| Glycobiology | Pages |
Differential recognition by proteins of [alpha]-galactosyl residues on endothelial cell surfaces
Introduction
Results
Discussion
Materials and methods
Acknowledgments
Abbreviations
References
Differential recognition by proteins of [alpha]-galactosyl residues on endothelial cell surfaces
The binding of proteins to cell surface carbohydrates contributes to cell-cell interactions in development, immunity, and various physiologic processes. Such interactions are presumably dictated not only by the chemical structure of the carbohydrate but also reflect the properties of the protein and the microenvironment in which the protein-carbohydrate interaction occurs. To explore the factors influencing the recognition of cell surface carbohydrates by proteins, the extent to which three classes of proteins-anti-Gal[alpha]1-3Gal IgM found in higher primates, Griffonia simplicifolia type I lectin, isolectin B4 (GS-IB4), and [alpha]-galactosidase-interact with Gal[alpha]1-3Gal on porcine cell surfaces was tested. Although the Gal[alpha]1-3Gal residues expressed on porcine endothelial cells and recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 were both sensitive to cleavage by [alpha]-galactosidase, the sites recognized by GS-IB4 were more sensitive to cleavage than sites recognized by anti-Gal[alpha]1-3Gal IgM. Cross-blocking studies on porcine cell surfaces revealed that a significant proportion of anti-Gal[alpha]1-3Gal IgM bound to sites not recognized by GS-IB4; however, anti-Gal[alpha]1-3Gal IgM and GS-IB4 recognized the same sites on solubilized membrane proteins and model compounds. These results suggest that features of the cell surface such as the three-dimensional arrangement of Gal[alpha]1-3Gal and characteristics of the protein such as size and valency play critical roles in specific interactions on cell surfaces.
Introduction
Cell adhesion molecules, lectins, glycosidases, and some antibodies interact specifically with complex carbohydrates expressed on the surface of cells in that each of these classes of proteins can recognize one carbohydrate but not another. Thus, the quantity on the cell surface of such molecules as phosphatidyl choline, stage-specific embryonic antigens, and the Lewis x saccharide are often determined by measuring the binding of antibodies or lectins (Linder, 1969; Platt et al., 1983; Mercolino et al., 1989; Snyder et al., 1994). Similarly, the presence of many blood group antigens which contain carbohydrate determinants, including A, B, H, N, T, Tk, Th, Tx, Tn, Cad, and Leb has been evaluated based on the binding of lectins (Bird, 1952; Etzler and Kabat, 1970; Ottensooser and Silberschmidt, 1953; Boyd et al., 1958; Bird, 1964; Howard, 1979; Hanfland and Graham, 1981; Springer, 1984; Blanchard et al., 1985; Delbaere et al., 1990). While the binding of carbohydrate-specific proteins to cell surface carbohydrates may be highly specific, it is sensitive to factors other than the specificity of the protein-carbohydrate interaction. For example, enzymes such as [alpha]-fucosidase, [alpha]-glucosidase, glycopeptidase F, and endoglycosidase H interact efficiently with carbohydrate residues only when they are removed from the cell surface or cleaved from the protein to which the carbohydrates are attached. Thus, the expression of an appropriate carbohydrate on the outer leaflet of the cell membrane does not by itself provide a sufficient basis to allow protein-carbohydrate interactions to occur. This limitation is generally recognized for the action of enzymes. However, with notable exceptions (Bird, 1978), the influence of factors other than specificity is not generally considered to contribute to the binding of lectin or antibody.
One cell surface carbohydrate of biological importance is Gal[alpha]1-3Gal. Gal[alpha]1-3Gal is expressed on the cells of all mammalian species except for humans, apes and Old World monkeys (Galili et al., 1987b). The higher primates such as man which do not express Gal[alpha]1-3Gal have naturally occurring antibodies specific for that structure. We have recently shown that antibodies against Gal[alpha]1-3Gal are singularly responsible for triggering the hyperacute rejection of porcine organ xenografts (Collins et al., 1995; Lin et al., 1997). 'Xenoreactive" natural antibodies against Gal[alpha]1-3Gal consist of a population of antibodies which is homogeneous with respect to specificity (Parker et al., 1996a) and functional avidity (Parker et al., 1994). The endothelial cell surface molecules which bear Gal[alpha]1-3Gal modifications consist primarily of members of the integrin family ([alpha]1,[alpha]v, [alpha]3/[alpha]5, [beta]1, and [beta]3) (Holzknecht and Platt, 1995; Lin et al., 1996), suggesting that the binding of antibodies to the cell surface might affect the physiology of those cells, and studying the interactions of proteins with these saccharides might yield information regarding the biological properties of the integrin cores.
The expression of Gal[alpha]1-3Gal and the amount of Gal[alpha]1-3Gal available for the binding of xenoreactive antibodies is generally determined by analyzing the binding of the Griffonia simplicifolia type I lectin, isolectin B4 (GS-IB4) (Thall and Galili, 1990; Galili et al., 1987b; Collins et al., 1994), which recognizes [alpha]-galactosyl moieties (Hayes and Goldstein, 1974; Wood et al., 1979). The binding of GS-IB4 accurately reflects the density of Gal[alpha]1-3Gal moieties on isolated glycoproteins (Thall and Galili, 1990). However, the number of sites on a cell recognized by GS-IB4 may not correspond to the number of sites recognized by anti-Gal[alpha]1-3Gal IgM (Alvarado et al., 1995), suggesting that there may be differences between the recognition of Gal[alpha]1-3Gal on a cell surface by anti-Gal[alpha]1-3Gal antibodies and recognition of the same carbohydrate by GS-IB4. The mechanisms underlying the differential binding of these two proteins have not been reported, yet understanding these differences could provide insight into the manner in which the interaction of proteins with cell surface carbohydrates occurs.
We tested the extent to which anti-Gal[alpha]1-3Gal IgM and GS-IB4 differ in the recognition of Gal[alpha]1-3Gal residues on porcine endothelial cell surfaces and on isolated porcine endothelial cell integrins. Our studies revealed that a significant fraction of Gal[alpha]1-3Gal residues bound by anti-Gal[alpha]1-3Gal IgM on a cell surface are not recognized by GS-IB4 and that some residues bound by GS-IB4 are not bound by anti-Gal[alpha]1-3Gal IgM. The results also suggest that residues which are relatively inaccessible can be bound by anti-Gal[alpha]1-3Gal IgM but not by GS-IB4 because of the smaller size of the binding domain of IgM (VC + VH domains; 25 kDa; Lin and Putnam, 1978) compared to the relatively globular GS-IB4 (tetrameric protein; 114 kDa) (Hayes and Goldstein, 1974; Wood et al., 1979) and the ability of IgM to bind more disperse residues than GS-IB4 because of the larger size and valency of the IgM molecule. On the other hand, the smaller size of GS-IB4 may allow the lectin to bind some residues on the cell surface which anti-Gal[alpha]1-3Gal IgM can not. The present findings indicate that steric factors involving rather long range interactions ([sim]25 Å or more) may dictate to a large extent the interactions of proteins with cell surface carbohydrates.
