|Home | About | Journals | Submit | Contact Us | Français|
Surface phosophoglycans such as lipophosphoglycan (LPG) or proteophosphoglycan (PPG) and glycosylinositol phospholipids (GIPLs) modulate essential interactions between Leishmania and mammalian macrophages. Phosphoglycan synthesis depends on the Golgi GDP-mannose transporter encoded by LPG2. LPG2-null (lpg2−) L. major cannot establish macrophage infections or induce acute pathology, whereas lpg2− L. mexicana retain virulence. lpg2− L. donovani has been reported to survive poorly in cultured macrophages but in vivo survival has not been explored. Herein we discovered that, similar to lpg2− L. major, lpg2− L. donovani promastigotes exhibited diminished virulence in mice, but persisted at consistently low levels. lpg2− L. donovani promastigotes could not establish infections macrophages and could not transiently inhibit phagolysosomal fusion. Furthermore, lpg2− promastigotes of L. major, L. donovani and L. mexicana were highly susceptible to complement mediated lysis. We conclude that phosphoglycan assembly and expression mediated by L. donovani LPG2 are important for promastigote and amastigote virulence, unlike L. mexicana but similar to L. major.
All Leishmania spp. are covered by a complex glycocalyx throughout the infectious cycle, and glycoconjugates are thought to be important factors promoting their virulence (reviewed in refs. Turco 1990; Ferguson 1999; Turco et al. 2001; Naderer et al. 2004). The promastigote glycocalyx is especially dense and contains high levels of phosphoglycans (PGs) comprised of polymeric [6Gal(β 1,4)Man(α1-PO4)-] disaccharide phosphate-based repeating units. The main structural distinction among PGs from various Leishmania spp. is the side chain sugar substitutions that branch off the disaccharide-phosphate backbone. Depending on the growth phase, the abundant promastigote surface glycolipid lipophosphoglycan (LPG) contains 15–30 PG repeating units, bearing an external capping oligosaccharide and anchored to the surface by a heptasaccharide glycosylphosphatidyl inositol anchor (reviewed in ref. Turco 1990). Proteophosphoglycans (PPGs) comprise a family of large proteins containing Ser-Thr rich regions to which the PG repeating units are covalently linked (reviewed in refs. Ilg 2000). Amastigotes lack significant LPG but retain PPG expression, and both stages express high levels of smaller glycosylinositol phospholipids (GIPLs) (Elhay et al. 1988; McConville et al. 1994). The GPI anchors of LPG and PPGS show varying degrees of structural identity or similarity to those of GIPLs and GPI-anchored proteins (reviewed in refs. Turco 1990; Ferguson et al. 1999; Turco et al. 2001; Naderer et al. 2004). Many studies have shown that purified (or synthetic) LPG, PPG and GIPLs have significant effects on parasite survival, attributed their ability to execute key steps of the infectious cycle such as suppression of host cell signaling and activation, and evasion of killing by activated complement (reviewed in refs. Descoteaux and Turco 1999; Naderer et al. 2004). However, one challenge to the interpretation of in vitro studies is that test glycoconjugate(s) are provided outside of the context of the parasite, leading to concerns about the effects of dosage and route of application (Beverley and Turco 1998). A further challenge is the fact that these PG-containing and GPI-anchored Leishmania glycoconjugates show varying extents of structural relatedness, raising concerns about their cross activity during in vitro assays.
A second approach to glycoconjugate function has thus been to generate mutants through either forward or reverse genetic approaches (Beverley and Turco 1998). Depending of the specific gene affected, individual or specific subsets of glycoconjugates can be defective in mutants. These living organisms permit assessment of glycoconjugate deficiency in the proper biological context. For example, L. major lpg1− mutants lack the galactofuranosyl transferase activity required for synthesis of the LPG core domain but other parasite surface components remain unchanged by this mutation (Sacks et al. 2000; Spath et al. 2000; Spath et al. 2003a). These LPG-deficient L. major exhibit increased susceptibility to complement- and oxidant-mediated toxicity, and show decreased survival in macrophage infections, although they retain the ability to suppress macrophage responses leading to NO and IL-12 production (Spath et al. 2000; Spath et al. 2003a; Spath et al. 2003b). As expected because LPG expression is greatly down-regulated in amastigotes (Turco and Sacks 1991; Moody et al. 1993), lpg1− L. major amastigotes are fully virulent (Spath et al. 2000).
