Leishmania major Glycosylation Mutants Require Phosphoglycans (lpg2 −) but Not Lipophosphoglycan (lpg1 −) for Survival in Permissive Sand Fly Vectors

Background Sand fly species able to support the survival of the protozoan parasite Leishmania have been classified as permissive or specific, based upon their ability to support a wide or limited range of strains and/or species. Studies of a limited number of fly/parasite species combinations have implicated parasite surface molecules in this process and here we provide further evidence in support of this proposal. We investigated the role of lipophosphoglycan (LPG) and other phosphoglycans (PGs) in sand fly survival, using Leishmania major mutants deficient in LPG (lpg1 −), and the phosphoglycan (PG)-deficient mutant lpg2 −. The sand fly species used were the permissive species Phlebotomus perniciosus and P. argentipes, and the specific vector P. duboscqi, a species resistant to L. infantum development. Principal Findings The lpg2 − mutants did not survive well in any of the three sand fly species, suggesting that phosphoglycans and/or other LPG2-dependent molecules are required for parasite development. In vitro, all three L. major lines were equally resistant to proteolytic activity of bovine trypsin, suggesting that sand fly-specific hydrolytic proteases or other factors are the reason for the early lpg2 − parasite killing. The lpg1 − mutants developed late-stage infections in two permissive species, P. perniciosus and P. argentipes, where their infection rates and intensities of infections were comparable to the wild type (WT) parasites. In contrast, in P. duboscqi the lpg1 − mutants developed significantly worse than the WT parasites. Conclusions In combination with previous studies, the data establish clearly that LPG is not required for Leishmania survival in permissive species P. perniciosus and P. argentipes but plays an important role in the specific vector P. duboscqi. With regard to PGs other than LPG, the data prove the importance of LPG2-related molecules for survival of L. major in the three sand fly species tested.