Results
Cleavage of [alpha]-galactosyl residues from the porcine cell surface by [alpha]-galactosidase
The fraction of sites on porcine endothelial cell membranes recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 which were sensitive to cleavage by [alpha]-galactosidase was determined. The binding of IgM and GS-IB4 to glutaraldehyde-fixed porcine cells was measured following treatment of the cells with [alpha]-galactosidase or with control buffer. As shown in Figure 1, following treatment of endothelial cells with 2 U/ml of [alpha]-galactosidase at 37°C, binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 decreased by greater than 95%, indicating that nearly all sites recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 are sensitive to cleavage by [alpha]-galactosidase.
Figure 1. Differential sensitivity to cleavage by [alpha]-galactosidase of sites on porcine cell surfaces recognized by anti-Gal[alpha]1-3Gal-IgM and sites recognized by GS-IB4. Porcine aortic endothelial cells were treated with various concentrations of [alpha]-galactosidase at 37°C for 30 min, and then the binding of anti-Gal[alpha]1-3Gal IgM or GS-IB4 to porcine cells was measured by enzyme linked assays. Samples were run in duplicate and the standard error is shown. The results of one representative experiment are shown. Similar results were obtained using IgM from two different individuals and with cultured aortic endothelial cells from four different pigs.
To determine whether anti-Gal[alpha]1-3Gal IgM and GS-IB4 attach to the same sites on endothelial cell surfaces, the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 to porcine endothelial cells treated with various concentrations of [alpha]-galactosidase was evaluated. Under conditions in which relatively low concentrations (0.05-0.2 U/ml) of [alpha]-galactosidase were used, the binding of IgM decreased to a lesser extent than the binding of GS-IB4 (Figure 1).
Time course of cleavage of [alpha]-galactosyl residues from the porcine cell surface by [alpha]-galactosidase
One potential explanation for the relative resistance of anti-Gal[alpha]1-3Gal IgM binding sites to [alpha]-galactosidase compared to GS-IB4 binding sites is that the binding of IgM might be limited sterically so that many potential binding sites are not utilized and therefore the cleavage of some Gal[alpha]1-3Gal residues on the cell surface occurs before an appreciable decrease in the binding of anti-Gal[alpha]1-3Gal IgM is observed. This possibility was tested by evaluating the sensitivity of the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 to cells treated with [alpha]-galactosidase for various periods of time at 25°C. As Figure 2 shows, treatment of cells with [alpha]-galactosidase for less than 2% of the time required for complete digestion resulted in a measurable decrease in the binding of anti-Gal[alpha]1-3Gal IgM. This result indicates that the greater resistance of IgM versus GS-IB4 binding to [alpha]-galactosidase is not due to a relative excess of anti-Gal[alpha]1-3Gal IgM binding sites, but rather may reflect some qualitative difference in the sites recognized by anti-Gal[alpha]1-3Gal IgM versus GS-IB4 or in the microenvironment of those sites.
Figure 2. Time course of digestion by [alpha]-galactosidase of Gal[alpha]1-3Gal residues recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on porcine endothelial cell surfaces. The number of sites recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 was determined by measuring the binding of saturating amounts of the antibody and the lectin to porcine cells in an enzyme linked assay. Samples were run in duplicate and the standard error is shown. A decrease in the fraction of sites recognized by anti-Gal[alpha]1-3Gal IgM and by GS-IB4 seen immediately following the addition of [alpha]-galactosidase indicates that the number of potential binding sites, or the density of Gal[alpha]1-3Gal residues, on the cell surface is not in excess. The presence of excess Gal[alpha]1-3Gal residues would have allowed no measurable decrease in the binding of IgM or lectin at early time points of digestion.
Table I
| Target surface | Density of protein binding | Density of Gal[alpha]1-3Gal on target surfacea(residues/cm2) | Gal[alpha]1-3Gal residues utilized | ||
| IgM (molecules/cm2) |
GS-IB4 (molecules/cm2) |
IgMb (residues/cm2) |
GS-IB4b (residues/cm2) |
||
| Porcine endothelial cells | 5.9 × 1011 | 1.9 × 1012 | 4.4 × 1013 | 2.7 × 1012 [6%]c | 4.8 × 1012 [11%]c |
| Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA on hydrophobic surface |
6.9 × 1011 | 1.2 × 1012 | 2.5 × 1013 | 3.1 × 1012 [12%]c | 3.0 × 1012 [12%]c |
| Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA on hydrophobic surface |
4.9 × 1011 | 8.1 × 1011 | 3.6 × 1013 | 2.2 × 1012 [6%]c | 2.0 × 1012 [6%]c |
bCalculated based on a valency of 4.5 for IgM and 2.5 for GS-IB4. See Figure 3 and Table II.
Partial cleavage of Gal[alpha]1-3Gal residues on the surface of porcine cells
To test whether some [alpha]-galactosyl residues might be recognized by anti-Gal[alpha]1-3Gal IgM but not by GS-IB4, conditions that would allow [alpha]-galactosidase to selectively digest the sites to which GS-IB4 bound were sought. GS-IB4 binding sites were far more sensitive to digestion than IgM binding sites when treatment with [alpha]-galactosidase was carried out under limiting conditions. For example, treatment of cells with 200 mU/ml of [alpha]-galactosidase for 120 min at 25°C eliminated all sites recognized by GS-IB4 but left 33% of the binding sites recognized by anti-Gal[alpha]1-3Gal IgM (Figure 2). More extensive cleavage of Gal[alpha]1-3Gal using 20 U/ml [alpha]-galactosidase at 25°C eliminated 100% of the binding of anti-Gal[alpha]1-3Gal IgM and 100% of the binding of GS-IB4, indicating that [alpha]-galactosidase could reach all sites recognized by anti-Gal[alpha]1-3Gal IgM and by GS-IB4 under the conditions used. These results demonstrate that some Gal[alpha]1-3Gal residues which can be recognized by anti-Gal[alpha]1-3Gal IgM cannot be recognized by GS-IB4 on the cell surface.
To determine which properties of anti-Gal[alpha]1-3Gal IgM and GS-IB4 might underlie the differential recognition of Gal[alpha]1-3Gal on endothelial cell surfaces, the relative importance of molecular mass, valency and shape in determining the interaction of proteins with Gal[alpha]1-3Gal on endothelial cell surfaces was tested. The binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 were compared with the binding of anti-Gal[alpha]1-3Gal IgG, which has size and valency similar to GS-IB4 and binding domains (VH + VL; Fab) similar to IgM. After partial digestion of endothelial cells with [alpha]-galactosidase, the binding of anti-Gal[alpha]1-3Gal IgG and GS-IB4 decreased more than the binding of anti-Gal[alpha]1-3Gal IgM. For example, after treatment of cells with 60 mU/ml of [alpha]-galactosidase for 2 h at 37°C, the binding of GS-IB4 and anti-Gal[alpha]1-3Gal IgG decreased by 83% but the binding of anti-Gal[alpha]1-3Gal IgM decreased by only 40% (data not shown). This result is consistent with the idea that, because of the larger size and valency, IgM engages in multivalent interactions under conditions where smaller molecules such as GS-IB4 and anti-Gal[alpha]1-3Gal IgG cannot.
