Characterization of a ricin-resistant mutant of Leishmania donovani that expresses lipophosphoglycan

Abstract

The abundant cell-surface lipophosphoglycan (LPG) of Leishmania parasites plays a central role throughout the eukaryote's life cycle. A number of LPG-defective mutants and their complementing genes have been isolated and have proven invaluable in assessing the importance of LPG and related glycoconjugates in parasite virulence. While ricin agglutination selection protocols frequently result in lpg− mutants, one  Leishmania donovani variant we isolated, named JABBA, was found to be lpg+. Procyclic (logarithmic) JABBA expresses significant amounts of a large-sized LPG, larger than observed from procyclic wild type but similar in size to LPG from wild type from metacyclic (stationary) phase. Structural analysis of the LPG from logarithmically grown JABBA by capillary electrophoresis protocols revealed that it averaged 30 repeat units composed of the unsubstituted Gal(β1,4)Man(α1)-PO₄ typical of wild-type L. donovani. Analysis of JABBA LPG caps indicated that 20% is branched trisaccharide Gal(β1,4)[Glc(β1,2)]Man and tetrasaccharide Gal(β1,4)[Glc(β1,2)Man(α1,2)]Man instead of the usual Gal(β1,4)Man and Man(α1,2)Man terminating caps. Consistent with these structural observations, analyses of the relevant glycosyltransferases in JABBA microsomes involved in LPG biosynthesis showed a 2-fold increase in elongating mannosylphosphoryltransferase activity and up-regulation of a β-glucosyltransferase activity. Furthermore, the caps of JABBA LPG are cryptic in presentation as shown by the loss of binding by the lectins, ricin, peanut agglutinin and concanavalin A and reduced accessibility of the terminal galactose residues to oxidation by galactose oxidase. These results indicate that LPG from JABBA is intriguingly similar to the larger LPG in wild-type parasites that arises following the differentiation of the non-infectious procyclic promastigotes to infectious, metacyclic forms.

Introduction

Protozoan parasites of the genus Leishmania afflict at least 12 million people worldwide (Kaye and Scott 2011). In the extracellular part of its digenetic life cycle, logarithmically growing, procyclic promastigotes bind to and replicate in the midgut of their sand fly vector. In a process called metacyclogenesis, the promastigotes differentiate into non-dividing, virulent metacyclic forms and detach from the gut wall for transmission when the sand fly takes a new blood meal. In the vertebrate host, the parasites infect macrophages where they differentiate into intracellular, non-motile amastigotes (de Assis et al. 2012).

The parasites produce a characteristic family of phosphoglycan (PG)-containing glycoconjugates that includes membrane-bound lipophosphoglycan (LPG) and proteophosphoglycan (mPPG), and secreted PG, PPG and acid phosphatase. In addition to LPG and mPPG, the dense glycocalyx surface of Leishmania contains various glycosylinositol phospholipids (GIPLs) and glycosylphosphatidylinositol-anchored proteins. The various glycoconjugates play crucial roles for the survival of the parasite throughout its life cycle (Turco and Descoteaux 1992; Beverley and Turco 1998; Descoteaux and Turco 1999; Liew et al. 1999), including ligand-mediated midgut attachment and protection against the action of digestive enzymes within the sand fly (Sacks et al. 1995, 2000; Pimenta et al. 1997; Jecna et al. 2013). In the mammalian host, PG-containing glycoconjugates have been implicated in essential steps such as complement activation, resistance to complement-mediated lysis, and macrophage invasion and survival (Turco and Descoteaux 1992; McConville et al. 1993; McConville and Ralton 1997; Aebischer et al. 2005; Ibraim et al. 2013).

The prototype of PG-glycoconjugates is LPG, having a basic structure composed of four domains (Figure 1A): (i) a 1-O-alkyl-2-lyso-phosphatidyl(myo)inositol anchor, (ii) a glycan core with the structure of Gal(α1,6)Gal(α1,3)Gal_f(β1,3)[Glc(α1-PO₄)-6]Man(α1,3)Man(α1,4) GlcN(α1,6), (iii) the PG domain consisting of Gal(β1,4)Man(α1)-PO₄ repeat units and (iv) an oligosaccharide-phosphate cap. During metacyclogenesis, LPG undergoes two major modifications: the number of repeat units doubles from ∼15 to ∼30 (McConville et al. 1992; Sacks et al. 1995) and the sugars attached to the PG domain are altered in a species-dependent manner (McConville et al. 1992; Mahoney et al. 1999; Dobson et al. 2010). In addition, LPG is absent in  Leishmania donovani in its amastigote-cell stage (McConville and Blackwell 1991).

Fig. 1.

Schematic diagram of the LPG domains and structure of LPG. (A) Wild-type LD4, (B) JABBA.

