|Home | About | Journals | Submit | Contact Us | Français|
Giardia lamblia is present in the intestinal lumen as a binucleate, flagellated trophozoite or a quadranucleate, immotile cyst. Here we used the plant lectin wheat germ agglutinin (WGA), which binds to the disaccharide di-N-acetyl-chitobiose (GlcNAc2), which is the truncated Asn-linked glycan (N-glycan) of Giardia, to affinity purify the N-glycomes (glycoproteins with N-glycans) of trophozoites and cysts. Fluorescent WGA bound to the perinuclear membranes, peripheral acidified vesicles, and plasma membranes of trophozoites. In contrast, WGA bound strongly to membranes adjacent to the wall of Giardia cysts and less strongly to the endoplasmic reticulum and acidified vesicles. WGA lectin-affinity chromatography dramatically enriched secreted and membrane proteins of Giardia, including proteases and acid phosphatases that retain their activities. With mass spectroscopy, we identified 91 glycopeptides with N-glycans and 194 trophozoite-secreted and membrane proteins, including 42 unique proteins. The Giardia oligosaccharyltransferase, which contains a single catalytic subunit, preferred N glycosylation sites with Thr to those with Ser in vivo but had no preference for flanking amino acids. The most-abundant glycoproteins in the N-glycome of trophozoites were lysosomal enzymes, folding-associated proteins, and unique transmembrane proteins with Cys-, Leu-, or Gly-rich repeats. We identified 157 secreted and membrane proteins in the Giardia cysts, including 20 unique proteins. Compared to trophozoites, cysts were enriched in Gly-rich repeat transmembrane proteins, cyst wall proteins, and unique membrane proteins but had relatively fewer Leu-rich repeat proteins, folding-associated proteins, and unique secreted proteins. In summary, there are major changes in the Giardia N-glycome with the differentiation from trophozoites to cysts.
Giardia lamblia is a deeply divergent protist, which causes diarrhea (1, 32). While Giardia is a tremendous problem in the developing world where hygiene is inadequate to block its transmission by the fecal-oral route, some two million Americans are infected each year with Giardia, which is present in streams and lakes or is transmitted in day-care centers (46).
Binucleate trophozoites of Giardia are motile by means of eight flagella and are adherent by means of a ventral disc (13, 41). Secreted and membrane proteins of Giardia are present in the endoplasmic reticulum (ER); in peripheral vesicles, which have features in common with lysosomes; and in the plasma membrane, which has variant-specific surface proteins (VSPs) that are rich in Cys residues (1, 15, 27, 28, 32, 35, 44, 45).
The infectious and diagnostic stage of Giardia is the quadranucleate cyst, the wall of which contains a unique β-1,3-linked GalNAc polymer (12). Three abundant cyst wall proteins (CWP1, CWP2, and CWP3), which are present in encystation-associated vesicles, have a series of Leu-rich repeats (LRRs) and a C-terminal Cys-rich domain (15, 33, 47). CWP1 is the target for diagnostic monoclonal antibodies to Giardia in clinical specimens. There is a complex set of membranes closely apposed to the cyst wall of Giardia (8). Cys-rich proteins resembling VSPs have been implicated in encystation (9), while numerous cysteine proteinases have been associated with either encystation or excystation (50, 54).
We are interested in Asn-linked glycans (N-glycans) and glycoproteins with N-glycans (the so-called N-glycome) of Giardia for many reasons. First, while the vast majority of eukaryotes (animals, fungi, and plants) synthesize N-glycans by means of a lipid-linked precursor containing 14 sugars (Glc3Man9GlcNAc2-PP-dolichol), Giardia makes a lipid-linked precursor with just two sugars (GlcNAc2-PP-dolichol) (16, 42). This is because Giardia lacks Asn-linked glycans (Alg) enzymes, which add Man and Glc to Asn-linked glycans (42). Although secondary loss is the most likely cause of the absence of Alg enzymes from Giardia and other protists, there is presently no good explanation for why so many medically important protists have no N-glycans (microsporidia and Theileria), very short N-glycans containing one or two sugars (Giardia and Plasmodium), or truncated N-glycans containing seven sugars (Entamoeba and Trichomonas) (42).
Second, the oligosaccharyltransferase (OST), which transfers the glycan from the dolichol-linked precursor to the Asn residues on nascent peptides, is composed of four to eight subunits in most eukaryotes but is composed of a single catalytic residue (STT3) in Giardia and Trypanosoma (21). Eukaryotic sequons (sites of N glycosylation) are specified by Asn-Xaa-Thr or Asn-Xaa-Ser (NxT or NxS), where Xaa cannot be Pro, and sequons containing Thr are more frequently occupied (5, 20, 22). The sequons of Campylobacter, which also has a single catalytic peptide in the OST, have a negatively charged residue at the −2 position (Asp- or Glu-Xaa-Asn-Xaa-Ser or -Thr) (24).