The number of binding sites utilized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on porcine cell surfaces
Another potential explanation for the apparent resistance of anti-Gal[alpha]1-3Gal IgM binding sites to [alpha]-galactosidase is that Gal[alpha]1-3Gal residues utilized by anti-Gal[alpha]1-3Gal IgM represent only a small subset of Gal[alpha]1-3Gal residues utilized by GS-IB4 and this subset is relatively insensitive to [alpha]-galactosidase digestion. That the residues utilized by anti-Gal[alpha]1-3Gal IgM are a small subset of residues utilized by GS-IB4 was tested by comparing the density of sites recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4. Under saturating conditions, [sim]5.9 × 1011 anti-Gal[alpha]1-3Gal IgM and [sim]1.9 × 1012 GS-IB4 bound per cm2 of porcine endothelial cell surface (Table I). Assuming that the valency of binding is the value of n which results in the best linear relationship between the initial relative rate of binding to a cell surface versus [S]n, where [S] is the relative total number of binding sites, anti-Gal[alpha]1-3Gal IgM bound with a valency of 4-5 whereas GS-IB4 bound with a valency of 2-3 (Figure 3). Using a valency of 4.5 for anti-Gal[alpha]1-3Gal IgM and a valency of 2.5 for GS-IB4, IgM bound 2.7 × 1012 Gal[alpha]1-3Gal residues/cm2 and GS-IB4 bound 4.8 × 1012 Gal[alpha]1-3Gal residues/cm2 on the endothelial cell surface (Table I). Thus, the idea that anti-Gal[alpha]1-3Gal IgM recognizes a small fraction of GS-IB4 binding sites is not tenable. However, this finding does not rule out the possibility that GS-IB4 binding sites are more accessible to [alpha]-galactosidase than anti-Gal[alpha]1-3Gal IgM binding sites. The finding that GS-IB4 bound to more Gal[alpha]1-3Gal residues than anti-Gal[alpha]1-3Gal IgM (Table 1) suggests that GS-IB4 may recognize some Gal[alpha]1-3Gal residues on the porcine cell surface not recognized by anti-Gal[alpha]1-3Gal IgM. Moreover, since the total number of galactose residues released by [alpha]-galactosidase far exceeded the number of Gal[alpha]1-3Gal residues recognized by either anti-Gal[alpha]1-3Gal IgM or GS-IB4 (Table I), [alpha]-galactosidase may reach Gal[alpha]1-3Gal residues not accessible to IgM or GS-IB4 or the requirements for cleavage by the enzyme may be less stringent than those for effective binding.
Figure 3. The valency of interaction between anti-Gal[alpha]1-3Gal IgM and GS-IB4 and porcine endothelial cells. The initial rate of binding was determined by measuring the binding of 1.3 µg/ml of antibody or 5.0 µg/ml of lectin after incubation with porcine cells for 10 min at 4°C. The rate of binding was then plotted against Sn (shown in the insets for n = 1), where S is the relative total number of binding sites determined by the amount of binding resulting from incubation of the cells with saturating amounts of anti-Gal[alpha]1-3Gal IgM or GS-IB4 and n is a positive real number greater than or equal to 1.0. The valency of binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 was taken to be the value of n which resulted in the best fit of the data (rate = k1 * Sn) to a line as judged by the highest value of the coefficient of determination (r2).
Cross-blocking by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on cultured porcine endothelial cells
The extent to which the binding sites utilized by anti-Gal[alpha]1-3Gal IgM might be the same as those utilized by GS-IB4 was determined by cross-blocking experiments. Anti-Gal[alpha]1-3Gal IgM and GS-IB4 were applied sequentially to endothelial cells under saturating conditions and binding of the second protein was measured. Saturating conditions were established by demonstrating that application of higher concentrations or sequential application of each did not result in additional binding (not shown). The anti-Gal[alpha]1-3Gal IgM in three human sera blocked 77%, 83%, and 95% of GS-IB4 binding (Figure 4A) and GS-IB4 blocked 45%, 53%, and 58% of anti-Gal[alpha]1-3Gal IgM binding to porcine cell surfaces (Figure 4B). Under the conditions used in this study, IgM (Parker et al., 1997) and GS-IB4 (data not shown) were irreversibly bound. Furthermore, there was no evidence of displacement of one protein by another, since the amount of each protein bound to the cell surface did not change during incubation with a second protein (Figure 5). These results suggest that although [ge]77% of the [alpha]-galactosyl sites recognized by GS-IB4 are also recognized by anti-Gal[alpha]1-3Gal IgM, only about half of the [alpha]-galactosyl sites recognized by anti-Gal[alpha]1-3Gal IgM are also recognized by GS-IB4. Although fixed cells were used to analyze specific binding to Gal[alpha]1-3Gal, the cross-blocking of GS-IB4 and xenoreactive IgM on unfixed porcine endothelial cells was indistinguishable from the cross-blocking on fixed cells.
Figure 4. Overlap of binding sites recognized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on cultured porcine endothelial cells. Porcine endothelial cells were blocked with saturating concentrations of (A) anti-Gal[alpha]1-3Gal IgM or (B) GS-IB4. Following the blocking step, blocked and unblocked cells were incubated with biotinylated GS-IB4 (A) or with human serum as a source of anti-Gal[alpha]1-3Gal IgM (B), and the binding to endothelial cells was measured by an enzyme linked assay. Samples were run in quadruplicate and the standard error is shown. (A) Anti-Gal[alpha]1-3Gal IgM blocked the binding of all but 15% of GS-IB4 (17%, 23%, and 5% in three sera tested) to porcine endothelial cell surfaces. (B) GS-IB4 blocked about one-half of the binding of IgM, with 42%, 55%, and 47% of binding remaining in three different sera tested. These data provide further evidence that a large fraction of [alpha]-galactosyl sites recognized by IgM are not recognized by GS-IB4 and suggest that some [alpha]-galactosyl sites recognized by GS-IB4 may not be recognized by IgM.
