H, mixed glycosidases and to protease treatment. This epitope was not detected on viruses grown in the presence of 2-deoxy-D-glucose or tunicamycin. An epitope was detected on VP6 which was sensitive to periodate oxidation. The blotted protein reacted with a glycan assay kit; yet the epitope was not affected by endoglycosidase H and was found on viruses grown in the presence of 2-deoxy-D-glucose or tunicamycin. These results suggest that VP7 and VP6 epitopes are carbohydrate dependent. The VP7 epitope contains an N-linked carbohydrate moiety in contrast to the VP6 epitope which appears to contain O-linked glycosyl units.

Characterization and purification of an immunosuppressive factor produced by a small cell lung cancer cell line.

The present study was undertaken to determine whether small cell lung cancer (SCLC) cell lines produce immunosuppressive factors and, if they do, to characterize the factors. The supernatants of SCLG cell lines, H69 and N857, inhibited not only the blastogenic response of human peripheral blood lymphocytes (PBL) to phytohemagglutinin or concanavalin A, but also the cytotoxic activity of lymphokine‐activated killer cells. Neither was inhibited by supernatants from non‐SCLC cell lines PC9, QG56, and A549. The immunosuppressive activity of H69 supernatant was stable upon heating to 56°C for 60 min, but labile when heated to 70°C for 10 min. The activity was abolished after dialysis at pH 2.0 or pH 11.0, but not at pH 4.5 or pH 9.0. Digestion with trypsin or proteinase eliminated the immunosuppressive activity, whereas treatment with neuraminidase, mixed glycosidase, DNase or RNase had no effect, suggesting that the immunosuppressive activity in H69 supernatant is due to a protein factor. This H69‐derived immunosuppressive factor was isolated by ion exchange chromatography using a gradient of 0.04 to 0.08 M NaCl solution. Gel filtration and sodium dodecyl sulfate‐polyacrylamide gel electrophoresis showed the factor to have molecular weights of 98 kD and 102 kD, respectively. These results suggest that SCLC cells produce a potent immunosuppressive factor which may account for the immune deficiency in SCLC patients.

Lymphocyte adhesion to cultured Peyer's patch high endothelial venule cells is mediated by organ-specific homing receptors and can be regulated by cytokines.

Adhesion of lymphocytes to high endothelial venule (HEV) cells is the first step in the migration of these cells from blood into lymph nodes and Peyer's patches (PP). In the present study, we isolated and cultured HEV cells from PP of the rat and assessed their capacity to interact with lymphocytes. Flow cytometric analysis with a rat HEV-specific mAb KJ-4 revealed that greater than 90% of the cultured cells were stained by the antibody. Furthermore, confluent monolayers of PP HEV cells retained the capacity to support the adhesion of lymphocytes from spleen, thoracic duct, and lymph nodes but not binding of immature cells from thymus and bone marrow, which are deficient in cells capable of binding to HEV in vivo. In addition, intraepithelial lymphocytes that preferentially migrated into mucosal lymphoid tissues were also enriched in cells that adhered to the endothelial monolayers. The binding process required energy, was calcium-dependent, and could be inhibited by cytochalasin D, trypsin, and mixed glycosidase. Interestingly, pretreatment of PP HEV cells with rTNF, IFN-gamma, or granulocyte-macrophage CSF significantly increased the endothelial adhesiveness for thoracic duct lymphocytes in a time- and dose-dependent manner. In contrast, stimulation of lymphocytes with phorbol ester or TNF resulted in the rapid modulation of the surface expression of the PP homing receptor and decrease in lymphocyte binding to normal or TNF-stimulated HEV cells. The adhesion of lymphocytes to normal or cytokine-stimulated HEV cells can be blocked by pretreatment of lymphocytes, but not HEV cells, with the PP homing receptor-specific 1B.2.6 antibody. Taken together, these experiments provide strong evidence that the interaction between lymphocytes and cultured HEV cells are mediated by adhesive mechanisms that regulate lymphocyte entry into PP in vivo and that cytokines can promote HEV adhesiveness for lymphocytes through increased expression of organ-specific ligands on HEV cells.