Introduction
The distribution of diseases caused by the protozoan parasite Leishmania is limited by the distribution of the sand fly vectors and their capacity to support parasite development. Survival of Leishmania parasites during bloodmeal digestion and their attachment to the midgut epithelium have been identified as two critical steps determining the vector competence. Based upon experimental tests of their ability to support development of wide or limited range of Leishmania species, sand flies have been classified as permissive or specific vectors [1]. According to previous investigations, there is a close evolutionary fit between Phlebotomus papatasi and P. sergenti with Leishmania major and L. tropica respectively, as other Leishmania species survive poorly in these sand fly hosts. In contrast, other sand flies tested (P. argentipes, P. halepensis, P. arabicus and Lutzomyia longipalpis) were broadly permissive to the development of different Leishmania parasites.
This classification is based on experimental studies and does not imply the constraints of natural transmissions (vector capacity). However, it reflects the vector competence of permissive sand flies for transmission of various parasites (for review see [1,2]).
Leishmania surface molecules have been strongly implicated in parasites survival within sand fly vectors. Leishmania promastigotes synthesise an abundance of glycoconjugates composed of polymeric units based upon a conserved Gal-Man-P phosphoglycan (PG) repeating unit for review see [3]. These include the membrane-attached glycosylphosphatidyl-inositol (GPI) anchored liposphophoglycan (LPG) and proteophosphoglycan (PPG), as well as secreted forms of PPGs and secretory acid phosphatases (sAPs). PGs have been implicated in the early survival of L. donovani within the bloodfed midgut [4], presumably by conferring resistance to, or by modulating the activity of digestive enzymes. The role of these molecules in sand fly interactions has been studied by biochemical methods using purified LPG, PPG and other molecules, although the shared PG determinants across molecules makes the assignment of function to specific PG classes problematic [5,6] Thus, complementary to studies using purified PGs, researchers have generated and studied the behaviour of mutants lacking specific LPG or PG biosynthetic genes, which in turn affect the synthesis of subsets of PG-bearing molecules. The lpg1 2 mutant lacks the LPG1 gene (GenBank accession no. AF234766) which encodes a galactofuranosyltransferase required for synthesis of the LPG glycan core, rendering such mutants specifically deficient in LPG alone [7]. In contrast, the Leishmania major lpg2 2 mutant lacks the LPG2 gene (GenBank accession no. AF350492) encoding a Golgi GDP-sugar transporter. As a result, these parasites fail to synthesize LPG and all other PGs [8,9]. Recent studies suggest that the lpg2 2 mutant may lack additional, as yet unidentified and likely rare glycoconjugates [10]. Importantly, the LPG and virulence phenotypes of the L. major lpg1 2 and lpg2 2 lines in both mammalian and sand fly infections were restored to WT following re-expression of the cognate gene, confirming the genetic specificity of the virulence defects [4,5,7,[9][10][11][12][13].
The genetic and biochemical approaches above have established that LPG, the dominant surface glycoconjugate of Leishmania promastigotes, mediates attachment to the midgut epithelium in P. papatasi, preventing the loss L. major parasites during blood meal excretion, as lpg1 2 parasites survive the initial stages of fly infection but are subsequently lost due to a failure to bind to a sand fly midgut LPG receptor [4,14,15]. Notably P. papatasi is considered to be a highly specific vector, in that other species of Leishmania are unable to establish infection in this sand fly [16][17][18][19]. In contrast, in P. arabicus and Lu. longipalpis, two species shown to be permissive vectors by virtue of their ability support development of various Leishmania species, LPG-deficient lpg1 2 L. major develop and produce mature infections [20]. From these data Myskova et. al hypothesised that LPG is required in specific vectors, while in permissive vectors Leishmania bind via an LPG independent mechanism.
In this work we further test this hypothesis by infections of three additional sand fly species with WT and mutant L. major. In addition, we attempted to assess the importance of LPG and other LPG2-dependent molecules in protection against proteolytic attack by exposing the mutant parasite lines to the action of bovine trypsin in vitro. Importantly, the three sand fly species used in the study are important vectors known to transmit Leishmania. Phlebotomus duboscqi is a vector of cutaneous leishmaniasis caused by L. major in sub-Saharan Africa [21,22]. It is a sister species of P. papatasi and belongs to the same subgenus. Unlike P. papatasi, some populations of P. duboscqi have been shown experimentally to support development of L. tropica [16]. Midgut glycosylation and the degree of permissivity of this species are unclear. We addressed the question of permissivity of P. duboscqi sand flies in this study by infecting them with L. infantum. The other two species used are permissive vectors transmitting parasites of L. donovani complex. Myskova et al. [20] demonstrated that both, P. argentipes and P. perniciosus posses midgut glycoproteins with HPA (Helix pomatia agglutinin, lectin with specificity to N-acetyl-D-galactosamine)binding epitopes. Phlebotomus argentipes is a vector of visceral anthroponotic leishmaniasis caused by Leishmania donovani in the Indian subcontinent [23]. In experimental conditions it supports development of L. donovani, L. amazonensis, L. major and L. tropica [4,19,24]. Phlebotomus perniciosus is a vector of Leishmania infantum in the western Mediterranean and in experimental conditions it supports the development of L. tropica (V.S. and P.V., unpublished results).

Sand fly colonies
Laboratory colonies of three sand fly species were used: Phlebotomus perniciosus (originally from Spain), P. argentipes (originally from India) and P. duboscqi (originally from Senegal). Colonies were maintained in conditions described previously [26]. Adults were maintained at 26uC and fed on 50% sucrose ad libitum.

Leishmania development in sand flies
Female sand flies (5-10 days old) were fed through a chick skin membrane with 4-5 day old promastigotes at cell density of 5610 6 (P. duboscqi infections with L. major) or 1610 6 promastigotes/ml (all other infections, including P. duboscqi with L. infantum) in heat inactivated rabbit blood (Bioveta, Ivanovice). Blood-engorged females were maintained at 26uC with access to cotton wool soaked in 50% solution of sugar in distilled water and sacrificed for microscopical examination and counting of parasites in the midgut 2 and 5 or 9 days post infection. Intensity of infection was graded as light (,100 parasites/gut), moderate (100-1000 parasites/gut) or heavy (.1000 parasites/gut) as described previously [27]. Phlebotomus perniciosus and P. duboscqi defecate between 75 and

