Novel Antibody-Lectin Enzyme-Linked Immunosorbent
http://www.100md.com
微生物临床杂志 2005年第3期
Institute of Pathology
Department of Neuroscience
Cancer Research Center, Case Western Reserve University School of Medicine, Cleveland, Ohio
Division of Neuropathology, University of Edinburgh, Edinburgh, United Kingdom
ABSTRACT
We used different anti-prion protein (anti-PrP) monoclonal antibodies to capture either full-length or truncated PrP species and then used biotinylated lectin to compare the nature of the glycans on bound PrP species present in control, sporadic Creutzfeldt-Jakob disease (sCJD), or variant CJD (vCJD) brains. When full-length PrP species in these three groups were compared, no significant difference in the binding of concanavalin A or Aleuria aurantia lectin was detected. However, the binding of Ricinus communis agglutinin I (RCA) to sCJD and vCJD samples was significantly increased. In contrast, when only truncated PrP species were compared, only vCJD samples had more RCA binding activity. Therefore, while most of the RCA binding activity in sCJD is restricted to the full-length PrP species, the RCA binding activity in vCJD is associated with truncated and full-length PrP species. Furthermore, the RCA binding activity in sCJD and vCJD samples is mostly associated with proteinase K-resistant PrP species, a known signature of infectious prion. Therefore, PrP species in sCJD and vCJD have dissimilar lectin immunoreactivity, which reflects differences in their N-linked glycans. These differences may account for the distinct phenotypes of sCJD and vCJD.
INTRODUCTION
According to the protein-only hypothesis, the central event in the pathogenesis of prion disease is the conversion of a normal cellular prion protein, PrPC, into an abnormal, pathogenic, scrapie isoform, PrPSc (21, 22). Accumulation of PrPSc in the brain is the cause of prion disease. The PrPC-to-PrPSc conversion is based on a change in conformation from a predominantly -helical structure that characterizes PrPC to a predominantly -sheet structure that characterizes PrPSc (14, 23). However, the molecular events that govern PrPC-to-PrPSc conversion are complex and not completely understood. One major consequence of the conformational change is that while PrPC is highly sensitive to digestion by proteinase K (PK), the conformational change renders the PrPSc relatively more resistant to PK digestion (1, 13). Therefore, resistant to PK digestion has been used as an in vitro diagnostic hallmark of all prion diseases.
It is generally believed that in the human brain, PrPC species are present in three glycoforms: diglycosylated, monoglycosylated, and unglycosylated. More recently, using a panel of anti-PrPC monoclonal antibodies (MAbs) that react with different regions of the PrPC, we found that the expression of PrPC in normal brain is far more complex than the three-band pattern previously seen with MAb 3F4 (15, 16, 32). By two-dimensional immunoblotting, we observed seven major species of PrPC, which could be subdivided into multiple subspecies (16). These PrP species were generated as a result of differences in N-linked glycosylation as well as the site of truncation (16). On the basis of this evidence, we concluded that most of the smaller PrP species are N-terminally truncated PrP species rather than unglycosylated or monoglycosylated PrP species. Furthermore, by separation of the full-length and truncated PrPC, the two populations were shown to be glycosylated differentially (16). We speculated that full-length and truncated PrPC might have different functions under physiological conditions and may play different roles in PrPSc formation (15, 17).
Earlier studies have demonstrated that the N-linked glycans in PrP are of the complex type, which is resistant to endoglycosidase H but sensitive to peptide-N-glycosidase F (PNGase F) (4, 6, 25, 30). The N-linked glycans contain terminal galactose, sialic acid, and fucose attached to the innermost N-acetyl-glucosamine cores (6). More-detailed glycostructures of PrP have come from studies using mass spectrometry with purified hamster and murine PrP species (25, 29, 30). It was found that both of the N-linked glycosylation sites can be glycosylated; over 30 glycostructures were identified (29, 30). Recently, a comparative study was done between normal PrPC and PrPSc from hamsters (25). It was found that PrPSc and PrPC contain similar glycostructures. However, PrPSc contained proportionally more branched glycostructures, such as bisected structures and tri- and tetra-antennary structures than PrPC (25).
The biological significance of glycosylation in PrPC function and the pathogenesis of prion diseases is not completely understood. We and others have reported that different PrPC glycoforms are present in different regions of the CNS (3, 11, 28). It has been speculated that the glycosylation in PrPC could determine the region of PrPSc targeting and accumulation (7). Therefore, the nature of the N-linked glycans in PrP might be closely associated with the lesion profiles and the typing of PrPSc strains. In the cell model, some glycosylated PrP mutants can reach the cell membrane; however, unglycosylated mutant PrP species tend to accumulate intracellularly and form PrPSc-like products (9, 24). PrPC without N-linked glycans has been reported to accumulate inside the cell in the Golgi instead of being exported to the cell membrane, and the unglycosylated PrP exhibited certain biochemical features characteristic of PrPSc (9, 24). In studies with N2aSc cells, a PrPSc-infected cell line, it has been reported that inhibition of N-linked glycan synthesis promotes the accumulation of PrPSc (31).
Over recent years, we have developed a large panel of anti-PrPC MAbs (10, 32). The epitopes of these MAbs spread from the N terminus to the C terminus of PrPC. In the studies described here, we used a novel antibody-lectin enzyme-linked immunosorbent assay (ELISA) to compare the natures of the N-linked glycans on different PrP species in control, sporadic Creutzfeldt-Jakob disease (sCJD), or variant CJD (vCJD) brain homogenates. We describe the use of different anti-PrP MAbs to capture either full-length or truncated PrP species in the brain homogenates and biotin-conjugated lectins to detect the presence of different N-glycans on bound PrP species. Our results clearly demonstrated that when compared to non-CJD controls, PrP species in sCJD and vCJD brains have different lectin immunoreactivities.
MATERIALS AND METHODS
Tissues. For controls, brain autopsies from the frontal cortex were obtained from individuals (n = 5) confirmed negative for prion disease. For sCJD, tissues were obtained from the frontal cortices of individuals with autopsy-confirmed cases (n = 5) with sporadic prion disease that were M/M at amino acid residue 129 and had the PrPSc type 2. For vCJD, frontal cortex was obtained from cases (n = 4) confirmed at autopsy. All samples from sCJD and vCJD cases had PK-resistant PrPSc, as demonstrated by immunoblotting with anti-PrP MAb 8H4 (not shown)
Preparation of brain homogenate. Human brain tissues were homogenized in 10 volumes of ice-cold lysis buffer (10 mM Tris, 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 5 mM EDTA [pH 8.0]) in the presence of 1 mM phenylmethylsulfonyl fluoride. If the homogenate was to be treated with PK, phenylmethylsulfonyl fluoride was omitted from the lysis buffer. After microcentrifugation at 3,000 rpm for 10 min, the supernatants were stored in aliquots at –80°C.
Anti-PrPC MAbs. The generation and characterization of anti-PrPC MAbs have been previously described in great detail (10, 32). MAb 8B4 and MAb 8H4 are immunoglogulin G1, and the epitopes recognized by these MAbs are diagrammatically presented in Fig. 1. While the epitope recognized by MAb 8H4 includes the second potential N-linked glycosylation site, MAb 8H4 reacts equally well with recombinant PrP and native PrPC (32). Thus, the presence of N-linked glycans does not impede the binding of MAb 8H4.
MAb-lectin ELISA. Initially, we carried out titration experiments to identify a concentration of the capture MAb at the linear portion of the PrP binding curve with a specific amount of brain homogenate. Therefore, the amount of plate-coated MAb is limiting, while the level of PrP species is in excess. This is to ensure that the amounts of captured PrP species from various samples are similar; thus, the differences in ELISA readings would not be due to quantitative difference in the amounts of captured PrP but would rather reflect the difference in lectin binding.