Third, Giardia N-glycans, which are predominantly composed of GlcNAc2, are not modified in the ER or Golgi apparatus (42). Indeed, Giardia has a single nucleotide sugar transporter for UDP-GlcNAc, which appears to be involved in the synthesis of glycolipids rather than glycoproteins (2). In contrast, GlcNAc, Gal, Fuc, and NANA are added from nucleotide-sugar donors to make complex metazoan N-glycans (6, 26).
Fourth, Giardia is missing the machinery for N-glycan-dependent quality control of protein folding and degradation, which is present in higher eukaryotes (3, 16, 51). However, Giardia has N-glycan-independent quality control of protein folding in the ER (chaperones, protein disulfide isomerases, and peptidyl-prolyl cis-trans isomerases) and of degradation (Der1, Ccd48, Np14, and Ufd1) (3, 23).
Fifth, wheat germ agglutinin (WGA), which recognizes GlcNAc polymers as well as sialic acid, binds to the surfaces of trophozoites and cysts and blocks encystation (30, 38). Because Giardia has none of the enzymatic machinery to make sialic acid or to transfer sialic acid from host glycoproteins (32; data not shown), WGA likely binds to Giardia N-glycans composed of GlcNAc2 (42). WGA-affinity columns have been used to dramatically enrich for serum glycoproteins (53).
Sixth, while many secreted and membrane proteins are predicted by the whole-genome sequencing of Giardia (32), it is not known how many of these proteins are present in trophozoites or cysts. These secreted and membrane proteins of Giardia are likely important for pathogenesis (e.g., through interaction with the host epithelium or cyst wall formation) and may include novel targets for anti-Giardia vaccines (1, 32, 37).
Four interrelated goals of this study were (i) to localize N-glycans of Giardia trophozoites and cysts, using fluorescent WGA, which binds to GlcNAc2; (ii) to use WGA affinity to enrich Giardia secreted and membrane glycoproteins with N-glycans and test the activity of selective enzymes; (iii) to use mass spectroscopy to identify glycopeptides containing occupied sequons and so determine the recipient peptide specificity of the single-subunit Giardia OST; and (iv) to use mass spectroscopy to identify the Giardia N-glycome in order to compare secreted and membrane proteins of trophozoites and cysts.
Giardia lamblia strains WB (genome project strain) and MR4 were grown axenically in TYI-S-33 medium supplemented with bile salts. Trophozoites were encysted by standard methods (14). Briefly, Giardia trophozoites were cultured in preencystation TYI-S-33 medium lacking bile salts; upon reaching ~60% confluence of adherent trophozoites, the preencystation medium was exchanged for TYI-S-33 medium supplemented with 10 mg/ml bile salts and 5 mM lactic acid hemi-calcium salt, pH 7.8. Nonadherent water-resistant cysts were isolated and washed twice with deionized water. For fluorescence microscopy, cysts of Giardia strain H3, which had been passaged through gerbils, were obtained from Waterborne Incorporated (New Orleans, LA). To further characterize in vitro- or gerbil-derived cysts, excystation was performed, using standard protocols (4).
Logarithmic-phase trophozoites were chilled to release adherent Giardia cells, concentrated by low-speed centrifugation, and washed in chilled phosphate-buffered saline (PBS). For labeling the surfaces of live trophozoites with an amine-reactive fluorescent probe, Giardia strain WB was placed in 1 ml of 200 mM carbonate-bicarbonate buffer (to maximize the number of deprotonated amino groups on the surface) and incubated with 20 μg/ml of Alexa Fluor 488 5-TFP (Molecular Probes) for 60 min on ice (13). Alexa Fluor-labeled Giardia cells were washed two times in PBS and then fixed for 10 min at 4°C in 2% paraformaldehyde in 100 mM phosphate (no saline), pH 7.4. Immediately prior to rinsing, Triton X-100 was added to a final concentration of 0.1% to permeabilize the organisms. Cells were gently pelleted by centrifugation and washed with PBS plus 2% bovine serum albumin (BSA). After being washed with PBS-BSA to quench free aldehydes, Giardia cells were colabeled with 20 μg/ml tetramethyl rhodamine isocyanate-conjugated WGA in PBS-BSA for 30 min at 4°C and then washed twice in PBS. The nuclei of Alexa Fluor- and WGA-labeled Giardia were labeled with 0.1 μg/ml DAPI (4′6-diamidino-2-phenylindole), and SlowFade antifade solution (Invitrogen) was added.
