Characterization of a midgut mucin-like glycoconjugate of Lutzomyia longipalpis with a potential role in Leishmania attachment

Data reported by Myšková and colleagues [9] suggested that a 45 kDa O-linked glycoprotein on the Lu. longipalpis midgut epithelium may function as a parasite ligand. Here we used a combination of biochemical, molecular and parasitological approaches to characterize the biological properties of this glycoconjugate. We provide evidence that the molecule with an apparent molecular weight 45 kDa on SDS-PAGE corresponds to a putative 19 kDa protein with unknown function detected in a midgut cDNA library of Lu. longipalpis [18]. The difference between the size of the native protein and predicted protein from cDNA library is assumed to be caused by a high degree of glycosylation. This is in agreement with the numerous predicted glycosylated serine and threonine residues found in the C- terminal part of the protein sequence, and also corresponds with the smeared appearance of the 45 kDa band on the gels or blots.

O-glycosylation prediction on the entire set of putative proteins identified in the midgut transcriptome of Lutzomyia longipalpis [18] did indeed find LuloG as the most heavily O-glycosylated protein. Beside LuloG, one more putative protein with extensive glycosylation was identified: ABV60335. It is a putatively secreted protein of 16.6 kDa with no other apparent conserved domains. No protein with a similar degree of O-glycosylation (a cluster of more than 40 putatively glycosylated Ser/Thr residues) was identified in the P. papatasi midgut transcriptome data [19].

An essential prerequisite for this GalNAc displaying molecule of permissive sand flies to function as a midgut receptor for Leishmania is its abundance and localization on midgut epithelial cells. Figure 7 shows that antibodies against the recombinant LuloG protein recognized round/spherical cells of dissected Lu. longipalpis midguts, typical of the luminal surface of epithelial midgut cells [17]. On the other hand, preimunne serum recognized mostly musculature, consisting of circular and longitudinal fibres, which is located on the external side of sand fly midgut [20].

Although we could not prove involvement of LuloG in Leishmania attachment by blocking of the native protein on the epithelial surface with anti-rLuloG during sand fly infections, we demonstrated a strong binding ability of rLuloG to whole Leishmania body by in vitro indirect imunofluorescence experiments. This suggests that the receptor for the sand fly glycoconjugate is localized on the entire surface of promastigotes, including the body and flagellum and is accessible to the sand fly midgut ligand. Our data correspond with the observations of Myšková and colleagues [9], where midgut lysates of the permissive sand fly P. halepensis bound to Leishmania promastigote surfaces, and this interaction was visualized by FITC labelled HPA, the lectin specific for GalNAc.

We can speculate about the reason for the in vivo blocking attachment results. First, anti-rLuloG antibodies could be proteolytically degraded during blood digestion in the sand fly midgut, which is commonly observed in oral immunotherapy used for mammals. Proteolytic enzymes involved in the degradation of orally administered immunoglobulins in humans include pepsin, trypsin, chymotrypsin, carboxypeptidase and elastase [21]. In P. papatasi and Lu. longipalpis midguts, trypsin-like proteases were described as the most abundant. Secondly, the effectivity of anti-rLuloG antibodies in Leishmania attachment could be limited due to the recombinant protein rLuloG being expressed in soluble form without a GPI anchor, which was used for antibody production. Important roles for GPI anchors in binding of antibodies was demonstrated by numerous studies where removal of the GPI lipid moiety can influence ligand binding properties, most likely due to conformational changes in the protein. Studies showed that antibodies raised against the soluble form of proteins sometimes do not react well with the membrane-bound protein on the living organism, reviewed in [22]. Seeing this unsuccesful inhibiton of Leishmania attachment in vivo, we performed ex vivo inhibition of Leishmania infantum binding to Lutzomyia longipalpis midgut by binding experiments. Details of the method are given by Wilson and coworkers [23]. Although we did not find significant reduction of attached parasites in the presence of anti rLuloG, again a tendency to lower parasite numbers was observed (data not shown).

Nevertheless, a key question is what kind of parasite molecule could interact with this glycoconjugate. We may speculate that the parasite receptor for the 45 kDa glycoconjugate is a lectin-like molecule with predicted specificity for GalNAc, mimicking the activity of HPA lectin. Lectin activities and heparin-binding activities were demonstrated in Leishmania promastigotes by various authors [24–28]. More recent studies suggest that 65 kDa and 55 kDa heparin-binding proteins from the surface of L. braziliensis promastigotes can be involved in parasite adhesion to Lu. longipalpis cells through heparan sulphate bridges [29–31]. These heparin-binding proteins were reported to possess metallo-proteinase like activity [31]. As they were not characterized by mass spectrometry, one cannot exclude the possibility that they are in reality identical with some isoforms of the major surface protease, leishmanolysin/gp 63 (reviewed in [32]).

We predict that the parasite receptor would be expressed at a reasonably high level, to ensure that specific binding to the midgut can occur despite the presence of other high copy surface molecules such as LPG. Reviewing current knowledge of the promastigote surface there is a variety of candidate molecules. These include leishmanolysin/gp63, PSA/gp46, mPPG, laminin-binding proteins or the already mentioned heparin-binding proteins, and maybe other unidentified proteins. A potential role for gp63 in parasite attachment has been previously proposed [33, 34]. These studies showed that the monoxenous trypanosomatids Herpetomonas samuelpessoai and Leptomonas species produce a metallopeptidase that has similar properties to Leishmania gp63 and contributes to parasite adhesion to Aedes aegypti guts or the Aedes albopictus cell line C6/36. These results are consistent with those obtained with Leishmania species; using an ex vivo by binding assay Jecna and colleagues [11] found that the gp63 of L. amazonensis gp63 down-regulated transfectants is functionally important for binding of its promastigotes to the Lu. longipalpis midgut.

In conclusion, while the L. major–P. papatasi work has established an important paradigm, recent evidence indicates that another mechanism of attachment could exist in permissive vectors. The present study has revealed interesting and novel properties of the “19 kDa midgut protein” from Lu. longipalpis midguts that can be briefly summarized as follows: (i) it is a mucin-like glycoconjugate with apparent weight of 45 kDa; (ii) it is located on the luminal surface of the midgut; and (iii) recombinant form rLuloG is able to bind to the surface of Leishmania cells. Nevertheless, additional studies will be necessary to prove the biological role of the “19 kDa midgut protein”, especially its role as a putative midgut ligand in permissive vectors.