Glycobiology

Glycobiology Advance Access originally published online on February 8, 2008
Glycobiology 2008 18(5):367-383; doi:10.1093/glycob/cwn014

This Article Abstract FREE Full Text (PDF) Supplementary Data OA All Versions of this Article: 18/5/367    most recent cwn014v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Disclaimer Google Scholar Articles by Manthri, S. Articles by Ferguson, M. A J PubMed PubMed Citation Articles by Manthri, S. Articles by Ferguson, M. A J Social Bookmarking          What's this?

© 2008 The Author(s)
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Deletion of the TbALG3 gene demonstrates site-specific N-glycosylation and N-glycan processing in Trypanosoma brucei

Sujatha Manthri², M Lucia S Güther², Luis Izquierdo², Alvaro Acosta-Serrano³ and Michael A J Ferguson^1,2

² The Division of Biological Chemistry and Drug Discovery, The Wellcome Trust Biocentre, College of Life Sciences, University of Dundee, Dundee DD1 5EH
³ The Wellcome Centre for Molecular Parasitology, Glasgow Biomedical Research Centre, University of Glasgow, Glasgow G12 8TA, Scotland, UK

---

¹ To whom correspondence should be addressed: Tel. +44-1382-384219; Fax +44-1382-348896; e-mail: m.a.j.ferguson{at}dundee.ac.uk

Received on January 5, 2008; accepted on February 3, 2008

	Abstract

Top Abstract Introduction Results Discussion Materials and methods Supplementary Data Funding Conflict of interest statement References

 
We recently suggested a novel site-specific N-glycosylation mechanism in Trypanosoma brucei whereby some protein N-glycosylation sites selectively receive Man₉GlcNAc₂ from Man₉GlcNAc₂-PP-Dol while others receive Man₅GlcNAc₂ from Man₅GlcNAc₂-PP-Dol. In this paper, we test this model by creating procyclic and bloodstream form null mutants of TbALG3, the gene that encodes the -mannosyltransferase that converts Man₅GlcNAc₂-PP-Dol to Man₆GlcNAc₂-PP-Dol. The procyclic and bloodstream form TbALG3 null mutants grow with normal kinetics, remain infectious to mice and tsetse flies, respectively, and have normal morphology. However, both forms display aberrant N-glycosylation of their major surface glycoproteins, procylcin, and variant surface glycoprotein, respectively. Specifically, procyclin and variant surface glycoprotein N-glycosylation sites that are modified with Man₉GlcNAc₂ and processed no further than Man₅GlcNAc₂ in the wild type are glycosylated less efficiently but processed to complex structures in the mutant. These data confirm our model and refine it by demonstrating that the biantennary glycan transferred from Man₅GlcNAc₂-PP-Dol is the only route to complex N-glycans in T. brucei and that Man₉GlcNAc₂-PP-Dol is strictly a precursor for oligomannose structures. The origins of site-specific Man₅GlcNAc₂ or Man₉GlcNAc₂ transfer are discussed and an updated model of N-glycosylation in T. brucei is presented.

Key words: ALG3 / mannosyltransferase / N-glycosylation / Trypanosoma brucei/variant surface glycoprotein

	Introduction

Top Abstract Introduction Results Discussion Materials and methods Supplementary Data Funding Conflict of interest statement References

 
The trypanosmatid Trypanosoma brucei is a parasitic protozoan organism that causes nagana in cattle and human African sleeping sickness. The organism undergoes a complex life cycle, involving major biochemical and morphological changes, between its mammalian host and tsetse fly vector. These changes include complete remodeling of the major cell surface coat molecules.

The tsetse midgut-dwelling procyclic form of T. brucei has a surface coat of 3 x 10⁶ polyanionic, rod-like, procyclin glycoproteins (Mowatt and Clayton 1987; Roditi et al. 1987, 1989; Richardson et al. 1988; Treumann et al. 1997) as well as other unidentified glycoproteins (Güther et al. 2006). In T. brucei strain 427, used in this study, the parasites contain (per diploid genome) two copies of the GPEET1 gene, encoding a procyclin with 6 Gly-Pro-Glu-Glu-Thr repeats; one copy each of the EP1-1 and EP1-2 genes, encoding EP1 procyclins with 30 and 25 Glu-Pro repeats, respectively; two copies of the EP2-1 gene, encoding EP2 procyclin with 25 Glu-Pro repeats; and two copies of the EP3-1 gene, encoding EP3 procyclin with 22 Glu-Pro repeats (Acosta-Serrano et al. 1999; Roditi and Clayton 1999). The EP1 and EP3 procyclins contain a single N-glycosylation site, at the N-terminal side of the Glu-Pro repeat domain, occupied exclusively by a conventional Endo-H-sensitive triantennary Man₅GlcNAc₂ oligosaccharide (Treumann et al. 1997). Neither EP2 nor GPEET procyclin is N-glycosylated but GPEET1 procyclin is phosphorylated on six out of seven Thr residues (Butikofer et al. 1999; Mehlert et al. 1999). GPEET and EP procyclins contain similar glycosylphosphatidylinositol (GPI) membrane anchors, based on the ubiquitous ethanolamine-P-6Man1-2Man1-6Man1-4GlcN1-6PI core (Ferguson 1999), where in this case, the phosphatidylinositol (PI) lipid is a 2-O-acyl-lyso-PI structure, substituted with large branched poly-N-acetyllactosamine structures that can terminate with 2-3-linked sialic acid residues (Ferguson et al. 1993; Treumann et al. 1997).

The bloodstream form of T. brucei has a surface coat of 5 x 10⁶ variant surface glycoprotein (VSG) homodimers (Mehlert, Richardson, et al. 1998). The VSG coat serves as a physical barrier to components of the host complement system and undergoes antigenic variation (Pays et al. 2004; Taylor and Rudenko 2006). There are many VSG genes and each encodes a GPI-anchored glycoprotein with one to three N-glycosylation sites (Mehlert, Richardson, et al. 1998). The cell line used in this study expresses VSG variant 221 (also known as MiTat1.2). VSG221 has a GPI anchor with the same core as the procyclins but with a dimyristoyl-PI component and carbohydrate side chains of between two and six Gal residues (Mehlert, Zitzmann, et al. 1998). VSG221 has two N-glycosylation sites: the Asn428 site, five residues from the GPI attachment site, is occupied mostly by Endo-H-sensitive oligomannose structures (Man_5–9GlcNAc₂), whilst the Asn263 site is occupied by small Endo-H-resistant biantennary structures ranging from Man_3–4GlcNAc₂ to GalGlcNAcMan₃GlcNAc₂ (Zamze et al. 1991).

Protein N-glycosylation in eukaryotes serves a wide variety of functions including signaling through interaction with lectins, protein stabilization, protease resistance, endocytic sorting functions, and protein folding (Varki 1993; Rudd and Dwek 1997; Helenius and Aebi 2004). In eukaryotes, a precursor for N-glycosylation is built up in the endoplasmic reticulum (ER) on the lipid carrier dolichol pyrophosphate (Dol-PP) (Burda and Aebi 1999). The glycan portion of this Dol-PP-oligosaccharide is transferred en bloc by the action of oligosaccharyltransferase (OST), generally during protein translation and sequestration into the lumen of the ER, to Asn residues within Asn-X-Ser/Thr sequons (Helenius and Aebi 2004). Processing of the precursor structure by glycosidase and glycosyltransferase enzymes within the ER and Golgi apparatus generates the final set of mature structures (Kornfeld R and Kornfeld S 1985; Schachter 2000).

