Specific Recognition of Leishmania major Poly- (cid:1) -galactosyl Epitopes by Galectin-9 POSSIBLE IMPLICATION OF GALECTIN-9 IN INTERACTION BETWEEN L. MAJOR AND HOST CELLS*

Leishmania parasites are the causative agents of leishmaniasis, manifesting itself in a species-specific manner. The glycan epitopes on the parasite are suggested to be involved in the Leishmania pathogenesis. One of such established species-unique glycan structures is the poly- (cid:1) -galactosyl epitope (Gal (cid:1) 1–3) n found on L. major , which can develop cutaneous infections with strong inflammatory responses. Interestingly, the polygalactosyl epitope of through its para-crine mediation of L. major -macrophage interaction further study is necessary to investigate the potential roles of galectin-9 as a Leishmania receptor.

Innate immunity plays a critical role in the protection against pathogen invasion. The recognition of invading pathogens initiates different innate immune responses, including elimination of pathogens, antigen presentation to initiate acquired immunity, and elicitation of signaling cascades, which regulate local/systemic immune responses (1). Some of the Ctype (calcium-dependent) lectins have been demonstrated to play critical roles as pathogen recognition molecules (2). For example, soluble collectins bind to pathogens resulting in their efficient removal (3). Another example is the membrane-associated dendritic cell-specific intercellular adhesion molecule-3grabbing nonintegrin (DC-SIGN). DC-SIGN binds to "self " glycans as well as pathogen-associated glycans on Mycobacterium tuberculosis, which results in the suppression of dendritic cell functions (4 -8). Thus, pathogen recognition by a lectin could induce distinct immune responses.
In contrast to those C-type lectins, S-type lectins, more recently termed galectins (9,10), have been considered as lectins that bind to self glycans because galectins bind preferentially to polylactosamine chains attached to the host cell glycoproteins/lipids (11,12). However, a number of pathogenic entities express glycoconjugates containing ␤-galactosides, either similar to those of the host cells or unique to pathogens, such as the polygalactosyl epitope on protozoa Leishmania major (13,14). Thus, questions such as whether some galectins can play roles as "nonself " or pathogen-recognizing molecules have recently emerged. Indeed we and others (15)(16)(17) suggest that galectin-3 can recognize such pathogens as Klebsiella pneumoniae and L. major.
Galectin molecules contain one or two carbohydrate recognition domains (CRDs), 1 which have a relatively similar tertiary structure with conserved peptide sequence elements, although they differ in the protein quaternary structure, i.e. the presentation of the CRDs in the molecules. The arrangement of the CRD(s) within the molecule is unique for each galectin, which are consequently classified into three types: prototype, tandemrepeat, and chimera type (18). Prototype galectins contain one CRD and exist as a monomer (such as galectin- 7) or dimer (such as galectin-1, -2), whereas tandem-repeat galectins, such as galectin-8 and -9, contain two CRDs connected by a short linker region. Galectin-3 is chimeric type galectin, composed of one CRD and a non-CRD domain, which regulates the multivalence status of galectin-3. Galectins have different affinities for the substituted lactosamine residues (11,12,19) because of the structural differences in detailed carbohydrate binding pockets (20 -22). Hence, it is possible that some glycoconjugates, including nonself pathogenic glycans, could be recognized by various galectins, whereas others could be relatively specific to a particular galectin (23). It has also been speculated that the differences in the galectin CRD arrangements influence the cross-linking mode of the galectin ligand and contribute to their functions (24 -26), although such a possibility remains to be investigated.
Recent studies demonstrate that some galectins act as immunomodulators and cell adhesion modulators (25,(27)(28)(29)(30)(31)(32)(33)(34)(35). We recently found that galectin-3 binds to L. major through L. major-specific poly-␤-galactosyl epitope (Gal␤1-3) n (16). Galectin-3 binds to L. major but not to the other species, L. donovani or L. mexicana, which do not express the polygalactosyl epitope (16). It has been proposed that the speciesspecific manifestations of Leishmania in human, ranging from fatal visceral infections to mild cutaneous lesions, are accounted for in part by the species-specific glycoconjugates (see Fig. 1) (36 -38). Interestingly, the L. major-specific polygalactosyl epitope is implicated in two stages of the Leishmania life cycle: the host-pathogen interactions (39,40) and the attachment of Leishmania to its sand fly vector (41,42). It is likely that such roles of the polygalactosyl epitopes are mediated by ␤-galactoside-binding proteins such as galectins. However, galectin-3 is hitherto the only lectin that has been shown to have an affinity for this unique epitope (16). Various galectins are expressed in dermal tissues (initial site of the Leishmania infection) which is composed of phagocytic cells, fibroblasts, and epithelial cells as well as emigrated lymphocytes. Thus, it is possible that other galectins recognize L. major. Because each galectin possesses a unique structure, binding of different galectins to the speciesspecific ligands could affect the different manifestations of the infections, especially the initial immune responses. However, it remains to be determined whether other galectins have any affinities to the L. major-specific glycan epitopes and whether interactions between the parasites and different galectins influence the development of immune responses in a manner different from that of galectin-3.