Because Giardia cysts are often very poorly permeable to lectins or antibodies even after being treated with nonionic detergent, Giardia cysts that had been fixed were frozen and thawed three times in order to produce an ice artifact prior to being incubated with WGA and DAPI. This method has been used previously to improve the internal morphology and labeling of Cryptosporidium sporocysts (18) and Entamoeba cysts (our unpublished data).
To label the plasma membrane with tetramethyl rhodamine isocyanate-WGA, live Giardia cells were incubated with WGA in PBS at 4°C, as described, and subsequently rinsed in PBS, fixed in paraformaldehyde (with 0.15% glutaraldehyde), and labeled with DAPI. To compare the localization of WGA with that of acidified vesicles, live Giardia cells were incubated for 90 min at 37°C with 100 nM LysoTracker Red DND-99 (Invitrogen), which is a fluorophor linked to a weak base that accumulates in acidified vesicles. To label the cyst wall, encysting Giardia were incubated with Waterborne Inc.'s Alexa Fluor 488-labeled monoclonal antibody, which is specific for CWP1 (33).
Slides were examined by three-dimensional multiple-wavelength fluorescence microscopy, using an Olympus IX70 microscope equipped for Deltavision deconvolution (Applied Precision). This system employs restoration as well as deconvolution techniques to provide resolutions up to four times greater than those of conventional light microscopes and is used to study the ultrastructure of intracellular structures such as the kinetochore (7, 31). Images were collected at 0.2-mm optical sections for the indicated wavelengths and were subsequently deconvolved, using softWoRx (Applied Precision). Data were examined as either optical sections or as projections of the entire stack.
Adherent logarithmic-phase Giardia and water-resistant cysts were harvested on ice, washed in PBS, and sonicated on an ice-water slurry in the presence of 0.1% Triton X-100 and EDTA-free Complete protease inhibitor cocktail (Roche). Insoluble material was removed by centrifugation (>12,000 × g), and soluble proteins were applied to a WGA-Sepharose column (EY Laboratories, Inc.) (53). Because there was nonspecific binding to this column, Giardia N-linked glycoproteins were selectively eluted with 50 mM tetra-N-acetylchitotriose rather than with SDS. Proteins eluted from the WGA column were electrophoresed on sodium dodecyl sulfate (SDS)-polyacrylamide gels containing a 4 to 20% gradient of acrylamide (Bio-Rad). In parallel lanes were (i) Giardia proteins solubilized with 0.1% Triton X-100 but not applied to the WGA column and (ii) Giardia proteins that were treated twice with 1,000 units of peptide:N-glycanase F (PNGaseF; New England Biolabs) for 9 h each at 37°C in NEB G7 phosphate buffer. SDS-polyacrylamide gels were fixed and stained with silver or Coomassie blue, and proteins were transferred to nitrocellulose membranes by electroporation, incubated with horseradish peroxidase-conjugated WGA, and developed with ECL chemiluminescent substrate (Pierce).
To demonstrate the ability of WGA affinity to enrich for enzymatic activity among the Giardia N-glycomes, trophozoite glycoproteins were prepared as described above, with the following differences. To isolate active proteases and acid phosphatases, we harvested Giardia trophozoites at 0°C in the absence of a protease inhibitor cocktail. Lectin affinity and sample handling were carefully performed to ensure that the material remained on ice or at 4°C. The resulting enriched glycoproteins were run on a zymogram gel containing 10% acrylamide and gelatin (Bio-Rad). Samples were prepared using a nonreducing zymogram sample buffer (Bio-Rad). The resulting gels were rinsed, incubated 1 h in zymogram renaturation buffer, and then incubated at 37°C in 50 mM acetate at pH 5.5 for 2 h. Proteolytic activity was indicated by negative Coomassie blue staining on the zymogram gel. The acid phosphatase activity of the enriched glycoproteins was assayed using an acid phosphatase assay kit (Sigma).