In most eukaryotes, the mature precursor used by OST is Glc₃Man₉GlcNAc₂-PP-Dol. However, genomic and experimental comparisons have shown that some lower eukaryotes do not possess all the ALG genes needed to make Glc₃Man₉GlcNAc₂-PP-Dol and that they transfer smaller glycans to protein (Parodi 1993; Samuelson et al. 2005). Differences in the compositions and donor specificities of eukaryotic OST complexes, which usually contain eight different subunits, have also been noted (Kelleher and Gilmore 2006; Kelleher et al. 2007).

Seminal work by Parodi and colleagues on several trypanosomatid parasites (excluding T. brucei) showed that protein N-glycosylation in these organisms is aberrant (reviewed in Parodi 1993). None of these organisms make Dol-P-Glc and so fail to make glucosylated Dol-PP-oligosaccharide precursors. The mature Dol-PP-oligosaccharide species used for transfer to protein vary according to species. For example, Trypanosoma conhorini, Trypanosoma dionisii, Leptomonas samueli, Herpetomonas samuelpessoai, and Herpetomonas muscarum utilize triantennary Man₉GlcNAc₂-PP-Dol; Crithidia fasciculata, Crithidia Harmosa, and Leishmania enriettii utilize biantennary Man₇GlcNAc₂-PP-Dol; Leishmania mexicana, Leishmania adleri, and Blastocrithidia culicus utilize biantennary Man₆GlcNAc₂-PP-Dol (Parodi et al. 1981; Parodi and Quesada-Allue 1982; Previato et al. 1986; de la Canal and Parodi 1987); and Trypanosoma cruzi, the causative agent of Chagas’ disease in the Americas, utilizes Man₉GlcNAc₂-PP-Dol during most of its life cycle but uses both Man₉GlcNAc₂-PP-Dol and Man₇GlcNAc₂-PP-Dol in its bloodstream trypomastigote stage (Doyle et al. 1986).

In a recent study, we showed that wild-type bloodstream form T. brucei utilizes both Man₉GlcNAc₂-PP-Dol and Man₅GlcNAc₂-PP-Dol to glycosylate the major surface coat glycoprotein VSG221 and, uniquely, does so in a site-specific manner (Jones et al. 2005). Thus, whereas Man₉GlcNAc₂-PP-Dol is used to glycosylate Asn428, Man₅GlcNAc₂-PP-Dol is used to glycosylate Asn263. In addition, characterization of VSG221 made by an -glucosidase II null mutant showed that the Asn263-linked Man₅GlcNAc₂ glycan serves exclusively on VSG221 as the acceptor for the parasite's unfolded glycoprotein glucosyltransferase (UGGT) (Jones et al. 2005). We further speculated that Man₅GlcNAc₂ is the obligate precursor of all Endo-H-resistant (i.e., small biantennary oligomannose and complex) structures whereas Man₉GlcNAc₂ is the obligate precursor of all conventional Endo-H-sensitive (i.e., Man_5–9GlcNAc₂) oligomannose structures.

To test and refine our model of site-specific N-glycosylation and N-glycan processing, we constructed bloodstream and procyclic form T. brucei null mutants of the TbALG3 gene and observed the effects of limiting the parasite to solely a Man₅GlcNAc₂-PP-Dol lipid-linked precursor.

	Results

Top Abstract Introduction Results Discussion Materials and methods Supplementary Data Funding Conflict of interest statement References

 
Identification and cloning of the TbALG3 gene
A BLASTp (Altschul et al. 1990) search of the geneDB T. brucei predicted proteins database with the predicted amino acid sequence of the yeast ALG3 gene (accession no. NP 009471) yielded a putative T. brucei Dol-P-Man-dependent 1-3 mannosyltransferase sequence encoded by gene Tb10.70.0260, referred to from now on as TbALG3. The genome strain of T. brucei (strain 927) is different from that used in this study (strain 427). The TbALG3 gene was, therefore, cloned and sequenced from strain 427. The consensus sequence (accession no. AM850907 [GenBank] ) has four single base polymorphisms compared to the strain 927 database sequence, resulting in only a single amino acid polymorphism (Y in place of C at position 278). Southern blot analysis using a TbALG3 ORF probe revealed a single fragment after restriction enzyme digestion by HindIII, SacII, MscI, PstI, and XhoI whereas BsrGI yielded two fragments as it cuts the gene in the center of the gene (supplementary Figure S1). These data indicate that the TbALG3 gene is present as a single copy per haploid genome.

The predicted 46.4 kDa TbALG3 multispanning membrane protein has 26.2%, 33.4%, 33.7%, and 34.7% identity with the Saccharomyces cerevisiae, Mus musculus, Drosophila melanogaster, and Arabidopsis thaliana ALG3 sequences, respectively. The TbALG3 protein sequence conforms to that of a member of the 3,4-mannosyltransferase superfamily composed of dolichol cycle ALG3 3-mannosyltransferases and GPI biosynthesis PIG-M 4-mannosyltransferase sequences (Oriol et al. 2002). Based on conserved amino acid motifs, TbALG3 can further be distinguished as an ALG3-type 3-mannosyltransferase (Oriol et al. 2002) (Figure 1). This confirms that the sequence patterns described in Oriol et al. (2002) for discriminating ALG3 genes from related sequences can be extended to this ancient and highly divergent eukaryote. The TbALG3 sequence contains, like those of S. cerevisiae, D. melanogaster and M. musculus, a putative C-terminal ER retention signal.

View larger version (107K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 1 Predicted amino acid sequence of TbALG3 and alignment with related sequences. Predicted amino acid sequences of ALG3 genes from Trypanosoma brucei (T. b), Arabidopsis thaliana (A. t), Saccharomyces cerevisiae (S. c), Drosophila melanogasta (D. m), and Mus musculus (M. m) were aligned using ClustalW. Residues in black boxes are conserved between all sequences while those in gray boxes indicate conservative changes between the sequences. Residues indicated by # and * are those that, according to Oriol et al. (2002), are common to the 3,4-mannosyltransferase superfamily and that define the ALG3 3-mannosylatransferase family, respectively. The C-terminal sequences in bold italics are putative ER-retention signals

 
Genes encoding proteins similar to TbALG3 were also found in the T. cruzi (gene Tc00.1047053510187.404) and Leishmania major (gene LmjF36.2040) databases with 48.5% and 44.5% predicted amino acid identity to TbALG3, respectively. Interestingly, the L. major sequence has 8- and 16-amino-acid inserts that are absent form the T. brucei and T. cruzi sequences.

Creation of TbALG3 null mutants
The two TbALG3 alleles were sequentially replaced by homologous recombination (Wirtz et al. 1999) with genes encoding puromycin acetyltransferase (PAC) followed by hygromycin phosphotransferase (HYG), for bloodstream form T. brucei, or PAC followed by blasticidin deaminase (BSD), for procyclic form T. brucei. Drug resistant clones were selected and characterized by Southern blotting of HindIII- and SacII-digested genomic DNA with a TbALG3 ORF probe followed by a β-tubulin probe, to ensure equal loading. A representative blot for one of the bloodstream form TbALG3 null mutants (TbALG3::PAC/TbALG3::HYG) is shown in Figure 2.