Figure 5. (A) Fraction of anti-Gal[alpha]1-3Gal IgM remaining after sequential incubation with anti-Gal[alpha]1-3Gal IgM and GS-IB4. Saturating amounts of anti-Gal[alpha]1-3Gal IgM and GS-IB4 were applied sequentially to cultured porcine endothelial cells. The amount of anti-Gal[alpha]1-3Gal IgM remained bound to the cells after the incubation with GS-IB4 (IgM/GS-IB4) was measured by ELISA and compared to the amount remained bound after sequential application of anti-Gal[alpha]1-3Gal IgM and PBS (IgM/PBS). The result indicates that there was no displacement of bound anti-Gal[alpha]1-3Gal IgM by GS-IB4 on cultured porcine endothelial cells. (B) Fraction of GS-IB4 remaining after sequential incubation with GS-IB4 and anti-Gal[alpha]1-3Gal IgM. Saturating amounts of GS-IB4 and anti-Gal[alpha]1-3Gal IgM were applied sequentially to cultured porcine endothelial cells. The amount of GS-IB4 remained bound to the cells after incubation with human serum as a source of anti-Gal[alpha]1-3Gal IgM (GS-IB4/IgM) was measured by enzyme-linked lectin assay and compared to the amount remained bound after sequential application of GS-IB4 and PBS (GS-IB4/PBS). The result indicates that there was no displacement of bound GS-IB4 by anti-Gal[alpha]1-3Gal IgM on cultured porcine endothelial cells. All samples were run in quadruplicate, and the standard error is shown.
Based on the findings that anti-Gal[alpha]1-3Gal IgM recognizes 2.7 × 1012 Gal[alpha]1-3Gal residues per cm2 of endothelial cells (Table I) and that half of the anti-Gal[alpha]1-3Gal IgM binding sites cannot be blocked by GS-IB4 (Figure 4B), [sim]1.3 × 1012 Gal[alpha]1-3Gal residues per cm2 of cultured endothelial cells are recognized by anti-Gal[alpha]1-3Gal IgM but not by GS-IB4. The density of residues recognized by anti-Gal[alpha]1-3Gal IgM but not by GS-IB4 was verified in experiments involving limited digestion of [alpha]-galactosyl residues (Figure 2), where the difference between the density of Gal[alpha]1-3Gal residues occupied by 60% of binding of IgM (1.62 × 1012 residues per cm2) and the density occupied by 10% of binding of GS-IB4 (0.48 × 1012 residues per cm2) represents the density of Gal[alpha]1-3Gal residues recognized by anti-Gal[alpha]1-3Gal IgM but not by GS-IB4. This density, 1.14 × 1012 Gal[alpha]1-3Gal residues per cm2 of endothelial cells, is very similar to the density obtained from the cross-blocking experiments.
While these data provide strong evidence that some Gal[alpha]1-3Gal residues are recognized by IgM but not by GS-IB4 and vice versa, the amount of overlap between the sites recognized by GS-IB4 and by anti-Gal[alpha]1-3Gal IgM cannot be ascertained precisely since the binding of one protein might affect the binding of a second protein without directly hindering the sites recognized by the second protein. For example, the binding of anti-Gal[alpha]1-3Gal IgM might aggregate unbound Gal[alpha]1-3Gal residues in such a way that some residues become inaccessible to GS-IB4. On the other hand, some residues normally inaccessible to GS-IB4 might be 'exposed" as a result of IgM binding.
Blocking of anti-Gal[alpha]1-3Gal IgM binding by GS-IB4 on solubilized porcine endothelial cell integrins
GS-IB4 recognizes some [alpha]-galactosyl structures not recognized by anti-Gal[alpha]1-3Gal IgM (Wood et al., 1979; Gautreau et al., 1993; Sandrin et al., 1993; Parker et al., 1996b). To test whether differences in the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 might reflect differences in specificity in this system, binding to isolated porcine integrin chains, known to be a major target of xenoreactive IgM (Holzknecht and Platt, 1995; Lin et al., 1996), was tested. As expected, treatment of isolated integrins with [alpha]-galactosidase abrogated the binding of IgM and GS-IB4 in solid phase assays (not shown). In contrast to what was observed using intact porcine endothelial cells (Figure 4), GS-IB4 blocked the binding of anti-Gal[alpha]1-3Gal IgM to solubilized membrane glycoproteins by greater than 82% (Figure 6). These findings demonstrate that anti-Gal[alpha]1-3Gal IgM and GS-IB4 recognize the same chemical structures produced by the pig and suggest that the difference in their binding to porcine endothelial cells does not reflect a difference in the specificity of antibody and lectin and that other features, such as the microenvironment of the cell surface, may be important in the differential recognition of [alpha]-galactosyl residues by these two proteins.
Figure 6. Cross-blocking by antibody and lectin on porcine membrane antigens in solution. Porcine endothelial cell integrins were isolated by affinity chromatography and biotinylated. The integrins were blocked with saturating concentrations of GS-IB4 and then incubated with human serum as a source of anti-Gal[alpha]1-3Gal IgM and precipitated with immobilized avidin. The amount of anti-Gal[alpha]1-3Gal IgM bound in the presence of GS-IB4 was then quantitated by an enzyme linked assay and compared to the amount bound to integrins not blocked with GS-IB4. Samples were run in quadruplicate and the standard error is shown. GS-IB4 blocked the binding of IgM by greater than 80%, indicating that, in contrast to observations using intact porcine cell surface, GS-IB4 can block most or all of the sites on solubilized porcine proteins which are recognized by IgM.
Binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 on model surfaces
To test the idea that the microenvironment in which Gal[alpha]1-3Gal residues are presented might influence the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4, cross-blocking experiments similar to those carried out on the cell surface were performed on Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on model surfaces. With Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on a hydrophobic surface, which presumably disrupts the native conformation of the core protein, thereby fully exposing the carbohydrate epitopes that would otherwise have been accessible by one protein but not by another, anti-Gal[alpha]1-3Gal IgM blocked almost completely the binding of GS-IB4 (Figure 7A) and GS-IB4 blocked almost completely the binding of anti-Gal[alpha]1-3Gal IgM (Figure 7B). This result suggested that anti-Gal[alpha]1-3Gal IgM and GS-IB4 recognize the same [alpha]-galactosyl residues on this surface. On the other hand, with Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on a hydrophilic surface, which presumably preserves the native tertiary structure of the core protein, anti-Gal[alpha]1-3Gal IgM blocked about 80% of the binding of GS-IB4 (Figure 8A), while GS-IB4 blocked only about 20% of the binding of anti-Gal[alpha]1-3Gal IgM (Figure 8B). These results suggest the possibility that steric factors such as the shape of a protein core on a surface might influence the binding of lectins and anti-carbohydrate antibodies and thus might play a role in the differential recognition of [alpha]-galactosyl residues by anti-Gal[alpha]1-3Gal IgM and GS-IB4.
Figure 7. Interaction of anti-Gal[alpha]1-3Gal IgM and GS-IB4 with Gal[alpha]1-3Gal-[beta]1-4GlcNAc-BSA immobilized on a hydrophobic surface. Immobilized Gal[alpha]1-3Gal-[beta]1-4GlcNAc-BSA was blocked with saturating concentrations of (A) anti-Gal[alpha]1-3Gal IgM or (B) GS-IB4. Blocked and unblocked surfaces were then incubated with biotinylated GS-IB4 (A) or with human serum as a source of anti-Gal[alpha]1-3Gal IgM (B), the binding of which was quantitated by an enzyme linked assay. Samples were run in quadruplicate and the standard error is shown. Antibody and lectin each blocked the binding of the other almost completely, in contrast to results obtained on intact cells (Figure 4).