Isolation of A Potent Cholesterol Nucleation–Promoting Activity From Human Gallbladder Bile: Role in the Pathogenesis of Gallstone Disease

Gallbladder bile contains nucleation-promoting activity that binds to concanavalin A. The activity was found in gallbladder bile from cholesterol gallstone patients but also in gallbladder bile from patients without stones and patients with pigment stones. Bile from patients with multiple cholesterol gallstones contained high concanavalin A-binding nucleation-promoting activity. The activity was much lower in bile samples from pigment stone patients, patients without stones and patients with a solitary cholesterol stone. Serum contained very little activity and no concanavalin A-binding nucleation-promoting activity could be demonstrated in gallbladder mucosa. This suggests that concanavalin A-binding nucleation promoter is produced in the liver or bile duct epithelium. The activity was fully resistant to digestion with pronase but was heat labile and could be destroyed by prolonged incubation with a mixed glycosidase preparation indicating that sugar residues are important for this activity. On a Superose 12 gel permeation column, promoting activity eluted in two major peaks at apparent molecular weights of 150 +/- 30 kD (n = 5) and less than 5 kD respectively. The mobility on the column was not influenced by pronase digestion. The factor with the higher molecular weight could be isolated further by polyacrylamide gel electrophoresis under nondenaturing conditions. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the apparent molecular weight of the glycoprotein was 130 kD. In conclusion, gallbladder bile contains nucleation-promoting activity that binds to concanavalin A. The activity is increased in bile from patients with multiple cholesterol gallstones and could therefore play an important role in the pathogenesis of gallstone disease.

Incorporation of UDP-[14C]Glucose into Xyloglucan by Pea Membranes

The water-insoluble 1,4-beta-linked products formed from UDP-[(14)C]glucose by pea membranes were dissolved in hot dimethyl-sulfoxide/paraformaldehyde and fractionated on columns of controlled pore glass beads calibrated with dextran standards. The products eluted with a peak size close to 70 kilodaltons in dextran equivalents. Similar elution profiles were obtained for products formed in brief or extended incubations and at high or low substrate concentrations. Methylation analysis indicated that only a few [(14)C]glucose units had been added to an endogenous acceptor to form this product. In the presence of UDP-xylose at concentrations equal to or less than UDP-[(14)C]glucose, incorporation from the latter was enhanced and the products elongated with time to a size range where the major components eluted between dextran 264 and 500 kilodaltons. Treatment with endo-1,4-beta-glucanase resulted in a mixture of oligosaccharides, including the xyloglucan subunit Glc(4)Xyl(3), which were hydrolyzed further by mixed glycosidases to labeled glucose and isoprimeverose (xylosyl-1,6-alpha-d-glucose). In pulse-chase experiments, the low molecular weight product formed from UDP-[(14)C]glucose alone was clearly a precursor for high molecular weight products formed subsequently in the presence of both UDP-glucose and UDP-xylose. It is concluded that the 1,4-beta-transglucosylation activity detected in these tests was due to an enzyme that is required for biosynthesis of the backbone of xyloglucan.

The activation of Dictyostelium discoideum alkaline phosphatase by carbohydrate binding proteins

Dictyostelium discoideum vegetative cell alkaline phosphatase was rapidly inactivated in the presence of 50 mM Tris–HC1, pH 7.4, when the incubation temperature was raised to 50 °C. In contrast, when extracts were preincubated in the presence of concanavalin A, alkaline phosphatase was rapidly activated when the temperature was raised to 50 °C. The alkaline phosphatase of culmination phase cell extracts was also inactivated at 50 °C, and although preincubation in the presence of concanavalin A reduced this inactivation, stimulation of activity was not observed. A mixed glycosidase preparation also activated the vegetative cell alkaline phosphatase and the binding of the glycosidases rather than their catalytic activity appeared to be the important event. In contrast, alkaline phosphatase from culmination phase cells was only marginally stimulated by glycosidase treatment. These results are discussed in relation to the unmasking of alkaline phosphatase activity that occurs during D. discoideum differentiation.Key words: concanavalin A, glycosidase, differentiation, Dictyostelium discoideum, alkaline phosphatase.