View larger version (39K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 2 Southern blot of DNA from wild-type cells and the TbALG3 null mutant. Panel A: schematic representation of the replacement of the TbALG3 alleles with drug resistance genes PAC and HYG. Panel B: restriction map of the TbALG3 locus. Panel C: southern blot of genomic DNA from wild-type (WT) cells and TbALG3 null mutant cells probed with a fluorescein labeled probe of the TbALG3 ORF. Lanes 1 and 2 contain DNA of WT cells that have been digested with HindIII (H) and SacII (S), respectively, and lanes 3 and 4 contain DNA from the null mutant digested with HindIII and SacII, respectively. Panel D: the same Southern blot membrane was probed with fluorescein labeled ß-tubulin probe to confirm equal loading of DNA. Lanes 1 and 3 contain a fragment of about 4 kb and lanes 2 and 4 yield a fragment of about 1 kb that hybridizes with the ß-tubulin probe. (The signals from the TbALG3 probe are still visible in lanes 1 and 2.)

 
The TbALG3 gene is nonessential to bloodstream form and procyclic form T. brucei
The in vitro growth rates of the TbALG3 null mutants of both life-cycle stages were indistinguishable from their parental cell lines (supplementary Figure S2). Similarly, scanning electron micrographs of the TbALG3 null mutants of both life-cycle stages were indistinguishable from their parental cell lines (supplementary Figure S3). Furthermore, bloodstream form TbALG3 null mutants were infective to mice and procyclic form TbALG3 null mutants were infective to tsetse flies, although the infectivity may have been marginally reduced (Figure 3). From these data we conclude that TbALG3 is a nonessential gene to the procyclic form and the disease-causing bloodstream form of T. brucei.

View larger version (27K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 3 Infectivity of wild-type and TbALG3 null mutant procyclic form parasites to tsetse flies. Tsetse flies were fed with infected bloodmeals containing wild-type (WT) or TbALG3 null mutants (ALG3). After 14 or 21 days, flies were dissected and midgut infections scored as heavy (black; more than 100 parasites per field), intermediate (dark gray; 20–100 parasites), weak (light gray; 1–19 parasites), and negative (no detectable parasites). The number of dissected flies is indicated on top of each group.

 
TbALG3 encodes a dolichol cycle 3-mannosyltransferase
Previous studies on dolichol-linked oligosaccharides and protein N-glycosylation in T. brucei have noted that although bloodstream form T. brucei make and utilize Man₉GlcNAc₂-PP-Dol (Zamze et al. 1991; Jones et al. 2005), the steady-state levels of species larger than Man₅GlcNAc₂-PP-Dol are extremely low in this life-cycle stage (Low et al. 1991). On the other hand, the Man₆GlcNAc₂-PP-Dol to Man₉GlcNAc₂-PP-Dol species are easier to observe in GDP-[³H]Man labeling experiments using the procyclic form cell-free system (Low et al. 1991; Leal et al. 2004). For this reason, we utilized procyclic form cell-free systems to assess the metabolic lesion caused by deletion of both TbALG3 alleles.

Washed membranes (cell-free systems) prepared from wild-type and TbALG3 null mutant procyclic form trypanosomes were incubated UDP-GlcNAc and GDP-[³H]Man, which labels both GPI and dolichol-PP-linked oligosaccharide precursors and their biosynthetic intermediates (Acosta-Serrano et al. 2004). To selectively analyze the dolichol-PP-linked oligosaccharide species, the labeled glycolipids were treated with mild acid and partitioned between butan-1-ol and water. The mild acid-labile pyrophosphate bonds of the Dol-PP-oligosaccharides result in the recovery of their radiolabeled free oligosaccharides in the aqueous phase, whereas the mild acid stable radiolabeled GPI species partition into the butan-1-ol phase. Following reduction of the radiolabeled free oligosaccharides with sodium borohydride and desalting, the resulting oligosaccharide alditols were analyzed by high-performance thin layer chromatography (HPTLC) and fluorography alongside authentic sodium borotritiide-reduced oligosaccharide alditol standards (Figure 4A). The fluorograph clearly shows that, as expected, the TbALG3 null mutant cell-free system is unable to produce oligosaccharides larger than Man₅GlcNAc₂. The radiolabeled free oligosaccharide alditols were further analyzed before and after digestion with Aspergillus saitoi Man1-2Man-specific -mannosidase (ASAM) (Figure 4B). The digestion patterns are consistent with the Man₅GlcNAc₂ structure being Man1-6(Man1-2Man1-2Man1-3)Manβ1-4GlcNAcβ1-4GlcNAc, as summarized in Figure 4C.

View larger version (33K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 4 Comparison of mild acid-released glycans from the Dol-PP-oligosaccharides of wild-type and TbALG3 null mutant cells. Panel A: radiolabeled oligosaccharides were released from GDP-[3H]mannose labeled cell free systems of wild-type (WT) and TbALG3 null mutant (TbALG3–/–) cells, reduced with NaBH4, and analyzed by HPTLC and fluorography. The positions of authentic NaB[3H]4-reduced oligosaccharide standards (names in bold italics) are shown on the left together with presumed descriptors for the remaining bands. The upper band is [3H]mannitol, derived from the release of [3H]Man from Dol-P-[3H]Man and its subsequent reduction to [3H]mannitol. Panel B: the NaB[3H]4-reduced Man9GlcNAc2 standard (lanes 1 and 2), and the wild type (lanes 3 and 4) and TbALG3 null mutant (lanes 5 and 6) reduced oligosaccharides were treated with (+) and without (–) A. saitoi 1-2-specific -mannosidease (ASAM), as indicated, and analyzed by HPTLC and fluorography. The conversion of the Man9GlcNAc2 standard to Man5GlcNAc2 is a positive control for the ASAM activity. The uppermost band running ahead of [3H]mannitol in lanes 4 and 6 is free [3H]Man released from the oligosaccharides by ASAM. Panel C: schematic representation of the ASAM-induced interconversions of the radiolabeled oligosaccharides seen in panel B.

 
Together, these data provide clear evidence that the largest Dol-PP-linked oligosaccharide made by the TbALG3 null mutant is a biantennary Man₅GlcNAc₂ species and confirm that the TbALG3 gene encodes a 1-3 mannosyltransferase respon- sible for the conversion of Man1-6(Man1-2Man1-2Man1-3)Manβ1-4GlcNAcβ1-4GlcNAc-PP-Dol (Man₅GlcNAc₂-PP- Dol) to Man1-3Man1-6(Man1-2Man1-2Man1-3)Man- β1-4GlcNAcβ1-4GlcNAc-PP-Dol (Man₆GlcNAc₂-PP-Dol).

Procyclin N-glycosylation is affected in TbALG3 null mutant procyclic form T. brucei
Procyclins were extracted from wild-type and TbALG3 null trypanosomes by differential solvent extraction and octyl-Sepharose hydrophobic interaction chromatography. The proteins were first analyzed by SDS–PAGE and periodate-Schiff staining for carbohydrate. There was a small but perceptible downward shift in the apparent molecular weight range of the intact procyclins in the TbALG3 null mutants compared to those in the wild type (Figure 5, lanes 1 and 3), whereas the two samples appeared similar after peptide N-glycosidase-F (PNGaseF) treatment (Figure 5, lanes 2 and 4).

