Lipophosphoglycan polymorphisms do not affect Leishmania amazonensis development in the permissive vectors Lutzomyia migonei and Lutzomyia longipalpis

Background Lipophosphoglycan (LPG) is a dominant surface molecule of Leishmania promastigotes. Its species-specific polymorphisms are found mainly in the sugars that branch off the conserved Gal(β1,4)Man(α1)-PO4 backbone of repeat units. Leishmania amazonensis is one of the most important species causing human cutaneous leishmaniasis in the New World. Here, we describe LPG intraspecific polymorphisms in two Le. amazonensis reference strains and their role during the development in three sand fly species. Results Strains isolated from Lutzomyia flaviscutellata (PH8) and from a human patient (Josefa) displayed structural polymorphism in the LPG repeat units, possessing side chains with 1 and 2 β-glucose or 1 to 3 β-galactose, respectively. Both strains successfully infected permissive vectors Lutzomyia longipalpis and Lutzomyia migonei and could colonize their stomodeal valve and differentiate into metacyclic forms. Despite bearing terminal galactose residues on LPG, Josefa could not sustain infection in the restrictive vector Phlebotomus papatasi. Conclusions LPG polymorphisms did not affect the ability of Le. amazonensis to develop late-stage infections in permissive vectors. However, the non-establishment of infection in Ph. papatasi by Josefa strain suggested other LPG-independent factors in this restrictive vector.

Several biochemical analyses of LPG revealed intra and inter-species polymorphisms in the sequence and composition of sugars attached to repeat units. Early studies determined that these interspecies variations were important for attachment in the vector and for virulence in the vertebrate hosts [3,4]. In the Old World species, intraspecific LPG variations were reported for Leishmania major [5,6], Leishmania tropica [7,8] and Leishmania donovani [9] and their ability to infect sand fly species (Phlebotomus papatasi, Phlebotomus sergenti/ Phlebotomus arabicus and Phlebotomus argentipes) has been reported. In the New World, purified LPG from 14 strains of Leishmania infantum displayed important LPG polymorphisms (type I: no side-chains branchingoff the repeat units; type II: one β-glucose residue branching-off the repeat units and; type III: 1-3 βglucose as side-chains). However, those LPG variations did not affect the interaction with Lutzomyia longipalpis [10,11]. Our preliminary data suggested qualitative LPG polymorphisms between two Le. amazonensis strains (PH8 and Josefa) based on antibody recognition, but they did not result in different activation in macrophages and/or sand fly infection [12].
Sand fly species can be divided into restrictive and permissive vectors depending on their ability to support development of various Leishmania species [13]. For example, Ph. papatasi, the restrictive vector, can sustain infection only with Le. major [4] and Leishmania turanica [14], two parasites with LPG terminated by βgalactosyl residues [15][16][17]. On the other hand, permissive vectors, like Lu. longipalpis and Lutzomyia migonei, can sustain infection of several Leishmania species; the former supports the development of Le. infantum, Le. amazonensis and Le. major (reviewed in [18]), whereas the latter supports the infection of Le. amazonensis, Leishmania braziliensis and Le. infantum [12,19,20].
The attachment of the parasite to a midgut receptor is a crucial event to avoid passage with the digested blood meal. The mechanisms of midgut attachment in restrictive vectors is LPG-dependent [21] and involves midgut galectin [22], while in permissive vectors it could be LPG-independent [23,24] or may involve midgut mucin-like proteins [25].
The main vector of Le. amazonensis is Lu. flaviscutellata. However, established laboratory colonies of this species are not available. For this reason, Lu. migonei has been used as a successful model for interaction with Le. amazonensis [19]. Preliminary studies using Le. amazonensis strains (PH8 and Josefa) showed that they could survive for several days inside Lu. migonei [12]. However, their ability to reach the stomodeal valve and accomplish metacyclogenesis was not evaluated.
As a part of a wider project on the glycobiology of New World species of Leishmania, we described a detailed biochemical characterization of Le. amazonensis LPGs and found intraspecific differences. The interaction of Le. amazonensis strains with permissive and restrictive vectors was performed. Additionally, the ability of the Josefa strain, bearing terminal galactose residues in its LPG, to survive in the restrictive vector Ph. papatasi was evaluated.