WGA-bound Giardia glycoproteins were run 0.5 cm into the running gel used for SDS-polyacrylamide gel electrophoresis, excised, digested in-gel with sequencing grade trypsin (43), and identified by liquid chromatography-tandem mass spectrometry (LC-MS-MS), using methods that we have used previously to identify the cyst proteins of Entamoeba (52). Briefly, reversed-phase chromatography was carried out using a nano-high-pressure liquid chromatography pump and autosampler (Surveyor and Micro AS, respectively; Thermo Finnigan, San Jose, CA) on a 10 cm by 100-micron-internal diameter MAGIC C18 reversed-phase capillary column (Michrom, Auburn, CA) at the Boston University Proteomics Core Facility or at the MIT Center for Cancer Research Biopolymers Laboratory. Peptides were separated, using gradients of 5% to 90% acetonitrile, over 30 to 120 min in the presence of 0.5% acetic acid (55). Peptides were analyzed using an LTQ ProteomeX ion trap mass spectrometer (Thermo Finnigan), and mass spectra were compared to tryptic digests of Giardia proteins predicted from whole-genome sequencing (32), using SEQUEST and GPM software (www.thegpm.org). All searches were conducted on a reverse database to ensure that the false-positive rate for protein identification was kept below 2%. Tryptic peptides with a SEQUEST XCorr score of >1.5, 2.5, or 3.5 for Z (the charge state of the peptide on the mass spectrometer) = 1, 2, or 3, respectively, and a peptide score of <0.05 were considered a match. GPM-based search scores were based on having log(e) values greater than the cutoff threshold of 2% false positives, typically a log(e) of <−10. Proteins with one or more high-scoring fully tryptic peptides were considered present. In addition, we used a neutral loss of 203.08 Da or 406.16 Da to identify glycopeptides with Asn-linked GlcNAc or GlcNAc2, respectively. The mass spectroscopy data have been submitted to the GiardiaDB website.
Two-dimensional protein gels were simulated from mass spectroscopy data, using the GPM, where the position of each protein was determined by its predicted pI and mass, not including posttranslational modifications, and the size of the spot was proportional to the number of observed ions corresponding to that protein. These two-dimensional gels highlighted the relative abundance of secreted and plasma membrane proteins (defined by either an N-terminal ER-targeting sequence or a transmembrane helix) (25, 36) versus that of nucleocytoplasmic proteins (defined by the absence of these features). The ratio of secreted to nucleocytosolic proteins in a given sample was determined by comparing the SEQUEST area/height results for the set of secreted proteins with the SEQUEST area/height results for the nucleocytosolic proteins.
Giardia proteins identified by mass spectroscopy were compared to annotated Giardia proteins at GiardiaDB and to proteins in the NR database at NCBI. Secreted and membrane proteins were sorted into ER, lysosomes, and plasma membranes based upon their previous characterization in Giardia or in other eukaryotes. Proteins identified by LC-MS-MS were characterized as unique Giardia proteins in cases where (i) there was no similarity to other proteins in the NR database and (ii) the predicted open reading frame was designated “hypothetical protein” in GiardiaDB or the NR database at NCBI.
Hierarchical clustering of the protein data was performed using the Gene Expression Pattern Analysis Suite (GEPAS; http://gepas.bioinfo.cipf.es/) to generate a heat map comparing the stage specificities of the proteins identified by mass spectrometry (17). Briefly, SEQUEST area/height results for all samples were normalized via a Z transformation and submitted as a single data set for cluster analysis. Complete linkage analysis was performed via GEPAS, using a weighted-pair group method (arithmetic averages), and distances were calculated using Spearman's correlation coefficient method. The resulting heat map (see Fig. S1 in the supplemental material) clearly groups samples as either cyst or trophozoite, and the stage-specific identification of individual proteins can be identified by color coding on the map.
An additional stage-specific comparison of identified proteins was performed by hand (see the Excel file in the supplemental material) by listing proteins by their average SEQUEST area/height results or by GPM (X! Tandem) ion intensity for all identified peptides. Stage specificity was estimated by calculating the ratio of the average abundance of proteins associated with trophozoites versus that of cysts (troph/cyst ratio). Proteins with a troph/cyst ratio less than 0.5 were estimated to be predominantly cyst associated, and proteins with a troph/cyst ratio greater than 2 were trophozoite associated.
The first goal here was to compare the binding of fluorescent WGA to that of Alexa Fluor, which binds evenly to the surfaces of the body and flagella of Giardia trophozoites (Fig. 1A and B) (13) or to that of LysoTracker, which is taken up into acidified vesicles by Giardia trophozoites (Fig. 1E and F) (11). The most important observations with three-dimensional high-resolution fluorescence microscopy included the following.