Author Summary
Phlebotomine sand flies are small blood-feeding insects, medically important as vectors of protozoan parasites of the genus Leishmania. Sand flies species can be divided roughly into two groups, termed specific or permissive, depending on their ability to support development of one or a few strains vs. a broad spectrum of these parasites. In this study, we explored the ability of two Leishmania major glycocalyx mutants to survive within these different types of vectors. The lpg1 2 mutant, which specifically lacks lipophosphoglycan (LPG), was able to survive normally in two permissive species, Phlebotomus argentipes and P. perniciosus, but was only able to survive within the specific species P. duboscqi for a limited time prior to dissolution of the peritrophic matrix. Consistent with its classification as a specific sand fly vector, P. duboscqi was not able to support development of L. infantum. The lpg2 2 L. major mutant, which is a broader mutant and lacks all phosphoglycans including LPG and proteophosphoglycans, was unable to survive in all the three vector species tested. This study extends the knowledge on the role of Leishmania major surface glycoconjugates to development in three important vector species and gives supporting evidence for the existence of an LPG-independent mechanism for survival in sand flies, as well as the importance of LPG2-dependent glycoconjugates in parasite survival.
Experiments were repeated twice. The x 2 test was used for comparison of infection rates (number of infected versus uninfected females) and intensities of infection (heavy, moderate light, zero) between the WT and the mutant lines using S-PLUS 2000 programme.

Parasite susceptibility to bovine trypsin
Promastigotes of a 4-day culture were washed in medium 199 (M 199), adjusted to the concentration of 3610 6 cells/ml of M 199 and exposed to bovine trypsin (13.500 BAEE units/ml) (Sigma) alone or bovine trypsin plus 6% human haemoglobin (Sigma; one BAEE unit will produce a D A 253 of 0.001 per min at pH 7.6 at 25 uC using benzoyl-L-arginine ethyl ester (BAEE) as substrate). In control groups, parasites were cultivated in M 199 alone. After 24 hours at 23uC, parasite numbers were determined by haemocytometer counting. Assay was performed in triplicate and the experiment was repeated twice. Data were evaluated statistically by means of ANOVA test.

Development of L. major in P. duboscqi
Phlebotomus duboscqi sand flies were infected with WT, lpg1 2 and lpg2 2 mutants of L. major in order to study the role of LPG and other PGs. On day 2 post-infection, no differences were observed between development of WT and the lpg1 2 line, with both showing very high rates of infection (97% and 93% respectively), with about 75% of heavy infections (Fig. 1). The lpg2 2 mutant survived less well however; their infection rate was significantly lower (74%; P,0.01), with only 24% of heavy infections.
In contrast to day 2, on day 9 post infection the lpg1 2 mutant survived much more poorly than WT, with only 25% of the flies retaining parasites, as opposed to 92% for WT (P,0.01). Moreover, the lpg1 2 line produced very few mature infections colonizing the stomodeal valve in only 10% of females. Notably the lpg2 2 line did not survive, as no lpg2 2 parasites were found in the midguts by day 9 (Fig. 1). Previous studies have established that phenotypes arising from the lpg1 2 and lpg2 2 mutants in diverse settings are specific, as they are reversed by complementation with the cognate gene, and thus do not arise as a result of nonspecific culture passage or other sources [4,7,9,13]. Thus these mutant data argue that, as seen previously in the specific sand fly P. papatasi, LPG is required for late but not early survival [4], while LPG2 is important for early survival and essential for late stage survival.

Development of L. major lines in P. argentipes
Similar to the results obtained in P. duboscqi, on day 2 no statistically significant differences were found between the WT and the lpg1 2 mutant, while the lpg2 2 mutant was severely impaired (Fig. 2). Very high infection rates (96%) were present in both the WT and lpg1 2 lines, with heavy infections in 70% and 50%, respectively. In contrast, lpg2 2 mutants were more severely affected, with infections seen in 62% of flies but with very low parasite loads (less than 100 Leishmania, except for 1 fly). The  Leishmania Mutants in Sand Flies www.plosntds.org differences in infection intensity and rate between the lpg2 2 and the WT were statistically highly significant (P,0.01).
Day-5 dissections revealed a continuation of the trends described above during late stage infections. Females infected with WT and lpg1 2 mutants retained high infection rate and high parasite loads (no statistically significant differences between the lines). Also the localizations of parasites were remarkably similar; WT and lpg1 2 parasites reached the stomodeal valve in 46% and 48%, respectively. In contrast, lpg2 2 line showed a remarkable decrease in infection rate, with only 18% of positive females (highly significant difference from the WT parasites, P,0.01) and no colonization of the stomodeal valve.

Development of L. major lines in P. perniciosus
On day 2, no significant differences were observed between the three lines, all of them survived well inside the peritrophic sac producing heavy infections in about 25% of females.
On day 9, lpg2 2 mutants were eliminated while lpg1 2 mutants developed similarly to the WT parasites (Fig. 3). WT and lpg1 2 lines developed mature infections colonizing the stomodeal valve with high parasite burdens in majority of females. In contrast, none of the lpg2 2 parasites were able to persist until day 9, suggesting that they were lost during defecation.