The role of PGs, which constitute major portions of both LPG and PPGs, was first probed genetically through studies of the lpg2− mutants which lack the Golgi GDP-mannose transporter LPG2 required for PG synthesis (Descoteaux et al. 1995; Ma et al. 1997; Spath et al. 2003b). Like lpg1- mutants, L. major lpg2− mutants show heightened defects in promastigote virulence, but unlike lpg1- mutants, they are unable to replicate as amastigotes in macrophages and/or cause acute pathology. Nonetheless, they persist indefinitely at a low level (Spath et al. 2003b) and they induce protective immunity in mice (Uzonna et al. 2004; Kebaier et al. 2006). The defect in L. major lpg2− amastigotes was attributed to the lack of PGs in this stage, the only known glycoconjugate synthetic defect in lpg2− amastigotes of L. major, L. donovani and L. mexicana (Descoteaux et al. 1995; Ilg et al. 2001; Goyard et al. 2003; Spath et al. 2003b), and. However, recent studies of L. major challenged this assumption, using a double mutant (lpg5A−/lpg5B−) lacking PGs through ablation of the UDP-Gal transporters encoded by LPG5A and LPG5B (Spath et al. 2004; Capul et al. 2007b; Capul et al. 2007a). Remarkably this mutant shows a virulence phenotype very similar to that of LPG-deficient lpg1− L. major, which affects only promastigote virulence whereas amastigotes retain full virulence (Capul et al. 2007a). Thus the loss of amastigote virulence in the L. major lpg2− mutant is probably not merely due to the absence of PGs. The lack of involvement of PGs in amastigote virulence agrees with results obtained with lpg2− L. mexicana which, unlike lpg2− L. major, retain amastigote virulence and induce disease despite the absence of PGs (Ilg et al. 2001).
In this work we focus on the role of LPG2 in the virulence of L. donovani. Similar to L. major, L. donovani resides in a ‘tight’ as parasitophorous vacuole, which differs from the ‘spacious’ vacuole occupied by L. mexicana (Castro et al. 2006). Curiously, whereas the forward genetic complementation methodology was first developed and then applied to the identification of LPG synthetic genes in LPG-deficient mutants in this species (Ryan et al. 1993; Beverley and Turco 1998), the use of these L. donovani mutants for functional studies of virulence in animal infections was compromised by several factors. One was the use of heavy mutagenesis and the lack of sexual crossing to generate isogenic lines, which was solved by the use of targeted replacement methods (Spath et al. 2000). A second problem was the tendency of some Leishmania species and especially those causing visceral disease to rapidly lose virulence during in vitro culture.
Previous studies have shown that some strains of lpg2− L. donovani are unable to establish infections in macrophages in vitro (Lodge et al. 2006). However, the ability of these strains to survive in animal hosts has not previously been reported. In this work we make use of the observation that virulence can be maintained in L. donovani lines adapted for cycling between growth as promastigotes and axenic amastigotes despite extensive in vitro culture (Goyard et al. 2003; Debrabant et al. 2004), and generated lpg2− null mutants in such a strain, the LdBob line of L. donovani (Goyard et al. 2003). This mutant (and its complemented control line) has allowed us, for the first time, to study the role of LPG2 in a virulent L. donovani background in survival in mice, and to compare the requirement with the contrasting observations reported in L. major and L. mexicana. The LdBob lpg2− line additionally allowed us to confirm lpg2− phenotypes previously described such as macrophage survival and transient inhibition of phagolysosomal fusion, and extend this to susceptibility to lysis by complement.