First, the binding of WGA to Giardia trophozoites was remarkably similar from one parasite to the next, so that all of the micrographs of individual Giardia organisms in Fig. Fig.11 are representative of the group. Second, WGA labeled the plasma membranes of nonpermeabilized Giardia trophozoites in a punctate pattern, where the flagella appeared less bright than the body of the Giardia parasite (Fig. (Fig.1C).1C). This punctate pattern of WGA, for which we have no explanation, was distinct from the even labeling of the Giardia surface with Alexa Fluor, which included brightly labeled flagella (Fig. 1A and B) (13). Third, WGA labeled perinuclear membranes and small vesicles, which were distributed throughout the cytosol (Fig. 1A and B). The distribution of WGA labeling was similar to that previously shown for ER markers BiP and protein disulfide isomerase (11, 28, 44). Fourth, WGA more strongly labeled the intracellular vesicles than the plasma membrane. This result is similar to our experience with antiretroviral lectin cyanovirin-N binding to Entamoeba N-glycans, which is greater on the internal membranes of permeabilized trophozoites than on the surface (29). Fifth, WGA bound to the same acidified vesicles in the body of Giardia trophozoites that accumulate LysoTracker (Fig. 1D to F) (11).
This result was shown in five ways. (i) Glycoproteins that bind WGA were dramatically enriched when solubilized proteins of Giardia trophozoites were bound to a WGA-affinity column and eluted with chitotetrose (Fig. (Fig.2B).2B). (ii) N-glycanase treatment of Giardia glycoproteins affinity purified with WGA entirely eliminated the binding of WGA (Fig. (Fig.2E).2E). (iii) WGA affinity dramatically increased the activity of Cys proteases in zymograms of Giardia proteins (Fig. (Fig.2C).2C). Similarly, WGA affinity increased the acid phosphatase activity of Giardia protein lysates by 10-fold (data not shown). (iv) WGA affinity dramatically increased the number and coverage of secreted and membrane proteins identified by mass spectroscopy. In the absence of WGA, most Giardia trophozoite proteins identified by LC-MS-MS were nucleocytosolic (Fig. (Fig.3A).3A). In contrast, after WGA-affinity purification, most proteins identified by LC-MS-MS were secreted or membrane proteins (Fig. (Fig.3B).3B). Please see below for a more-complete description of the Giardia N-glycome.
(v) WGA-affinity chromatography and mass spectroscopy identified 91 glycopeptides containing either N-linked GlcNAc2 or GlcNAc (Fig. (Fig.4;4; see the Fasta file in the supplemental material). In contrast, in the absence of WGA, no glycopeptides with N-linked GlcNAc2 or GlcNAc were identified. In some relatively abundant glycoproteins, where many peptides were identified, either 4 of 4 potential sites of N glycosylation were occupied (Gl50803_8360, a sphingomyelin-like phosphodiesterase), 2 of 2 were occupied (Gl50803_16380, a cathepsin), or 3 of 4 were occupied (Gl50803_9780, a unique secreted trophozoite protein).
All of the Giardia N-glycans identified by mass spectroscopy were present in sequons (Asn-Xaa-Thr or Asn-Xaa-Ser, where Xaa can be any amino acid except Pro), as previously shown for N-glycans of higher eukaryotes (22) (Fig. (Fig.4A).4A). There was no preference by the Giardia OST for a negatively charged amino acid at position −2, as has been shown for the single-subunit OST of bacteria (Fig. (Fig.4B)4B) (24). Sixty-six percent of the occupied Giardia sequons contained Thr, while 48% of the potential sites of N-linked glycosylation contained Thr (Fig. (Fig.4C).4C). This result suggests moderate selection (~40% increase) in vivo for sequons with Thr over those with Ser by the Giardia OST, which contains a single catalytic subunit and transfers GlcNAc2 (5, 20-22, 42).
We compared the data for Giardia with recent data for Caenorhabditis elegans, where 1,191 unique occupied sequons were identified by lectin chromatography and mass spectroscopy of tryptic peptides (Fig. (Fig.4C)4C) (19). There was a modest increase in selection (~20% increase) in vivo for sequons with Thr over those with Ser by the Caenorhabditis OST, which contains 8 subunits and transfers a 14-sugar N-glycan (16, 21, 42).
The WGA affinity-purified glycoproteins (N-glycome) of Giardia trophozoites included 169 secreted and plasma membrane proteins (Fig. (Fig.3,3, ,5,5, and and66 and Table Table1;1; see the Fasta and Excel files in the supplemental material). The 20 most-abundant N-glycome proteins of Giardia trophozoites were each at least 1% of the total proteins identified and together comprised 45% of the proteins identified (Table (Table1).1). Of these 20 glycoproteins, all but 2 were enriched at least twofold more in trophozoites than in cysts (Fig. (Fig.66 and Table Table11).