Figure 8. Interaction of anti-Gal[alpha]1-3Gal IgM and GS-IB4 with Gal[alpha]1-3Gal-[beta]1-4GlcNAc-BSA immobilized on a hydrophilic surface. Immobilized Gal[alpha]1-3Gal-[beta]1-4GlcNAc-BSA was blocked with saturating concentrations of (A) anti-Gal[alpha]1-3Gal IgM or (B) GS-IB4. Blocked and unblocked surfaces were then incubated with biotinylated GS-IB4 (A) or with human serum as a source of anti-Gal[alpha]1-3Gal IgM (B), the binding of which was quantitated by an enzyme linked assay. Samples were run in quadruplicate and the standard error is shown. Anti-Gal[alpha]1-3Gal IgM blocked [sim]80% of binding of GS-IB4 and GS-IB4 blocked [sim]20% of binding of anti-Gal[alpha]1-3Gal IgM, results similar to cross-blocking on endothelial cell surfaces. Inset, Fraction of GS-IB4 remaining after sequential incubation with GS-IB4 and anti-Gal[alpha]1-3Gal IgM. The extent to which anti-Gal[alpha]1-3Gal IgM displaces GS-IB4 bound to immobilized Gal[alpha]1-3Gal-[beta]1-4GlcNAc-BSA on a hydrophilic surface was determined by ELISA as described in Figure 5. The amount of GS-IB4 remained bound after sequential incubation with GS-IB4 and human serum, as a source of anti-Gal[alpha]1-3Gal IgM (IgM), was essentially the same as the amount that remained bound after sequential incubation with GS-IB4 and PBS (PBS). The result indicates that there was no displacement of bound GS-IB4 by anti-Gal[alpha]1-3Gal IgM on immobilized Gal[alpha]1-3Gal-[beta]1-4GlcNAc-BSA.
As a second approach to testing the idea that the manner in which [alpha]-galactosyl residues are presented within the microenvironment might influence the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4, the percent of total Gal[alpha]1-3Gal residues utilized by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on each of the two model surfaces was compared (Table I). The density of Gal[alpha]1-3Gal residues on a given surface was determined by quantitation of galactose released by [alpha]-galactosidase. The valency of binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 was assumed to be similar to that observed on porcine cells (see above). Approximately 12% of the total Gal[alpha]1-3Gal residues on Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on a hydrophobic surface were recognized by anti-Gal[alpha]1-3Gal IgM and by GS-IB4. On the other hand, only [sim]6% of the total Gal[alpha]1-3Gal residues on Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on a hydrophilic surface were recognized by anti-Gal[alpha]1-3Gal IgM and by GS-IB4. Thus, with Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA adsorbed on a hydrophobic surface, Gal[alpha]1-3Gal residues are more accessible for the binding of lectin or antibody, perhaps because the core protein, BSA, is in a denatured or 'flattened" conformation. On the other hand, Gal[alpha]1-3Gal residues on Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA adsorbed on a hydrophilic surface are less accessible to lectin and antibody, perhaps because BSA retains a more native or globular form. These results further support the idea that the shape of the protein core to which the carbohydrate residues are attached may influence the topography of Gal[alpha]1-3Gal residues presented on these model surfaces.
Discussion
The interactions of proteins with carbohydrates on a cell surface are influenced by the local environment. For example, Gahmberg and Hakomori, using normal and transformed 3T3 and NIL cell lines (Gahmberg and Hakomori, 1973), and Steck and Dawson, using human erythrocytes (Steck and Dawson, 1974), found that galactose oxidase interacts more efficiently with some galactose residues than others. We have probed the complexity of endothelial cell surfaces using anti-Gal[alpha]1-3Gal IgM and GS-IB4, the binding of which are thought to have biological and evolutionary implications (Castronovo et al., 1989; Galili et al., 1987b; Rother et al., 1995). The results indeed provide strong support for the idea that the topography of Gal[alpha]1-3Gal on a surface can result in dramatic differences in the binding of antibody and lectin to chemically identical residues, some topologies favoring the binding of antibody but not lectin, other topologies favoring the binding of lectin but not antibody. Thus, biological responses associated with protein-carbohydrate interactions may be significantly conditioned by the microenvironment of the cell surface.
One factor which might account for differential binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 with a cell surface is specificity. The fine specificities of anti-Gal[alpha]1-3Gal IgM and GS-IB4 differ in that anti-Gal[alpha]1-3Gal antibodies bind more strongly to Gal[alpha]1-3Gal than to other [alpha]-galactosyl structures and do not bind Gal[alpha]1-4 structures or blood group B (Gal[alpha]1-3-(Fuc[alpha]1-2)Gal) to any extent (Galili et al., 1984, 1987a; Good et al., 1992; Sandrin et al., 1993; Parker et al., 1996a), while GS-IB4 binds to all [alpha]-galactosyl structures (Wood et al., 1979; Hayes and Goldstein, 1974). However, two lines of evidence suggest that differential binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 to cell surfaces does not reflect different specificities. First, in contrast to results seen on the cell surface, antibody and GS-IB4 effectively blocked the binding of each other to soluble porcine membrane antigens. Second, antibody and lectin exhibited differential binding to a synthetic glycoconjugate, much as they do to porcine cells.
Structural differences between anti-Gal[alpha]1-3Gal IgM and GS-IB4 (Table II) that are not related to specificity must therefore account for differences in the binding of these proteins to a cell surface. One possibility is that IgM may bind to Gal[alpha]1-3Gal moieties not accessible to GS-IB4 because IgM has an effective binding unit (VH +VL; 25 kDa) (Lin and Putnam, 1978) which is smaller than the binding 'unit" of GS-IB4 (likely the entire globular protein of 114 kDa) (Hayes and Goldstein, 1974; Wood et al., 1979). This possibility is suggested by the finding that some sites on the Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized in a way that probably preserves the globular structure of BSA but not on those in a presumably denatured or 'flattened" conformation, are recognized by anti-Gal[alpha]1-3Gal IgM but not by GS-IB4. Another possibility is that IgM, having more binding sites and a greater size and flexibility than GS-IB4, may be able to bind more effectively than GS-IB4 to relatively disperse [alpha]-galactosyl residues. Evidence to support this idea stems from the observation that anti-Gal[alpha]1-3Gal IgG binds to porcine cells in a manner similar to GS-IB4 but not to anti-Gal[alpha]1-3Gal IgM. Thus, the smaller size of the IgM binding domain as well as the larger valency, size, and flexibility of the entire IgM molecule might facilitate binding of that molecule under conditions in which GS-IB4 does not bind. On the other hand, the smaller size and shorter distance between binding subunits might facilitate the binding of GS-IB4 under other conditions.