Here we show that galectin-9 can bind to L. major in a species-specific manner through L. major-specific poly-␤-galactose (Gal␤1-3) 1ϳ4 . Even though that both galectin-3 and -9 recognize L. major, only galectin-9 promotes L. major-macrophage interaction. Thus the data suggest the distinctive roles of galectin-9 in the L. major-specific pathogenesis of leishmaniasis.

MATERIALS AND METHODS
Reagents-Chemicals and other reagents were obtained from Sigma unless specified otherwise. Antibodies against galectin-1, -3, and -9 were raised in our laboratories by injecting recombinant galectins to rabbits (16,33).
For some experiments, the promastigote population was separated into procyclic-and metacyclic-rich fractions using Ficoll gradients as published by Spath and Beverley (44). Briefly, stationary phase cultures of Leishmania (ϳ10 9 parasites) were centrifuged at 5,000 ϫ g for 10 min at room temperature and resuspended in 3 ml of serum-free Dulbecco's modified Eagle's medium (Invitrogen). The cell suspensions were then loaded onto a Ficoll gradient composed, from the bottom, of 2 ml of 20%, 2 ml of 15%, 2 ml of 10%, and 2 ml of 5% Ficoll diluted in 4ϫ serum-free medium 199. The gradient was next centrifuged at 1,300 ϫ g for 10 min at room temperature. The metacyclic promastigotes were recovered on the top of the 10% Ficoll layer, and the procyclic ones were recovered in the 15% Ficoll layer. The parasites in those fractions were washed twice with RPMI 1640 complemented with 25 mM HEPES and used immediately for the experiments.
To prepare galectin-immobilized columns, the recombinant galectins were dissolved in 0.2 M NaHCO 3 , pH 8.3, containing 0.5 M NaCl and 0.1 M lactose and coupled to HiTrap NHS-activated columns (Amersham Biosciences) according to the manufacturer's instructions. After washing and deactivating excess active groups by ethanolamine, the galectin-immobilized agarose beads were removed from the cartridge. Each gel matrix was suspended in PBS containing 1 mM EDTA, and the slurry was packed into a stainless steel column (4.0 ϫ 10 mm; bed column, 0.126 ml, GL Sciences, Inc.).
Alexa 488-labeled galectins were prepared according to the manufacturer's instructions (Molecular Probes) with a slight modification. Briefly, recombinant galectin in PBS containing 5 mM lactose and 20 mM HEPES was first transferred into the manufacturer-supplied bottle containing Alexa 488 carboxylic acid, succinimidyl ester, dilithium salt (Molecular Probes), and incubated at room temperature for 1 h. After terminating the labeling reaction by adding Tris-HCl, pH 7.5 (final concentration 20 mM), the reaction mixture was first applied to a PD10 column (Amersham Biosciences) to remove lactose and excess dye, and the void fraction was collected to obtain Alexa 488-labeled galectin. Alexa 488-labeled galectin was purified further by asialofetuin-agarose as published before (32).
Galectin Binding to L. major Parasites-Leishmania parasites (1 ϫ 10 7 ) were incubated with recombinant galectin (1 ϳ 2 M) in 250 l of buffer consisting of 125 l of serum-free RPMI 1640 containing 25 mM HEPES and 125 l of PBS in the presence or absence of 100 mM lactose at 4°C for 30 min as described previously (16). When the binding assays were performed in the presence of lactose, the sodium chloride concentration in the PBS was adjusted to maintain appropriate osmolarity, i.e. 317 mosmol/liter. After the incubation, parasite-free supernatants were obtained by spinning at 4,500 ϫ g for 7 min. The supernatants were fractionated by SDS-PAGE, and proteins in the gels were stained with Coomassie Brilliant Blue. Alternatively, fractionated proteins on the SDS-polyacrylamide gels were transferred to the nitrocellulose filters, and galectin was detected by the appropriate antibody against galectin.