The most-abundant category of trophozoite glycoproteins was that of enzymes associated with lysosomes in other organisms, which include serine, cysteine, and alanyl peptidases, as well as phosphodiesterases, phospholipases, and phosphatase (Fig. (Fig.3,3, ,5,5, and and66 and Table Table1).1). Some of these enzymes have been localized to peripheral vesicles of Giardia, which were labeled by WGA in this study (48, 49).
The second-most-abundant glycoproteins in the Giardia N-glycome were those involved in the N-glycan-independent quality control of protein folding in the ER lumen, which include protein disulfide isomerases and chaperones (Hsp70 and Hsp90) (Fig. (Fig.3,3, ,5,5, and and66 and Table Table1)1) (3, 11, 23, 28, 44). Folding-associated proteins, which make up 17% of the trophozoite N-glycome, were markedly decreased in cysts, where they make up just 4% of the N-glycome (Fig. (Fig.55 and and6)6) (see below). The abundance of folding-associated proteins in Giardia trophozoites is consistent with the heavy labeling by WGA of perinuclear membranes and small cytosolic vesicles (Fig. (Fig.1),1), where Hsp70 (also known as BiP) and protein disulfide isomerases have been localized (11, 28, 44). As noted in the introduction, Giardia is missing the set of proteins involved in N-glycan-dependent quality control of protein folding (3).
The third-most-abundant glycoproteins in the Giardia N-glycome were membrane proteins, which included multiple LRR proteins and Cys-rich repeat proteins, including VSPs (Fig. (Fig.3,3, ,5,5, and and66 and Table Table1)1) (1, 9, 32, 35). While Cys-rich repeat proteins were also abundant in cysts, LLR proteins made up a larger percentage of the trophozoite N-glycome (12%) than of the cyst N-glycome (5%) (Fig. (Fig.55 and and66).
With WGA-affinity chromatography and mass spectroscopy, we identified peptide sequences for 33 previously “hypothetical proteins,” which we now refer to as unique proteins (Fig. (Fig.3,3, ,5,5, and and66 and Table Table1;1; see the Fasta and Excel files in the supplemental material) (32). In trophozoites, the most-abundant unique proteins contained an N-terminal signal peptide but no transmembrane helices or glycosylphosphatidylinositol anchors (10, 25, 36). Unique secreted proteins accounted for 17% of the N-glycome of Giardia trophozoites, including a glycoprotein (Gl50803_9780) that is 6% of the N-glycome of trophozoites and relatively specific to trophozoites (Table (Table1;1; see the Fasta file in the supplemental material).
Unique transmembrane proteins included a novel family of Gly-rich repeat (GRREAT) proteins, which contain an N-terminal signal peptide and a C-terminal transmembrane helix (see Fig. Fig.8).8). Because the GRREAT proteins were more abundant in the N-glycome of cysts than in that of trophozoites, these unique transmembrane proteins are described below.
The walls of encysting Giardia were densely labeled with a monoclonal antibody to CWP1 (Fig. (Fig.7)7) (33). WGA (Fig. (Fig.7)7) bound strongly to the surfaces of encysted Giardia, which were closely apposed to the cyst wall, where numerous membranes had been localized by electron microscopy (8). In contrast, WGA bound less strongly to perinuclear membranes, peripheral vesicles, and the cyst wall itself. Giardia cysts isolated from the intestines of infected gerbils, which readily excysted in vitro, consistently showed four intact nuclei with DAPI staining (Fig. 7B and C) (4, 54; data not shown). In contrast, numerous Giardia cysts made in vitro often showed degenerative changes, which included WGA binding to structures that had pulled away from the cyst wall, loss of nuclear integrity, and/or failure to excyst in vitro (Fig. (Fig.7A).7A). The tendency for Giardia cysts made in vitro to lose their viability after the cyst wall is formed has been noted previously (14).
The WGA affinity-purified glycoproteins (N-glycome) of Giardia cysts made in vitro included 157 secreted and plasma membrane proteins (Fig. (Fig.55 and and66 and Table Table2;2; see Fig. S1 and the Excel file in the supplemental material). The 20 most-abundant N-glycome proteins of Giardia cysts were each at least 0.8% of the total proteins identified and together comprised 40% of the proteins identified (Table (Table1).1). The relative paucity of cyst wall proteins (CWP1, CWP2, and CWP3) (15, 33, 47) likely results from (i) the difficulty in releasing these proteins from the cyst walls with nonionic detergent and (ii) the relative paucity of trypsin cleavage sites in these proteins, which produce fewer usable peptides for mass spectroscopy.