Studies reported here suggest that differential recognition of [alpha]-galactosyl groups by IgM and GS-IB4 may involve structural features separated by greater than 25 Å, such as the overall shape of the glycoprotein, rather than features derived from primary and secondary structure, such as the specificity of recognition, which are manifest at shorter distances. These factors may influence the accessibilities of the carbohydrate epitopes to the antibodies and the lectin. For example, in cross-blocking studies, the cell surface was mimicked by immobilized neoglycoconjugate in a putatively globular form, but not by neoglycoconjugate in a more denatured form or by porcine membrane glycoproteins in a soluble form.
Table II
| Molecular weight (Da) |
Specificity | Number of binding sites per molecule |
Number of binding sites utilizedc |
|
| [alpha]-Galactosidase (from coffee bean) |
26,000 | [alpha]-Galactose | - | - |
| GS-IB4 | 114,000 | [alpha]-Galactose | 4 | 2-3 |
| IgM | 950,000 | Gal[alpha]1-3Gala | 10 | 4-5 |
| IgG | 157,000 | Gal[alpha]1-3Galb | 2 | 2d |
Although the primary focus of this work was on understanding of factors which affect the binding of proteins to cell surface carbohydrates, the interactions of IgM with porcine endothelial cell surfaces is of particular importance in the field of transplantation because this interaction initiates a rejection process which constitutes the major barrier to transplanting porcine organs into humans (Perper and Najarian, 1966; Perper and Najarian, 1967; Linn et al., 1968; Chavez-Peon et al., 1971; Merkel et al., 1971; Moberg et al., 1971; Platt et al., 1990a,b; Rose et al., 1991; Geller et al., 1992; Platt and Holzknecht, 1994). One implication of our findings is that binding of GS-IB4 does not accurately predict the binding of IgM to a cell surface. Of greater importance, however, is the possibility that our findings may help to explain the phenomenon of 'accommodation." We originally described accommodation in ABO-incompatible allografts where, following the temporary removal of anti-donor antibodies, an allograft would continue to function free of rejection even after the anti-donor antibodies returned to the circulation (Alexandre et al., 1987; Chopek et al., 1987; Platt et al., 1990c). Accommodated allografts contain little or no bound antibody even though the target antigen continued to be expressed. A similar process may occur in xenotransplants (Platt et al., 1990c,1991). The mechanism of accommodation has not been critically tested; however, the present findings suggest as one explanation that there may occur changes in topography of the cell surface rendering the antigenic carbohydrate inaccessible to the offending antibodies. To the extent that accommodation will prove essential to the clinical application of xenotransplantation (Platt et al., 1990c), our findings may help direct further inquiry in this field.
Materials and methods
Materials
Tissue culture plates were obtained from Becton Dickinson Labware (Lincoln Park, NJ). Nunc-Immuno Maxisorp and Polysorp polystyrene plates were purchased from VWR Scientific (Marietta, GA). Gal[alpha]1-3Gal and Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA were from Dextra Laboratories, Ltd. (Reading, UK). Biotinylated, fluorescein labeled, and unlabeled GS-IB4 were purchased from Vector Laboratories, Inc. (Burlingame, CA). N-Octylglucoside and [alpha]-galactosidase (EC 3.2.1.22; from green coffee beans) were obtained from Boehringer Mannheim (Indianapolis, IN). Sulfosuccinimidyl-6-(biotinamido) hexanoate was obtained from Pierce (Rockford, IL). Alkaline phosphatase conjugated, affinity isolated goat antibodies specific for human µ-chain were obtained from Sigma (St. Louis, MO). Mouse monoclonal IgG2b antibodies specific for human IgM (clone MB-11) were from Sigma. Slide-A-Lyzer Cassettes (10,000 Da molecular weight cut-off), for dialysis of samples with small volumes (0.5-3.0 ml), were purchased from Pierce (Rockford, IL).
Human serum was used as a source of anti-Gal[alpha]1-3Gal IgM antibodies. The serum was obtained from individuals with well-characterized concentrations of anti-Gal[alpha]1-3Gal natural antibodies (Parker et al., 1994) and low or undetectable amounts of anti-Gal[alpha]1-3Gal IgG (Yu et al., 1996) in their serum. Serum from a baboon which had been the recipient of a pig organ xenograft was used as a source of anti-Gal[alpha]1-3Gal IgG. The xenograft had been transplanted 16 days and removed 11 days prior to obtaining the serum. Serum was incubated at 56°C for 30 min to inactivate complement prior to use. In order to maximize the binding of IgG to porcine endothelial cell surface, the serum was diluted to 0.6% with phosphate-buffered saline (PBS) and reduced with 10 mM DTT to depolymerize IgM molecules and eliminate their binding to the cells (Yu et al., 1996).
Porcine aortic endothelial cells
Porcine aortic endothelial cells were explanted from porcine aortae and cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 2.0 mM l-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin as described previously (Ryan and Maxwell, 1986; Platt et al., 1990b). The cells express Gal[alpha]1-3Gal, thought to be the predominant structure on porcine cells recognized by xenoreactive natural antibodies (Galili et al., 1987b; Thall and Galili, 1990; Collins et al., 1994; Holzknecht and Platt, 1995). For analysis of protein binding, the endothelial cells were cultured in 96-well plates (Platt et al., 1990b,d). After the endothelial cells reached confluence (3-7 days), they were fixed with 0.1% glutaraldehyde and frozen at -80°C until needed. Despite the potential effects of culture and fixation on the endothelial cells, this treatment was used because it allows quantitative assay of protein binding and because cells treated in this way resist morphological changes which might otherwise be induced by the conditions used in this study.
Purification of porcine endothelial cell integrins
Extracts from porcine aortic endothelial cell membranes were prepared as described previously (Platt and Holzknecht, 1994). Porcine integrins were purified from the endothelial cell membrane extracts using immobilized collagen as an affinity matrix as described previously (Holzknecht and Platt, 1995). Collagen columns adsorb most endothelial cell integrins, including integrins [beta]1, [alpha]v, [alpha]3, and [alpha]5 (Holzknecht and Platt, 1995). The partially purified integrins (0.5-1.0 mg/ml) were biotinylated using a 20:1 molar ratio of sulfosuccinimidyl-6-(biotinamido)-hexanoate to protein in 100 mM sodium phosphate, 150 mM NaCl, 50 mM n-octylglucoside, pH 8.5. The reaction was carried out at 25°C for 30 min. The biotinylated integrins were then dialyzed overnight against Tris-buffered saline with 50 mM n-octylglucoside and stored at -80°C until use.