P. duboscqi is refractory to L. infantum
In order to test the degree of permissivity of P. duboscqi, females of this species were infected with L. infantum, a parasite that is not transmitted by P. duboscqi in nature. Promastigotes were able to survive inside the peritrophic sac during digestion of the bloodmeal but they were not able to persist beyond defecation of the blood remnants. On day 2 post infective bloodmeal, 91% of the flies (11 of 12) were Leishmania positive while on day 8, no parasites were found in any female tested (n = 15) (data not shown). These results demonstrate that P. duboscqi is refractory to L. infantum.

P. duboscqi midgut glycosylation
As detected by western blotting with lectins, P. duboscqi midgut lysate displays molecules that bind Con A and PSA, lectins detecting terminal mannose residues of glycans (Fig. 4). Controls with inhibitory sugar (250mM methyl-mannopyranoside) confirmed the specificity of lectin reactions (data not shown). In contrast, HPA, RCA and SBA reactions were negative indicating absence of b-galactose or N-acetyl-D-galactosamine residues in the midgut glycoproteins (Fig. 4). The lectin binding profile is similar to that previously observed in specific sand fly vectors P. papatasi and P. sergenti. In contrast, midgut lysates of all permissive sand fly species tested to date contain N-acetyl-D-galactosamine displaying glycoconjugates as detected by lectin affinity blotting [20].

Effect of bovine trypsin on Leishmania
The in-vitro growth of L. major promastigotes of WT, lpg1 2 and lpg2 2 lines in M 199 medium was not affected either by bovine trypsin (13.500 BAEE units/ml) or by bovine trypsin plus 6% human hemoglobin.
No significant differences in numbers of viable parasites were observed between the groups in any of the three lines tested (P = 0.84).