All strains studied were derivatives of the L. donovani strain 1S2D (MHOM/SD/62/1S-CL2D) clonal line LdBob, which were grown alternately as amastigotes and promastigotes in serum-containing medium specific for each form as described (Goyard et al. 2003). Amastigotes were cultivated at 37ºC, 5% CO2, and promastigotes were grown at 26ºC. Parasites were converted between forms every 3 weeks.
Previously we described the homozygous LPG2 knockout line (formally, Δlpg2::HYG/Δ::HYG), referred to as lpg2− in this work, and a complemented derivative bearing an episomal LPG2 construct (formally, Δlpg2::HYG/Δlpg2::HYG [pX63NEO-LPG2]), referred to as lpg2−/+LPG2(e) here (Goyard et al. 2003). An integrated version of the complemented line was generated as follows: first, the L. donovani LPG2 gene was excised by XhoII digestion from pXG63HYG-LPG2 [strain B1544 (Descoteaux et al. 1995)], and inserted into either the SmaI or BglII sites of pIR1SAT, yielding the constructs pIR1SAT-LdLPG2(a) and pIR1SAT-LdLPG2(a) (strains B5041 and B5043, respectively). These constructs were digested with SwaI and the targeting fragment was isolated and introduced into L. donovani lpg2− by electroporation (Robinson and Beverley 2003). Clonal lines expressing the transfected construct were identified following plating on M199 media containing 5 μg/ml nourseothricin. All transfectant lines [Δlpg2::HYG/Δlpg2::HYG SSU:IR1SAT-LdLPG2(a) or LdLPG2(b), respectively] bear LPG2 integrated into the ribosomal RNA locus, where they are stably expressed at high levels from the rRNA promoter (Robinson and Beverley 2003). All lines showed good differentiation and similar levels of LPG expression, and are referred to here as lpg2−/+LPG2(i).
Complement sensitivity tests were performed as described (Spath et al. 2003a). Briefly, promastigotes in logarithmic or stationary phase growth were exposed to 2% fresh human serum for 30 min in the presence of propidium iodide, whose uptake by permeabilized cells was quantitated by flow cytometry.
Promastigote or amastigote proteins were separated on reducing 9% denaturing SDS polyacrylamide gels, transferred to Nytran, and blocked in 5% milk/PBS/0.01% Tween-20. Filters were incubated with the following antisera: CA7AE, specific for unsubstituted [6Gal(β 1,4)Man(α1-PO4)-] phosphoglycan repeats (Goyard et al. 2003); polyclonal antiserum generated in sheep to purified L. chagasi GP63 (1:10,000), or monoclonal antibody to α-tubulin (AB-1, 0.1 μg/ml, Oncogene, San Diego, CA) (Yao et al. 2002)
Leishmania mannans were extracted by chloroform:methanol:water extraction, and fluorophore labeled. Mannans were separated by fluorophore-assisted carbohydrate electrophoresis (FACE) as described (Capul et al. 2007a).
Female BALB/c mice (age 4–6 weeks) were purchased from Harlan Laboratories (Indianapolis, IN). Mice were infected with 107 stationary phase parasites belonging to different lines intravenously through the tail vein. At specified time points, animals were euthanized, organs were removed, and parasite loads were determined microscopically from touch preparations and organ weights as described (Stauber 1958; Wilson et al. 1987). All animal procedures were approved by the University of Iowa and the Iowa City VA institutional animal care and use committees.
Bone marrow cells (BMMs) from BALB/c mouse femurs were cultured at 37°C, 5% CO2 in RP-10 (RPMI with 10% heat-inactivated fetal calf serum, 2mM L-glutamine, 100 U of penicillin/ml, and 50 μg of streptomycin/ml [GIBCO, Carlsbad, CA.]) containing 20% L929 cell culture supernatant (American Tissue Type Collection, Manassas, VA) as a source of macrophage colony-stimulating factor. After 7 to 9 days, differentiated adherent macrophages were detached with 2.5 mg trypsin/ml plus 1 mM EDTA (GIBCO) (Coligan et al. 2007).