Quantitation of binding of anti-Gal[alpha]1-3Gal IgM to porcine cells
The binding of human IgM to porcine endothelial cells was evaluated by enzyme-linked immunosorbent assay as previously described (Platt et al., 1990b). Endothelial cells cultured and prepared in 96-well plates as described above were blocked with 1% BSA in PBS for 1 h and washed with PBS. The cells were then incubated with a source of human anti-Gal[alpha]1-3Gal IgM for 3 h at 4°C, washed three times with PBS, and incubated for 1 h at 25°C with affinity purified alkaline phosphatase-conjugated goat antibodies specific for human µ-chain. The cells were then washed four times with PBS, and 100 µl of a developing solution consisting of p-nitrophenyl phosphate in a 100 mM diethanolamine buffer was added to each well. The absorbance at 405 nm (A405) was determined using an EL 340 Bio Kinetics Reader (Bio-Tek Instruments; Winooski, VT). The binding of IgM specific for Gal[alpha]1-3Gal was taken to be that binding eliminated by the addition of maximally inhibitory concentrations of soluble Gal[alpha]1-3Gal or by treatment of the cells with [alpha]-galactosidase as described below. Typically, 50-90% of the IgM which bind to the cultured endothelial cells utilize Gal[alpha]1-3Gal (Parker et al., 1996b), and some of the antibodies which bind to sites other than Gal[alpha]1-3Gal do so either nonspecifically or utilizing determinants not present in vivo (Parker et al., 1996b). The amount (µg/ml) of anti-Gal[alpha]1-3Gal IgM bound to the cell surface was quantitated using standards as described previously (Parker et al., 1994).
Quantitation of binding of anti-Gal[alpha]1-3Gal IgM to immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA
The amount of anti-Gal[alpha]1-3Gal IgM which bound to immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA was determined by quantitating, using enzyme-linked immunosorbent assay (ELISA), the amount of anti-Gal[alpha]1-3Gal IgM extracted from the immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA. Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA was immobilized on Nunc Polysorp plates and on Nunc Maxisorp plates by coating the plates with 50 µl of 20 µg/ml Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA in PBS for a minimum of 3 h. Nunc Polysorp plates bind proteins through hydrophobic interactions, presumably disrupting the native structure, while Maxisorp plates interact with proteins via hydrophilic forces, presumably preserving native tertiary structure. After the plates were coated with Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA, they were washed three times with PBS and blocked with 1% BSA. Fifty microliters per well of 2.5-50% human serum were then incubated for 3 h at 4°C on these two surfaces. The plates were washed three times with cold (4°C) PBS to remove unbound IgM. The bound anti-Gal[alpha]1-3Gal IgM was extracted with 20 mM Gal[alpha]1-3Gal in PBS (pH 7.4) at 4°C for 2 h. Fifty microliters per well of the extracted IgM and standards were applied to 96-well polystyrene plates (Nunc Polysorp) for quantitation by ELISA. The Polysorp plates were previously coated with 75 µl of 30 µg/ml anti-human IgM antibodies at 4°C for overnight, washed three times with PBS, and blocked with 1% BSA in PBS. The extracted IgM and standards were applied to the Polysorp plates for 3 h at 4°C. The plates were then washed with PBS, incubated with alkaline phosphatase-conjugated anti-human IgM antibodies, washed again with PBS, and reacted with the developing solution as described above.
Quantitation of binding of GS-IB4 to porcine cells
The binding of GS-IB4 to cultured porcine aortic endothelial cells was evaluated using an enzyme-linked lectin assay. Endothelial cells cultured and fixed in 96-well plates as described above were blocked with 1% BSA in PBS for 1 h, washed with PBS and then incubated for 3 h at 4°C with biotinylated GS-IB4 diluted in PBS containing 1 mM CaCl2 and 1 mM MgCl2. The cells were then washed three times with PBS and incubated with alkaline phosphatase-conjugated avidin for 1 h. The cells were then washed four times with PBS and reacted with the developing solution as described above. The binding of GS-IB4 to Gal[alpha]1-3Gal was taken to be that binding which was eliminated by treatment of the cells with [alpha]-galactosidase as described below.
The amount of GS-IB4 which bound to cultured endothelial cells was determined by quantitating, using enzyme-linked lectin assay, the amount of GS-IB4 extracted from the endothelial cells. Endothelial cells were incubated with 50 µl/well of 1.25-50 µg/ml of biotinylated GS-IB4 for 3 h at 4°C and then washed three times with cold (4°C) PBS to remove unbound GS-IB4. Biotinylated GS-IB4 bound to the endothelial cells was extracted for 2 h at 4°C with a solution containing 0.5 M [alpha]-CH3 galactose, 20 mM EDTA, and 1%BSA in PBS (pH 7.4). The amount of GS-IB4 extracted was quantitated by enzyme-linked lectin assay using 96-well polystyrene plates (Nunc Polysorp) coated with 75 µl of 35 µg/ml of avidin at 4°C for overnight. Fifty microliters per well of the extracted biotinylated GS-IB4 and of solutions containing known concentrations (4-20 ng/ml) of biotinylated GS-IB4 were applied to the avidin-coated plates for 3 h at 4°C. The plates were then washed three times with PBS, incubated with alkaline phosphatase-conjugated avidin for 1 h, washed four times with PBS, and reacted with the developing solution as described above.
Quantitation of binding of GS-IB4 to immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA
The amount of GS-IB4 which bound to immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA was determined by quantitating, using an enzyme-linked lectin assay, the amount of GS-IB4 extracted from the immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA. Fifty microliters per well of 1.25-50 µg/ml of biotinylated GS-IB4 were incubated with Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on Polysorp plates and with Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on Maxisorp plates for 3 h at 4°C. Unbound GS-IB4 was removed by washing the plates three times with cold (4°C) PBS. Biotinylated GS-IB4 bound to these surfaces was then extracted for 2 h at 4°C with a solution containing 0.5 M [alpha]-CH3 galactose, 20 mM EDTA, and 1% BSA in PBS (pH 7.4). The amount of extracted GS-IB4 was quantitated by enzyme-linked lectin assay as described above.
Treatment of porcine aortic endothelial cells with [alpha]-galactosidase
Porcine aortic endothelial cells were digested with [alpha]-galactosidase which, as previously described (Collins et al., 1994), is specific for [alpha]-galactosyl residues on porcine endothelial cell surfaces. Endothelial cells cultured in 96-well plates were fixed with glutaraldehyde, blocked with 1% BSA in PBS, and washed as described above. A solution containing 0.0 U to 1.0 U of [alpha]-galactosidase in 50 µl of 100 mM NaCl, 50 mM sodium acetate, pH 5.0 was added to each well and incubated at 37°C for 3 h unless otherwise indicated. For experiments involving quantitation of galactose released by [alpha]-galactosidase, [alpha]-galactosidase was dialyzed for 1 h against 100 mM NH4Ac, pH 5.0 just prior to use, and the reaction was carried out in 100 mM NH4Ac, pH 5.0. The reaction was judged to be complete when further digestion resulted in no further decrease in the binding of xenoreactive IgM or GS-IB4 and when soluble Gal[alpha]1-3Gal[beta]1-4GlcNAc did not inhibit the binding of xenoreactive natural antibodies or GS-IB4. Following treatment with the enzyme, plates were washed six times with PBS and blocked for 1 h with 1% BSA in PBS.