Leishmania is generally diploid and lacks a manipulative sexual cycle (Panton et al. 1991; Cruz et al. 1993), rendering mutant creation difficult since both alleles of a single gene have to be altered. Mutation recovery is extremely low (10⁻⁷), even after heavy mutagenesis (Iovannisci and Ullman 1984; Kaur et al. 1988; King and Turco 1988). LPG offers a good target for mutagenesis, however, because screening mutants defective in LPG is relatively straightforward. Generating lpg⁻ mutants is feasible due to the availability of excellent selection tools, such as the lectin ricin agglutinin, which recognizes terminal β-linked galactose residues, and the PG-specific monoclonal antibody CA7AE, which binds the Man-PO₄-Gal-Man-PO₄ portion of repeat units (Tolson et al. 1989). Previously, four lpg⁻ mutants were successfully isolated using this selection method (Descoteaux et al. 1995, 1998, 2002; King and Turco 1988). Functional complementation of the mutants with a cosmid library of Leishmania DNA resulted in the isolation of four complementing genes involved in LPG synthesis (Ryan et al. 1993; Beverley and Turco 1995, 1998; Descoteaux et al. 1995, 2002), which can be divided into two groups according to their mode of action (Ryan et al. 1993; Beverley and Turco 1998). Two biosynthetic genes, LPG1 and LPG4A, complemented the mutants R2D2 and JEDI, respectively. LPG1 encodes a putative galactofuranosyltransferase involved in galactofuranose addition in the glycan core of LPG (Ryan et al. 1993), while LPG4A is required for the ‘elongating’ mannosylphosphoryltransferase activity for elongation of PG chains (Descoteaux et al. 1998; Xu et al. unpublished data). Two compartmentalization/biogenesis genes LPG2 and LPG3 were isolated by functional complementation of the PG mutants C3P0 and OB1, respectively. LPG2 encodes for a Golgi-specific GDP-Man transporter necessary for PG synthesis, resulting in the lack of the PG and cap domains in the LPG of C3PO (Descoteaux et al. 1995; Ma et al. 1997; Hong et al. 2000). LPG3 encodes a putative chaperone of the GRP94/HSP90 family (Descoteaux et al. 2002) that is necessary for the first Gal incorporation in the PG glycoconjugates.

While most of the lpg− mutants that have been reported have phenotypes consistent with loss-of-function alleles, we report an interesting mutant that is lpg+. This mutant named JABBA was isolated based on resistance of L. donovani to agglutination by ricin agglutinin. In this manuscript, we show that JABBA expresses a hyper-phosphoglycosylated LPG. Complicating this mutant is the unexpected presence of glucose substitutions in the LPG cap. Thus, JABBA appears to synthesize a “metacyclic” version of the L. donovani LPG and may represent a “gain-of-function” mutant with the induction of glucose substitutions.

Results

Analysis of LPG from L. donovani and JABBA by SDS–PAGE

Our laboratory pioneered the generation of Leishmania mutants by negative selection using lectin agglutination (Descoteaux et al. 1998; King and Turco 1988). Most L. donovani mutants selected based on resistance to agglutination by ricin (Descoteaux et al. 1995, 1998, 2002; King and Turco 1988) were found to be deficient in LPG synthesis and expression. In remarkable contrast, one of the ricin-resistant mutants, named JABBA, that we obtained in our selection protocol does express considerable amounts of LPG (30 pmol/10⁷ cells) versus wild-type L. dovovani (LD4) (100 pmol/10⁷ cells). To assess the relative sizes of LPG extracted from wild type and JABBA, LPG was extracted from logarithmically (procyclic) and stationary phase (metacyclic) parasites and subjected to gel electrophoresis. As expected, all of the LPGs migrated as broad bands consistent with their polydisperse nature. LPG from logarithmic and stationary phase wild-type migrated as previously observed with the stationary phase LPG having a larger average molecular weight (Sacks et al. 1995), whereas the LPG from JABBA revealed substantially increased sizes after extraction from both phases of growth (Figure 2).

Fig. 2.

Representative SDS–PAGE of purified LPG from logarithmic phase and stationary phase wild-type LD4 and JABBA parasites. Following electrophoresis, the gel was stained with Stains-all solution. Each panel is from a different gel, but representative of the migratory differences seen on all.

Determination of the number of Gal-Man repeating units in JABBA LPG

An assessment of the identity of the repeat units (Figure 1) of the JABBA LPG was obtained upon mild acid hydrolysis of LPG, treatment with alkaline phosphatase to dephosphorylate the sugars, and the dephosphorylated LPG substituents were resolved by fluorophore-assisted carbohydrate electrophoresis (FACE) analysis along with appropriate standards (Figure 3). Along with additional data (not shown), the sole repeat units in the LPG from JABBA were found to be the typical 6Gal(β1,4)Man(α1)-PO₄ disaccharides identical to LD4 LPG (Sacks et al. 1995).

Fig. 3.

Fluorophore-assisted carbohydrate electrophoresis of the dephosphorylated repeat units of LPG. The band from JABBA co-migrating with the Gal(β1,4)Man disaccharide from wild-type LD4 repeat unit was determined to be identical based on β-galactosidase digestion and the linkage was confirmed by partially methylated alditol acetates by GC–MS.