Ninety-six-well ELISA plates (Bio-Rad) were coated with affinity-purified MAb 8H4 (40 ng) or MAb 8B4 (20 ng) in 100 μl at 37°C for 3 h and blocked with 3% bovine serum albumin in phosphate-buffered saline (PBS) at 4°C overnight. To decrease the background, the plates were incubated with 0.2 ml of oxidation buffer (100 mM sodium periodate, 50 mM citric acid, [pH 4.0])/well for 15 min. Plates were then washed (four times) with washing buffer (PBS with 0.05% Tween 20). Two hundred micrograms of total brain proteins in 100 μl of PBS was added to the plate in duplicate and incubated at room temperature for 3 h. After four additional washes with washing buffer, 100 μl of different biotinylated lectins was added at a concentration of 2 μg/ml in washing buffer and incubated at room temperature for 2 h. The lectins were detected with a strepavidin-conjugated horseradish peroxidase (PharMingen, San Diego, Calif.). The lectins used include Aleuria aurantia lectin (ALL), specific for (-1,3)- and (-1,6)-linked fucose; concanavalin A (ConA), specific for -linked mannose; and Ricinus communis agglutinin I (RCA I), specific for galactose (Vector Laboratories, Burlingame, Calif.). Optical density values at 405 nm were measured with ABTS [2,2-azino-di-(3-ethylbenzthiazoline sulfonic acid] ammonium salt solution] (Roche Diagnostic, Indianopolis, Ind.) with an automated plate reader (Molecular Dynamics). The results presented were the average of duplicates and all experiments were repeated at least three times.
To examine the binding specificity of RCA to PrP, a competitive ELISA was also performed: glucose, sucrose, galactose and lactose (Sigma, St. Louis, Mo.) of different concentrations were included to compete with plate-immobilized PrP for the binding of biotinylated RCA.
Treatment of brain homogenates with PNGase F or guanidine hydrochloride (GdHCl). For PNGase F treatment, each brain preparation was mixed with 0.1 volume of 10x denaturing buffer (20 mM Tris, 150 mM NaCl, 2 mM EDTA, 10% mercaptoethanol, 5% sodium dodecyl sulfate [pH 7.5]) and heated for 10 min at 95°C. PNGase F digestion was carried out at 37°C for 1 h with the addition of 1% N40 and recombinant PNGase F (New England BioLabs) with a total of 1,000 U per 50 μl of reaction mixture, as recommended by the supplier.
To prepare denatured PrP species, each brain homogenate (4 μl of the 10% brain homogenate) was incubated with an equal volume of 4 M GdHCl (Sigma, St. Louis, Mo.) at room temperature for 10 min. The treated brain homogenate was then diluted with PBS to a final GdHCl concentration of 0.05 M. At this concentration, the effect of GdHCl on the interaction between PrP and anti-PrP MAb is minimal.
RESULTS
Anti-PrPC MAb-lectin ELISA: full-length PrP species from sCJD and vCJD cases have more RCA immunoreactivity. In an anti-PrPC MAb-lectin ELISA, an immobilized anti-PrPC MAb was used to capture the PrP species in the homogenate, and a biotin-conjugated lectin was then used to react with the bound PrP species. The epitopes recognized by the MAbs used in this study, MAb 8B4 and 8H4, are presented diagrammatically in Fig. 1. When immobilized on the plate, MAb 8B4 captures only the full-length PrP species (16). On the other hand, MAb 8H4 captures all described PrP species including the full-length PrP and the N-terminally truncated PrP species (16).
We first sought to confirm that the captured PrP species indeed reacted with lectin in a manner that was specific for both PrP and the sugar moiety. We demonstrated that the binding of biotinylated lectins (AAL, ConA, and RCA) to MAb 8B4-captured PrP species is PrP dosage dependent and proportional to the amount of PrP immobilized onto the plate (Fig. 2A). Furthermore, the binding of RCA to MAb 8B4-captured PrP was efficiently inhibited by appropriate sugars such as galactose and lactose, but not glucose and sucrose (Fig. 2B). Therefore, the binding of RCA is specific for galactose residues on PrP. Similar studies were performed to verify the specificity of AAL and ConA (data not shown).
Human brain homogenate contains both full-length PrP species and multiple truncated PrP species (16). Initially, we compared the lectin binding activity of the full-length PrP species present in brain homogenates from non-CJD controls (n = 4) and sCJD (n = 5) and vCJD (n = 4) donors, using immobilized MAb 8B4 to capture only the full-length PrP species.
The various lectins used in this study and their major binding specificities are shown in Table 1. All three groups of homogenates (non-CJD, sCJD, and vCJD) have comparable AAL and ConA binding activities (Table 1). In contrast, we consistently observed significantly more RCA binding activity in the full-length PrP species from sCJD and vCJD donors than the full-length PrP species from non-CJD controls (Table 1). These results provide evidence that the N-linked glycans on full-length PrP species present in sCJD and vCJD brains may be different from those that are present on full-length PrP species in non-CJD brain.
We verified that the assay indeed detects N-linked glycans on PrP species by removing the N-linked glycans with PNGase F prior to performing the MAb-lectin ELISA. After PNGase treatment, brain homogenates from all three groups lost their lectin immunoreactivity (Table 2). This experiment provided conclusive evidence that the anti-PrPC MAb-lectin ELISA is specific for the N-linked glycans. We also depleted the full-length PrP species in the brain homogenates with MAb 8B4-conjugated beads prior to the MAb 8B4-lectin ELISA. Removal of the full-length PrP species also reduced the binding of all three lectins to background levels (results not shown). Collectively, these results provide strong evidence that indeed the MAb 8B4-lectin ELISA detects N-linked glycans present on full-length PrP species.
The major target of RCA is galatose. Full-length PrP species from sCJD and vCJD may have more galatose residues. Alternatively, due to conformational differences between PrP species, more galatose residues on sCJD and vCJD may be exposed. To test these possibilities, each brain homogenate was treated with a denaturing agent, GdHCl, prior to carrying out the MAb-lectin ELISA. After GdHCl treatment, both the ConA and ALL binding activities were significantly increased in all three groups (Table 2). In contrast, identical treatment did not significantly alter the binding profiles of RCA. RCA immunoreactivity remained significantly higher in sCJD and vCJD samples than in non-CJD control samples. These results favor the interpretation that the reason full-length PrP species in sCJD and vCJD have more RCA immunoreactivity is that they have more galatose residues.
RCA immunoreactivity in sCJD sample is restricted to the full-length PrP species, while in vCJD it is associated with both full-length and truncated PrP species. Subsequently, we investigated whether RCA immunoreactivity is also present in the truncated PrP species in sCJD and vCJD brains. Brain homogenates from non-CJD, sCJD, and vCJD donors were first immunoprecipitated with MAb 8B4-conjugated beads to deplete the full-length PrP species as previously described by us (16). In these experiments, the efficiency of depletion was verified by immunoblotting the depleted sample with MAb 8H4. After immunoprecipitation with MAb 8B4 beads, the full-length PrP species was not detectable in control, sCJD, and vCJD brain homogenates (Fig. 3A). The depletion of full-length PrP species was also confirmed by conventional capture ELISA, using MAb 8B4 as the capture antibody and MAb 8H4 as a detector. After depletion with MAb 8B4 beads, PrP immunoreactivity was reduced to background levels in all samples (Fig. 3B). Furthermore, before depletion, the levels of PrP immunoreactivity in all samples were comparable, indicating that similar amounts of full-length PrP species were captured and that the capture antibody in the assay was indeed the limiting factor.
After depletion of the full-length PrP species, RCA immunoreactivity associated with the truncated PrP species in control and sCJD samples was reduced to levels that were indistinguishable (Table 3). On the other hand, removal of the full-length PrP species from the vCJD samples did not reduce RCA immunoreactivity. RCA binding activity in vCJD samples remained significantly higher than in the control and sCJD homogenates. The increase in RCA binding ranged from 300 to 450% in three different experiments. Accordingly, RCA binding activity in sCJD homogenate was mainly associated with full-length PrP species. In contrast, RCA binding activity in vCJD homogenate was associated with full-length and truncated PrP species.
A large proportion of the RCA immunoreactivity is associated with PK-resistant PrP species in sCJD and vCJD homogenates. One feature that distinguishes PrPC from PrPSc is that while PrPC is sensitive to PK, the pathogenic PrPSc is resistant (1, 13). Consequently, we investigated whether RCA immunoreactivity in sCJD and vCJD samples was associated with PK-sensitive or PK-resistant PrP species. Brain homogenates from controls (n = 4), sCJD (n = 4), and vCJD (n = 4) samples were treated in vitro with PK as described in Materials and Methods. The presence of PK-resistant PrP species in sCJD and vCJD brain homogenates but not in non-CJD brain homogenates was confirmed by immunoblotting the PK-treated brain homogenates with MAb 8H4 (results not shown).