For flow cytometric analysis, Leishmania were incubated as above with Alexa 488-labeled galectin for 15 min at 4°C. Unbound galectin was removed by a brief wash with ice-cold PBS twice, and Leishmania were fixed in 2% formaldehyde containing PBS for 10 min at 4°C and analyzed for the relative fluorescent intensity on a FACScalibur flow cytometer (BD Biosciences, San Jose, CA) as described previously (32).
Polygalactosyllactose Derivatives and Pyridylaminated (PA) Oligosaccharides-Polygalactosyllactose derivatives were purified from the milk of the tammar wallaby (Macropus eugenii; kindly provided by Dr. Michael Messer, Department of Biochemistry, University of Sydney, Australia) as described previously (47)(48)(49)(50). Briefly, 4 volumes of chloroform/methanol mixture (2:1, v/v) were added to the wallaby milk. After being thoroughly mixed and separated by centrifugation at 4°C and 3,500 ϫ g for 30 min, the methanol/water upper fraction containing oligosaccharides was obtained, and solvents were removed by evaporation followed by freeze-drying. The oligosaccharides (ϳ100 mg) were fractionated by gel filtration on a Sephadex G-25 column (100 ϫ 2.5 cm, Amersham Biosciences), and polygalactosyllactose derivatives were purified further by gel filtration on a Bio-Gel P-4 (100 ϫ 2.5 cm, Bio-Rad), as published before. The purified oligosaccharides were labeled with PA following the manufacturer's instructions (Takara Shuzo, Tokyo, Japan). The purity of PA-labeled oligosaccharides was confirmed as described previously (51).
Frontal Affinity Chromatography (FAC) Analysis-FAC analysis was carried out as described previously (51)(52)(53). Briefly, PA-oligosaccharide was dissolved in PBS containing 1 mM EDTA and applied to the column through a 2-ml sample loop connected to the Rheodyne 7725 injector.
The sample loop and the column were immersed in a 20°C water bath. The flow rate was controlled by a Shimadzu LC-10Advp pump at 0.25 ml/min. Elution of PA-oligosaccharide from the column was monitored by a Shimadzu PR10AxL fluorescence detector at 400 nm (excitation at 320 nm). The elution volume of PA-oligosaccharide of interest (V f ) was determined as described previously. The elution volume of PA-oligosaccharide, which has no affinity to galectins (V 0 ), was determined by using PA-rhamnose.
The total amount of immobilized galectin in the column, B t is first determined by using PA-lacto-N-fucopentaose I (Fuc␣1-2Gal␤1-3GlcNAc␤1-3Gal␤1-4Glc OPA ) using the equation described previously, where K d is the dissociation constant between interacting biomolecules, B t is the total amount of immobilized ligand, [A] 0 is the initial concentration of the PA-oligosaccharide of interest (A), V f is the elution volume of A, and V 0 is the elution volume of PA-rhamnose, which has no affinity toward galectins. As in the assays employed in this study, [A] 0 , the initial concentration of the PA-oligosaccharides, was 10 nM, which was negligibly small compared with K d . Thus, V f approached the maximum value, V f , which is independent of [A] 0 , and we used the following equation to obtain the values of K d of each galectin toward a polygalactosyllactose derivative.
Infection of Macrophages with L. major Parasites-Stationary phase L. major promastigotes at a cell concentration of 10 7 cells/ml were labeled with cell tracker green (4.65 g/ml, Molecular Probes) following the manufacturer's instruction. J774A.1 macrophages (2 ϫ 10 5 cells) were incubated in 0.5 ml of RPMI 1640 and 25 mM HEPES with the labeled parasites (5 ϫ 10 6 cells) in the absence or presence (1.6 M) of galectins and incubated for 30 min at 37°C. Following the incubation, parasites that were not associated with macrophages were removed by washing three times with RPMI 1640 and 25 mM HEPES, and cells were then lysed in PBS containing 1% Triton X-100. Fluorescenceassociated macrophages, which corresponds to the number of L. major bound to or internalized by macrophages, were measured using a fluorometric plate reader (PerSeptive Biosystems).