View larger version (56K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 5 SDS–PAGE and peroidate-Schiff staining of wild-type and TbALG3 null mutant procyclins before and after PNGaseF digestion. Lanes 1 and 2 contain wild-type procyclins before and after PNGaseF treatment, respectively. Lanes 3 and 4 contain the TbALG3 null mutant procyclins before and after PNGaseF treatment, respectively. The positions of molecular weight markers are shown on the left.

 
To gain a more detailed insight, procyclins from wild-type and TbALG3 null mutant cells were treated in three different ways, as described in Acosta-Serrano et al. (1999) and Leal et al. (2004), and analyzed by negative-ion mode MALDI-TOF. The three treatments were as follows: (i) dephosphorylation with cold aqueous HF alone (aq. HF), to remove their heterogeneous GPI anchors and reveal the masses of their full amino acid sequence plus N-glycans; (ii) dephosphorylation with cold aqueous HF followed by de-N-glycosylation with PNGaseF (aq. HF + PNGaseF), to reveal the masses of their full amino acid sequences; (iii) dephosphorylation with cold aqueous HF followed by mild acid hydrolysis of Asp-Pro peptide bonds with trifluoroacetic acid (aq. HF + TFA), to reveal the masses of their diagnostic C-terminal peptide domains. The aforementioned treatments and their products are summarized at the top of Figure 6 and Table I and the MALDI-TOF spectra are shown in Figure 6A–F.

View larger version (32K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 6 The MALDI-TOF mass spectra of treated wild-type and TbALG3 null mutant procyclins. Procyclins from wild-type cells (panels A, C, and E) and TbALG3 null mutant cells (panels D, E, and F) were analyzed by negative-ion MALDI-TOF mass spectrometry after treatment with aq. HF and mild acid (panels A and B), aq. HF and PNGaseF (panels C and D), and aq. HF alone (panels E and F). A schematic representation of the products of these treatments for wild-type EP1 procyclin is shown above the spectra.

 

View this table: [in this window] [in a new window]   Table I Procyclin species observed in Figure 6 by negative-ion MALDI-TOF

 
The aq. HF/TFA data are the simplest to interpret, since they show the diagnostic C-terminal domain peptide fragments that define which of the procyclin variants are present in the sample (Acosta-Serrano et al. 1999). From these data, we can see that the wild-type cells are expressing a mixture of EP3, EP1-2 and EP1-1 procyclins, whereas the TbALG3 null mutant is expressing only EP3 and EP1-2 (Figure 6A and B). Such changes in procyclin expression are known in procyclic form T. brucei (Treumann et al. 1997; Morris et al. 2002; Vassella et al. 2004).

The aq. HF/PNGaseF data gave the expected result for the wild-type procyclins, i.e., the EP3, EP1-2, and EP1-1 full-length peptides were observed (Figure 6C). However, the TbALG3 null mutant data were unusual in that, although the EP3 full-length peptide is clearly seen (at m/z 8489), the majority of the EP1-2 procyclin was observed as homodimers at m/z 18496 or as EP1-2/EP3 heterodimers at m/z 17777 (Figure 6D). A trace of EP3 homodimer is also seen at m/z 177041. Why the procyclins from the TbALG3 null mutant, and in particular EP1-2, should have this propensity to appear as dimers by MALDI-TOF is unclear. None of the procyclins have cysteine residues, ruling out disulfide formation.

The aq. HF only data gave the expected result for the wild-type procyclins, i.e., the EP3, EP1-2, and EP1-1 full-length peptides plus conventional triantennary Man₅GlcNAc₂ N-linked glycans were observed at m/z 9707, 10417, and 11519, respectively (Figure 6E). On the other hand, the aq. HF only data for the TbALG3 null mutant procyclins were strikingly different. In this case, a major ion at m/z 8492, corresponding to non-N-glycosylated EP3 procyclin was seen together with a cluster of three ions at m/z 9384, 9748, and 10119 that would correspond to EP3 procyclin containing Hex₃HexNAc₂, Hex₄HexNAc₃, and Hex₅HexNAc₄ N-linked glycans (Figure 6F). Again, in this spectrum, the EP1-2 procyclin appears predominantly as dimers in nonglycosylated and glycosylated form, containing the same Hex₃HexNAc₂, Hex₄HexNAc₃, and Hex₅HexNAc₄ structures. The combinations of possible dimer glycoforms give rise to the relatively complex peak pattern between m/z 17764 (the unglycosylated EP3/EP1-2 heterodimer) and 20106 (the EP1-2 homodimer with one Hex₅HexNAc₄ N-linked glycan) (Table I).

The identities of the N-linked glycans in the TbALG3 null mutant procyclins were analyzed by releasing them with PNGaseF followed by permethylation and positive-ion MALDI-TOF mass spectrometry. Ions corresponding to the [M + Na]⁺ ions of three major species of composition Hex₃HexNAc₂, Hex₄HexNAc₃, and Hex₅HexNAc₄ (together with trace amounts of Hex₃HexNAc₃, Hex₄HexNAc₄, and Hex₆HexNAc₅) were observed (Figure 7A), consistent with the assignments in Table I and Figure 6. The same sample of permethylated glycans was analyzed by electrospray tandem mass spectrometry (ES-MS/MS). Ions at 821.9 and 1047.0, corresponding to [Hex₄HexNAc₃ + 2Na]²⁺ and [Hex₅HexNAc₄ + 2Na]²⁺, respectively, were selected and subjected to collision-induced dissociation; the resulting product-ion spectra contained intense m/z 486 ions, characteristic of N-acetyllactosamine (LacNAc) containing species (Figure 7B, C). The lack of product ion at m/z 935 in Figure 7B, corresponding to a LacNAc₂ fragment ion, suggests that the Hex₅HexNAc₄ species is a conventional Galβ1-4GlcNAcβ1-2Man1-6(Galβ1-4GlcNAcβ1-2Man1-3)Manβ1-4GlcNAcβ1-4GlcNAc (NA2-type) complex structure with one LacNAc unit on each branch.

View larger version (30K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 7 Mass spectrometric analyses of PNGaseF released and permethylated glycans from TbALG3 null mutant procyclins. Panel A: positive-ion MALDI-TOF of the released and permethylated glycans, revealing a set of [M+Na]+ ions. Panel B: ES-MS/MS product ion spectrum of an m/z 821.9 [M+2Na]2+ ion (corresponding to the m/z 1620.6 [M+Na]+ ion in panel A). Panel C: The ES-MS/MS product ion spectra of an m/z 1047.0 [M+2Na]2+ ion (corresponding to the m/z 2069.9 [M+Na]+ ion in panel A). The proposed structures and fragment patterns of the glycans are shown in the panel insets.

 
In summary, these data show that whereas wild-type procyclic form T. brucei normally transfer Man₉GlcNAc₂ oligosaccharides to their procyclin glycoproteins and process these to conventional triantennary Man₅GlcNAc₂ oligomannose structures (Treumann et al. 1997; Acosta-Serrano et al. 1999), there is significant under-N-glycosylation and aberrant processing to biantennary complex structures when biantennary Man₅GlcNAc₂-PP-Dol is the only donor available to the OST(s) present in that life-cycle stage.