Since PK treatment eliminates the MAb 8B4 epitope, MAb 8H4 was used as the capture reagent in MAb-lectin ELISA experiments. As expected, PK treatment completely eliminated ConA and RCA immunoreactivity in all four non-CJD control samples (results not shown). However, identical PK treatment did not significantly alter the binding of ConA to either sCJD or vCJD homogenates (Fig. 4). On the other hand, PK treatment significantly increased the binding of RCA to sCJD as well as the vCJD samples (Fig. 4). The increases in RCA binding ranged from 37% (sample sCJD-3) to 363% (sample vCJD-4). The increase in RCA immunoreactivity was most likely due to the elimination of PrPC in the brain homogenates by PK digestion.
Therefore, most if not all of the RCA immunoreactivity in sCJD and vCJD brain homogenates is associated with PK-resistant PrP species, a known signature of PrPSc.
DISCUSSION
In this paper, we describe four new findings that not only have implications for the pathogenesis of human prion diseases, but may also provide the basis for the development of a differential diagnostic test for sCJD and vCJD. First, we found that the nature of the N-linked glycans detected on full-length PrP species from sCJD and vCJD brains are different from those present on full-length PrP species from non-CJD brains. This conclusion is based on our findings that full-length PrP species from sCJD and vCJD samples reacted more strongly to the lectin RCA. Second, the increase in RCA binding is not caused by conformational differences between PrP species but more likely reflects a quantitative difference in the levels of galatose residue. Third, while the increase in RCA immunoreactivity in the sCJD brain is exclusively associated with full-length PrP species, the increase in RCA binding activity in vCJD by contrast is associated with both full-length and truncated PrP species. Fourth and most importantly, the increase in RCA immunoreactivity in sCJD and CJD brains is mostly associated with PK-resistant PrP species, a signature of PrPSc. These differences may account for the phenotypic differences, such as disease onset, histopathology, and disease duration, between sCJD and vCJD.
Three lines of evidence support our interpretation that the MAb-lectin ELISA indeed detects N-linked glycans present on PrP. (i) Removal of the N-linked glycans with PNGase completely eliminated lectin immunoreactivity (Table 2). (ii) The lectin does not react with bacterially produced recombinant PrP, which lacks the N-linked glycans (results not shown). (iii) It does not react with human brain homogenates in which the PrP species have been first depleted with anti-PrP MAb 8H4 affinity chromatography (results not shown).
The binding specificities of lectins have been investigated extensively since the 1940s. In general, each lectin has a binding specificity similar to the binding of antibody to its antigen. With few exceptions, most lectins interact with the nonreducing, terminal glycosyl groups of polysaccharide and glycoprotein chain ends. An exception is ConA, which binds to terminal 2-O- mannopyranosyl residues as well as -mannose in the internal-core oliogosaccharide structure. AAL binds preferentially to terminal fucose that is linked (1-6) to N-acetylglucosamine or to fucose that is linked (1-3) to N-acetyllactosamine-related structures. RCA-1 reacts preferentially to oligosaccharides ending in -galactose, much more weakly to -galactose, and only marginally to N-acetygalactoseamine.
In our earlier studies with affinity-purified PrP species from non-CJD brains, we found that the PrP species bound ConA and AAL the most, followed by Sambucus nigra lectin and then RCA (16). In this study, we observed more RCA binding activity in the full-length PrP species from sCJD and vCJD brains than non-CJD controls. We concluded that full-length PrP species from sCJD and vCJD have more galactose residues. This interpretation is supported by our findings that treatment with a denaturing agent, GdHCl, did not significantly alter the binding profiles of RCA. After GdHCl treatment, both sCJD and vCJD samples still had higher levels of RCA immunoreactivity than similarly treated non-CJD controls. However, a more-detailed biochemical analysis of the purified full-length PrP species will be required to verify this point. Interestingly, while treatment with GdHCl did not alter the binding profiles of RCA, it did increase binding of ConA and AAL. These results suggest that treatment with GdHCl exposed many of the ConA and AAL binding sites on full-length PrP species that are normally unavailable for binding.
It has been known for decades that infectious prions can be propagated in susceptible hosts in a strain-specific manner, similar to that of conventional pathogens (2, 26). Most recent studies have suggested that each prion strain represents a PrPSc molecule with a unique conformation (20, 27). It has also been recognized that the size of the human PrPSc fragment that is resistant to PK treatment varies depending on the type of prion disease. Essentially, two major strains of PrPSc have been identified in human prion disease, based on the electrophoretic mobility of the main PrPSc fragment generated by PK digestion (18, 19). Type 1 migrates at 21 kDa and type 2 migrates at 19 kDa, following gel electrophoresis (18). More recently, it has been suggested that at least six human prion strains can be identified, based on their electrophoretic mobility (8). Interestingly, all vCJD patients belong to the type 2 disease group and are homozygous with M/M at residue 129. Either people with M/M are more susceptible to vCJD or these individuals have a much shorter incubation period.
We found that full-length PrP species from sCJD and vCJD donors bound more RCA. Unlike sCJD, however, the RCA binding activity in vCJD is associated with full-length and truncated PrP species. In this study, we divided PrP species into either full-length or truncated species for simplicity. We reported previously that while non-CJD control brain homogenate has only one full-length PrP species, there are at least three full-length PrP species in sCJD samples (15). Whether the RCA binding activity in sCJD and vCJD is associated with the same subpopulations of full-length PrP species is not known. We also do not know whether the RCA binding activity in vCJD brain is present in one or more subpopulations of the truncated PrP species.
PrPSc tend to be insoluble in certain detergents (5). We analyzed the supernatant fraction of detergent solubilized brain homogenates. Therefore, one may argue that our findings actually reflect differential recovery of PrPSc in the supernatant fractions from sCJD and vCJD samples, rather than revealing molecular differences in the structures of the N-linked glycans on PrP species. We consider this scenario unlikely, because in the MAb-lectin ELISA, the capture antibody was always the limiting factor. Therefore, comparable amounts of PrP species in different samples were captured (Fig. 2B). Again, this issue can be resolved with a detailed biochemical analysis of the purified individual PrP species.
Our observations that PrP species in sCJD and vCJD brains have different lectin immunoreactivity provide suggestive evidence that the strain of PrPSc which causes sCJD is distinct from the strain of PrPSc that causes vCJD. Our MAb-lectin ELISA results are also in accord with our recent findings using two-dimensional immunoblotting, in which we found that PrPSc species from sCJD and vCJD patients have distinct N-linked glycans (T. Pan, et al., unpublished data). Furthermore, since most of the RCA immunoreactivity is associated with PK-resistant PrP species, we conclude that the detectable RCA immunoreactivity must be present on PrPSc, a signature of prion diseases. Interestingly, our findings that human PrPSc and PrPC have different glycans are in good agreement with an earlier report by Manuelidis and her colleagues (12). Using highly purified scrapie-associated fibrils from scrapie-infected hamster brains and lectins in immunoblots, these investigators found that RCA but not ConA preferentially reacted with the scrapie-associated fibrils (12).
In summary, our results suggest that the nature of the glycans and the PrP species may dictate the phenotypes of prion diseases or the strains of prion. The unique glycan may allow the PrPSc strain to target a subpopulation of PrPC to convert. It is also possible that the infectivity of different strains of prion may be associated with different PrP species, such as full-length versus truncated PrP species. It has been suggested that each strain of prion represents a PrPSc molecule with a distinct conformation (20, 27). The results presented here suggest that the nature of the N-linked glycans may contribute to this difference and provide a mechanism by which prion strains diversify. So far, we have only studied sCJD and vCJD cases. All the cases were homozygous with methionine in position 129 and have type 2 diseases. Whether our observations could be generalized to other CJD diseases, such as those that are heterozygous at 129 or with type 1 diseases, or to inherited prion diseases remains to be determined. Furthermore, it has been speculated that vCJD most likely represents the transmission of the disease from bovine spongiform encephalopathy-affected cattle to humans. It will be important to determine whether our observations made with vCJD brain homogenates could also be observed in brain homogenates from bovine spongiform encephalopathy-affected cattle. Finally, all the current in vitro diagnostic tests for prion diseases in humans are time consuming and nonquantitative. The assay described here is simple and quantitative. It may provide the basis for the development of a differential test for human prion diseases and the typing of different prion strains.
ACKNOWLEDGMENTS
We thank Geoff Barnard, Neil Greenspan, and Michael Lamm for careful reading and insightful discussion of the manuscript. We also thank Pierluigi Gambetti for support and the National Prion Disease Pathology Surveillance Center for providing the normal controls and sCJD samples.
This work is supported in part by NIH grant NS-045981-01 and U.S. Army grant NP0-20030 (M.-S.S.).
REFERENCES
Bolton, D. C., M. P. McKinley, and S. B. Prusiner. 1982. Identification of a protein that purifies with the scrapie prion. Science 218:1309-1311.