Human mononuclear cells were isolated from the peripheral blood of normal healthy donors by density sedimentation in Ficoll-Hypaque (Sigma Chemical Co., St. Louis, MO). Monocytes were separated by adherence to six-well plates (flow cytometry) for 2 to 3 h at 37°C and 5% CO2 in RP-10.
5.0 × 105 macrophages were allowed to adhere to coverslips in 24-well plates and infected with opsonized promastigotes or amastigotes at a multiplicity of infection (MOI) of 5:1 as described (Rodriguez et al. 2006). The infection was synchronized by centrifugation (3 min, 330 × g, 4 ºC) and infected macrophages were incubated in 5% CO2 at 37°C. Extracellular parasites were removed by rinsing macrophages 30 min post-infection. At each time point, coverslips were fixed and stained with Diff-Quik (Fisher Scientific).
Promastigotes were incubated in 5 μM carboxyfluorescein diacetate succinimidyl ester (CFSE) for 30 min in a 37°C water bath, washed, and suspended at 2 × 107/ml in Hanks balanced salt solution as described(Chang et al. 2007). Adherent monocytes were infected with fluorescent promastigotes at a 5:1 MOI, synchronized by centrifugation, and incubated at 37°C, 5% CO2 for 15 min. Five μM of dihydroethidium (DHE, Molecular Probes) were added, and after an additional 10 min at 37°C, promastigotes were removed by rinsing and phagocytes were detached in citric saline (0.135 M KCl, 0.015 M Na citrate). Cells were fixed in 2% paraformaldehyde in phosphate-buffered saline (EMS, Hatfield, PA) and analyzed on a Becton Dickinson FACSCalibur equipped with a 488 nm argon laser (BD Biosciences, San Diego, CA). Fluorescence was monitored at 480/30 nm (CFSE) and 580/42 nm (DHE), and Ten thousand events were examined. Data were analyzed using Cell Quest or Flow Jo software (BD Biosciences).
Five × 105 macrophages on 12 mm coverslips were infected with promastigotes stained with carboxy-fluorescein diacetate succinimidyl ester(Chang et al. 2007). Lysosomes were pre-labeled by incubation of macrophages in 0.8 μg of Mr 10,000 TRITC-conjugated dextran/ml (Molecular Probes) for 16 hrs, followed by a rinse and chase in RP-10 without dextran for an additional 30 min. At the time of infection the endosomal pathway was marked in positive control macrophages by addition of 10,000 Mr fluorescein-labelled dextran (Molecular Probes) at the time of “infection”. Infected or control cells were fixed in 2% paraformaldehyde (30 min), permeabilized in 0.2% Triton X-100 (15 min), incubated in 50 mM glycine (15 min), and blocked in 5% non-fat dry milk/PBS (30 min). Macrophages were incubated with 50:1 goat anti-LAMP-1 (Santa Cruz) followed by 200:1 Alexa fluor 647 (blue) donkey anti-goat Ig at room temperature for 1 h. After rinsing in PBS and mounting with Vectashield H-1000 (Vector Labs, Burlingame, CA), slides were examined on a Zeiss 510 laser scanning confocal microscope, and captured using the LSM 510 version 3.2 software. Microscopic studies were performed at the University of Iowa Central Microscopy Research Facility.
Previously we reported the generation and properties of an lpg2− mutant in the LdBob line of L. donovani, able to differentiate between the promastigote and amastigote stage in vitro when cultured under appropriate conditions (Goyard, et al., 2003), along with complemented derivatives where LPG2 expression was restored by transfection with an episomal LPG2 expression vector [lpg2−/+LPG2(e)]. BALB/c mice were infected i.v. with WT, lpg2−, or lpg2−/+LPG2(e) stationary phase L. donovani, and at appropriate times animals were euthanized and parasite loads were quantified by microscopy (Fig. 1). As observed previously, WT parasites multiplied most quickly in the livers (Fig.1B), peaking at 4–6 weeks and declining thereafter, whereas infection began slowly but progressed in the spleens of infected animals (Kaye and Farrell, 2002, Wilson and Weinstock, 1996). In contrast, lpg2− infections showed considerably fewer parasites at all time points in both organs (Fig. 1). In the liver, lpg2− parasites were found at 4 % of WT at the peak of the infection at week 4 (Fig. 1). Similarly, in the spleen lpg2− parasites were present at 3.5 % of WT levels at week 10 of infection. Importantly, restoration of LPG2 expression partially restored parasite replication in both livers and spleens [Fig. 1; lpg2−/+LPG2(e)], although the effect was variable in different experiments (compare Figs. 1A,B with a second experiment shown in Fig. 1C).