Binding of the delta endotoxin from Bacillus thuringiensis to brush-border membrane vesicles of the cabbage butterfly (Pieris brassicae).

The insecticidal delta endotoxin of Bacillus thuringiensis was labeled with iodine-125. Brush-border membrane vesicles, prepared from the midgut epithelium of Pieris brassicae larvae, known to be highly susceptible to the toxin, and from a non-target tissue: the small intestine of rat, were examined for binding of 125I-toxin. The toxin was bound specifically only to insect vesicles. Its binding to the insect membrane system was competitively inhibited by 127I-toxin and non-iodinated toxin, whereas the binding of the 125I-toxin to the mammalian membrane system was not affected by unlabeled toxin. Vesicles of P. brassicae possess two individual binding-site populations for iodinated toxin with dissociation constants of 46 nM and 490 nM. The Hill coefficients of both sites were approximately 1 and the binding capacities were 0.2 pmol and 30 pmol/mg vesicle protein for the high and the low-affinity sites respectively. The estimation of the dissociation constant for non-iodinated toxin, using a competition experiment, revealed only one binding-site population which possessed a dissociation constant of 235 nM. It is concluded that this is the binding site for the native toxin. This site was sensitive towards treatment with proteases or mixed glycosidases. It is suggested that it is a protein or a glycoprotein.

Bacterial adhesion to intravenous cannulae: influence of implantation in the rabbit and of enzyme treatments

Comparison was made of the adhesion of Staphylococcus epidermidis, Staphylococcus aureus, Escherichia coli, Klebsiella aerogenes and Pseudomonas aeruginosa to six types of intravascular cannula material. Adhesion to materials removed from rabbit tissues did not differ significantly between types of material or between bacterial species. In contrast, major differences were found when unimplanted materials were examined; the overall rank order of adhesiveness of bacteria to unimplanted materials (S. epidermidis greater than P. aeruginosa greater than S. aureus much greater than K. aerogenes greater than E. coli) was highly significant (F = 13.0, P less than 0.0005), and although no single material was consistently least attractive to all micro-organisms, FEP-Teflon and PTFE-Teflon showed significantly lower overall affinity for bacteria than other materials (P less than 0.001); all species showed a significant preference for a silicone polymer (P less than 0.0005). The nature of the bacterial surface structures responsible for adhesion were investigated by the actions of pronase and mixed glycosidase, which produced significant respective decreases and increases in adhesion of staphylococci to unimplanted materials; their effects on the Gram-negative bacilli were less consistent.

Characterization of the Cryptosporidium antigens from sporulated oocysts of Cryptosporidium parvum.

The antigenic constituents of sporulated Cryptosporidium parvum oocyst antigens were characterized with antisera from mice immunized against C. parvum. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by silver staining defined the major proteins. Six of seven lectins used recognized as many as 15 bands. The lectins concanavalin A, Dolichos biflorus, and wheat germ agglutinin showed strong activity against the same eight bands with molecular weights ranging from 72,000 to greater than 100,000. An enzyme-linked immunosorbent assay was used to detect antibody to C. parvum. Antibody binding was significantly decreased by heat and enzymatic treatment with trypsin, protease, and mixed glycosidases. C. parvum antigens were further defined by the reactivity of immune sera with a C. parvum sonicate preparation separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose paper. Antisera from orally infected mice consistently recognized four antigens with molecular weights ranging from 72,000 to greater than 100,000. These antigens also bound concanavalin A. Treatment of the antigen preparation with mixed glycosidases reduced the reactivity of antisera with most antigens with molecular weights greater than 60,000. The data suggest that the antigenic composition of C. parvum is complex and that carbohydrates alone or in association with lipids or proteins may be important in the immune response to C. parvum.