Experimental infections of sand flies
Sand fly females were infected through the chick-skin membrane on a mixture of promastigotes and heatinactivated rabbit blood; the final concentration of parasites was 1 × 10 6 promastigotes/ml. The experiments were conducted with six sand fly-Leishmania a Purified LPGs were subjected to mild acid hydrolysis to depolymerize the repeat units and cap structures. Water-soluble fractions were partitioned using 1-butanol, treated with alkaline phosphatase (15 mM Tris buffer, pH 9.0, 1 unit, 16 h, 37°C) and desalted by passage through a two-layered column of AG50W-X12 over AG1-X8. The desalted repeat units were subject to FACE analysis and enzymatic treatments with β-glucosidase, β-galactosidase to CE analysis. Additionally, the repeat units were subjected to strong acid hydrolysis (2 M trifluoroacetic acid, 3 h, 100°C) and FACE assays to access monosaccharide composition. b Six combinations of Le. amazonensis strains (PH8 and Josefa) with Lu. longipalpis, Lu. migonei and Ph. papatasi were performed. Those were evaluated on days 1 and 5-6 post-infection (PI) for intensity, localization and morphometry combinations as depicted in Fig. 1b: Lu. migonei-PH8, Lu. migonei-Josefa, Lu. longipalpis-PH8 and Lu. longipalpis-Josefa, Ph. papatasi-PH8 and Ph. papatasi-Josefa. Terminal β-galactosyl residues were determinant for Le. major attachment to PpGalec in Ph. papatasi [22]. Since those sugars were also present in Josefa LPG, the ability of this strain to sustain infection in this vector was also checked compared to glucose-containing LPG from the PH8 strain.
Blood-engorged females were separated, maintained at 26°C and dissected on days 1 and 5-6 post-infection (PI). Individual guts were placed into a drop of saline and examined microscopically for the localization and intensity of Leishmania infections. Parasite loads were graded according to [23] as light (< 100 parasites per gut), moderate (100 to 1000 parasites per gut) and heavy (> 1000 parasites per gut). The experiments were repeated two times using the same vector/strain combinations. Data were evaluated statistically by means of the Fisher's exact or Chi-square (χ 2 ) tests using SPSS statistics version 23 software.

Morphometry
Parasite smears from midguts of the three sand fly species infected with Le. amazonensis strains were obtained on days 1 and 5-6 PI. The midguts were carefully dissected using fine needles and each part was separated and respective parasites counted. Slides were fixed with methanol, stained with Giemsa, examined under a light microscope with an oil-immersion objective and photographed with an Olympus D70 camera. For morphometry, body length, body width and flagellar length of 240 randomly selected promastigotes from six midgut smears were measured for each sand fly species and time interval using Image-J software. The morphological forms were distinguished based on the criteria of Sádlová et al. [34] and Rogers et al. [35]: (i) short nectomonads: body length < 14 μm and flagellar length < 2 times body length; (ii) long nectomonads: body length ≥ 14 μm; (iii) metacyclic promastigotes: body length < 14 μm and flagellar length ≥ 2 times body length. Data were evaluated statistically by analysis of variance using SPSS statistics version 23 software.