To determine the number of repeat units in the structurally heterogeneous LPG, quantification of the average number of repeat units was performed via a capillary electrophoresis-based protocol that we developed and published earlier (Barron and Turco 2006). The protocol involves deamination of LPG with nitrous acid to effect cleavage of glucosaminyl-inositol linkage in LPG (Figure 1) forming 2,5-anhydromannose at the reducing end of the delipidated PG. Then a strong acid hydrolysis of the deaminated LPG and CE of the monosaccharides enables a calculation of the ratio of mannose (derived from the Gal(β1,4)Man repeat disaccharides) to 2,5-anhydromannose as the indication of the number of repeat units. As expected for the wild-type, the LD4 LPG was confirmed as having an average of 15 and 30 repeat units derived from LPG extracted from logarithmic and stationary phase parasites, respectively. Consistent with the results from SDS–PAGE, LPG from logarithmically grown JABBA was found to have an average 30 repeat units while from stationary cells the average number was 47 (Table I).

Table I.

Average number of repeating units per molecule of LPG in L. donovani and JABBA parasites based on the growth phase

Parasite	Growth phase	Average repeats
LD4	Logarithmic	15a
	Stationary	30
JABBA	Logarithmic	30a
	Stationary	47

^aThe P-value using Student's t-test for these two values is <0.0001.

Structural analysis of JABBA LPG-terminal cap isoforms

The cap isoforms from LPG were generated by mild acid hydrolysis, purified by ionic exchange and resolved by Dionex HPLC. The major isoform of caps from wild-type LPG was determined as previously reported (Greis et al. 1992; Thomas et al. 1992): Gal-Man-Man with lesser amounts of Man-Man-Man and Gal-Man-Man-Man and the disaccharides Gal-Man and Man-Man (Figure 4A).

Fig. 4.

Dionex HPLC of cap isoforms of LPG. The cap isoforms of wild-type LD4 and JABBA were isolated and resolved by Dionex HPLC. Above the peaks are structures that were able to be identified following their isolation and characterization. In the profile from JABBA cap isoforms, the Gal-Man and Glc-Man isoforms could not be resolved by Dionex HPLC. (A) Wild-type LD4; (B) JABBA.

Importantly, the main isoforms from logarithmic phase JABBA LPG were chromatographically distinct compared with the tri- and tetrasaccharide caps, previously characterized in wild-type L. donovani (Greis et al. 1992; Thomas et al. 1992). The cap mixture from JABBA LPG was treated with β-glucosidase, and the unique tri- and tetrasaccharide peaks were eliminated (data not shown), suggesting that β-glucose was responsible for the chromatographic shift (Figure 4B) in retention times. Both tri- and tetrasaccharide cap isoforms were also sensitive to β-galactosidase treatment, but resistant to α-mannosidase digestion, showing the suggested structure of the two unique isoforms to be the branched Galβ[Glcβ]Man and Galβ[Glcβ-Man]Man. Cap isoforms from JABBA grown in stationary phase did not possess glucose residues as shown by the lack of glucose in strong acid hydrolyzates of the caps and by resistance of the caps to β-glucosidase digestion (data not shown).

The glycosidic linkage verification was performed by derivatizing the logarithmic phase cap isoforms to their permethylated alditol acetates and analysis of the adducts by GC–MS. Disaccharides were analyzed as a mixture while the individual tri- and tetrasaccharides were analyzed independently. Cap isoforms from wild-type LD4 LPG were as expected with disaccharides Gal(β1,4)Man and Man(α1,2)Man and the trisaccharides Man(α1,2) [Man(α1,4)]Man and Gal(β1,4)[Man(α1,2)]Man and the tetrasaccharide Man(α1,2)[Gal(β1,4)]Man(α1,2)Man. For the logarithmic JABBA cap isoforms, the structures were found to be the unique glucosylated isoform structures Glc(β1,2)Man as a disaccharide, Gal(β1,4)[Glc(β1,2)]Man as a trisaccharide, and Gal(β1,4)[Glc(β1,2)Man(α1,2)]Man as the tetrasaccharide, along with the non-glucosylated isoforms found in wild-type LD4 LPG (Table II). GC–MS of partially methylated alditol acetates from a bulk sample of JABBA caps from LPG is shown in Figure 5. The results indicated that β1,2-glucose residues are found on ∼20% of total cap isoforms in the logarithmic phase JABBA LPG.

Table II.

Cap isoforms and relative percentages found on logarithmic JABBA LPG

Oligosaccharide	Cap structures	Range (%) of total samplea
Disaccharides	Gal(β1,4)Man	50–80
	Man(α1,2)Man
	Glc(β1,2)Man
Trisaccharides	Gal(β1,4)[Man(α1,2)]Man	10–46
	Gal(β1,4)[Glc(β1,2)]Man
Tetrasaccharide	Gal(β1,4)[Glc(β1,2)Man(α1,2)]Man	0–4

^aCalculations were based on area percentages from capillary electrophoresis and permethylation experiments.

Fig. 5.