Human T-cell leukaemia-associated cell surface glycoprotein gp37: Studies with three monoclonal antibodies and a rabbit antiserum

Three monoclonal antibodies (mAbs), termed SN2, SN2a and SN2b, were used in the present work to study a human T-cell leukemia-associated cell surface glycoprotein, GP37. Strong specificity of mAbs SN2, SN2a and SN2b for T leukemia cells was demonstrated by radioimmunoassay and fluorescence-activated cell sorter (FACS) analysis. GP37 was not detected on normal human peripheral blood lymphocytes, purified normal T-cells, normal thymocytes nor normal bone marrow cells. Furthermore, GP37 was barely detectable on phytohemagglutinin (PHA)- and Concanavalin A (Con A)-activated T-cells. The results indicate clinical utility of these mAbs. Competitive binding experiments show that the epitopes recognized by SN2 and SN2a are sufficiently close to each other to allow complete reciprocal inhibition of binding whereas the epitopes recognized by SN2 and SN2b are less close to allow only partial reciprocal binding inhibition. The biochemical nature of antigenic determinants defined by these mAbs was studied by treating T leukemia cells with trypsin, chymotrypsin, thermolysin, neuraminidase and mixed glycosidases. The results suggest that the antigenic determinants defined by these mAbs all consist of the protein moiety of the glycoprotein GP37. No significant antigenic modulation was observed when T leukemia cells were reacted with SN2. In a sequential immunoprecipitation experiment, a 125I-labeled leukemia antigen preparation was first treated with a rabbit anti-T leukemia antiserum. The latter had been prepared by immunizing a rabbit with a partially purified human T leukemia antigen preparation and showed a good specificity for T leukemia cells. Subsequent treatment of the labeled antigen preparation with SN2 showed that SN2 antigen had been precleared. Thus, both mouse mAb SN2 and the rabbit anti-T leukemia antiserum react with the same GP37 molecule.

Use of mixed glycosidase for the removal of mucus from the rat nasal epithelium in SEM studies

AbstractThe surface layer of mucus, which obscures the epithelium from view, is a major obstacle when performing scanning electron microscopic studies of the nasal mucosa. Samples treated with a 1% mixed glycosidase solution for 1 or 2 minutes, followed by agitation, removed most of the mucus from the conchae without damaging the underlying epithelium. Removal of mucus allows the complete evaluation of the underlying epithelium in normal animals and the localization and characterization of lesions in animals exposed to nasal toxicants.

Mixed glycosidase pretreatment reduces nonspecific binding of antibodies to frozen tissue sections.

Indirect immunohistochemical studies of frozen mouse tissues with mouse monoclonal antibodies yield, in general, suboptimal results primarily because of indiscriminate binding of secondary antibody to all mouse immunoglobulins, i.e., to the monoclonal reagent and to endogenous immunoglobulin nonspecifically trapped in the tissue. To reduce this nonspecific staining, frozen sections of mouse kidney were treated enzymatically. Optimal results were obtained following a 2 hr treatment with 20 mg/ml of mixed glycosidases (MG). This treatment reduced the nonspecific background staining of the interstitial spaces and blood vessels, but did not affect the reactivity of structurally bound immunoglobulin G (IgG) in the glomeruli or alter the reactivity of mouse renal tissue to the monoclonal antibody that recognizes an oligosaccharide antigenic determinant (SSEA-1). Eluates from enzyme-treated frozen tissue sections contained normally immunoreactive IgG in the form of dimers. These data indicate that MG treatment of frozen sections could be safely used to reduce the content of nonstructurally bound immunoglobulins in frozen tissues and thus improve the visualization of specific monoclonal antibody binding.

Monoclonal antibodies to carcinoembryonic antigen: a systematic analysis of antibody specificities by using related normal antigens and evidence for allotypic determinants on carcinoembryonic antigen.