Dramatic differences in the N-glycome of cysts from that of trophozoites were shown by two-dimensional reconstruction of the mass spectroscopy data of a representative experiment (Fig. (Fig.6)6) and by hierarchical clustering of numerous experiments (see Fig. S1 in the supplemental material). These results, which are summarized in pie charts in Fig. Fig.5,5, are detailed in the Excel file in the supplemental material. Compared to trophozoites, cysts had fewer LRR proteins, folding proteins, and unique secreted proteins (Fig. (Fig.55 and and66 and Tables Tables11 and and2).2). In contrast, cysts were relatively enriched in GRREAT proteins (see next paragraph), cyst wall proteins, and unique transmembrane proteins. In addition, while the N-glycomes of trophozoites and cysts had similar overall percentages of lysosomal enzymes, numerous of the proteases and phospholipases were specific for one stage or the other (Fig. (Fig.55 and and66 and Table Table11).
Among the set of unique Giardia cyst transmembrane proteins were GRREAT proteins, which are type 1 membrane proteins that contain a large region of relatively degenerate Gly-rich repeats (GRR) (Fig. (Fig.8).8). GlGRREAT1 (Gl50803_114210), which comprised ~9% of the N-glycome of Giardia cysts, is 1,088-amino-acids long and shares a C-terminal Cys-rich domain adjacent to the transmembrane helix with other GRREAT proteins.
While the binding of WGA to Giardia cyst walls and the inhibition of encystation by WGA have been shown previously (30, 38), this is the first demonstration that WGA binds to its very short N-glycans, which contains just two GlcNAc; the first fine localization of WGA on the surfaces and interiors of trophozoites and cysts; the first use of WGA to identify large groups of trophozoite and cyst glycoproteins with N-glycans (the N-glycome); and the first demonstration of the abundance of GRREAT proteins in cysts.
This is also the first time that a large number of occupied sequons (sites of N-linked glycans) have been identified from any of the protists. This is the first demonstration that an OST with a single catalytic subunit has the same preference for sequons containing Thr over those containing Ser as the OSTs of metazoa and fungi that contain 7 to 8 peptides (6, 16, 20-22). While the Giardia OST resembles that of bacteria in that both have a single catalytic peptide, the Giardia OST uses the eukaryotic sequon rather than the bacterial sequon, which contains a negative charge in the −2 position (24).
Although N-glycans are added to sequons on many different glycoproteins and all of the sequons on a given Giardia glycoprotein may be occupied, it is not clear what the very short N-glycans of Giardia are doing, as they do not participate in the N-glycan-dependent quality control of protein folding or degradation (3). While there is positive selection for sites of N-linked glycosylation in secreted and membrane proteins of eukaryotes with N-glycan-dependent quality control of protein folding, there is no positive selection for sites of N-linked glycosylation in secreted proteins and membrane proteins of Giardia and other protists (e.g., Plasmodium, Cryptosporidium, and Toxoplasma), which lack N-glycan-dependent quality control of protein folding (our unpublished data). We speculate that GlcNAc2 on Giardia N-glycans stimulates the innate immune system, as has recently been described for chitin (40). We are presently attempting to knock down expression of Giardia Alg enzymes and STT3, which are crucial for the synthesis of N-glycans (16, 21, 42), to determine the effect on protist viability, phenotype, and differentiation.
While WGA affinity has been used previously to enrich serum glycoproteins containing sialic acid (53), this is the first time that WGA has been used to enrich glycoproteins with very short N-glycans containing GlcNAc2. The advantages of using lectin affinity are as follows. (i) It dramatically enriches the secreted and membrane glycoproteins of Giardia, many of which are likely important for pathogenesis. (ii) Glycoproteins may be released with chitooligosaccharides, so that enzymes are still active and proteins are still in their native forms for subsequent immunization. (iii) Multiple repeats of the same experiment can be performed quickly, so that low-abundance glycoproteins may be detected and better estimates of the relative abundances of proteins can be made. (iv) Both trophozoites and cysts appear to add N-glycans to many glycoproteins, as shown by mass spectroscopy and three-dimensional high-resolution fluorescence microscopy. Further, there is no evidence in any organism that Alg genes, which encode enzymes that make N-glycan precursors, show stage-specific expression (16). In addition, there is no subsequent modification of Giardia N-glycans (42), which might be affected by differentiation. (v) Identification of numerous tryptic peptides makes it possible to shift numerous Giardia proteins from “hypothetical” status to “unique protein” status. (vi) Lectin affinity also works very well to enrich the secreted and membrane proteins of Entamoeba, Trichomonas, and Cryptosporidium (our unpublished data).