GC–MS of partially methylated alditol acetates from a bulk sample of JABBA caps from LPG. GC retention times were determined using Gal(β1,4)Man, Man, Gal and Cellobiose standards. 2-Man and 2,4-Man were identified matching wild-type GC retention times and MS spectra. Inositol is used as an internal standard. T-Gal, terminal galactose; T-Man, terminal mannose; T-Glc, terminal glucose; 2-Man, 2-linked mannose; 4-Man, 4-linked mannose; 2,4-Man, 2,4-linked mannose.

Analysis of the core-PI anchor of JABBA LPG

The lipid anchor of the JABBA LPG was examined by first treating the LPG with nitrous acid deamination that removes the lipid from the PG portion (Orlandi and Turco 1987) The lipid was analyzed by TLC and found to be chromatographically identical to the lipid anchor from wild-type LPG (data not shown). Similarly, the glycan-core region was evaluated by depolymerizing the PG portion to mild acid hydrolysis to remove the repeat units and caps. The glycan core was subjected to analysis by capillary electrophoresis along with the authentic glycan core from wild-type LPG and both were identical in retention times (data not shown). Thus, these results indicated that, as expected, there is no alteration in glycan core–PI of the JABBA and wild-type LPGs and that the basis of the resistance of JABBA to ricin agglutination was likely to be found in the repeat units and/or cap isoforms.

Analysis of JABBA LPG by lectin agglutination

The JABBA clone was initially selected after mutagenesis by their resistance to agglutination by ricin, a bivalent lectin that binds terminal β-linked galactose residues (Wu et al. 1988) The inability of LPG on the surface of JABBA to provide ricin-binding sites despite having terminal β-linked galactose residues might be explained by masking of its cap structure similar to observations previously reported concerning conformational changes with LPG from metacyclic forms of wild-type L. donovani (Sacks et al. 1995). To determine whether the loss in lectin binding to the parasite surface was likely the result that the terminal capping sugars are poorly exposed on the surface of JABBA, we conducted studies examining the availability of LPG cap binding with several lectins and with accessibility to the cap to oxidation by galactose oxidase.

To assess terminally exposed sugars of LPG, live promastigotes were used, instead of purified LPG, in the lectin-binding studies since the LPG on the promastigotes' surface would be in its native state. Furthermore, since LPG is known to undergo developmental changes in structure as a function of growth phase, promastigotes grown in logarithmic and stationary phase were used. As shown in Figure 6, in all instances of exposure of JABBA promastigotes to the terminally exposed β-galactose-binding lectins ricin or peanut agglutinin (PNA) or to the α-mannose-binding lectin Con A resulted in a consistent 2- to 3-fold resistance to agglutination compared with stationary phase wild type. Since we have shown that the major compositional distinctions between wild-type LD4 and JABBA LPG reside in the capping sugars, these results are consistent with the notion of the loss of exposure of terminal sugars on the JABBA LPG.

Fig. 6.

Lectin-mediated agglutination analyses of whole parasites. Logarithmic phase and stationary phase parasites were incubated with ricin (A), Con A (B) or PNA (C) for 30 min. The percent of parasites resistant to agglutination to the respective lectins is shown. The P-value of <0.01 is determined using Student's t-test for differences among the data points, with the exception of logarithmic and stationary phase JABBA exposed to Con A.

Accessibility of terminal galactose molecules on JABBA LPG to galactose oxidase

To further evaluate the availability of LPG-terminal sugars for protein interactions, promastigotes were treated with galactose oxidase. If terminal β-linked galactose residues are accessible, the enzyme oxidizes the 6-carbon of such terminal galactose residues to an aldehyde group. Accordingly, promastigotes harvested at logarithmic and stationary phases of growth were subjected to galactose oxidase treatment for 1 h and then LPG was extracted and purified. The cap isoforms of LPG were generated by mild acid hydrolysis, separated from the core-repeat unit mixture by 1-butanol : water partitioning followed by anionic exchange chromatography, and then subjected to strong acid hydrolysis to obtain monosaccharides. The cap monosaccharides were resolved by capillary electrophoresis where the percentage of galactose remaining in galactose oxidase treated and untreated controls was normalized to the amount of mannose. The results are shown in Table III as percentage of galactose oxidized by galactose oxidase. As previously reported (Sacks et al. 1995), the terminal β-linked galactose residues of the LPG cap of wild-type LD4 promastigotes were found to be significantly inaccessible to the enzyme when the parasites were grown to stationary phase (Table III) compared with those harvested in logarithmic phase. Similar experiments with JABBA resulted in an even more reduction in terminal β-linked galactose residues of the LPG cap obtained in both logarithmic cells and substantially more so in parasites obtained in stationary phase where only 8% of LPG capping galactose residues were accessible to oxidation. Thus, surface treatment of JABBA with galactose oxidase resulted in extremely poor oxidation compared with wild type, suggesting that the terminal β-linked galactose residues on the neutral, capping sugars of the JABBA LPG were inaccessible to the enzyme.

Table III.