Given the attenuation of lpg2− L. donovani replication in mouse infections, we studied its survival in murine bone marrow macrophages. In these studies, WT stationary phase L. donovani enter macrophages successfully and after a slight decrease in numbers after 24 hr, replicated about 2-fold over the next ~72 hr (Fig. 2). In contrast, while lpg2− L. donovani were taken up by macrophages at a similar level as WT (Fig. 2; 2 hr time point), they were rapidly destroyed with their numbers declining to less than 5.2 % of WT by 96 hr. However and unlike the mouse infections experiments, one of the control lines in which LPG2 was restored on an extrachromosomal plasmid did not restore macrophage survival, i.e. the lpg2−/+LPG2(e) line was destroyed as efficiently as lpg2− (Fig. 2). This was similar to results obtained in another recent studying which episomal vector-based rescue of another PG deficient mutant (lacking the Golgi UDP-Gal nucleotide sugar transporters encoded by the LPG5A and LPG5B genes) was observed in mouse but not macrophage infections (Capul et al. 2007a). There we found that restoration of PG synthesis using a strong integrating LPG5A+LPG5B expression vector fully restored both mouse and macrophage virulence. Thus we hypothesized that the lower levels episomal LPG2 arising from the episomal expression vectors might account for the failure to rescue. This idea meshed well with our results with lpg2− L. donovani, as previous data showed that the lpg2−/+LPG2(e) line did not fully restore LPG expression back to WT levels either (Goyard et al. 2003).
We therefore generated several new transfectant restoration lines termed lpg2−/+LPG2(i), where LPG2 expression was restored using a strong rRNA-integrating expression vector (pIR1SAT; Methods). These lpg2−/+LPG2(i) lines synthesized WT levels of LPG and/or PPG, in stationary phase promastigotes, tested both before (Fig. 3A) and subsequent to mouse infections (Fig. 4B), and in axenic amastigotes (Fig. 3B). Curiously, restoration of LPG synthesis was not identical to WT in level or pattern in log phase promastigotes (Fig. 3A). However, we deemed this acceptable since virulence studies were carried out with infectious stationary phase promastigotes, which appeared normal in PG expression. Mannan levels were similar in WT and lpg2− L. donovani (Fig. 4A). In agreement with previous studies in L. major and L. mexicana lpg2−, GP63 levels were similar in WT, lpg2− and lpg2−/+LPG2(i) promastigotes (Fig. 4B).
In animal tests, the results with the lpg2−/+LPG2(i) line resembled the lpg2−/+LPG2(e) line in showing partial but significant rescue of the lpg2− phenotype (Fig. 1D). Now when tested in bone marrow macrophage infections, lpg2−/+LPG2(i) infected and replicated with kinetics indistinguishable from WT, whereas lpg2− and the lpg2−/+LPG2(e) lines entered but were rapidly destroyed as before (Fig. 2). These data established that the macrophage virulence defect of the lpg2− line was specifically attributable to the absence of LPG2. They also stress the importance of achieving full restoration of LPG2 expression, in a form that is maintained in the absence of drug pressure, in the restoration lines.