Bruce, M., H. Fraser, P. Mcbride, J. Scott, and A. Dickinson. 1992. The basis of strain variations in scrapie, p. 497-508. In S. B. Prusiner (ed.), Prion diseases in humans and animals. Ellis Horwood, London, United Kingdom.
DeArmond, S. J., Y. Qiu, H. Sanchez, P. R. Spilman, A. Ninchak-Casey, D. Alonso, and V. Daggett. 1999. PrPc glycoform heterogeneity as a function of brain region: implications for selective targeting of neurons by prion strains. J. Neuropathol. Exp. Neurol. 58:1000-1009.
Endo, T., D. Groth, S. B. Prusiner, and A. Kobata. 1989. Diversity of oligosaccharide structures linked to asparagines of the scrapie prion protein. Biochemistry 28:8380-8388.
Gabizon, R., M. P. McKinley, and S. B. Prusiner. 1987. Purified prion proteins and scrapie infectivity copartition into liposomes. Proc. Natl. Acad. Sci. USA 84:4017-4021.
Haraguchi, T., S. Fisher, S. Olofsson, T. Endo, D. Groth, A. Tarentino, D. R. Borchelt, D. Teplow, L. Hood, A. Burlingame, et al. 1989. Asparagine-linked glycosylation of the scrapie and cellular prion proteins. Arch. Biochem. Biophys. 274:1-13.
Hecker, R., A. Taraboulos, M. Scott, K. M. Pan, S. L. Yang, M. Torchia, K. Jendroska, S. J. DeArmond, and S. B. Prusiner. 1992. Replication of distinct scrapie prion isolates is region specific in brains of transgenic mice and hamsters. Genes Dev. 6:1213-1228.
Hill, A. F., S. Joiner, J. D. Wadsworth, K. C. Sidle, J. E. Bell, H. Budka, J. W. Ironside, and J. Collinge. 2003. Molecular classification of sporadic Creutzfeldt-Jakob disease. Brain 126:1333-1346.
Lehmann, S., and D. A. Harris. 1997. Blockade of glycosylation promotes acquisition of scrapie-like properties by the prion protein in cultured cells. J. Biol. Chem. 272:21479-21487. (Erratum, 273:5988, 1998.)
Li, R., T. Liu, B.-S. Wong, T. Pan, M. Morillas, W. Swietnicki, K. O'Rourke, P. Gambetti, W. K. Surewicz, and M.-S. Sy. 2000. Identification of an epitope in the C-terminus of normal prion protein whose expression is modulated by binding events in the N-terminus J. Mol. Biol. 301:567-574.
Liu, T., T. Zwingman, R. Li, T. Pan, B. S. Wong, R. B. Petersen, P. Gambetti, K. Herrup, and M.-S. Sy. 2001. Differential expression of cellular prion protein in mouse brain as detected with multiple anti-PrP monoclonal antibodies. Brain Res. 896:118-129.
Manuelidis, L., S. Valley, and E. E. Manuelidis. 1985. Specific proteins associated with Creutzfeldt-Jakob disease and scrapie share antigenic and carbohydrate determinants. Proc. Natl. Acad. Sci. USA 82:4263-4267.
Oesch, B., D. Westaway, M. Walchli, M. P. McKinley, S. B. Kent, R. Aebersold, R. A. Barry, P. Tempst, D. B. Teplow, and L. E. Hood. 1985. A cellular gene encodes scrapie PrP 27-30 protein. Cell 40:735-746.
Pan, K. M., M. Baldwin, J. Nguyen, M. Gasset, A. Serban, D. Groth, I. Mehlhorn, Z. Huang, R. J. Fletterick, F. E. Cohen, and S. B. Prusiner. 1993. Conversion of alpha-helices into beta-sheets features in the formation of the scrapie prion proteins. Proc. Natl. Acad. Sci. USA 90:10962-10966.
Pan, T., M. Colucci, B. S. Wong, R. Li, T. Liu, R. B. Petersen, S. Chen, P. Gambetti, and M.-S. Sy. 2001. Novel differences between two human prion strains revealed by two-dimensional gel electrophoresis. J. Biol. Chem. 276:37284-37288.
Pan, T., R. Li, B. S. Wong, T. Liu, P. Gambetti, and M.-S. Sy. 2002. Heterogeneity of normal prion protein in two-dimensional immunoblot: presence of various glycosylated and truncated forms. J. Neurochem. 81:1092-1101.
Pan, T., B. S. Wong, T. Liu, R. Li, R. B. Petersen, and M.-S. Sy. 2002. Cell-surface prion protein interacts with glycosaminoglycans. Biochem. J. 368:81-90.
Parchi, P., R. Castellani, S. Capellari, B. Ghetti, K. Young, S. G. Chen, M. Farlow, D. W. Dickson, A. A. Sima, J. Q. Trojanowski, R. B. Petersen, and P. Gambetti. 1996. Molecular basis of phenotypic variability in sporadic Creutzfeldt-Jakob disease. Ann. Neurol. 39:767-778.
Parchi, P., A. Giese, S. Capellari, P. Brown, W. Schulz-Schaeffer, O. Windl, I. Zerr, H. Budka, N. Kopp, P. Piccardo, S. Poser, A. Rojiani, N. Streichemberger, J. Julien, C. Vital, B. Ghetti, P. Gambetti, and H. Kretzschmar. 1999. Classification of sporadic Creutzfeldt-Jakob disease based on molecular and phenotypic analysis of 300 subjects. Ann. Neurol. 46:224-233.
Peretz, D., M. R. Scott, D. Groth, R. A. Williamson, D. R. Burton, F. E. Cohen, and S. B. Prusiner. 2001. Strain-specified relative conformational stability of the scrapie prion protein. Protein Sci. 10:854-863.
Prusiner, S. B. 1982. Novel proteinaceous infectious particles cause scrapie. Science 216:136-144.
Prusiner, S. B. 1998. Prions. Proc. Natl. Acad. Sci. USA 95:13363-13383.
Prusiner, S. B., M. P. McKinley, K. A. Bowman, D. C. Bolton, P. E. Bendheim, D. F. Groth, and G. G. Glenner. 1983. Scrapie prions aggregate to form amyloid-like birefringent rods Cell. 35:349-358.
Rogers, M., A. Taraboulos, M. Scott, D. Groth, and S. B. Prusiner. 1990. Intracellular accumulation of the cellular prion protein after mutagenesis of its Asn-linked glycosylation sites. Glycobiology 1:101-109.
Rudd, P. M., T. Endo, C. Colominas, D. Groth, S. F. Wheeler, D. J. Harvey, M. R. Wormald, H. Serban, S. B. Prusiner, A. Kobata, and R. A. Dwek. 1999. Glycosylation differences between the normal and pathogenic prion protein isoforms. Proc. Natl. Acad. Sci. USA 96:13044-13049.
Safar, J. 1990. Scrapie isoform of scrapie-associated protein. J. Am. Vet. Med. Assoc. 196:1681.
Safar, J., F. E. Cohen, and S. B. Prusiner. 2000. Quantitative traits of prion strains are enciphered in the conformation of the prion protein. Arch. Virol. Suppl. 16:227-235.
Somerville, R. A., and L. A. Ritchie. 1990. Differential glycosylation of the protein (PrP) forming scrapie-associated fibrils. J. Gen. Virol. 71:833-839.
Stahl, N., M. A. Baldwin, D. B. Teplow, L. Hood, B. W. Gibson, A. L. Burlingame, and S. B. Prusiner. 1993. Structural studies of the scrapie prion protein using mass spectrometry and amino acid sequencing. Biochemistry 32:1991-2002.
Stimson, E., J. Hope, A. Chong, and A. L. Burlingame. 1999. Site-specific characterization of the N-linked glycans of murine prion protein by high-performance liquid chromatography/electrospray mass spectrometry and exoglycosidase digestions. Biochemistry 38:4885-4895.
Winklhofer, K. F., U. Heller, A. Reintjes, and J. Tatzelt. 2003 Inhibition of complex glycosylation increases the formation of PrPsc. Traffic 4:313-322.