The percentage galactose available to galactose oxidase based on capillary electrophoresis quantification

Parasite	Growth stage	Galactose available (%)a
LD4	Logarithmic	>85
	Stationary	35
JABBA	Logarithmic	31
	Stationary	8

^aAverage was determined using seven experiments.

Assessment of initiating mannosylphosphoryltransferase, elongating mannosylphosphoryltransferase and β-glucosyltransferase activities

The structural analysis of the LPG from JABBA indicated an approximate doubling in the number of 6Gal(β1,4)Man(α1)-PO₄ repeat units and a surprising glucosylation of the caps. To examine the possible explanation for the increase in repeat unit number, we measured the relative activities of the iMPT, which uses Gal(α1,6)Gal as acceptor and is responsible for the initiating the assembly of the repeat units onto the glycan core–PI anchor, and eMPT, which uses the Gal(β1,4) Man(α1)-PO₄ repeats as acceptor and is responsible for elongating the repeat units. Based on the structural information, it would be reasonable to expect that iMPT would be similar in activity whereas eMPT would have greater activity in JABBA compared with wild-type LD4. The results of the iMPT assay showed a slight increase in activity from extracts of JABBA compared with wild-type LD4 (Figure 7, panel A). The eMPT activity from JABBA extracts was significantly increased (75 pmol/mg protein/h) compared with extracts from WT (28 pmol/mg protein/h) (Figure 7, panel B). This >2-fold increase in eMPT activity is in accord with the doubling in the number of repeat units in the JABBA LPG.

Fig. 7.

Comparative iMPT and eMPT activities in microsomal preparations of wild-type LD4 and JABBA. Microsomal preparations from wild-type LD4 (white bars) and JABBA (black bars) were assayed for iMPT and eMPT activities as described in Materials and Methods. Panel A, iMPT activity; Panel B, eMPT activity. The P-value of <0.05 is determined using Student's t-test for differences among the data points in the eMPT assay.

The unexpected presence of glucosylated cap isoforms on JABBA LPG indicated that a β-glucosyltransferase was present and active. Although wild-type does not have a glucose-containing cap, microsomal preparations did show minimal β-glucosyltransferase activity (Figure 8). In vitro extracts of JABBA had twice the β-glucosyltransferase activity compared with wild-type preparations (Figure 8). Product characterization of the in vitro assay did indicate that β-glucosyltransferase activity of JABBA was likely responsible for the expression of β-glucose-containing cap isoforms, as the in vitro product was sensitive to both β-galactosidase and β-glucosidase.

Fig. 8.

Comparative β-glucosyltransferase activities in microsomal preparations of wild-type LD4 and JABBA. Microsomal preparations from wild-type LD4 (white bar) and JABBA (black bar) were assayed for β-glucosyltransferase activities as described in Materials and Methods. The P-value of <0.05 is determined using Student's t-test for differences among the data points.

Discussion

Lectin-resistant mutants of Leishmania and their genetic complementation have been successful in identifying a number of genes in LPG biosynthesis (Huang and Turco 1993; Ma et al. 1997; Descoteaux et al. 1998, 2002). JABBA is the first reported lectin-selected mutant that not only expresses LPG but also expresses a larger-sized version of this surface glycoconjugate and, furthermore, displays gain of function in its cap isoforms. Our results in this study are that the JABBA LPG consists of an average of 30 repeat units when the parasite is grown in logarithmic phase and 47 in stationary phase, along with the additional cap structures likely precludes an easy molecular explanation of the phenotype, but allows for a unique perspective on LPG in the lifecycle of the parasite (Figure 1, panel B).

The cap structure differences involve the presence of a β2-glucose residue on both the typical Gal(β1,4)Man(α1)-PO₄ and Man(α1,2) [Man(α1,4)]Man(α1)-PO₄ cap isoforms forming the branched trisaccharide Gal(β1,4)[Glcβ]Man(α1)-PO₄ and tetrasaccharide Gal(β1,4)[Glc(β)Man(α1,2)]Man(α1)-PO₄. Glucose residues present in LPG was first reported in L. mexicana LPG (Ilg et al. 1992). Subsequently, β-glucose-containing domains in LPG are found in LPG expressed by L. donovani (Indian isolate) and L. chagasi (Brazilian isolate) with Glc(β1,3)Gal(β1,4)Man(α1)-PO₄ repeat units and cap (Mahoney et al. 1999; Soares et al. 2002). Unlike the Indian and Brazilian isolates, the glucose residues is added to the mannose residues of JABBA LPG caps in a β1,2-configuration. This conclusion was obtained by permethylation experiments showing the cap isoforms maintain the wild-type linkages of Gal(β1,4)Man and Man(α1,2)Man while introducing a new linkage position via the β-glucose residue. Microsomal preparations from JABBA were found to be active for the β-glucosyltransferase responsible for the addition of the β-glucose residues of the cap isoforms. Although in vitro assays with wild type did possess β-glucosyltransferase activity, β-glucose residues are not found in the cap isoforms of LPG in vivo, suggesting posttranslational regulation to inhibit the enzyme in whole parasites.