Analysis of the Binding Properties of Galectin-1 and -3 toward Polygalactosyllactose Derivatives by FAC-
We have recently developed an assay system to assess the binding affinity of various galectins for oligosaccharides by employing FAC analysis with fluorescence-tagged oligosaccharides (53). Thus, we next established a system to search for lectins that can bind to and distinguish L. major from other Leishmania species through recognition of the poly-␤-galactosyl epitope. To achieve this goal, galectin-1, which is ubiquitously expressed in various tissues, and galectin-3 were first used to assess the affinities for the epitope. For the FAC analysis, we used purified poly-␤galactosyllactose instead of L. major-derived lipophosphoglycans (LPGs) because the LPGs prepared from L. major are a mixture of various lengths of the polygalactosyl epitope (Gal␤1-3) n , ranging from one to three repeats (54) (R in Fig. 1). Various lengths of poly-␤-galactosyllactose derivatives (Gal␤1-3) n Gal␤1-4Glc, were purified from the milk of the tammar wallaby (48,49), and those derivatives were labeled with PA residues. One of the PA-polygalactosyllactose derivatives dissolved in PBS-EDTA at a concentration of 10 nM was applied continuously to a column containing either galectin-1 or -3 agarose. Elution profiles of PA-labeled sugars from galectin-1 and -3 are shown collectively in Fig. 2. As control, PA-rhamnose, which does not show any affinity for galectins, was also applied to the column.
As shown in Fig. 2, K d values of galectin-3 for (Gal␤1-3) n lactose (3.6 ϳ 6.1 M) were in the range where galectin-3 can establish stable association with the given oligosaccharides because the K d values of galectin-3 for an established ligand, biantennary N-linked oligosaccharide, was 1.4 M (11, 53). The FAC analysis also indicates that the binding affinity of galectin-3 increases additively in proportion to the number of Gal␤1-3 residues in ␤-polygalactosyllactose. Given the finding that the number of ␤-linked galactose residues substituted on the repeating saccharide unit on LPGs varies from one to three (54), those data obtained by FAC analysis are consistent with our previous result that galectin-3 binds to L. major (16). The number of Gal␤1-3 residues that are required for the binding by galectin-3 was not determined by our previous data. Galectin-3 K d values for (Gal ␤1-3) 1ϳ4 obtained by FAC analysis suggest that galectin-3 can bind to all L. major-specific galactosyl side chains regardless of the number of ␤-galactose residues on the repeating units.
In contrast to galectin-3, the K d values of galectin-1 toward any of the polygalactosyllactose derivatives were significantly higher (ϳ60-fold) than those of galectin-3. Because the K d values of galectin-1 for one of the established oligosaccharide ligand, bianntenary N-linked oligosaccharide, was 7.6 M (53), these data suggest that galectin-1 does not have a significant affinity for (Gal␤1-3) n lactose.
Establishment of a Novel System to Identify L. major-recognizing Galectins-We have demonstrated previously that galectin-3 binds to L. major parasites, in vivo, through recognition of the polygalactose epitope (16). Consequently, L. donovani parasites that do not contain this epitope (Fig. 1), are not recognize by galectin-3 (16). The K d values of galectin-1 for the polygalactose epitopes determined by FAC analysis are ϳ60 times higher than those of galectin-3, raising the possibility that galectin-1 cannot bind to L. major in vivo. Thus, whether the K d values of galectins correlate closely with their affinity for L. major in vivo was next studied. Galectin-1 or -3 at a concentration of 1 M was incubated with Leishmania at 4°C for 30 min. After incubation, the unbound galectin was Specific Recognition of L. major by Galectin-9 separated from Leishmania promastigotes by centrifugation. The amounts of galectin in the unbound fractions were analyzed by SDS-PAGE followed by protein staining with Coomassie Brilliant Blue or by Western blotting with antibodies against the corresponding galectin. Under the condition in which galectin-3 binds to L. major (16), the majority of galectin-1 added to the binding assay remained in the supernatant, suggesting that galectin-1 does not bind to L. major (data not shown). Neither galectin-1 nor -3 was associated with L. donovani (data not shown), suggesting that galectin-1 does not bind to the type of ␤-galactoside, Gal␤1-4Man-, which is found on most of the Leishmania species (Fig. 1). The data from these analysis suggest that the K d values obtained by FAC analysis with defined poly-␤-galactosyl polymers are correlated closely with the affinities of galectins toward L. major parasites in vivo. Thus, FAC analysis can be employed to search host galectins that can bind/recognize pathogens of interest.