We assessed the susceptibility of logarithmic and infective stage stationary phase lpg2− mutants to lysis by complement (fresh human serum). Prior report of this assay indicates that lysis of Leishmania is mediated exclusively by complement (Spath et al. 2003a). Both growth phases of lpg2− L. donovani were susceptible to complement under these conditions, as seen by their uptake of propidium iodide in the presence by not absence of serum, whereas WT L. donovani were not lysed (Fig. 5, bottom panels). These results were similar to those obtained with the lpg2− L. major controls studied previously (Spath et al. 2003a) (Fig. 5, top panel). Interestingly, both growth phases of lpg2− but not WT L. mexicana promastigotes were highly susceptible to complement-mediated lysis (Fig. 5, middle panels). These data suggest that LPG2 may similarly contribute to L. mexicana promastigote virulence by mediating resistance to lysis by complement in human serum, presumably through PG synthesis.
We previously developed a flow cytometry-based method for measuring mononuclear phagocyte oxidative response to L. chagasi promastigote phagocytosis, following oxidation of dihydroethidium, which primarily detects the products of the NADPH oxidase(Chang et al. 2007). Application of this methodology here showed that lpg2− L. donovani induced somewhat higher oxidant production than WT or lpg2−/+LPG2(i) upon infection of human mononuclear phagocytes (1.7 fold; Fig. 6). These data suggest that PGs and/or other LPG2-dependent molecules may assist in limiting oxidant levels produced by phagocytes (Chan et al. 1989; Spath et al. 2003a).
Studies of lpg1− and lpg2− mutants L. major and L. donovani have shown that LPG is responsible for the transient delay in the ability of infective promastigotes to delay fusion of phagosomes with lysosomes (Desjardins and Descoteaux 1997; Dermine et al. 2000; Spath et al. 2003a; Spath et al. 2003b). To assess phagolysosomal fusion, we used confocal microscopy to assess the degree of overlapping parasite and lysosomal markers after phagocytosis by murine bone marrow macrophages (only the 2 hr time point is shown in Fig. 7). As previously reported, the ability to delay phagosome-lysosome fusion was greatly reduced in lpg2− promastigotes, but returned to WT levels following restoration of LPG2 expression (Fig. 7, Fig. 8).
In this study we explored the effects of LPG2 ablation on the virulence of L. donovani, a parasite causing human visceral leishmaniasis in the Old World. Using mutants on the LdBob strain background, our studies confirm earlier reports that lpg2− L. donovani are deficient in PG-containing glycoconjugates (Descoteaux et al. 1995), resulting in a defect in its ability to delay phagosome-lysosome fusion, and a failure to replicate in isolated macrophages (Lodge and Descoteaux 2005; Lodge et al. 2006). Importantly, our work extends these observations by providing the key biological finding that lpg2− promastigotes induce very low levels of infection in spleens and livers of mice. These results indicate that the virulence phenotype of lpg2− L. donovani is more similar to L. major and unlike L. mexicana, and may contribute to our understanding of the mechanism of persistence.
Our studies show that the virulence defects of the lpg2− L. donovani were specific, as restoration of LPG2 expression to the lpg2− mutant fully or partially restored virulence in the tests performed. Curiously, while restoration of LPG2 expression by either episomal or integrating expression vectors yielded similar partial restoration of virulence in animal infections (Fig. 1), only the integrated LPG2 rescue line (lpg2−/+LPG2(i) showed restoration of macrophage survival when tested in BMMs in vitro (Fig. 2). This was similar to results obtained in another recent studying which episomal vector-based rescue of another PG deficient mutant (lacking the Golgi UDP-Gal nucleotide sugar transporters encoded by the LPG5A and LPG5B genes) was observed in mouse but not macrophage infections (Capul et al. 2007a). As in that study, we found that integration of LPG2 into the ribosomal RNA locus resulted in a more complete restoration of glycoconjugate expression (Fig. 3). Thus for some reason the in vitro macrophage assay is more sensitive to the imperfect re-expression of LPG2 whereas the mouse assay, in which expression is allowed over a longer period of time, allows differences between knockout and add-back to emerge. It remains to be seen the exact reason for these differences.