Discussion
The classic studies of Sacks and co-workers established a paradigm for the role of LPG in the survival of L. major and L. donovani in sand flies, mediated through binding of LPG to the sand  fly midgut [4,19]. In the case of L. major this interaction is now known to be mediated by the P. papatasi lectin PpGal ( [15]). However in 2007 we reported the occurrence of LPG-independent sand fly survival, importantly only in sand fly species now termed 'permissive' as defined by their ability in experimental tests to support the development of a wide range of Leishmania species. In contrast, previous studies of LPG-dependency were now associated with 'selective' sand fly species (again defined by experimental tests as supporting the development of a narrow range of Leishmania species and/or isolates) ( [20]). Given the implications of this new paradigm, it was important to garner additional data testing its validity by examining additional permissive and selective sand fly species, using the same well characterized LPG mutants studied previously. Additionally we have culled the literature and summarized the available data pertinent to this model (Table 1).
We found that L. major mutants specifically lacking LPG remain able to develop in the permissive vectors P. perniciosus and P. argentipes at levels resembling those of wild type parasites, with full midgut development and colonization of the stomodeal valve. These data suggest that in P. perniciosus and P. argentipes the LPG is required neither for parasite protection against digestive enzymes nor for midgut binding. LPG-independent development was previously reported for L. major in the permissive sand fly vectors Lu. longipalpis and P. arabicus [20], and L. mexicana development in Lu. longipalpis [29]. These data confirm and extend the results obtained in sand fly infections with all LPG-deficient Leishmania carried out by various laboratories to date ( Table 1). Collectively these data provide strong support for the role of LPG in specific but not permissive sand fly vectors.
Within this data set we could only identify one potential exception, involving a study of the permissive vector P. argentipes, where the L. donovani LPG-deficient mutant line known as the 'R2D2', also defective in LPG1 expression [30], did not survive when examined on day-5 post-infection [19]. Notably R2D2 was obtained following heavy mutagenesis and selection for LPGdeficiency, unlike the lpg1 2 L. major which was generated following precise gene targeting procedures [31]. It is well established in the genetics literature that mutagenesis frequently results in off-target deleterious effects. Our previous work established that the phenotypic alterations in the lpg1 2 and lpg2 2 arose specifically from alterations in these genes, as restoration of LPG1 and LPG2 function returned the phenotype to WT [7,9]. In contrast, R2D2 failed this test, as restoration of LPG1 expression to R2D2 only weakly restored both LPG and survival in P. argentipes [4]. In contrast to the permissive vectors, the development of L. major lpg1 2 mutants was severely impaired in the specific vector P. duboscqi. Although the early infections were similar to those of the WT parasites, there was a substantial decrease in the lpg1 2 infections rate after defecation of the bloodmeal. In very few females the lpg1 2 mutants produced mature late stage infections. Our results extend those reported in the study by Boulanger et al. [32] performed with a small number of sand flies. Similar results with L. major lpg1 2 mutants in P. duboscqi were recently obtained by Secundino et al. (submitted) ( Table 1). Our additional experiment confirmed that P. duboscqi is not permissive to L. infantum development and can therefore be classified as a specific vector. Moreover, lectin affinity blotting revealed that unlike P. perniciosus and P. argentipes, there are no N-acetyl-D-galactosamine-displaying epitopes in P. duboscqi midgut (Fig 4). Such glycoconjugates have been suggested as potential Leishmania ligands in the midgut of permissive vector species [20]. In conclusion, this study gives supporting evidence to the present distinction of sand flies into categories based on their susceptibility to various Leishmania species [1]. Together with the results of Myskova et al. [20] and Rogers et al. [29], our studies of L. major lpg1 2 development suggest the presence of a an LPG-independent parasite-binding mechanism within the midgut of permissive sand flies.
Unlike LPG-deficient lpg1 2 mutants, PG-deficient lpg2 2 mutants additionally were impaired in early development in sand fly and unable to survive at all stages in all sand fly species tested. In P. argentipes and P. duboscqi parasites of this line are severely impaired as early as day 2 post infection. For P. duboscqi, these data stand in line with those of Boulanger et al. [32]. A similar finding was described for PG-deficient mutants of L. donovani in P. argentipes by Sacks et al. [4]. Moreover, Secundino et al. (submitted) have recently made similar observations in P. duboscqi originating from Mali and in Lu. longipalpis (see Table 1).
It has been hypothesized that parasite death in the predefecated sand fly midgut is attributable to digestive enzymes and that the phosphoglycans other than LPG confer resistance to the proteolytic attack [33], specifically phosphoglycans dependent on the activity of LPG2. While LPG2 has been suggested to affect synthesis of other glycoconjugates beyond PGs, this possibility was excluded through studies of a second PG-deficient mutant, defective due to a lack of the UDP-Gal transporters LPG5A and LPG5B by gene targeting, showing that it is also unable to survive the late stages of P. duboscqi infection (Secundino et al., submitted). Thus, it is likely that the defects in lpg2 2 infection of midguts described here and previously arise primarily through loss of PGs other than LPG, potentially the PPGs common to all species, or sAPs which occur in L. major albeit to lesser extents than in other Leishmania species [34].
In this work we also attempted to assess the importance of LPG2-dependent molecules by exposing parasites to the action of a proteolytic enzyme in vitro. As trypsin-like proteases were described as the most abundant digestive enzymes in both P. papatasi and Lu. longipalpis midgut after bloodfeeding [35,36], bovine panceratic trypsin was chosen for these experiments. The bovine enzyme used shares all the conserved amino acid residues that influence the substrate specificity with sand fly midgut trypsin-like molecules. The lpg2 2 promastigotes lacking surface PGs did not prove to be more vulnerable to trypsin activity than the WT and lpg1 2 parasites whose resistance to trypsin has been previously reported [13]. These results could argue against a role of PGs in conferring resistance of promastigotes to the trypsin-like digestive enzymes in the sand fly gut. However, in light of the studies of Secundino et al (submitted), we think it more likely that bovine trypsin is not a good model for the activity and/or properties of all the proteolytic contents of the sand fly midgut. Most importantly, Secundino et al showed that inhibition of tryptic and other proteolytic activity in the midgut does in part rescue the survival of lpg2 2 parasites, although they do not exclude the possibility that also other factors contribute to the inability of the lpg2 2 to survive within the bloodmeal. Sand fly immunity has been shown to play a major role in the control of bacterial and parasitic infections [32] and potentially that molecules such as antimicrobial peptides secreted to the midgut lumen could contribute to the destruction of the lpg2 2 mutant parasite.
In summary, this study demonstrates that an LPG-independent mechanism of attachment of Leishmania is a feature common to permissive sand fly species. It also proves the importance of LPG2dependent molecules in the survival of L. major in various sand fly vectors. Moreover, it brings an evidence that P. duboscqi is not able to support development of L. infantum and therefore can be classified as a specific vector.