While the cap structures were novel, the identity of the repeat units were unchanged in JABBA compared with wild-type LPG consisting solely of Gal(β1,4)Man(α1)-PO₄ disaccharide units. The doubling in the number of repeat units, along with the presence of glucose-terminating caps, was totally unexpected. To gain information for the most likely possibility for this change of the number of repeat units, eMPT and iMPT activity assays were conducted and compared. The eMPT activity was more than doubled compared with that of wild type, consistent with the increase the number of repeat units in LPG. The iMPT activity was slightly increased in JABBA extracts, but not nearly as substantial as eMPT activity (Carver and Turco 1991, 1992; Mengeling et al. 1997; Descoteaux et al. 1998).

The larger size of JABBA LPG is intriguingly similar to the phenomenon of differentiation of the non-infectious procyclic promastigotes to infectious, metacyclic forms. It has been proposed that a doubling repeat units from the procyclic number of ∼15 to the metacyclic number of ∼30 causes a conformational change that in LPG, resulting in a cryptic cap (Sacks et al. 1995). Consequently, the conformational change prevents the terminal sugar residues from being recognized by βGal- and αMan-binding lectins, such as peanut agglutinin and concanavalin A.

As was previously demonstrated in metacyclic wild-type parasites, lectin and galactose oxidase accessibility were used to determine whether the caps of LPG in its native conformation in whole cells of JABBA were cryptic and, thus, whether or not the additional cap isoforms found in logarithmic phase contributed to the ricin resistance of the parasite. While JABBA was significantly greater in resistance to agglutination by ricin, the greatest resistance was in stationary phase, suggesting that the number of repeat units was the basis of masking of the caps and the resistance to agglutination. Furthermore, JABBA was more resistant to two other surface-binding lectins: the β-galactose-binding lectins PNA and α-mannose-binding lectin Con A. Galactose oxidase treatment of whole cells was another critical experiment to assess the actual inavailability of the cap isoforms on the cell surface. Thus, the loss of exposed terminal sugars on LPG was clearly demonstrated in the current studies by the reduction of lectin binding and enzyme accessibility to the JABBA parasites.

The masking of LPG caps been previously reported in other situations, both in determining the relative structure (Sacks et al. 1995) and in NMR work showing LPG as a molecule is both helical and flexible (Homans et al. 1992). Since a similar conformation situation alone would have resulted in the selection of a ricin-resistant mutant, it is unclear why JABBA also terminates its LPG caps with glucose. Nevertheless, JABBA appears to synthesize a “metacyclic” version of the L. donovani LPG and appear to represent a “gain-of-function” mutant with the induction of glucose substitutions. Whether it is also a “general” metacyclic mutant blocked in the metacyclic stage or specifically deregulated in controlling the number of PG chains is speculative and remains to be investigated. Although the molecular basis of the conformational modification of metacyclic LPG is unknown, this and previous data (Sacks et al. 1995) suggest or reinforce that a cryptic cap in the metacyclic phase is likely to occur (Sacks and Kamhawi 2001).

Materials and methods

Materials

Chemicals and reagents were obtained as follows: Bacto-Brain Heart Infusion from Difco Laboratories (Mauston, WI); alkaline phosphatase ( Escherichia coli), β-glucosidase (almond), β-galactosidase (E. coli), α-mannosidase (Jack Bean) from Sigma; fetal bovine serum from Atlanta Biologicals; 8-aminopyrene-1,3,6-trisulfonate (APTS) labeling from Ab-Sciex; C18 Sep-Pak Columns from Waters; Protease Inhibitor from Roche; and BCA reagent kit from Pierce.

Parasites and mutagenesis

Leishmania donovani (MHOM/SD/00/1S-2D) and JABBA were maintained in M199 with 10% fetal bovine serum as described previously (Kapler et al. 1990). Parasites were selected and mutagenized as done previously (King and Turco 1988). The concentration of cells used in ricin selection was between 2 × 10⁶ and 2 × 10⁷ cells/mL.

Extraction and purification of LPG

JABBA and LD4 promastigotes were harvested at 5 × 10⁵ to 1 × 10⁶ parasites/mL from brain heart infusion supplemented with heme and adenosine (Orlandi and Turco 1987). Parasites were extracted using the method as detailed previously (Proudfoot et al. 1996), followed by purification of LPG by chromatography on octyl-Sepharose using a 20–60% propanol gradient (Proudfoot et al. 1996). Fractions were spotted on silica TLC plates and charred using orcinol for identification. For stationary phase, promastigotes were harvested at >2 × 10⁷ parasites/mL and LPG was extracted via the modified Folch method as described previously, purified by chromatography on phenyl-Sepharose and elution with Solvent E (McConville et al. 1990; Orlandi and Turco 1987).