VSG N-glycosylation is affected in TbALG3 null mutant bloodstream form T. brucei
VSG glycoprotein was purified (Cross 1984) in its GPI-specific phospholipase C (GPI-PLC) cleaved soluble form (sVSG) (Ferguson et al. 1985) from wild-type and TbALG3 null mutant cells. Analysis by SDS–PAGE and Coomassie blue staining showed that whereas sVSG prepared from the wild-type cells appeared as a single 54 kDa band, sVSG prepared from the TbALG3 null mutant appeared as a doublet (Figure 8). To assess whether this change in the VSG pattern was due to changes in glycosylation, the wild-type and TbALG3 null mutant sVSGs were analyzed by mass spectrometry.

View larger version (61K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 8 SDS–PAGE of sVSG221 from bloodstream form wild-type and TbALG3 null mutant parasites. Soluble form sVSG221 was purified from wild-type cell cells (lane 1) and TbALG3 null mutant cells (lane 2), subjected to SDS–PAGE, and stained with Coomassie blue. The positions of molecular weight standards are indicated on the left.

 
First, the upper and lower bands of the TbALG3 null mutant sVSG were excised separately from a gel, digested with trypsin, and the resulting peptides were analyzed by MALDI-TOF/TOF. Both bands were clearly identified as VSG221 (data not shown), the same VSG variant as the wild-type cells, suggesting that the differences in the SDS–PAGE pattern are indeed due to the difference in posttranslational modifications of the same sVSG variant.

Next, the positive-ion ES-MS spectra of the intact glycoproteins were collected and deconvolved using a Bayesian protein reconstruction method. The wild-type sVSG spectrum (Figure 9A) was the same as that described previously for this sVSG variant (Jones et al. 2005; Urbaniak et al. 2006) and the compositions of the isobaric glycoforms, resulting from structural heterogeneity at the two N-glycosylation sites (Zamze et al. 1991) and in the GPI anchor (Mehlert, Richardson, et al. 1998), are described in Table II. The TbALG null mutant sVSG spectrum (Figure 9B) showed two groups of sVSG glycoforms, consistent with the SDS–PAGE result. The lower molecular weight group of glycoforms has masses consistent with the C-terminal Asn428 N-glycosylation site being unoccupied (Table II). The higher molecular weight group of glycoforms, with both N-glycosylation sites occupied, has significantly higher N-acetylhexosamine to hexose ratios than the wild-type glycoforms, suggesting that the N-glycans contain LacNAc structures at one or both sites (Table II). These data provided the first direct evidence that deletion of the TbALG3 gene affects VSG glycosylation.

View larger version (41K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 9 sVSG221 glycoform analysis by ES-MS. Samples of sVG221 from wild-type cells (panels A and C) and TbALG3 null mutant cells (panels B and D) grown the absence (panels A and B) or presence (panels C and D) of -mannosidase inhibitors were analyzed by the positive-ion ES-MS and deconvolved spectra of the various isobaric glycoforms were generated. The compositions of these glycoforms are given in Table II.

 

View this table: [in this window] [in a new window]   Table II Isobaric glycoforms of sVSG221 detected by ES-MS

 
Wild-type and TbALG3 null mutant sVSG were also prepared from trypanosomes that had been grown in the presence of a cocktail of -mannosidase inhibitors (kifunensine, swainsonine, and deoxymannojirimycin that, between them, can inhibit all known ER and Golgi -mannosidase activities) and these too were analyzed by ES-MS. This experiment was performed to confirm that the TbALG3 null mutant cannot transfer structures larger than Man₅GlcNAc₂ that might then be processed down to Man₅GlcNAc₂ by ER and/or Golgi -mannosidases. The spectrum for the wild-type sVSG made in the presence of -mannosidase inhibitors (MI) is similar to that previously described (Jones et al. 2005) with a general upward shift in glycoform size (Table II) due to substantial inhibition of -mannosidase trimming of both the internal and C-terminal N-glycans (Figure 9C). The effect of MI on the TbALG3 null mutant sVSG was more complex, with a general increase in size for the low molecular weight group of glycoforms and a general decrease in size for the high molecular weight group of glycoforms (Figure 9D).

To ascertain more precisely the changes induced by the deletion of TbALG3 gene (and the effects of MI) on sVSG glycosylation, the N-glycans present at each of the two N-glycosylation sites and the GPI-anchor structures attached to the C-terminal serine residue were determined by ES-MS and ES-MS/MS analysis of Pronase digests. The Pronase glycopeptides were also re-analyzed after digestion with Man1-2Man-specific Aspergillus saitoi -mannosidase (ASAM).

From these data, summarized in Figure 10, we can conclude that (i) glycosylation of the internal Asn263 residue is largely unchanged by the deletion of the TbALG3 gene. (ii) The dramatic increase in the Man₅-containing glycans attached to the Asn263 site when wild-type and TbALG3 null VSG is made in the presence of the -mannosidase inhibitor cocktail confirms that biantennary Man₅GlcNAc₂ is the precursor of all of the structures found at this glycosylation site. (The presence of some smaller structures suggests that the -mannosidase block was slightly incomplete in these experiments.) (iii) Although ASAM digestion converts most of the biantennary Man₅GlcNAc₂ and Man₄GlcNAc₂ structures to Man₃GlcNAc₂, some of the structures at Asn263 contain six Hex residues that are not digested by ASAM. This suggests that the Hex₆HexNAc₂ glycan attached to Asn263 is Man1-6(Glc1-3Man1-2Man1-2Man1-3)Manβ1-4GlcNAcβ1-4GlcNAc. This glycan species that occupies about 10% of the Asn263 site in the wild-type VSG has not previously been noted in mature VSG molecules. On the other hand, similar residual Glc is quite abundant in the N-linked glycans of L. major and L. mexicana promastigote surface protease glycoprotein, also known as gp63, (Olafson et al. 1990; Ilg et al. 1994). (iv) The most striking change in VSG glycosylation upon TbALG3 deletion is seen at the C-terminal N-glycosylation site where the mixture of conventional triantennary oligomannose Man_5-9GlcNAc₂ glycans is replaced by complex biantennary structures. (v) Some of the VSG molecules in the TbALG3 null mutant fail to be N-glycosylated at the C-terminal Asn428 site. (vi) When grown in the presence of -mannosidase inhibitors, the complex structures at Asn428 in the TbALG3 null mutant are replaced by atypical hybrid structures, suggesting that the usual processing of biantennary Man₅GlcNAc₂ after transfer to Asn428 is the addition of a LacNAc unit to the 1-6 arm of Man₅GlcNAc₂, either after or during the removal of the two 1-2-linked Man residues from the 1-3 arm, and the subsequent addition of up to a further 3 LacNAc units. If the two outer 1-2-linked Man residues on the 1-3 arm are retained (as in the VSG from -mannosidase inhibitor treated cells), elaboration appears to be limited to the addition of a single LacNAc unit to the 1-6 arm. (vii) There are no qualitative and only minor quantitative changes to the VSG GPI anchor glycan side chains in the TbALG3 null mutant.

View larger version (21K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 10 Representations of wild-type and TbALG3 null mutant VSG glycoforms. The principle glycans at each site of wild-type (WT) and TbALG3 null (ALG3 null) VSGs from untreated cells (panel A) and -mannosidase inhibitor treated cells (panel B) are indicated, together with their approximate relative percentages. The glycans indicated above and marked (+ASAM) are those found after A. saitoi Man1-2Man-specific -mannosidase treatment.