Cross-blocking by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on cultured porcine endothelial cells
To determine whether the binding sites recognized by anti-Gal[alpha]1-3Gal IgM were also recognized by GS-IB4, cross-blocking experiments were performed using a modification of the enzyme-linked assay described above. Porcine endothelial cells prepared as described above were blocked for 1 h with 1% BSA in PBS, washed with PBS, and incubated with a saturating concentration of GS-IB4 (50 µg/ml) for 3-6 h at 4°C. Excess GS-IB4 was washed away with cold (4°C) PBS, and the cells were incubated with a saturating concentration of anti-Gal[alpha]1-3Gal IgM (>20 µg/ml) or PBS for 1 h at 4°C. The cells were then washed three times with PBS and the binding of IgM was measured as described above. The signal generated by binding of alkaline phosphatase-conjugated goat antibodies specific for human µ-chain directly to GS-IB4 was subtracted to yield the signal from the binding of IgM.
To determine whether the binding sites recognized by GS-IB4 were also occupied by anti-Gal[alpha]1-3Gal IgM, porcine endothelial cells were blocked for 1 h with 1% BSA in PBS, washed with PBS, and incubated with a saturating concentration of anti-Gal[alpha]1-3Gal IgM (>20 µg/ml) for 3-6 h at 4°C as determined previously (Parker et al., 1994). Excess anti-Gal[alpha]1-3Gal IgM was removed by washing with cold PBS at 4°C, and the cells were incubated with a saturating concentration of biotinylated GS-IB4 (50 µg/ml) or PBS for 1 h at 4°C. The cells were then washed three times with PBS, and the binding of GS-IB4 to sites not blocked by anti-Gal[alpha]1-3Gal IgM was measured as described above. The signal generated by binding of alkaline phosphatase-conjugated avidin directly to anti-Gal[alpha]1-3Gal IgM was subtracted to yield the signal from the binding of GS-IB4.
Blocking of anti-Gal[alpha]1-3Gal IgM binding by GS-IB4 on solubilized porcine endothelial cell integrins
To determine whether anti-Gal[alpha]1-3Gal IgM and GS-IB4 recognize the same residues on a purified protein, the binding of anti-Gal[alpha]1-3Gal IgM to solubilized porcine integrins was measured before and after the integrins were blocked with saturating amounts of GS-IB4. Porcine integrins (3 µg/ml) conjugated with biotin were incubated with 10 mg/ml of GS-IB4 for 3 h at 4°C, after which anti-Gal[alpha]1-3Gal IgM (final concentration = 2.4 µg/ml) was added to the solution and incubated for another 30 min at 4°C. The mixture was then incubated for 4 h with immobilized avidin prepared as described below. The amount of anti-Gal[alpha]1-3Gal IgM precipitated by this procedure was determined using alkaline phosphatase-conjugated goat antibodies specific for human µ-chain as described above.
Immobilized avidin used in these experiments was prepared by incubating 50 µl/well of 40 µg/ml of avidin diluted in PBS to 96-well Maxisorp plates for at least 3 h. The plates were then blocked by incubation for at least 1 h with 1% BSA and 0.5% Tween in PBS. Following the blocking step, the plates were washed twice with cold PBS and stored at -80°C until use.
Cross-blocking by anti-Gal[alpha]1-3Gal IgM and GS-IB4 on immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA
To determine whether the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 depends on the shape of the protein core to which Gal[alpha]1-3Gal is attached, the binding of anti-Gal[alpha]1-3Gal IgM and GS-IB4 to denatured glycoproteins was compared with the binding to native glycoproteins. For this comparison, cross-blocking experiments like those performed on cultured porcine endothelial cells were carried out using Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA immobilized on Polysorp plates, which bind proteins through hydrophobic interactions, presumably disrupting the native structure, and Maxisorp plates, which interact with proteins via hydrophilic forces, presumably preserving native tertiary structure. These plates were coated with 50 µl of 20 µg/ml Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA in PBS for a minimum of 3 h. The plates were then washed three times with PBS and blocked with 1% BSA as described above. Cross-blocking experiments were performed on these plates as described above.
Quantitation of [alpha]-galactosyl residues on cultured porcine aortic endothelial cells and on immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA
To determine the total number of [alpha]-galactose residues on porcine endothelial cells and on immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA which were accessible to [alpha]-galactosidase, the [alpha]-galactosyl residues released from cells and from immobilized glycoconjugate using [alpha]-galactosidase were quantitated by high performance liquid chromatography (HPLC). To this end, 96-well plates with cultured fixed cells or immobilized Gal[alpha]1-3Gal[beta]1-4GlcNAc-BSA were washed six times with 150 µl/well of 100 mM NH4Ac, pH 5.0 for 6-18 h per wash. Plates were then treated with 70 µl/well of 0.8 U/ml of [alpha]-galactosidase, as described above. Samples from 54 wells were pooled (volume about 3.6 ml) and concentrated to 200 µl using an ISS Enprotech Univapo 100 H concentrator centrifuge, frozen, and lyophilized overnight using a Virtis 12SL lyophilizer. The samples were dissolved in deionized water and chromatographed isocratically using a Doinex CarboPac PA1 (4 × 250 mm) anion exchange column and a mobile phase of 16 mM NaOH at a flow rate of 1.4 ml/min. Galactose was detected by pulsed amphomeric detection using an ESA Coulochem II electrochemical detector equipped with a gold electrode and quantitated based on a standard curve generated using known quantities of galactose.
Acknowledgments
We thank Kathy Morris and Wendy Foley for their assistance in preparation of the manuscript and Dr. Charles W. Hoopes for helpful discussions. This work was supported by grants from the NIH (HL50985 and HL52297).
Abbreviations
GS-IB4, Griffonia simplicifolia type I lectin, isolectin B4; BSA, bovine serum albumin; PBS, phosphate-buffered saline; DTT, dithiothreitol; HPLC, high performance liquid chromatography.
References
4To whom correspondence should be addressed at: Department of Surgery, Duke University Medical Center, Box 2605, Durham, NC 27710
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 2 Apr 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. A. Koch, A. Kanazawa, R. Nishitai, B. E. Knudsen, K. Ogata, T. B. Plummer, K. Butters, and J. L. Platt Intrinsic Resistance of Hepatocytes to Complement-Mediated Injury J. Immunol., June 1, 2005; 174(11): 7302 - 7309. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Koch, Z. I. Khalpey, and J. L. Platt Accommodation: Preventing Injury in Transplantation and Disease J. Immunol., May 1, 2004; 172(9): 5143 - 5148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Xu, D. Yin, B. Naziruddin, L. Chen, A. Stark, Y. Wei, Y. Lei, J. Shen, J. S. Logan, G. W. Byrne, et al. The In Vitro and In Vivo Effects of Anti-Galactose Antibodies on Endothelial Cell Activation and Xenograft Rejection J. Immunol., February 1, 2003; 170(3): 1531 - 1539. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||