Gel electrophoresis of LPG

Approximately 10 µg of LPG was subjected to electrophoresis on an SDS–PAGE gel for 1 h at 120 V using 5 and 12% polyacrylamide for the stacker and main gel, respectively. The gel was rinsed three times for 10 min in 25% isopropanol and left rocking for 16 h. The isopropanol was removed, and the gel was placed in Stains-all solution (64.5% water, 25% isopropanol, 5% formamide, 5% Stains-all 1% 1.5 M Tris–HCl, pH 8.8) for 2 h (Bahr et al. 1993). The gel was destained in 40% ethanol.

Preparation of repeat units, caps and glycan core of LPG

LPG was depolymerized by mild acid hydrolysis with 0.02 N HCl for 10 min at 60°C. The hydrolyzate was dried under nitrogen and then evaporated three times with toluene. A 1 : 1 water-saturated butanol partition was used to separate the glycan core–lipid from the water-soluble repeat units and caps (McConville et al. 1987). The butanol extract was further purified over a Sep-Pak C18 column, dried, subjected to nitrous acid deamination to cleave the glycan core and lipid, and applied to chromatography on phenyl-Sepharose (Hardy and Townsend 1994). The collected water portion after butanol partition was applied to a column of AG1-X2 (acetate). Caps were eluted with 50% methanol and repeat units were eluted with 0.3 M NaCl in 50% methanol.

Dionex HPLC separation of caps

Samples were subjected to DX-500 HPLC (Dionex) using a CarboPac PA1 column (4 × 250 mm) with ED40 electrochemical detection as previously described (Descoteaux et al. 1998; Coelho-Finamore et al. 2011).

Flurophore-assisted carbohydrate electrophoresis

Oligosaccharide FACE analysis was performed as previously described (Lehrman and Gao 2003; Gao et al. 2013).

Thin-layer chromatography of lipid

Thin-layer chromatography of the nitrously deaminated lipid was performed as described previously (Orlandi and Turco 1987) using silica gel 60 TLC plates in a CHCl₃ : CH₃OH : 4.2N NH₄OH (9 : 7 : 2) system. The plate was visualized in an iodine chamber. Additional controls used on the plate included phosphatidylinositol, phosphatidylcholine and previously purified alkyl-lyso-phosphatidylinositol in addition to L. donovani and JABBA purified lipids.

Preparation of LPG for repeat unit quantification

LPG samples (125 µg) were prepared as previously described (Barron and Turco 2006) with the following modifications. LPG was subjected chemically deaminated by nitrous acid (Ferguson et al. 1985). The lipid portion was removed from the samples by passage of the sample through C18 Sep-Pak cartridges equilibrated in water. The aqueous-soluble material was dried by vacuum centrifugation and subjected to strong acid hydrolysis in 2 N trifluoroacetic acid for 3 h at 100°C. The multiple PO₄-Gal linkages in LPG are refractory to hydrolysis under these strong acidic conditions. The hydrolyzates were dried using vaccum centrifugation and then labeled in a vial in the dark of 4 µl for 90 min at 55°C with the fluorophore APTS (0.2 M) containing 1 M sodium cyanoborohydride, which labels the reducing termini of saccharides by reductive amination. The reaction was stopped by adding of 46 µl of water, diluting the sample 10-fold. Prior to injection in CE, the samples were further diluted in water at a ratio of 39 : 1.

Capillary electrophoresis for repeating unit quantification

APTS-derivatized monosaccharides were separated using a P/ACE MDQ Capillary Electrophoresis system as described previously (Barron and Turco 2006). For quantification, mannose and anhydromannose were used for standard curves and found to have a ratio of sensitivity to be 1.5–1.

GCMS analysis of partially methylated alditol acetates

The structure of the cap isoforms was determined using partially methylated alditol acetates on the microscale following the previously published protocol (Pettolino et al. 2012). The mass spectra were acquired by the University of Kentucky Mass Spectrometry Facility using a Shimadzu GCMS-QP500 equipped with a GC-17A gas chromatograph. Electron ionization mass spectra were recorded at 70 eV, scanning m/z 45–550 at 1 scan/s. The GC column was an Agilent DB-5 ms 30 × 0.25 mm with helium as the carrier gas. The column oven temperature program was 50°C with a 1 min initial hold to 280°C at 10°C/min with an injector temperature of 250°C and an interface temperature of 280°C.

Enzymatic treatments

Alkaline phosphatase treatment was done using 1 U in 1 mM Tris–HCl, pH 9, for 12 h at 37°C (Turco et al. 1987). β-Glucosidase treatment was performed for 12 h using 1 U in 200 mM ammonium acetate at 37°C (Mahoney et al. 1999). α-Mannosidase treatment was done for 12 h using 0.1 M sodium acetate, pH 4.5, with 1 U at 37°C.

Lectin agglutination assays

Agglutination assays using ricin, PNA or Con A were performed on logarithmic growing (2 × 10⁶ cells/mL) or cells grow in stationary phase (2 × 10⁷ cells/mL). A total of 5 × 10⁶ cells were incubated and rocked for 30 min in 5 mL of M199 containing ricin (10 μg/mL), PNA (100 μg/mL) or Con A (100 μg/mL). Cells (10 μL) were placed on a hemacytometer and free cells were counted (Andrade and Saraiva 1999).