 
Analysis of other glycoproteins by lectin blotting
As described above, sVSG analysis by mass spectrometry showed that the glycans at the C-terminal (Asn428) N-glycosylation site were changed from oligomannose type to complex type by deletion of the TbALG3 gene, resulting in an increase in the number of terminal galactose residues on the glycopeptides. We wanted to investigate other glycoproteins to see if this shift from oligomannose-type to complex-type glycans was a general effect of TbALG3 deletion.

Wild-type and TbALG3 null mutant bloodstream form trypanosomes were purified and subjected to osmotic shock, which leads to the GPI-PLC-mediated release of about 90% of cellular VSG as sVSG (Cross 1984; Ferguson et al. 1985). The remaining VSG-depleted wild-type and TbALG3 null mutant cell ghosts were washed and solubilized in 2% SDS, 8 M urea (Atrih et al. 2005) and equivalent amounts were subjected to SDS–PAGE and Western blotting with horseradish peroxidase conjugated concanavalin A (ConA) in the presence and absence of -methyl-mannose inhibitor. The TbALG3 null mutant lysate clearly binds less well to ConA than the wild-type lysate (Figure 11A). ConA is a lectin that binds with highest affinity to oligomannose and hybrid glycans containing a Man1-3(Man1-6)Man1-motif and less well to complex biantennary-type glycopeptides (Ohyama et al. 1985; Brewer and Bhattacharyya 1986; Naismith and Field 1996). The reduction in ConA binding to the TbALG3 null mutant glycoproteins is consistent with the expected general ablation of oligomannose (i.e., triantennary Man_5-9GlcNAc₂) structures. The same lysates were subjected to blotting with ricin, a lectin that binds terminal Gal residues, except this time approximately three times more wild type than TbALG3 lysate was loaded. Taking this into account, it can be seen that the TbALG3 null mutant lysate glycoproteins have significantly greater ricin binding capacity (Figure 11B).

View larger version (50K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 11 TbALG3 null mutant cell lysates show altered lectin binding. Lysates of washed trypansome cell ghosts were subjected to SDS–PAGE and transferred to nitrocellulose membranes. Panel A: lanes 1 and 2 are Ponceau-stained blots of wild-type and TbALG3 null mutant cell lysates, respectively. Lanes 3 and 4 are the same blots probed with biotin-conjugated ConA lectin. Lanes 5 and 6 are similar blots probed with ConA in the presence of -methyl mannose inhibitor (a control to demonstrate that the lectin binding is carbohydrate specific). The positions of molecular weight markers are indicated on the left. Panel B: as in panel A except that approximately three times more wild type than TbALG3 null mutant cell lysate was loaded and HRP-conjugated ricin and a galactose/lactose mixture were used as the lectin probe and lectin inhibitor, respectively.

 
Taken together, these data are consistent with a general upregulation of Gal-terminating complex glycans at the expense of conventional oligomannose structures in the TbALG3 null bloodstream form trypanosomes.

	Discussion

Top Abstract Introduction Results Discussion Materials and methods Supplementary Data Funding Conflict of interest statement References

 
Deletion of the TbALG3 gene in bloodstream and procyclic form T. brucei has no significant effect on their growth, morphology or the infectivity of the parasite to mice and tsetse flies, respectively. We may, therefore, conclude that the Dol-P-Man-dependent 1-3 mannosyltransferase encoded by TbALG3 is not useful drug target against African trypanosomiasis. However, the retention of ALG3, ALG12, and ALG9 genes in T. brucei indicates that the ability to make Man₉GlcNAc₂-PP-Dol presumably has some selective advantage for the organism and we cannot rule out a role for TbALG3 in other life-cycle stages or in parasite differentiation processes.

The glycosylation phenotypes of the null mutant parasites supports the model of site-selective N-glycosylation, originally proposed by Bangs et al. (1988) and elaborated in Jones et al. (2005). That model predicts that preventing the synthesis of dolichol cycle intermediates beyond Man₅GlcNAc₂-PP-Dol should have little or no effect on the glycosylation and subsequent processing of the Asn263 site of VSG221, since this site normally receives Man₅GlcNAc₂ from Man₅GlcNAc₂-PP-Dol in wild-type cells. This is indeed the case, apart from an increase (from approximately 10 to 19 mol%) in Glc₁Man₅GlcNAc₂ at this site. The latter effect may be due to the aberrant- and underglycosylation that is observed at the C-terminal Asn428 site, perhaps causing greater activity of UGGT in the UGGT/calreticulin/-glucosidase II-mediated refolding and quality control cycle.

The underglycosylation of the Asn428 site of VSG221 in the TbALG3 null mutant is significant because it provides evidence that the OST(s) that normally recognize this sequon have a preference for Man₉GlcNAc₂-PP-Dol and that, when this donor is absent, they utilize Man₅GlcNAc₂-PP-Dol relatively poorly. Exactly the same phenomenon is seen in the procyclic form of the parasite, where the procyclins that normally receive Man₉GlcNAc₂ are similarly underglycosylated. These observations support a refined model (Figure 12) whereby two different classes of OST activity exist in T. brucei: one, that we will refer to as TbOST-1, that recognizes sites like Asn263 in VSG221 and preferentially utilizes Man₅GlcNAc₂-PP-Dol and another, that we will refer to as TbOST-2, that recognizes sites like Asn428 in VSG221 and those of EP3 and EP1-2 procyclin and preferentially utilizes Man₉GlcNAc₂-PP-Dol. The genome sequencing supports this model in so far as T. brucei has three OST catalytic subunit STT3 genes (although, intriguingly, it lacks candidate genes for all other known OST subunits [Kelleher and Gilmore 2006]). The three TbSTT3 genes are found in a tandem array at the end of chromosome 5 and two of them encode almost identical proteins. We speculate that the unique TbSTT3 utilizes exclusively either Man₅GlcNAc₂-PP-Dol or Man₉GlcNAc₂-PP-Dol and that one or both of the two similar TbSTT3s utilize the other donor. In this model, the different VSG glycosylation sites are predicted to recruit the appropriate TbOST, and thus dictate the glycan transferred to that site. The fact that posttranslational N-glycosylation of Endo-H-sensitive (but not Endo-H-resistant) sites has been observed for two different VSG variants (Ferguson et al. 1986; Bangs et al. 1988) also provides support for this model and further suggests that TbOST-1 is associated with co-translational glycosylation whereas TbOST-2 is associated with posttranslational glycosylation. Testing of this revised model awaits the construction and characterization of TbSTT3 mutants and a (glyco)proteome-wide analysis of the occupancy of T. brucei N-glycosylation sites with Endo-H-sensitive and resistant glycans to identify the flanking peptidic features that are recognized by TbOST-1 and TbOST-2. This work is currently in progress in our laboratory.