Our findings emphasize some of the challenges posed by virulence tests in Leishmania. First and as indicated above, available Leishmania expression vectors lead to over-expression of LPG2 from multicopy episomes or at still higher levels following integration into the rRNA locus ((Kapler et al. 1990; Misslitz et al. 2000). A similar previous study of the HSP100 gene (Hubel et al. 1997) showed that Leishmania major Hsp100 is required chiefly in the mammalian stage of the parasite, and that over expression can be deleterious in various ways. Perhaps LPG2 when highly overexpressed in L. donovani is detrimental to virulence in vivo, although this phenomenon was not seen with studies of LPG2 in L. major (Spath, et al., 2003 b). Second, some variation in virulence restoration may be attributable to the well known tendency of these parasites to lose virulence during culture and/transfection (Cruz et al. 1993), rendering some transfectants less virulent for this reason rather than the intended genetic manipulation. Thus we and other investigators have adopted the operational criterion that successful restoration of virulence may nonetheless sometimes be incomplete or variable amongst different clonal transfectants, for the reasons cited above. Thus because of the tendency of Leishmania to lose virulence spontaneously, when examining independent clonal lines, negative results are meaningless, as long as one is able to achieve a successful phenotypic rescue with at least some of them.
Given the similar biochemical consequences of LPG2 ablation between L. major, L. donovani and L. mexicana, it is not immediately evident why the effects on virulence differ so greatly between L. major/L. donovani versus L. mexicana. These findings emphasize the well known concept that virulence determinants often play different roles in different stages of pathogens with complex life cycles. One possible explanation is that there could be as yet undetected differences in the LPG2-dependent glyco-repertoire between these species. Whereas abundant Leishmania glycoconjugates have been extensively studied, potentially less abundant but functionally active molecules could remain to be discovered. Indeed, a similar explanation has been advanced to account for the fact that PG-deficient L. major mutants generated through ablation of the UDP-Gal transporters LPG5A and LPG5B show normal virulence as amastigotes, although they remain attenuated as promastigotes like lpg1− and lpg2− mutants (Capul et al. 2007a).
A second explanation to account for the differences between L. major/L. donovani and L. mexicana invokes fundamental differences in the nature of the parasitophorous vacuole inhabited by the parasites. The parasitophorous vacuole harboring intracellular L. mexicana is relatively ‘spacious’, and typically houses multiple organisms, whereas the parasitophorous vacuoles occupied by L. major and L. donovani are considered to be ‘tight’, with the parasite and host membranes in close proximity, and typically housing only a single organism (Castro et al. 2006). It is possible that the species specific differences in parasite survival reflect different interactions of LPG and/or other LPG2-dependent metabolites with these two types of vacuoles. Indeed, there are several studies which suggest that the fusogenicity of the vacuole differs between the Leishmania species (Courret et al. 2002). As the effects of LPG on fusion of the L. mexicana vacuole has been less studied than the L. donovani vacuole, more specific hypotheses cannot yet be drawn.
Parasitophorous vacuoles harboring virulent L. donovani promastigotes show a transiently decreased ability to fuse with lysosomes for a period of up to 24 hr following infection. Studies of lpg1− L. donovani and L. major in the literature suggest this effect is mediated exclusively by LPG, according to studies with lpg1− mutants which specifically lack LPG but retain PPGs and other known glycoconjugates (Dermine et al. 2000; Spath et al. 2003a). As predicted, the LPG-deficient lpg2− L. donovani studied here was impaired in its ability to inhibit phagolysosomal fusion (Figs. 7, ,8).8). Amastigotes, in contrast to promastigotes, lack LPG and reside within highly fusogenic vacuoles (Alexander and Russell 1992; Lang et al. 1994; Naderer et al. 2004).