Zanusso, G., D. Liu, S. Ferrari, I. Hegyi, X. Yin, A. Aguzzi, S. Hornemann, S. Liemann, R. Glockshuber, J. C. Manson, P. Brown, R. B. Petersen, P. Gambetti, and M.-S. Sy. 1998. Prion protein expression in different species: analysis with a panel of new mAbs. Proc Natl Acad Sci USA 95:8812-8816.(Tao Pan, Ruliang Li, Boon)
Department of Neuroscience
Cancer Research Center, Case Western Reserve University School of Medicine, Cleveland, Ohio
Division of Neuropathology, University of Edinburgh, Edinburgh, United Kingdom
ABSTRACT
We used different anti-prion protein (anti-PrP) monoclonal antibodies to capture either full-length or truncated PrP species and then used biotinylated lectin to compare the nature of the glycans on bound PrP species present in control, sporadic Creutzfeldt-Jakob disease (sCJD), or variant CJD (vCJD) brains. When full-length PrP species in these three groups were compared, no significant difference in the binding of concanavalin A or Aleuria aurantia lectin was detected. However, the binding of Ricinus communis agglutinin I (RCA) to sCJD and vCJD samples was significantly increased. In contrast, when only truncated PrP species were compared, only vCJD samples had more RCA binding activity. Therefore, while most of the RCA binding activity in sCJD is restricted to the full-length PrP species, the RCA binding activity in vCJD is associated with truncated and full-length PrP species. Furthermore, the RCA binding activity in sCJD and vCJD samples is mostly associated with proteinase K-resistant PrP species, a known signature of infectious prion. Therefore, PrP species in sCJD and vCJD have dissimilar lectin immunoreactivity, which reflects differences in their N-linked glycans. These differences may account for the distinct phenotypes of sCJD and vCJD.
INTRODUCTION
According to the protein-only hypothesis, the central event in the pathogenesis of prion disease is the conversion of a normal cellular prion protein, PrPC, into an abnormal, pathogenic, scrapie isoform, PrPSc (21, 22). Accumulation of PrPSc in the brain is the cause of prion disease. The PrPC-to-PrPSc conversion is based on a change in conformation from a predominantly -helical structure that characterizes PrPC to a predominantly -sheet structure that characterizes PrPSc (14, 23). However, the molecular events that govern PrPC-to-PrPSc conversion are complex and not completely understood. One major consequence of the conformational change is that while PrPC is highly sensitive to digestion by proteinase K (PK), the conformational change renders the PrPSc relatively more resistant to PK digestion (1, 13). Therefore, resistant to PK digestion has been used as an in vitro diagnostic hallmark of all prion diseases.
It is generally believed that in the human brain, PrPC species are present in three glycoforms: diglycosylated, monoglycosylated, and unglycosylated. More recently, using a panel of anti-PrPC monoclonal antibodies (MAbs) that react with different regions of the PrPC, we found that the expression of PrPC in normal brain is far more complex than the three-band pattern previously seen with MAb 3F4 (15, 16, 32). By two-dimensional immunoblotting, we observed seven major species of PrPC, which could be subdivided into multiple subspecies (16). These PrP species were generated as a result of differences in N-linked glycosylation as well as the site of truncation (16). On the basis of this evidence, we concluded that most of the smaller PrP species are N-terminally truncated PrP species rather than unglycosylated or monoglycosylated PrP species. Furthermore, by separation of the full-length and truncated PrPC, the two populations were shown to be glycosylated differentially (16). We speculated that full-length and truncated PrPC might have different functions under physiological conditions and may play different roles in PrPSc formation (15, 17).
Earlier studies have demonstrated that the N-linked glycans in PrP are of the complex type, which is resistant to endoglycosidase H but sensitive to peptide-N-glycosidase F (PNGase F) (4, 6, 25, 30). The N-linked glycans contain terminal galactose, sialic acid, and fucose attached to the innermost N-acetyl-glucosamine cores (6). More-detailed glycostructures of PrP have come from studies using mass spectrometry with purified hamster and murine PrP species (25, 29, 30). It was found that both of the N-linked glycosylation sites can be glycosylated; over 30 glycostructures were identified (29, 30). Recently, a comparative study was done between normal PrPC and PrPSc from hamsters (25). It was found that PrPSc and PrPC contain similar glycostructures. However, PrPSc contained proportionally more branched glycostructures, such as bisected structures and tri- and tetra-antennary structures than PrPC (25).
The biological significance of glycosylation in PrPC function and the pathogenesis of prion diseases is not completely understood. We and others have reported that different PrPC glycoforms are present in different regions of the CNS (3, 11, 28). It has been speculated that the glycosylation in PrPC could determine the region of PrPSc targeting and accumulation (7). Therefore, the nature of the N-linked glycans in PrP might be closely associated with the lesion profiles and the typing of PrPSc strains. In the cell model, some glycosylated PrP mutants can reach the cell membrane; however, unglycosylated mutant PrP species tend to accumulate intracellularly and form PrPSc-like products (9, 24). PrPC without N-linked glycans has been reported to accumulate inside the cell in the Golgi instead of being exported to the cell membrane, and the unglycosylated PrP exhibited certain biochemical features characteristic of PrPSc (9, 24). In studies with N2aSc cells, a PrPSc-infected cell line, it has been reported that inhibition of N-linked glycan synthesis promotes the accumulation of PrPSc (31).
Over recent years, we have developed a large panel of anti-PrPC MAbs (10, 32). The epitopes of these MAbs spread from the N terminus to the C terminus of PrPC. In the studies described here, we used a novel antibody-lectin enzyme-linked immunosorbent assay (ELISA) to compare the natures of the N-linked glycans on different PrP species in control, sporadic Creutzfeldt-Jakob disease (sCJD), or variant CJD (vCJD) brain homogenates. We describe the use of different anti-PrP MAbs to capture either full-length or truncated PrP species in the brain homogenates and biotin-conjugated lectins to detect the presence of different N-glycans on bound PrP species. Our results clearly demonstrated that when compared to non-CJD controls, PrP species in sCJD and vCJD brains have different lectin immunoreactivities.
MATERIALS AND METHODS
Tissues. For controls, brain autopsies from the frontal cortex were obtained from individuals (n = 5) confirmed negative for prion disease. For sCJD, tissues were obtained from the frontal cortices of individuals with autopsy-confirmed cases (n = 5) with sporadic prion disease that were M/M at amino acid residue 129 and had the PrPSc type 2. For vCJD, frontal cortex was obtained from cases (n = 4) confirmed at autopsy. All samples from sCJD and vCJD cases had PK-resistant PrPSc, as demonstrated by immunoblotting with anti-PrP MAb 8H4 (not shown)
Preparation of brain homogenate. Human brain tissues were homogenized in 10 volumes of ice-cold lysis buffer (10 mM Tris, 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 5 mM EDTA [pH 8.0]) in the presence of 1 mM phenylmethylsulfonyl fluoride. If the homogenate was to be treated with PK, phenylmethylsulfonyl fluoride was omitted from the lysis buffer. After microcentrifugation at 3,000 rpm for 10 min, the supernatants were stored in aliquots at –80°C.
Anti-PrPC MAbs. The generation and characterization of anti-PrPC MAbs have been previously described in great detail (10, 32). MAb 8B4 and MAb 8H4 are immunoglogulin G1, and the epitopes recognized by these MAbs are diagrammatically presented in Fig. 1. While the epitope recognized by MAb 8H4 includes the second potential N-linked glycosylation site, MAb 8H4 reacts equally well with recombinant PrP and native PrPC (32). Thus, the presence of N-linked glycans does not impede the binding of MAb 8H4.
MAb-lectin ELISA. Initially, we carried out titration experiments to identify a concentration of the capture MAb at the linear portion of the PrP binding curve with a specific amount of brain homogenate. Therefore, the amount of plate-coated MAb is limiting, while the level of PrP species is in excess. This is to ensure that the amounts of captured PrP species from various samples are similar; thus, the differences in ELISA readings would not be due to quantitative difference in the amounts of captured PrP but would rather reflect the difference in lectin binding.
Ninety-six-well ELISA plates (Bio-Rad) were coated with affinity-purified MAb 8H4 (40 ng) or MAb 8B4 (20 ng) in 100 μl at 37°C for 3 h and blocked with 3% bovine serum albumin in phosphate-buffered saline (PBS) at 4°C overnight. To decrease the background, the plates were incubated with 0.2 ml of oxidation buffer (100 mM sodium periodate, 50 mM citric acid, [pH 4.0])/well for 15 min. Plates were then washed (four times) with washing buffer (PBS with 0.05% Tween 20). Two hundred micrograms of total brain proteins in 100 μl of PBS was added to the plate in duplicate and incubated at room temperature for 3 h. After four additional washes with washing buffer, 100 μl of different biotinylated lectins was added at a concentration of 2 μg/ml in washing buffer and incubated at room temperature for 2 h. The lectins were detected with a strepavidin-conjugated horseradish peroxidase (PharMingen, San Diego, Calif.). The lectins used include Aleuria aurantia lectin (ALL), specific for (-1,3)- and (-1,6)-linked fucose; concanavalin A (ConA), specific for -linked mannose; and Ricinus communis agglutinin I (RCA I), specific for galactose (Vector Laboratories, Burlingame, Calif.). Optical density values at 405 nm were measured with ABTS [2,2-azino-di-(3-ethylbenzthiazoline sulfonic acid] ammonium salt solution] (Roche Diagnostic, Indianopolis, Ind.) with an automated plate reader (Molecular Dynamics). The results presented were the average of duplicates and all experiments were repeated at least three times.