Characterization of LPG repeats units
Purified LPGs from Le. amazonensis strains were differentially recognized by the mAbs CA7AE and LT22 (Fig. 2a-d). As shown in Fig. 2a-b, the LPG from the PH8 strain and respective controls (LD4 and Mongi) were recognized by CA7AE and LT22. However, a different recognition profile was observed for the Josefa strain since its LPG was recognized by LT22 (Fig. 2d) but not by CA7AE (Fig. 2c), indicating the presence of sidechains branching-off the repeat units.
Confirming the previous immunoblotting profiles, both Le. amazonensis strains displayed a distinct oligosaccharide profile of their neutral repeat units (Fig. 3a). The repeat units of the PH8 strain exhibited higher disaccharide content, represented by Gal-Man (G 2 position) that is common to all LPGs, which explains their reactivity against CA7AE (Fig.  2a). The FACE analysis also revealed up to 1 and 2 side-chains in their structures (Fig. 3a). On the other hand, the repeat units of the Josefa strain showed up to 1 to 3 side-chains, and no Gal-Man disaccharide was detected (Fig. 3a). This profile elucidates the observed lack of recognition by CA7AE (Fig. 2c) and the positive reaction against LT22 (Fig. 2d), suggesting that most, if not all, repeating units contained side-chains.
The monosaccharide profile of the PH8 strain revealed galactose and mannose, common to all LPGs, and high content of glucose (Fig. 3b). These data confirmed the presence of glucose as side-chains, as previously indicated by immunoblotting. The Josefa strain, however, exhibited only galactose and mannose as monosaccharides (Fig. 3b). Confirming the FACE analysis, the dephosphorylated repeat units from the PH8 strain consisted of a di-(67%), a tri-(30%) and a tetrasaccharide (3%) (Fig. 4a). The trisaccharide and tetrasaccharide were susceptible to treatment with β-glucosidase (Fig. 4b). These data are consistent with the structure of the trisaccharide as Glc(β)Gal(β1,4)Man and the tetrasaccharide as Glc 2 (β)Gal(β1,4)Man. The LPG of the Josefa strain showed a small amount of di-(9%), an abundance of tri-(29%) and tetra-(57%), and again a small amount of pentasaccharide (5%) (Fig. 4c). The minimal disaccharide content supports the non-reactivity by CA7AE (Fig. 2c). Interestingly, this LPG was not susceptible to βglucosidase treatment (Fig. 4d), indicating that another sugar is terminating the repeat units. After βgalactosidase treatment, we observed the disappearance of side-chains confirming the presence of β-galactoses (Fig. 4e). These results confirmed that both strains possess intraspecific polymorphisms (Fig. 5).

Sand fly infections and morphometry
The development of Le. amazonensis strains was studied in Lu. migonei, Lu. longipalpis and Ph. papatasi on days 1 and 5-6 PI. During the early phase of infection (on day 1 PI), the infection rates were fully comparable in all parasite-vector combinations; more than 82% of females were infected and moderate or heavy intensities of infections were observed in majority of females. Parasites were in the endoperitrophic space within the blood meal surrounded by the PM. On days 5-6 PI, both Le. amazonensis strains could establish heavy late stage infections with the colonization of the stomodeal valve in permissive vectors Lu. longipalpis and Lu. migonei. In contrast, in Ph. papatasi the infections were lost during defecation of bloodmeal remnants (Figs. 6 and 7).
Since no infection was developed in Ph. papatasi after defecation, morphological analysis was performed on Le. amazonensis parasites derived from Lu. migonei and Lu. longipalpis infections on days 1 and 5-6 PI. Both strains accomplished metacyclogenesis in both vectors. Interestingly, Lu. longipalpis infections with PH8 strains resulted in a higher percentage of metacyclic forms compared to Lu. migonei by days 5-6 PI (P < 0.0001) (Fig. 8).