The disadvantages of lectin affinity are as follows. (i) It misses nucleocytosolic proteins, which include many of the enzymes important for Giardia to adapt to its anaerobic environment and for synthesis of precursors of its cyst wall polymer (12, 34). We have recently used two-dimensional liquid chromatography and many mass spectroscopic runs to identify peptides from >1,400 total proteins of Giardia, which have also been deposited at the GiardiaDB website. (ii) The method does not work well with secreted and membrane proteins that lack sequons or that have sequons that are not N glycosylated. Conversely, glycoproteins that contain many sequons or glycoproteins that are heavily glycosylated are overrepresented in mass spectroscopic data. (iii) WGA binds weakly or not at all to N-glycans containing a single GlcNAc. This is not a problem in Giardia, in which most of the N-glycans contain GlcNAc2, but it is a problem in Plasmodium falciparum, in which many of the N-glycans contain a single GlcNAc (42; our unpublished data). (iv) The method does not work if the secreted or membrane proteins are not solubilized with nonionic detergent, which is likely the case for cyst wall proteins (15, 33, 47). (v) The method does not discriminate between plasma membrane proteins and those located in the ER or lysosomes. (vi) The makeup of a given sample is calculated based either upon MS-MS peak area/height analysis or on peptide coverage. These methods provide a useful estimate of protein abundance but cannot determine the absolute values for protein composition in a sample. (vii) Similarly, hierarchical clustering, which compares the normalized abundances of proteins from different samples (cyst versus trophozoite), gives the same weight to both minor and major protein constituents. Due to this normalization, the resulting heat maps tend to dramatically overstate the differences between samples due to variability in the runs on a particular mass spectrometer, while at the same time identifying significant clusters of proteins that are stage specific. (viii) There is no direct correlation between protein abundance and protein importance, although it is often the case that monoclonal antibodies identify vaccine candidates that are abundant (e.g., VSPs of Giardia or Gal/GalNAc lectins of Entamoeba) (35, 37, 39). (ix) These mass spectroscopic methods, like microarray data, generate large lists of proteins, each of which calls for further investigation and better characterization.
The composition of the N-glycome of Giardia trophozoites was remarkable for the large numbers of lysosomal enzymes and ER enzymes involved in the N-glycan-independent quality control of protein folding. In contrast, the results here, which focus on N-glycans that are added in the ER lumen and do not appear to be subsequently modified, shed little light on glycosylation in the Golgi apparatus and do not address other functions of the Golgi apparatus (29, 42, 45, 49).
The large number of unique secreted and membrane glycoproteins of Giardia identified here likely includes novel vaccine candidates. A possible advantage of these unique Giardia proteins is that they likely do not belong to a large gene family that has variations in expression, as is the case for Giardia VSPs (35). It is likely that WGA-affinity methods may be used to collect Giardia glycoproteins in their native states for immunization and for the screening of monoclonal antibodies (37).
The Giardia cysts made in vitro in this study do not excyst nearly as well as those isolated from the intestines of gerbils, so it is likely that our mass spectroscopic data here will not perfectly reflect proteins of in vivo Giardia cysts. Further, the conditions we used for encystation did not produce well-synchronized cysts, so we used only end points to describe the transition from trophozoites to cysts. With these caveats and those described above for the WGA-affinity methods, the composition of the N-glycome of Giardia cysts was remarkable for the relative paucity of ER folding proteins and unique secreted proteins and for the abundance of encystation-associated proteases, unique transmembrane proteins, and GRREAT family proteins. These results support previous experimental demonstrations of the importance of proteases (50, 54) and Cys-rich proteins (9) in encystation. These results are also consistent with the idea that secretion is less important to encysting Giardia than is development of membranous structures (8). Identification of VSPs in Giardia cysts is consistent with a previous demonstration of encystation-associated VSP-like proteins (9). Of great interest is the function of GRREAT proteins, which are very abundant in Giardia cysts and less abundant in Giardia trophozoites but have no homologs in other organisms.
This work was supported by NIH grants AI048082 (to J.S.) and GM31318 (to P.W.R.). Support for D.M.R. was provided by the Training Program in Host Pathogen Interactions (T32 AI052070).
We thank Dick Cook of MIT for some mass spectroscopy data.
Published ahead of print on 26 September 2008.
†Supplemental material for this article may be found at http://ec.asm.org/.