Binding Properties of Various Galectins to PA-labeled Poly-␤-galactosyllactose Derivatives-The affinities of other galectins for the poly-␤-galactosyl epitopes were next examined by FAC analysis. As shown in Table I, among galectins tested here, the K d values of galectin-9 for all tested poly-␤-galactosyllactose residues (Gal␤1-3) 1ϳ4 were comparable with those of galectin-3, which binds to L. major parasites in vivo. The binding affinities of galectin-9 for poly-␤-galactosyllactose residues were synergistically enhanced with close correlation to the number of Gal␤1-3 units (Table II). Thus, the results suggest that galectin-9 has K d values that are in the range to show affinity for L. major-containing poly-␤-galactosides (Gal␤1-3) 1ϳ3 . Galectin-1, -2, and -7 had relatively higher K d values, suggesting that those galectins cannot stably bind to those oligosaccharides in vitro and likely to L. major in vivo. Although the K d value of galectin-8 toward (Gal␤1-3) 4 lactose was comparable with the K d value of galectin-3, galectin-8 did not appear to have significant affinity for the (Gal␤1-3) 1ϳ2 lactose, which are found as the major glycoconjugates of L. major, LPG.
Binding of Galectin-9 to Leishmania Parasites-Binding of galectin-9 to L. major was next investigated. Galectin-9 (0.5 M, 250 l) was incubated with L. major or L. donovani (1 ϫ 10 7 cells) for 30 min at 4°C. As shown in Fig. 3A, galectin-9 was detected readily by polyclonal antibody against galectin-9 in the parasite-free medium when incubated with L. donovani. In contrast, incubation with L. major resulted in very little galectin-9 in the parasite-free medium (Fig. 3A), indicating that galectin-9 binds to L. major. Those data suggest that galectin-9 can distinguish L. major from L. donovani.
Polygalactosyl residues are synthesized by a L. major-specific enzyme, ␤1,3-galactosyltransferase (41,55,56). Recently, Butcher et al. (41) established a L. major mutant, called Spock, which lacks the activity of the ␤1,3-galactosyltransferase in the background of L. major Friedlin. Consequently, the phosphoglycan domain of Spock LPG is not modified with terminal polygalactosylation and exhibits structural similarity with L. donovani LPG (41). To study directly whether galectin-9 binds to L. major-specific polygalactose epitopes, the interaction of galectin-9 with L. major Friedlin or with its mutant Spock was next investigated. When galectin-9 was incubated with these parasites at 4°C, a significantly reduced amount of galectin-9 was found in the incubation supernatant exposed to L. major Friedlin wild type as expected. In contrast, there was no obvious absorption by the L. major Spock mutant (Fig. 3A). Thus, the data suggest that galectin-9 binds to L. major through recognition of L. major-specific polygalactose epitopes, consistent again with the results obtained by FAC analyses.
The interactions between galectins and Leishmania promastigotes were further estimated by analyzing the binding of Alexa 488-labeled galectin-3 or galectin-9 to the parasites by flow cytometric analysis. As shown in Fig. 3B, significant levels of both galectin-3 and galectin-9 were associated with L. major, consistent with the results obtained by the measurement of galectin absorption by L. major in vivo (Fig. 3A (16)) and FAC analysis (Fig. 2, Table I, and Table II). In contrast, no significant galectin-3 was found on the L. donovani surface, also confirming our previous observation that galectin-3 shows no affinity for L. donovani surface glycoconjugates (16). We observed a small shift in flow cytometry analysis with L. donovani incubated with galectin-9, possibly because of the existing slight interaction between galectin-9 and L. donovani. Nevertheless, significantly lower levels of galectin-9 were associated with L. donovani than with L. major, demonstrating that galectin-9 does not exhibit high affinity for L. donovani glycoconjugates.