To examine the binding specificity of RCA to PrP, a competitive ELISA was also performed: glucose, sucrose, galactose and lactose (Sigma, St. Louis, Mo.) of different concentrations were included to compete with plate-immobilized PrP for the binding of biotinylated RCA.
Treatment of brain homogenates with PNGase F or guanidine hydrochloride (GdHCl). For PNGase F treatment, each brain preparation was mixed with 0.1 volume of 10x denaturing buffer (20 mM Tris, 150 mM NaCl, 2 mM EDTA, 10% mercaptoethanol, 5% sodium dodecyl sulfate [pH 7.5]) and heated for 10 min at 95°C. PNGase F digestion was carried out at 37°C for 1 h with the addition of 1% N40 and recombinant PNGase F (New England BioLabs) with a total of 1,000 U per 50 μl of reaction mixture, as recommended by the supplier.
To prepare denatured PrP species, each brain homogenate (4 μl of the 10% brain homogenate) was incubated with an equal volume of 4 M GdHCl (Sigma, St. Louis, Mo.) at room temperature for 10 min. The treated brain homogenate was then diluted with PBS to a final GdHCl concentration of 0.05 M. At this concentration, the effect of GdHCl on the interaction between PrP and anti-PrP MAb is minimal.
RESULTS
Anti-PrPC MAb-lectin ELISA: full-length PrP species from sCJD and vCJD cases have more RCA immunoreactivity. In an anti-PrPC MAb-lectin ELISA, an immobilized anti-PrPC MAb was used to capture the PrP species in the homogenate, and a biotin-conjugated lectin was then used to react with the bound PrP species. The epitopes recognized by the MAbs used in this study, MAb 8B4 and 8H4, are presented diagrammatically in Fig. 1. When immobilized on the plate, MAb 8B4 captures only the full-length PrP species (16). On the other hand, MAb 8H4 captures all described PrP species including the full-length PrP and the N-terminally truncated PrP species (16).
We first sought to confirm that the captured PrP species indeed reacted with lectin in a manner that was specific for both PrP and the sugar moiety. We demonstrated that the binding of biotinylated lectins (AAL, ConA, and RCA) to MAb 8B4-captured PrP species is PrP dosage dependent and proportional to the amount of PrP immobilized onto the plate (Fig. 2A). Furthermore, the binding of RCA to MAb 8B4-captured PrP was efficiently inhibited by appropriate sugars such as galactose and lactose, but not glucose and sucrose (Fig. 2B). Therefore, the binding of RCA is specific for galactose residues on PrP. Similar studies were performed to verify the specificity of AAL and ConA (data not shown).
Human brain homogenate contains both full-length PrP species and multiple truncated PrP species (16). Initially, we compared the lectin binding activity of the full-length PrP species present in brain homogenates from non-CJD controls (n = 4) and sCJD (n = 5) and vCJD (n = 4) donors, using immobilized MAb 8B4 to capture only the full-length PrP species.
The various lectins used in this study and their major binding specificities are shown in Table 1. All three groups of homogenates (non-CJD, sCJD, and vCJD) have comparable AAL and ConA binding activities (Table 1). In contrast, we consistently observed significantly more RCA binding activity in the full-length PrP species from sCJD and vCJD donors than the full-length PrP species from non-CJD controls (Table 1). These results provide evidence that the N-linked glycans on full-length PrP species present in sCJD and vCJD brains may be different from those that are present on full-length PrP species in non-CJD brain.
We verified that the assay indeed detects N-linked glycans on PrP species by removing the N-linked glycans with PNGase F prior to performing the MAb-lectin ELISA. After PNGase treatment, brain homogenates from all three groups lost their lectin immunoreactivity (Table 2). This experiment provided conclusive evidence that the anti-PrPC MAb-lectin ELISA is specific for the N-linked glycans. We also depleted the full-length PrP species in the brain homogenates with MAb 8B4-conjugated beads prior to the MAb 8B4-lectin ELISA. Removal of the full-length PrP species also reduced the binding of all three lectins to background levels (results not shown). Collectively, these results provide strong evidence that indeed the MAb 8B4-lectin ELISA detects N-linked glycans present on full-length PrP species.
The major target of RCA is galatose. Full-length PrP species from sCJD and vCJD may have more galatose residues. Alternatively, due to conformational differences between PrP species, more galatose residues on sCJD and vCJD may be exposed. To test these possibilities, each brain homogenate was treated with a denaturing agent, GdHCl, prior to carrying out the MAb-lectin ELISA. After GdHCl treatment, both the ConA and ALL binding activities were significantly increased in all three groups (Table 2). In contrast, identical treatment did not significantly alter the binding profiles of RCA. RCA immunoreactivity remained significantly higher in sCJD and vCJD samples than in non-CJD control samples. These results favor the interpretation that the reason full-length PrP species in sCJD and vCJD have more RCA immunoreactivity is that they have more galatose residues.
RCA immunoreactivity in sCJD sample is restricted to the full-length PrP species, while in vCJD it is associated with both full-length and truncated PrP species. Subsequently, we investigated whether RCA immunoreactivity is also present in the truncated PrP species in sCJD and vCJD brains. Brain homogenates from non-CJD, sCJD, and vCJD donors were first immunoprecipitated with MAb 8B4-conjugated beads to deplete the full-length PrP species as previously described by us (16). In these experiments, the efficiency of depletion was verified by immunoblotting the depleted sample with MAb 8H4. After immunoprecipitation with MAb 8B4 beads, the full-length PrP species was not detectable in control, sCJD, and vCJD brain homogenates (Fig. 3A). The depletion of full-length PrP species was also confirmed by conventional capture ELISA, using MAb 8B4 as the capture antibody and MAb 8H4 as a detector. After depletion with MAb 8B4 beads, PrP immunoreactivity was reduced to background levels in all samples (Fig. 3B). Furthermore, before depletion, the levels of PrP immunoreactivity in all samples were comparable, indicating that similar amounts of full-length PrP species were captured and that the capture antibody in the assay was indeed the limiting factor.
After depletion of the full-length PrP species, RCA immunoreactivity associated with the truncated PrP species in control and sCJD samples was reduced to levels that were indistinguishable (Table 3). On the other hand, removal of the full-length PrP species from the vCJD samples did not reduce RCA immunoreactivity. RCA binding activity in vCJD samples remained significantly higher than in the control and sCJD homogenates. The increase in RCA binding ranged from 300 to 450% in three different experiments. Accordingly, RCA binding activity in sCJD homogenate was mainly associated with full-length PrP species. In contrast, RCA binding activity in vCJD homogenate was associated with full-length and truncated PrP species.
A large proportion of the RCA immunoreactivity is associated with PK-resistant PrP species in sCJD and vCJD homogenates. One feature that distinguishes PrPC from PrPSc is that while PrPC is sensitive to PK, the pathogenic PrPSc is resistant (1, 13). Consequently, we investigated whether RCA immunoreactivity in sCJD and vCJD samples was associated with PK-sensitive or PK-resistant PrP species. Brain homogenates from controls (n = 4), sCJD (n = 4), and vCJD (n = 4) samples were treated in vitro with PK as described in Materials and Methods. The presence of PK-resistant PrP species in sCJD and vCJD brain homogenates but not in non-CJD brain homogenates was confirmed by immunoblotting the PK-treated brain homogenates with MAb 8H4 (results not shown).
Since PK treatment eliminates the MAb 8B4 epitope, MAb 8H4 was used as the capture reagent in MAb-lectin ELISA experiments. As expected, PK treatment completely eliminated ConA and RCA immunoreactivity in all four non-CJD control samples (results not shown). However, identical PK treatment did not significantly alter the binding of ConA to either sCJD or vCJD homogenates (Fig. 4). On the other hand, PK treatment significantly increased the binding of RCA to sCJD as well as the vCJD samples (Fig. 4). The increases in RCA binding ranged from 37% (sample sCJD-3) to 363% (sample vCJD-4). The increase in RCA immunoreactivity was most likely due to the elimination of PrPC in the brain homogenates by PK digestion.