View larger version (24K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 12 Model of the site-specific protein N-glycosylation and N-glycan processing in T. brucei. Panel A illustrates the proposed existence of two distinct oligosaccharyltransferase activities, one (TbOST-1) that cotranslationally transfers Man5GlcNAc2 to specific glycosylation sites (e.g., Asn263 of VSG221) and the other (TbOST-2) that preferentially transfers Man9GlcNAc2 posttranslationally to any remaining sites (e.g., Asn428 of VSG221 and procyclin sites). Panel B illustrates the processing of Man9GlcNAc2 and Man5GlcNAc2 in VSG221 and procyclin in wild-type trypanosomes. The structures in boxes are those that have been experimentally observed at the respective glycosylation sites in the mature glycoproteins. The inability of T. brucei to process Man9GlcNAc2 beyond triantennary Man5GlcNAc2 is thought to be due to unusual specificity of TbGnT-I and the absence of Golgi -mannosidase II. Panel C illustrates the processing of Man5GlcNAc2 in VSG221 and procyclin in TbALG3 null mutant trypanosomes. Structures in boxes are those that have been experimentally observed at the respective glycosylation sites in the mature glycoproteins. The +/– symbol indicates inefficient oligosaccharide transfer of Man5GlcNAc2 by TbOST-2 that normally transfers Man9GlcNAc2. The processing of Man5GlcNAc2 to complex biantennary structures at Asn428 of VSG221 supports the model that sites that receive Man5GlcNAc2 are destined to be processed to complex structures in T. brucei.

 
The other key result for this study is the processing of Man₅GlcNAc₂ (i.e., aberrantly transferred to "Man₉GlcNAc₂-sites" in the TbALG3 null mutants) to complex biantennary structures on both Asn428 of VSG221 and the N-glycosylation sites of EP3 and EP1-2 procyclin. This supports the notion (Jones et al. 2005) that Man₅GlcNAc₂ is destined for processing into complex structures and, importantly, suggests that there is nothing inherent in those sites that prevent their processing to complex structures.

This, in turn, begs the question as to why sites that normally receive Man₉GlcNAc₂ can only be processed to other Endo- H-sensitive oligomannose structures (i.e., triantennary Man_5–9GlcNAc₂). The answer may lie partly in the absence of a putative T. brucei Golgi -mannosidse II gene. Golgi -mannosidse II normally removes the 1-3 and 1-6-linked Man residues from the 6-arm of the trimannosyl core to provide a substrate for elaboration by GnT-II into complex structures (Schachter 2000). However, this alone does not explain why GnT-I activity could not convert triantennary Man₅GlcNAc₂ into conventional hybrid structures. However, whereas eukaryotic GnT-I enzymes typically work on triantennary Man₅GlcNAc₂, analysis of ConA-resistant pro- cyclic mutants (Hwa and Khoo 2000) strongly suggest that TbGnT-I can only operate on Man1-3Man1-6(Man1-3)- Manβ1-4GlcNAcβ1-4GlcNAc (Man₄GlcNAc₂) and Man1– 6(Man1–3)Manβ1–4GlcNAcβ1–4GlcNAc (Man₃GlcNAc₂) structures. Combining these observations leads to the model shown in Figure 12 that satisfies most or all of the current experimental data. For example, it explains why the effects of TbALG3 and TbALG12 deletion are similar with respect to procyclin underglycosylation (because TbOST2 prefers Man₉GlcNAc₂-PP-Dol) yet different with respect to the final processed glycan structures (Leal et al. 2004). In the TbALG12 null mutant (and in ConA-resistant mutants), the procyclin-linked Man₇GlcNAc₂ structure can be processed by ER 1-2 mannosidases to Man1-3Man1-6(Man1-3)Manβ1-4Glc- NAcβ1-4GlcNAc (Man₄GlcNAc₂) and further elaborated on the 3-arm of the trimannosyl core by a LacNAc unit (Acosta-Serrano et al. 2000; Hwa and Khoo 2000; Leal et al. 2004) whereas in the TbALG3 mutant, the Man1-6(Man1-2Man1-2Man1-3)Manβ1-4GlcNAcβ1-4GlcNAc (Man₅GlcNAc₂) structure attached to procyclin is processed by ER 1-2 mannosidases to Man1-6(Man1-3)Manβ1-4GlcNAcβ1-4GlcNAc (Man₃GlcNAc₂) that can receive LacNAc units on both arms of the trimannosyl core.

In summary, the revised model presented here serves to highlight fundamental differences in the mechanisms of protein N-glycosylation and glycan processing between T. brucei and its mammalian hosts. Further study may reveal therapeutically exploitable differences.

	Materials and methods

Top Abstract Introduction Results Discussion Materials and methods Supplementary Data Funding Conflict of interest statement References

 
Cultivation of trypanosomes
Bloodstream form T. brucei strain 427 that has genetically been modified to express T7 polymerase and the tetracycline repressor protein, referred to as wild type for convenience, were cultured in the HMI-9 medium (Hirumi H and Hirumi K 1994) with the addition of 2.5 µg/mL G418 at 37°C in a 5% CO₂ incubator. For some experiments the trypanosomes were grown for 48 h in the presence of a cocktail of -mannosidase inhibitors as previously described (Jones et al. 2005). Procyclic T. brucei strain 427 clone 29.13, referred to as wild-type cells, were cultured in SDM-79 (Brun and Schonenberger 1979) in the presence of 15% fetal bovine serum (FBS) and 7.5 mg/L haemin (Wirtz et al. 1999) together with 15 µg/mL G418 and 50 µg/mL hygromycin in a 28°C incubator. The concentrations of drugs used for the selection of mutants were 4 µg/mL for hygromycin and 0.1 µg/ mL for puromycin for bloodstream form cells and 10 µg/ mL for blasticidin and 1 µg/mL for puromycin for procyclic form cells.

DNA manipulation
Plasmid DNA was purified using Miniprep or Maxiprep kits (Qiagen, Crawley, UK). Gel extraction and cleanup were performed with a Qiaquick kit (Qiagen). Genomic DNA was isolated from 5 x 10⁸ bloodstream form and procyclic form cells using DNAzol (Helena Biosciences, Gateshead, UK).

Cloning and sequencing of the TbALG3 ORF
The 1214 bp ORF was amplified by PCR from genomic DNA using Pfu polymerase with 5'-CCCAAGCTTATGGAGTAT- CGGTGGCTG-3' forward and 5'-CGGGATCCCTACTTTC- CCCTTTTCACAG-3' reverse primers using the cycling parameters 94°C for 5 min, 30 cycles of 94°C for 30 s, 52°C for 2 min, and 68°C for 30 s, followed by 10 min extension at 68°C. Four separate PCR products were purified, cloned into pCR-Blunt II TOPO (Invitrogen, Paisley, UK), and sequenced twice in both directions.

Gene replacement constructs
To generate the T. brucei gene replacement cassettes, 500 bp 5'-UTR and 500 bp 3'-UTR immediately adjacent to the TbALG3 ORF were amplified by PCR from T. brucei genomic DNA using Pfu polymerase with primers 5'-ATAAGAA- TGCGGCCGCTCATTTTTTTGATTGGTTACCTC-3' and 5'-GTTTAAACTTACGGACCGTCAAGCTTTCCTCACTGGT- GGGGAC-3' for the 5'-UTR, and primers 5'-GACGGT- CCGTAAGTTTAAACGGATCCGCAACCGGAACAAGTG- ATC-3' and 5'-ATAAGAATGCGGCCGCCCCCGACGCGAC- CCAA-3' for the 3'-UTR. The products from the above PCR reactions were joined together in a further PCR via a short BamHI–PmeI–HindIII linker region contained within the primers and a NotI site (underlined) at each end. This PCR product was ligated into pGEM-5Zf(+) (Promega) via the NotI sites. Antibiotic resistance markers were ligated between the BamHI and HindIII restriction sites, to produce constructs containing puromycin acetyltransferase (PAC), hygromycin phosphotransferase (HYG) and blasticidin deaminase (BSD) resistance genes. After sequencing, plasmids were purified and digested with NotI. The digestion products were precipitated and washed in 70% ethanol, dissolved in sterile distilled water, and used for electroporation to sequentially replace TbALG3 alleles by homologous recombination (Wirtz et al. 1999).