Eleven hybridoma clones producing monoclonal antibodies to carcinoembryonic antigen (CEA) were prepared by fusions of the mouse myeloma cell line P3-X63-Ag8-U1 with splenocytes of BALB/c mice immunized with highly purified CEA. Their specificities were systematically analyzed by radioimmunoassay for reactivity with CEA and four related normal antigens: nonspecific cross-reacting antigen (NCA), nonspecific cross-reacting antigen-2 (NCA-2), normal fecal antigen-1 (NFA-1), and normal fecal antigen-2 (NFA-2). Antibody-antigen profiles revealed four groups of hybridomas. The Group I antibodies from two clones reacted with sites shared among CEA, NCA, NCA-2, and NFA-2. The Group II antibody from one clone recognized an epitope common to CEA, NCA-2, and NFA-2. The Group III antibodies from five clones reacted with sites commonly residing on CEA, NCA-2, NFA-1, and NFA-2, whereas antibodies of Group IV from three clones bound only CEA. Mutual inhibition assays between antibodies in the respective groups for CEA binding further revealed that at least eight different epitopes can be identified on the CEA molecule: two in Group I, one in Group II, three in Group III, and two in Group IV, respectively. Three monoclonal antibodies in Group IV appeared to be specific for CEA, but they did not react with any of five purified CEA preparations other than that used for immunization. In 13 sera from 50 various cancer patients with elevated CEA levels, however, we detected the CEA molecule reactive with these antibodies, indicating that the epitopes recognized by the Group IV antibodies may be of allotypic characters. These epitopes were unrelated to main blood group antigens and did not show any organ specificity. All of the eight epitopes identified in this study were found to be resistant to neuraminidase, mixed glycosidases, and Smith degradation (SI-stage) as determined by a competitive radioimmunoassay, suggesting that they reside on the protein moiety of the CEA molecule. Six epitopes were sensitive to reduction and alkylation, but two other epitopes (one in Group II and the other in Group III) were resistant to these treatments. This result indicates that the latter two epitopes are conformation-independent.

The role of glycosylation on haemagglutination and immunological reactivity of rubella virus.

Treatment of purified rubella virus with mixed glycosidases resulted in loss of haemagglutinating (HA) activity but the capacity to combine with haemagglutination inhibiting (HI) and other antibodies was retained. Electrophoretic analysis revealed a greater effect on VPI than VPII.

Polymorphism of Human Thyroxine-Binding Globulin

T4-binding globulin (TBG) consists of a single polypeptide chain containing 4 oligosaccharide units with an average of 10 terminal sialic acid residues. Isoelectric focusing (IEF) of TBG purified from pooled human plasma showed 4 main bands isoelectric at pH 4.2, 4.3, 4.5, and 4.6. Removal of 85% of carbohydrates by treatment with mixed glycosidases reduced these bands to 2, at pH 5.4 and 5.7. This residual microheterogeneity was not related to protein-ampholyte complexes, since it was still present in 8 M urea. It also did not represent an equilibrium mixture of interchangeable conformations, since each band obtained in the first dimension IEF gave a single spot when rerun in a second dimension. Therefore, the residual microheterogeneity of TBG after removal of carbohydrates can be attributed to variation in amino acid composition. Since genetic polymorphism of TBG was recently demonstrated, we further investigated whether the residual microheterogeneity was genetically determined. Plasma samples from 20 white donors and 17 black donors were labeled with [125I]T4 and submitted to IEF, followed by autoradiography. TBG-1, found in white donors and most black donors, showed the same 4 bands as TBG purified from pooled plasma. Two less frequent phenotypes were found in black individuals: TBG-2, with 4 bands at approximately pH 4.25, 4.45, 4.55, and 4.7; and TBG-1,2, having all of the bands present in TBG-1 and TBG-2. Electrophoretically homogeneous preparations of TBG of each type were obtained from 100 ml plasma; after deglycosylation, TBG-1 revealed 2 bands isoelectric at pH 5.4 and 5.7, TBG-2 had 2 bands at pH 5.7 and 5.9, and TBG-1,2 had 3 bands at pH 5.4, 5.7, and 5.9. The same TBG bands were found after neuraminidase treatment of whole plasma from the same donors. These data demonstrate two kinds of TBG polymorphism. The first is found in deglycosylated TBG from individual donors and is probably due to variation in amino acid composition. The second, also unrelated to the carbohydrate moiety, is a genetic polymorphism found in blacks.