Therefore, most if not all of the RCA immunoreactivity in sCJD and vCJD brain homogenates is associated with PK-resistant PrP species, a known signature of PrPSc.
DISCUSSION
In this paper, we describe four new findings that not only have implications for the pathogenesis of human prion diseases, but may also provide the basis for the development of a differential diagnostic test for sCJD and vCJD. First, we found that the nature of the N-linked glycans detected on full-length PrP species from sCJD and vCJD brains are different from those present on full-length PrP species from non-CJD brains. This conclusion is based on our findings that full-length PrP species from sCJD and vCJD samples reacted more strongly to the lectin RCA. Second, the increase in RCA binding is not caused by conformational differences between PrP species but more likely reflects a quantitative difference in the levels of galatose residue. Third, while the increase in RCA immunoreactivity in the sCJD brain is exclusively associated with full-length PrP species, the increase in RCA binding activity in vCJD by contrast is associated with both full-length and truncated PrP species. Fourth and most importantly, the increase in RCA immunoreactivity in sCJD and CJD brains is mostly associated with PK-resistant PrP species, a signature of PrPSc. These differences may account for the phenotypic differences, such as disease onset, histopathology, and disease duration, between sCJD and vCJD.
Three lines of evidence support our interpretation that the MAb-lectin ELISA indeed detects N-linked glycans present on PrP. (i) Removal of the N-linked glycans with PNGase completely eliminated lectin immunoreactivity (Table 2). (ii) The lectin does not react with bacterially produced recombinant PrP, which lacks the N-linked glycans (results not shown). (iii) It does not react with human brain homogenates in which the PrP species have been first depleted with anti-PrP MAb 8H4 affinity chromatography (results not shown).
The binding specificities of lectins have been investigated extensively since the 1940s. In general, each lectin has a binding specificity similar to the binding of antibody to its antigen. With few exceptions, most lectins interact with the nonreducing, terminal glycosyl groups of polysaccharide and glycoprotein chain ends. An exception is ConA, which binds to terminal 2-O- mannopyranosyl residues as well as -mannose in the internal-core oliogosaccharide structure. AAL binds preferentially to terminal fucose that is linked (1-6) to N-acetylglucosamine or to fucose that is linked (1-3) to N-acetyllactosamine-related structures. RCA-1 reacts preferentially to oligosaccharides ending in -galactose, much more weakly to -galactose, and only marginally to N-acetygalactoseamine.
In our earlier studies with affinity-purified PrP species from non-CJD brains, we found that the PrP species bound ConA and AAL the most, followed by Sambucus nigra lectin and then RCA (16). In this study, we observed more RCA binding activity in the full-length PrP species from sCJD and vCJD brains than non-CJD controls. We concluded that full-length PrP species from sCJD and vCJD have more galactose residues. This interpretation is supported by our findings that treatment with a denaturing agent, GdHCl, did not significantly alter the binding profiles of RCA. After GdHCl treatment, both sCJD and vCJD samples still had higher levels of RCA immunoreactivity than similarly treated non-CJD controls. However, a more-detailed biochemical analysis of the purified full-length PrP species will be required to verify this point. Interestingly, while treatment with GdHCl did not alter the binding profiles of RCA, it did increase binding of ConA and AAL. These results suggest that treatment with GdHCl exposed many of the ConA and AAL binding sites on full-length PrP species that are normally unavailable for binding.
It has been known for decades that infectious prions can be propagated in susceptible hosts in a strain-specific manner, similar to that of conventional pathogens (2, 26). Most recent studies have suggested that each prion strain represents a PrPSc molecule with a unique conformation (20, 27). It has also been recognized that the size of the human PrPSc fragment that is resistant to PK treatment varies depending on the type of prion disease. Essentially, two major strains of PrPSc have been identified in human prion disease, based on the electrophoretic mobility of the main PrPSc fragment generated by PK digestion (18, 19). Type 1 migrates at 21 kDa and type 2 migrates at 19 kDa, following gel electrophoresis (18). More recently, it has been suggested that at least six human prion strains can be identified, based on their electrophoretic mobility (8). Interestingly, all vCJD patients belong to the type 2 disease group and are homozygous with M/M at residue 129. Either people with M/M are more susceptible to vCJD or these individuals have a much shorter incubation period.
We found that full-length PrP species from sCJD and vCJD donors bound more RCA. Unlike sCJD, however, the RCA binding activity in vCJD is associated with full-length and truncated PrP species. In this study, we divided PrP species into either full-length or truncated species for simplicity. We reported previously that while non-CJD control brain homogenate has only one full-length PrP species, there are at least three full-length PrP species in sCJD samples (15). Whether the RCA binding activity in sCJD and vCJD is associated with the same subpopulations of full-length PrP species is not known. We also do not know whether the RCA binding activity in vCJD brain is present in one or more subpopulations of the truncated PrP species.
PrPSc tend to be insoluble in certain detergents (5). We analyzed the supernatant fraction of detergent solubilized brain homogenates. Therefore, one may argue that our findings actually reflect differential recovery of PrPSc in the supernatant fractions from sCJD and vCJD samples, rather than revealing molecular differences in the structures of the N-linked glycans on PrP species. We consider this scenario unlikely, because in the MAb-lectin ELISA, the capture antibody was always the limiting factor. Therefore, comparable amounts of PrP species in different samples were captured (Fig. 2B). Again, this issue can be resolved with a detailed biochemical analysis of the purified individual PrP species.
Our observations that PrP species in sCJD and vCJD brains have different lectin immunoreactivity provide suggestive evidence that the strain of PrPSc which causes sCJD is distinct from the strain of PrPSc that causes vCJD. Our MAb-lectin ELISA results are also in accord with our recent findings using two-dimensional immunoblotting, in which we found that PrPSc species from sCJD and vCJD patients have distinct N-linked glycans (T. Pan, et al., unpublished data). Furthermore, since most of the RCA immunoreactivity is associated with PK-resistant PrP species, we conclude that the detectable RCA immunoreactivity must be present on PrPSc, a signature of prion diseases. Interestingly, our findings that human PrPSc and PrPC have different glycans are in good agreement with an earlier report by Manuelidis and her colleagues (12). Using highly purified scrapie-associated fibrils from scrapie-infected hamster brains and lectins in immunoblots, these investigators found that RCA but not ConA preferentially reacted with the scrapie-associated fibrils (12).
In summary, our results suggest that the nature of the glycans and the PrP species may dictate the phenotypes of prion diseases or the strains of prion. The unique glycan may allow the PrPSc strain to target a subpopulation of PrPC to convert. It is also possible that the infectivity of different strains of prion may be associated with different PrP species, such as full-length versus truncated PrP species. It has been suggested that each strain of prion represents a PrPSc molecule with a distinct conformation (20, 27). The results presented here suggest that the nature of the N-linked glycans may contribute to this difference and provide a mechanism by which prion strains diversify. So far, we have only studied sCJD and vCJD cases. All the cases were homozygous with methionine in position 129 and have type 2 diseases. Whether our observations could be generalized to other CJD diseases, such as those that are heterozygous at 129 or with type 1 diseases, or to inherited prion diseases remains to be determined. Furthermore, it has been speculated that vCJD most likely represents the transmission of the disease from bovine spongiform encephalopathy-affected cattle to humans. It will be important to determine whether our observations made with vCJD brain homogenates could also be observed in brain homogenates from bovine spongiform encephalopathy-affected cattle. Finally, all the current in vitro diagnostic tests for prion diseases in humans are time consuming and nonquantitative. The assay described here is simple and quantitative. It may provide the basis for the development of a differential test for human prion diseases and the typing of different prion strains.
ACKNOWLEDGMENTS
We thank Geoff Barnard, Neil Greenspan, and Michael Lamm for careful reading and insightful discussion of the manuscript. We also thank Pierluigi Gambetti for support and the National Prion Disease Pathology Surveillance Center for providing the normal controls and sCJD samples.
This work is supported in part by NIH grant NS-045981-01 and U.S. Army grant NP0-20030 (M.-S.S.).
REFERENCES
Bolton, D. C., M. P. McKinley, and S. B. Prusiner. 1982. Identification of a protein that purifies with the scrapie prion. Science 218:1309-1311.