Interaction of Galectins with L. major Metacyclic and Procyclic Promastigotes-L. major LPG structures are modified during the differentiation of promastigotes from a less infectious procyclic form (in logarithmic growth phase) to a highly infectious metacyclic form (in stationary growth phase). Detailed structure analyses of LPG of the different stages of L. major reveal that the average lengths of poly-␤-galactose are relatively shorter in metacyclic promastigotes than in procyclic ones; 39% of the LPG side chains of metacyclic promastigotes are (Gal␤1-3) 1-3 , whereas in procyclic promastigotes they represent 81% of the LPG side chains (57). As summarized in Table II, FAC analysis revealed some interesting differences between galectin-3 and -9 in binding to the longer form of polygalactose; the affinity constant of galectin-9 for polygalactose epitope is increased synergistically with the number of  3. Binding of galectins to Leishmania promastigotes. A, L. major-specific disappearance of galectin-9 through L. major-specific polygalactosyl epitopes. Galectin-9 (0.5 M, 250 l) was incubated with various species of Leishmania promastigotes for 30 min at 4°C. After incubation, the supernatants were separated from the parasites by centrifugation. Galectin-9 in parasite-free fractions was detected by Western blotting with anti-galectin-9 antibody. Galectin-9, which was incubated without Leishmania, is shown as control. B, binding of galectin-3 and galectin-9 to L. major promastigotes. Leishmania promastigotes were incubated with Alexa 488-labeled galectin-3 or galectin-9 for 15 min at 4°C. In controls, the treatment with Alexa 488-labeled galectin is omitted. Relative fluorescent intensity is shown on abscissa and cell numbers on the ordinate.
repeats, possibly suggesting that the affinities of galectins for different stages of promastigotes are different. We have been using the stationary phase of promastigotes, which are in fact the mixture of procyclic and metacyclic promastigotes. Thus, we next investigated the effect of Leishmania promastigote differentiation on the binding of galectins. Procyclic and metacyclic promastigote-rich fractions were obtained by Ficoll partition. As shown in Fig. 4, galectin-3 was absorbed equally by both procyclic (47% compared with galectin-3 absorbed in the presence of lactose) and metacyclic forms of L. major promastigotes (45%), indicating that galectin-3 binds to both forms of Leishmania promastigotes through its lectin activity. Galectin-9 also showed affinity to both L. major metacyclic and procyclic promastigotes, although galectin-9 was more absorbed by procyclic promastigotes (92% compared with galectin-9 absorbed in the presence of lactose) than by metacyclic one (43%). In summary, whereas galectin-9 may bind to L. major procyclic promastigotes in preference to the metacyclic promastigotes, both galectin-3 and -9 can recognize procyclic and metacyclic forms of L. major.
Promotion of the Interaction between L. major and Macrophages by Galectin-9 -Leishmania are obligate intracellular parasites, and thus the invasion of macrophages is a critical part of the parasite life cycle. Because various galectins are implicated in cell-cell interactions (25), we next investigated whether the abilities of galectin to recognize L. major could lead to the promotion of L. major interaction with macrophages. To quantify the interaction between L. major and macrophages, we used CellTracker green-labeled L. major. When CellTracker green, a membrane-permeable dye, is taken up inside parasites, the dye is transformed into a cell-impermeable fluorochrome and is therefore not transferred to adjacent cells. Thus, the fluorescent values found in macrophages reflect the cell number of parasites associated with macrophages. Macrophage cell line J774A.1 were infected with L. major promastigotes labeled with CellTracker green in the absence or presence of galectins during a period of 30 min. After incubation, parasites that were not associated with macrophages were removed by washing, and the fluorescence intensity associated with macrophages was measured to calculate the percentage of Leishmania bound to macrophages. As shown in Fig. 5, in the absence of galectin, 10% of added L. major parasites were found to be associated with macrophages possibly through macrophage endogenous mannose/ fucose receptors or through complement receptor type 3 (58,59). Addition of galectin-3 did not affect the interaction between parasites and macrophages. In contrast, in the presence of exogenously added galectin-9, the association of L. major with macrophages was promoted 2-fold. Thus, the data demonstrate that galectin-9 but not galectin-3 can promote L. major interaction with macrophages, suggesting the differential roles of galectins in leishmaniasis. DISCUSSION Our previous paper (16) and the present results obtained by classical biological assays with galectins and by FAC analysis demonstrate that galectin-3 and -9 recognize L. major but not L. donovani through L. major-specific poly-␤-galactosyl epitopes. These findings have several implications in the understanding of the pathogenesis of leishmaniasis and the L. major life cycle.