The functional consequences of LPG-dependent transient phagolysosomal fusion inhibition many be multifold. Initial studies assumed the delay served merely to provide sufficient time for promastigotes to differentiate into amastigotes (Desjardins and Descoteaux 1997). However, mechanistic studies of the lpg1− mutants showed a delay in fusion does not necessarily promote parasite survival (Spath et al. 2003a). Recent studies suggest that LPG alters phagolysosomal biogenesis profoundly, one consequence of which in L. donovani infections is the formation of a periphagosomal barrier of F-actin which hinders the ability of the NADPH oxidase to assemble on the phagocyte membrane (Dermine et al. 2005; Lodge and Descoteaux 2005; Lodge et al. 2006). Consistently, stationary phase WT induce significant levels of oxidants, but the levels are heightened during infection with LPG-deficient mutants (Fig. 6). Transient inhibition of phagolysosome fusion presumably allows the parasite to avoid exposure to physically separate itself from toxic lysosomal contents. Inhibition may also play a role in other processes, for example in modulating early interactions of macrophages with other arms of the immune system (Lodge and Descoteaux 2005). The summation of lower oxidant generation, residence in an unfused vacuole surrounded by an actin barrier that may prevent access of oxidants, and resistance to oxidant-mediated toxicity conferred by the LPG and PPGs of WT promastigotes (Spath et al. 2003a; Lodge et al. 2006) likely accounts for the lowered survival of lpg2-compared to WT L. donovani in host macrophages. The reciprocal relationship of LPG and the oxidative stress may have been even more remarkable if the infectious metacyclic forms of WT promastigotes had been used, since their LPGs contain approximately twice the number of the oxidant-scavenging repeating units compared to procyclic promastigote LPGs (Chan et al. 1989). There is unfortunately no method that could obtain metacyclic lpg2- parasites comparable to this control.
Previous studies implicated LPG as a major determinant promoting parasite resistance to lysis by complement. There are important differences between the lytic arm of the complement pathway in murine versus human hosts, with serum from the former exhibiting considerably weaker lytic activity than in other mammals (Ong and Mattes 1989). Thus while sensitive to lysis by human complement, LPG-deficient Leishmania are resistant to lysis by murine complement, and no alterations in the course of L. major infection are seen in comparisons between C5-deficient and normal mice (Spath et al. 2003a). During the current study, we showed for the first time that as expected lpg2− L. donovani was exquisitely sensitive to complement-mediated lysis (Fig. 5). Furthermore and despite its lack of in vivo attenuation in mouse models, the L. mexicana lpg2− mutant was similarly susceptibility to lysis by human complement. Because of the differences between complement of mice and humans, it is possible that the sensitivity of lpg2− mutants is of little consequence in mouse infections, whereas in human hosts the complement susceptibility of lpg2− mutant promastigotes of all three species of Leishmania could result in attenuation.
In summary, our studies emphasize several important differences in the effects of LPG2 on the virulence of three Leishmania species, revealing an unexpected complexity and divergence in outcomes. We have focused in this work on the consequences for virulence as measured by parasite replication and/or pathology. Future studies may focus on immune responses to WT vs. lpg2− L. donovani in comparison to WT vs. lpg2− L. major (Uzonna et al. 2004; Kebaier et al. 2006). It may be of interest to perform quantitative comparisons of L. major and L. donovani lpg2− mutant infections. If the numbers of persistent lpg2− L. donovani are indeed greater than lpg2− L. major footpads (typically ~103) as suggested in Figure 1 (>108 and >106 in liver and spleen, respectively) (typically ~103; Spath et al. 2003b; Spath et al. 2004), this could reflect differences in the infected cell types and/or the size or cell numbers between these organs and footpad infection sites. Alternatively, this could reflect fundamental differences between the biology of the two parasite species, manifesting in vivo but less apparent in macrophage infections in vitro (Fig. 2). These studies will have some practical value, as parasites showing ‘persistence without pathology’ such as lpg2− L. major and L. donovani hold promise as live attenuated vaccine lines (Spath et al. 2003b; Uzonna et al. 2004).
This work is supported in part by NIH grants AI045540 and AI067874 (MEW), a VA Merit Review grant (MEW), NIH grant AI31078 (SMB and SJT), and NIH training grants T32 AI07511 and T32 AI007260-22 (support of UG). We thank Tamara Barron for technical assistance and members of our laboratories for discussions. The authors are also grateful to Jian Shao at the University of Iowa Central Microscopy Facility for technical assistance.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.