Bruce, M., H. Fraser, P. Mcbride, J. Scott, and A. Dickinson. 1992. The basis of strain variations in scrapie, p. 497-508. In S. B. Prusiner (ed.), Prion diseases in humans and animals. Ellis Horwood, London, United Kingdom.
DeArmond, S. J., Y. Qiu, H. Sanchez, P. R. Spilman, A. Ninchak-Casey, D. Alonso, and V. Daggett. 1999. PrPc glycoform heterogeneity as a function of brain region: implications for selective targeting of neurons by prion strains. J. Neuropathol. Exp. Neurol. 58:1000-1009.
Endo, T., D. Groth, S. B. Prusiner, and A. Kobata. 1989. Diversity of oligosaccharide structures linked to asparagines of the scrapie prion protein. Biochemistry 28:8380-8388.
Gabizon, R., M. P. McKinley, and S. B. Prusiner. 1987. Purified prion proteins and scrapie infectivity copartition into liposomes. Proc. Natl. Acad. Sci. USA 84:4017-4021.
Haraguchi, T., S. Fisher, S. Olofsson, T. Endo, D. Groth, A. Tarentino, D. R. Borchelt, D. Teplow, L. Hood, A. Burlingame, et al. 1989. Asparagine-linked glycosylation of the scrapie and cellular prion proteins. Arch. Biochem. Biophys. 274:1-13.
Hecker, R., A. Taraboulos, M. Scott, K. M. Pan, S. L. Yang, M. Torchia, K. Jendroska, S. J. DeArmond, and S. B. Prusiner. 1992. Replication of distinct scrapie prion isolates is region specific in brains of transgenic mice and hamsters. Genes Dev. 6:1213-1228.
Hill, A. F., S. Joiner, J. D. Wadsworth, K. C. Sidle, J. E. Bell, H. Budka, J. W. Ironside, and J. Collinge. 2003. Molecular classification of sporadic Creutzfeldt-Jakob disease. Brain 126:1333-1346.
Lehmann, S., and D. A. Harris. 1997. Blockade of glycosylation promotes acquisition of scrapie-like properties by the prion protein in cultured cells. J. Biol. Chem. 272:21479-21487. (Erratum, 273:5988, 1998.)
Li, R., T. Liu, B.-S. Wong, T. Pan, M. Morillas, W. Swietnicki, K. O'Rourke, P. Gambetti, W. K. Surewicz, and M.-S. Sy. 2000. Identification of an epitope in the C-terminus of normal prion protein whose expression is modulated by binding events in the N-terminus J. Mol. Biol. 301:567-574.
Liu, T., T. Zwingman, R. Li, T. Pan, B. S. Wong, R. B. Petersen, P. Gambetti, K. Herrup, and M.-S. Sy. 2001. Differential expression of cellular prion protein in mouse brain as detected with multiple anti-PrP monoclonal antibodies. Brain Res. 896:118-129.
Manuelidis, L., S. Valley, and E. E. Manuelidis. 1985. Specific proteins associated with Creutzfeldt-Jakob disease and scrapie share antigenic and carbohydrate determinants. Proc. Natl. Acad. Sci. USA 82:4263-4267.
Oesch, B., D. Westaway, M. Walchli, M. P. McKinley, S. B. Kent, R. Aebersold, R. A. Barry, P. Tempst, D. B. Teplow, and L. E. Hood. 1985. A cellular gene encodes scrapie PrP 27-30 protein. Cell 40:735-746.
Pan, K. M., M. Baldwin, J. Nguyen, M. Gasset, A. Serban, D. Groth, I. Mehlhorn, Z. Huang, R. J. Fletterick, F. E. Cohen, and S. B. Prusiner. 1993. Conversion of alpha-helices into beta-sheets features in the formation of the scrapie prion proteins. Proc. Natl. Acad. Sci. USA 90:10962-10966.
Pan, T., M. Colucci, B. S. Wong, R. Li, T. Liu, R. B. Petersen, S. Chen, P. Gambetti, and M.-S. Sy. 2001. Novel differences between two human prion strains revealed by two-dimensional gel electrophoresis. J. Biol. Chem. 276:37284-37288.
Pan, T., R. Li, B. S. Wong, T. Liu, P. Gambetti, and M.-S. Sy. 2002. Heterogeneity of normal prion protein in two-dimensional immunoblot: presence of various glycosylated and truncated forms. J. Neurochem. 81:1092-1101.
Pan, T., B. S. Wong, T. Liu, R. Li, R. B. Petersen, and M.-S. Sy. 2002. Cell-surface prion protein interacts with glycosaminoglycans. Biochem. J. 368:81-90.
Parchi, P., R. Castellani, S. Capellari, B. Ghetti, K. Young, S. G. Chen, M. Farlow, D. W. Dickson, A. A. Sima, J. Q. Trojanowski, R. B. Petersen, and P. Gambetti. 1996. Molecular basis of phenotypic variability in sporadic Creutzfeldt-Jakob disease. Ann. Neurol. 39:767-778.
Parchi, P., A. Giese, S. Capellari, P. Brown, W. Schulz-Schaeffer, O. Windl, I. Zerr, H. Budka, N. Kopp, P. Piccardo, S. Poser, A. Rojiani, N. Streichemberger, J. Julien, C. Vital, B. Ghetti, P. Gambetti, and H. Kretzschmar. 1999. Classification of sporadic Creutzfeldt-Jakob disease based on molecular and phenotypic analysis of 300 subjects. Ann. Neurol. 46:224-233.
Peretz, D., M. R. Scott, D. Groth, R. A. Williamson, D. R. Burton, F. E. Cohen, and S. B. Prusiner. 2001. Strain-specified relative conformational stability of the scrapie prion protein. Protein Sci. 10:854-863.
Prusiner, S. B. 1982. Novel proteinaceous infectious particles cause scrapie. Science 216:136-144.
Prusiner, S. B. 1998. Prions. Proc. Natl. Acad. Sci. USA 95:13363-13383.
Prusiner, S. B., M. P. McKinley, K. A. Bowman, D. C. Bolton, P. E. Bendheim, D. F. Groth, and G. G. Glenner. 1983. Scrapie prions aggregate to form amyloid-like birefringent rods Cell. 35:349-358.
Rogers, M., A. Taraboulos, M. Scott, D. Groth, and S. B. Prusiner. 1990. Intracellular accumulation of the cellular prion protein after mutagenesis of its Asn-linked glycosylation sites. Glycobiology 1:101-109.
Rudd, P. M., T. Endo, C. Colominas, D. Groth, S. F. Wheeler, D. J. Harvey, M. R. Wormald, H. Serban, S. B. Prusiner, A. Kobata, and R. A. Dwek. 1999. Glycosylation differences between the normal and pathogenic prion protein isoforms. Proc. Natl. Acad. Sci. USA 96:13044-13049.
Safar, J. 1990. Scrapie isoform of scrapie-associated protein. J. Am. Vet. Med. Assoc. 196:1681.
Safar, J., F. E. Cohen, and S. B. Prusiner. 2000. Quantitative traits of prion strains are enciphered in the conformation of the prion protein. Arch. Virol. Suppl. 16:227-235.
Somerville, R. A., and L. A. Ritchie. 1990. Differential glycosylation of the protein (PrP) forming scrapie-associated fibrils. J. Gen. Virol. 71:833-839.
Stahl, N., M. A. Baldwin, D. B. Teplow, L. Hood, B. W. Gibson, A. L. Burlingame, and S. B. Prusiner. 1993. Structural studies of the scrapie prion protein using mass spectrometry and amino acid sequencing. Biochemistry 32:1991-2002.
Stimson, E., J. Hope, A. Chong, and A. L. Burlingame. 1999. Site-specific characterization of the N-linked glycans of murine prion protein by high-performance liquid chromatography/electrospray mass spectrometry and exoglycosidase digestions. Biochemistry 38:4885-4895.
Winklhofer, K. F., U. Heller, A. Reintjes, and J. Tatzelt. 2003 Inhibition of complex glycosylation increases the formation of PrPsc. Traffic 4:313-322.
Zanusso, G., D. Liu, S. Ferrari, I. Hegyi, X. Yin, A. Aguzzi, S. Hornemann, S. Liemann, R. Glockshuber, J. C. Manson, P. Brown, R. B. Petersen, P. Gambetti, and M.-S. Sy. 1998. Prion protein expression in different species: analysis with a panel of new mAbs. Proc Natl Acad Sci USA 95:8812-8816.(Tao Pan, Ruliang Li, Boon)