Southern blotting
T. brucei genomic DNA, prepared from 100 mL cultures of bloodstream form and 10 mL cultures of procyclic form cells, was digested with the various restriction enzymes overnight and resolved on a 0.8% agarose gel and transferred on to a Hybond-N positively charged membrane (GE Healthcare, Amersham, UK). The membrane was hybridized overnight in ULTRA-HYB (Ambion) at 42°C with the fluorescein-labeled TbALG3 ORF probe (CDP-Star random prime labeling kit, GE Healthcare). After two 20-min washes (once with 2 x SSC, 0.1% SDS and once with 0.2 x SSC, 0.1% SDS) the probe was detected by using the CDP-Star detection system.

Tsetse fly infections
Pupae of Glossina morsitans were obtained from Institute of Zoology, Slovak Academy of Science (Bratislava, Slovakia). Newly hatched (teneral) flies were fed with an infected bloodmeal, which consisted of 10⁷ parasites mixed with washed defribinated horse blood (containing 10% FBS). Infected flies were fed with bloodmeals every 2–3 days. After 2 weeks or 3 weeks, midguts were isolated from infected flies and disrupted by mechanical force in cold SDM-79 containing 10% FBS. Isolated parasites from individual midguts were kept on ice until counted on a haemocytometer.

Preparation and radiolabeling of trypanosome cell free systems
Trypanosome membranes were prepared from wild-type cells and TbALG3 null mutant cells as described by Masterson et al. (1989). Cell lysates were thawed on ice and washed twice with a HKML buffer (50 mM HEPES, pH 7.4, 25 mM KCl, 5 mM MgCl₂, 1 µg of leupeptin/mL, 0.1 mM TLCK) and resuspended at a concentration of 1 x 10⁹/mL in 2 x incorporation buffer (HKML buffer with 5 mM MnCl₂ and 1 mM DTT). The membranes (2 x 10⁷ cell equivalents) were transferred to another tube containing GDP-[3,4-³H]Man and 2 mM UDP-GlcNAc for 5 min at 30°C. The reaction was chased with 1 mM nonradioactive GDP-Man for 20 min at 30°C. Reactions were terminated by adding chloroform and methanol to reach a final CHCl₃:CH₃OH:H₂O ratio of 10:10:3 (v/v/v) and sonicated in a sonicating water bath for 10 min. After overnight extraction, the chloroform/methanol/water extract was centrifuged and the supernatant dried under a stream of nitrogen and partitioned between butan-1-ol and water, as described in Acosta-Serrano et al. (2004). The labeled glycolipid products were treated with 100 µL of 0.1 M TFA, 98°C, 15 min, cooled on ice and extracted twice with butan-1-ol to remove labeled GPIs. The aqueous phase, containing the acid-released glycans from the Dol-PP-oligosaccharides, were dried under a stream of nitrogen and reduced with 100 µL of 0.5 M NaBH₄, 3 h, room temperature. Excess reductant was destroyed by adding 100 µL 1 M acetic acid and the reduced labeled glycans were desalted by application to a column of 0.2 mL AG50 (H⁺ form) and elution with 0.8 mL water. The eluate was freeze-dried and residual boric acid was removed by coevaporation (twice) with 250 µL methanol, 5% acetic acid and twice with 250 µL of methanol. Residual acetic acid was removed by coevaporation with 50 µL of toluene. The radiolabeled glycans were dissolved in 40% propanol and applied to silica gel 60 high-performance thin layer chromatography plates (Merck, Nottingham, UK) and developed with butan-1-ol/ethanol/water (4:3:3, v/v/v). The plate was sprayed with En³Hance and [³H]-labeled products were detected by fluorography by exposing the Kodak-X-Omat X-ray film at –80°C.

Purification and MALDI-TOF analysis of procyclins
The procyclins were purified from 5 x 10¹⁰ freeze-dried procyclic form trypanosomes by organic solvent extraction followed by octyl-Sepharose chromatography (Ferguson et al. 1993; Treumann et al. 1997; Mehlert et al. 1999). Samples (approx. 250 pmole) of octyl-Sepharose-purified procyclins were dried and treated with 50 µL of ice-cold 50% aqueous hydrogen fluoride (aq. HF) for 24 h at 0°C to cleave the GPI anchor ethanolamine-phosphate bond. Some preparations were further treated with 50 µL of 40 mM trifluoroacetic acid, 100°C for 20 min to cleave Asp-Pro bonds and remove N-glycosylated N-termini. Some aq. HF-treated procyclin samples were deglycosylated with 15 units of peptide N⁴(N-acetyl-β-glucosaminyl) asparagine amidase F (PNGase F) (Roche, Burgess Hill, UK) in 2 µL of 0.25 M sodium phosphate buffer (pH 7.5). The samples were dried and redissolved in 5 µL of 0.1% trifluoroacetic acid and aliquots (1 µL) of each sample were mixed with 1 µL of 10 mg/mL -cyano-4-hydroxycinnamic acid as the matrix acid in 50% acetonitrile, 0.1% trifluoroacetic acid and analyzed by linear-mode negative-ion MALDI-TOF on an ABI Voyager DE-STR instrument. The accelerating voltage was 25,000 V, and the grid voltage was set at 93% with an extraction time delay of 400 ns. Data were collected manually at 500 shots per spectrum with the laser intensity set at 2000 V.

sVSG isolation
Soluble form VSG was isolated from 2 x 10⁸ bloodstream form T. brucei cells as described in Cross (1975, 1984). Briefly, cells were chilled on ice for 10 min, centrifuged at 2500 x g for 10 min, washed in an isotonic buffer and resuspended in 300 µL of 10 mM sodium phosphate buffer, pH 8.0, containing 0.1 mM TLCK, 1 µg/mL leupeptin, and 1 µg/mL aprotinin. After 5 min at 37°C, the mixture was cooled on ice and centrifuged (14,000 x g, 5 min). The supernatant was applied to a small (0.2 mL) DE52 anion exchange column equilibrated in a 10 mM sodium phosphate buffer, pH 8.0 and eluted with 0.8 mL of the same buffer. The entire column eluate (containing about 100 µg sVSG) was concentrated and diafiltered with water on a Microcon YM-10 concentrator (Millipore, Watford, UK) and recovered in 100 µL of water.

ES-MS analysis of intact VSG
sVSG obtained from small scale VSG purification was diluted to 0.05 µg/µL in 50% methanol, 1% formic acid, loaded into nanotips (Micromass-type F). The positive-ion electrospray mass spectra were recorded on the ABI QSTAR-XL system with tip potential of 900 V and declustering potential of 60 V. Data were deconvolved by using the ABI Analyst software and using the Bayesian protein reconstruct program.

ES-MS and ES-MS/MS analysis of Pronase glycopeptides
Aliquots of sVSG (approximately 50 µg in 50 µL water) were mixed with 5 µL of 1 M ammonium bicarbonate and 10 µL of 1 mg/mL Pronase in 5 mM calcium acetate and incubated at 37°C for 36 h. The Pronase glycopeptides were purified on Envicarb graphitized carbon microcolumns which were prepared as follows: the contents of an Envicarb car