Discussion
Leishmania amazonensis, a member of the Leishmania mexicana complex, is one of the most studied Leishmania species. It is an excellent model for immunology, molecular biology and chemotherapy. This species is often associated with treatment failure, being naturally resistant to the first line chemotherapy drugs [36]. However, studies involving Le. amazonensis glycoconjugates and their interaction with sand fly vectors are still scarce.
The LPGs are implicated in a variety of functions in the sand flies including attachment to a microvilli receptor in Old World Ph. papatasi (restrictive) [37]. Early in vitro studies reported the role of interspecies LPG polymorphisms in the specificity during sand fly-Leishmania interactions [4]. In this study, Ph. papatasi midguts were recognized only by Le. major phosphoglycan (PG), whereas those from Ph. argentipes (permissive) were recognized by PGs from several species. Later, several in vitro studies demonstrated that those models are consistent while using natural vector-Leishmania pairs not only in Old World but also in New World vectors. For example, the interaction with PGs from New World Le. infantum and Le. braziliensis adversely affected parasite attachment in permissive Lu. longipalpis and restrictive Lutzomyia intermedia/Lu. whitmani [10,[38][39][40]. However, in vitro system limitations appear while using unnatural sand fly-Leishmania combinations including the attachment of Le. braziliensis (New World) promastigotes to Ph. papatasi (Old World) midguts [41]. To circumvent such limitations, in this study we have performed experiments using in vivo models with previously tested amazonensis [12,19,20].
In the New World species of Leishmania, LPG variations were only studied in Le. infantum. In this species, polymorphisms in the glucose levels were not important for in vivo infectivity to Lu. longipalpis [11]. However, this sugar was important for in vitro interaction of Le. infantum (PP75 strain) with the midguts of Lu. longipalpis [10].
Consistent with our previous observations, a more detailed biochemical analysis of LPG using FACE and CE from Le. amazonensis strains revealed important intraspecific polymorphisms. The LPG of the PH8 strain    β-glucose residues as side-chains whereas the Josefa strain had β-galactose residues (Fig. 5). The βglucose residues are commonly found in the New World species of Leishmania including Le. mexicana [42], Le. infantum [10] and Le. braziliensis [31]. Those β-glucose residues in Le. infantum procyclic were downregulated after metacyclogenesis resulting in loss of in vitro interaction with the midgut of Lu. longipalpis [10]. Surprisingly, the occurrence of β-galactose residues as sidechains in the Josefa strain of Le. amazonensis was detected for the first time in a New World Leishmania species. This sugar is often observed in the Old World Leishmania species including Le. major [15], Le. tropica /Leishmania aethiopica [5,7] and Le. turanica [17]. Unlike Le. major and Le. turanica, Le. amazonensis Josefa strain with β-galatosylated LPG could not survive inside Ph. papatasi. This phenomenon could be explained by other LPG-independent mechanisms, probably related to the distance of the sand fly-Leishmania pair used (Old vs New World) or other molecules involved. Together with LPG, glycoprotein 63 (GP63) was also evaluated in Le. mexicana/Le. amazonensis -Lu. longipalpis attachment in vivo [43] or in vitro [44]. A recent study demonstrated that both glycoconjugates could be determinant for inhibiting in vitro parasite attachment in Lu. longipalpis and Lu. intermedia [40].
Several studies have already described the differentiation of Le. amazonensis, Le. braziliensis and Le. infantum in New World sand fly species including Lu. longipalpis and Lu. migonei [12,19,20,45]. Consistent with those observations, both Le. amazonensis strains could survive defecation of permissive vectors Lu. longipalpis and Lu. migonei, establish late-stage infections, and colonize anterior midgut reaching the stomodeal valve. This is a strong indication that the parasite could attach and migrate towards the mouth parts for subsequent transmission. More importantly, both strains accomplished metacyclogenesis. This morphological transformation was more pronounced in Lu. longipalpis infected with the PH8 strain, probably due to the higher permissiveness of this species. This sand fly can sustain a wide variety of pathogens, including viruses, non-Leishmania protozoans and even helminths (reviewed in [18]).

Conclusions
In combination with previous studies, the LPG polymorphism in Le. amazonensis did not affect infection of the three sand fly species tested. To our knowledge, the biochemical data obtained from the Josefa strain represents the first description of a galactosylated LPG in a New World Leishmania species. However, these residues were not sufficient for survival in Ph. papatasi.