In the initial stages of Leishmania infections, extracellular, flagellated Leishmania (called promastigotes) are transmitted to humans through the bite of sand flies. The promastigotes are engulfed by macrophages where they differentiate into nonflagellated Leishmania amastigotes and multiply (36,60). Among the several Leishmania species, L. donovani infections result in a visceral form of the disease (lethal if not treated) with limited inflammatory response at the primary site of infection (36). On the other hand, L. major infections develop in dermal tissues as cutaneous lesions (often nonlethal), with massive recruitment of leukocytes to the site of infection, suggesting that the cutaneous forms of the disease are correlated with a vigorous inflammatory response (36). It has been proposed that the difference between L. major and L. donovani in the induction of initial immune responses is accounted for partly by variation in the surface molecules of these Leishmania species, such as the phosphoglycan structure of LPG and by the interaction of such species-specific molecules with tissue macrophages or various cells present at the site of infection (13,38). Further, various reports suggest that the L. major-specific poly-␤-galactosyl epitopes are implicated in the progression of immune responses against Leishmania infections (13,39,40,42,54), although the host receptors/lectins have remained unclarified. As our data in this and the previous paper (16) suggest that galectin-3 and galectin-9 can distinguish L. major from L. donovani, galectin-3 and 9 are two such host factors, which can be implicated in the species-specific tropism of the disease. It has been established that galectin-1, -3, and -9 play different roles in immune responses (24,25). In the case of the roles of galectin-3 and -9 in immune responses to L. major infections, our published results and the present data demonstrate that galectin-9 plays a different role than galectin-3. Among the various immunomodulative activities of galectin-3, the immunological importance of surface galectin-3-glycoprotein lattices has been demonstrated recently (61). The surface galectin-3 lattices can robustly restrict the lateral mobility of surface receptors, raising the threshold for ligand-dependent receptor clustering and signal transduction (27,61,62). We found previously that these immunosuppressive galectin-3 lattices on leukocytes are destroyed upon galectin-3 recognition of L. major during macrophage infection. Thus, it is proposed that through this disappearance of galectin-3 lattices, L. major infection decreases the threshold for the initiation of signal transduction pathways that bias the immune response toward strong local inflammatory responses, which is observed in L. major infection (16). In the case of galectin-9, our data suggest that galectin-9 acts as a L. major-specific tagging protein or an adhesion molecule so that macrophages can recognize L. major more efficiently. It has been known that some galectins, which can cross-link different cells expressing their oligosaccharide ligands (28,34,35), act as adhesion molecules even though galectins are soluble proteins (32,63,64). Galectin-9 contains two CRDs linked by a peptide linker and is thus intrinsically divalent (33,46,65). This multivalence of galectin-9 could, as suggested by our data, cross-link L. major to mononuclear phagocytic cells and work as a macrophage receptor. Galectin-9 is expressed abundantly in T lymphocytes, dendritic cells, lymphatic epithelial cells, as well as various lymphatic tissues, and its expression and secretion increase substantially in antigen-activated peripheral blood mononuclear cells (46,65,66). A few days after the infection, Leishmania are drained into lymph nodes where activated lymphocytes likely secrete galectin-9, and it is possible that galectin-9 influences the development of manifestation through its paracrine mediation of L. major-macrophage interaction although further study is necessary to investigate the potential roles of galectin-9 as a Leishmania receptor.