Galactose oxidase treatment

Parasites (1 × 10⁹) were treated with 3 U of galactose oxidase in 500 μL of phosphate-buffered saline at 25°C for 1 h as previously described (Sacks et al. 1995). LPG was then purified using the Folch method (Orlandi and Turco 1987; McConville et al. 1990). The cap isoforms were isolated and then subjected to strong acid hydrolysis (2 N trifluoroacetic acid for 3 h at 100°C) and the monosaccharides were analyzed by capillary electrophoresis. A 75 µM bare fused silica capillary was used at 20 kV equilibrated in a 15 mM sodium borate, 2.5% methanol solution for 10 min.

Preparation of microsomal membranes

Microsomal membranes were prepared as previously described (Mengeling et al. 1997), with the following modifications: cultures of parasites (200 mL) were centrifuged at 2700 rpm for 7 min, and resuspended in 10 mL of lysis buffer containing 100 mM HEPES, pH 7.4, 50 mM KCl, 1 mM EDTA, pH 8.0, 10% glycerol and protease inhibitor. The suspension was subjected to nitrogen cavitation using a Parr bomb for 30 min at 1500 psi on ice. Subsequent centrifugations took place at 4°C for 3000 × g, 10,000 × g for 10 min and 100,000 × g for 1 h. The supernatant was removed, and the pellet was resuspended in membrane buffer containing 100 mM HEPES, pH 7.4, 50 mM KCl, 1 mM TLCK and 1 μg/mL leupeptin using a homogenizer. The final protein concentration was determined with a BCA reagent kit.

Initiating mannosylphosphoryltransferase assay

The iMPT in vitro assay was performed according to our previously reported assay (Mengeling et al. 1997) with the following modifications. The assay mixture contained 1 mg protein from microsomes, 15% glycerol, 2.5 mM thioglycerol, 10 mM MnCl₂, 5 mM MgCl₂, 4 mM dithiothreitol, 800 μM GDP-Man and 27 nanocuries/4 mM stachyose. The mixture was incubated for 4 h at 30°C and terminated on ice. Product purification was done as previously (Mengeling et al. 1997) and radioactivity was counted via scintillation.

Elongating mannosylphosphoryltransferase assay

The eMPT assay involved the preparation of soluble proteins from the microsomal preparations. These soluble microsomal proteins were prepared with final concentrations of 50 mM HEPES, pH 7.5, 150 mM KCl, 10 mM MnCl₂, 5 mM MgCl₂, 20% glycerol, 1 μg/mL leupeptin and 0.5% dodecyl-maltoside and 3 mg of microsomes. The microsomal mixture was rotated for 1 h at 4°C and then centrifuged at 100,000 × g for 1 h. The supernatant collected was used for the eMPT assay.

The eMPT activity was quantified as done previously (Carver and Turco 1991) with the following modifications using an assay mixture of 1 mM ATP, 0.5 mM dithiothreitol, 2 mM GDP, 10 mM MnCl₂, 5 mM MgCl₂, 225 nanocuries of [³H]-GDP-Man at 200 µM and 75 µL solubilized microsomal protein. An exogenous acceptor (25 µg) of JEDI LPG (Descoteaux et al. 1998), possessing a single Gal-Man-PO₄ repeat on the glycan core–PI, was used. The assay mixture was incubated for 1 h at 28°C. The reaction was terminated with 0.5 M EDTA and the radioactive product was purified via chromatography on phenyl-sepharose. The radioactive product was determined by scintillation counting.

β-Glucosyltransferase assay

The glucosyltransferase assay was quantified as done previously using 0.5 µC of 34 µM UDP-Glc (Mahoney and Turco 1999). LPG was extracted, and total radioactivity was counted via scintillation. For product identification, the product from nonradioactive β-glucosyltransferase assays was desalted on a mixed bed column (2 mL) of AG-50W and AG1-X8 and dried. One aliquot of the product was treated with mild acid as described previous, then desalted in the same fashion. Another aliquot was treated with mild acid hydrolysis (0.2 N HCl, 20 min, 60°C) and dephosphorylated with alkaline phosphatase for 12 h and desalted. The samples were then reductively labeled with APTS and sodium cyanoborohydride in THF overnight in the dark at room temperature. Product obtained after mild acid and alkaline phosphatase was also analyzed by exoglycosidase treatments. For analysis by oligosaccharide capillary electrophoresis, injection pressures were 5 psi for 5 s (Soares et al. 2004).

Conflict of interest statement

None declared.

Funding

This work was supported by NIH Grant AI020941 (to S.M.B. and S.J.T.).

Abbreviations

APTS, 8-aminopyrene-1,3,6-trisulfonate; FACE, fluorophore-assisted carbohydrate electrophoresis; LD4, L. dovovani; LPG, lipophosphoglycan; mPPG, membrane-bound proteophosphoglycan; PG, phosphoglycan; PNA, peanut agglutinin