In the sand fly, proliferating Leishmania procyclic promastigotes, which are less infectious, are retained inside the midgut through attachment to the epithelial cell layers. Once differentiated into more infectious metacyclic promastigotes, the parasites lose their ability to adhere to the midgut, thus permitting the release of metacyclic Leishmania for subsequent transmission into the human host through bites. Thus, the procyclic parasite adhesion to the midgut is important to maintain the Leishmania infectious status in its vector. It has been suggested that the alteration in the side chain glycan structures of LPG repeating units during L. major metacyclic differentiation accounts for the regulation of the stage-specific interaction between L. major procyclic promastigotes and its sand fly vector Phlebotomus papatasi midgut epithelial cells (41,42). When L. major procyclic promastigotes are differentiated into metacyclic, the number of longer poly-␤-galactosyl (Gal␤1-3) n side units/repeating unit are decreased, and the substitution of (Gal␤1-3) n for arabinose (Ara␣1-2) at the nonreducing terminal is increased. It is yet unclarified which structural changes are implicated in the adhesion. Nevertheless, such recognition should be mediated by ␤-galactosidebinding lectins, although the identities of the lectins expressed in the sandfly midgut remain unknown so far. Because galectins are found in many phyla ranging from sponges, Caenorhabditis elegans, Drosophila, to human (67, 68), our present data suggesting that galectin-9 can promote L. major-cell in-teraction through its binding to polygalactosyl epitope raise the strong possibility that the Leishmania vector, sand fly, may utilize some galectin-like molecules for the retention of L. major within its intestine. In fact, a recent review introduces preliminary data suggesting that galectin-like molecules are present in the library of the midgut of the sand fly vector for L. major, P. papatasi (69). Thus, whereas galectin-like molecules are likely to play critical roles for the retention of L. major procyclic promastigotes in the midgut, it is not yet clear from our data how galectin-like molecules of the sand fly can distinguish between procyclic and metacyclic L. major because human galectin-3 and galectin-9 bind to both stages of L. major. Because each mammalian galectin is suggested to recognize distinctive sets of surface receptors in vivo, it is intriguing to propose that either sand fly galectins have a unique carbohydrate binding pocket that can distinguish the LPG differences, or the expression of the galectin-like molecule on the surface of the midgut can be regulated in the sand fly midgut.
In recent years, analyses under dynamic reaction equilibrium by employing the measurement of surface plasmon resonance (SRP) have contributed significantly to the understanding of oligosaccharide-lectin interactions. SRP analyses studying the interaction between selectins and their glycoprotein ligands (70,71) and between galectin-3 and laminin (72) have clearly demonstrated some of the precise nature of these oligosaccharide-lectin interactions under dynamic reaction equilibrium. SPR signals are changed in proportion to the surface molecular mass changes, which occur when soluble molecules bind to the molecules immobilized to the sensor chip. Thus, the tested soluble molecule (analyte) has to have a relatively large molecular mass for the analysis of oligosaccharidelectin interaction by SRP. Therefore, oligosaccharides of interest rather than lectins have to be covalently coupled onto the sensor chips because the molecular masses of oligosaccharides are generally smaller than proteins. Thus, the analysis by SRP may not be suitable to study the interaction between lectins of interest and multiple sets of oligosaccharides to search for lectin ligand candidates. The FAC system that we employed here has certain advantages for this purpose. The interaction between oligosaccharides and lectins can still be measured under equilibrium without using a large amount of novel oligosaccharides. Because FAC analysis can be performed with immobilized lectins rather than immobilized oligosaccharides, multiple sets of oligosaccharides can be screened. Analysis of the binding properties of galectins employing FAC system enables us to screen the potential galectin candidates for the L. major-specific epitopes systematically. Our data demonstrate that the K d values of galectins for PA-polygalactosyl epitopes correlate closely to their abilities to bind to Leishmania parasites in vivo. Thus, FAC analysis is able to search for whether certain novel or unique oligosaccharides expressed on pathogens or host cells can be potential galectin ligands with a relatively small amount of the oligosaccharides. Using this technique, we were able to find unique oligosaccharides such as poly-␤-galactose epitope which can bind to a galectin. Recent advances in the determination of rare oligosaccharide structures and in the discovery of glycosyltransferases that synthesize those rare glycans began to reveal various unique carbohydrate structures that play crucial roles in cellular communication during nonself pathogen recognition or other cellular activities (6,(73)(74)(75). Therefore, employment of FAC analysis would broaden the horizons to find such specific lectin ligands easily in the future with limited amounts of novel oligosaccharides.
In conclusion, our data suggest that galectin-9 but not galectin-1 can recognize L. major through L. major-specific poly-␤-galactose epitopes. Although galectin-3 and -9 can bind to L. major, galectin-9 but not galectin-3 promotes the interaction of L. major with macrophages, suggesting that galectin-9 acts as a macrophage L. major-specific receptor. Recent reports demonstrate that galectin-3 and -9 are secreted actively from inflammatory activated macrophages/dendritic cells and T lymphocytes, respectively (32,33,76) and act as novel immunomodulators for the development of innate immune responses (28,32,34,35). Because the L. major-specific poly-␤-galactose epitope is implicated in the species-specific strong local immune response observed in L. major-induced manifestation, our data raise the possibility that galectin-3 and -9 act as two host factors that can influence the development of L. majorspecific strong initial immune responses in distinctive manners.