Pre-B cell receptor acts as a selectivity switch for Galectin-1 at the pre-B cell surface

Galectins are glycan binding proteins translating the sugar-encoded information of cellular glycoconjugates into many physiological activities including immunity, cell migration, and signaling. During early B lymphocytes (BL) development at the pre-B cell stage, BL express the pre-B cell receptor (pre-BCR) and are supported by mesenchymal stromal cells secreting Galectin-1 (Gal-1). Gal-1 interacts with glycosylated receptors from stromal and pre-B cell surfaces but also with the pre-BCR through a direct carbohydrate-independent contact. How this interaction might interplay with the glycan-decoding function of Gal-1 is unknown. Here, we investigated Gal-1 binding to cell surface ligands using NMR spectroscopy on native membranes. We showed that pre-BCR regulates Gal-1 binding to specifically target α2,3-sialylated receptors on pre-B cells. Upon pre-BCR interaction, dynamic changes resulted in additional contacts with α2,3-sialylated glycans converting Gal-1 from an exo- to an endo-type lectin. Remarkably, this selectivity switch is able to promote pre-B cell survival. Altogether, we shed light on a new mechanism allowing fine-tuning of Galectin specificity at the cell surfaces.


Introduction
Glycans are essential for life and play many different fundamental roles in nearly all biological processes 1 . Their tremendous diversity is key to store chemical information at the cell surfaces, called « glycocode », which is translated into cellular responses by binding to specific proteins mainly known as lectins 2,3 . Among lectins, Galectins are a highly conserved family of fifteen known proteins (Gal-1 to Gal-15), defined by their ability to bind βgalactosides through a conserved carbohydrate recognition domain (CRD) which includes the carbohydrate binding site (CBS) 4 . These lectins, found in many cell types 5 , are well-known regulators of cell responses and mediate a wide variety of functions such as cell binding, migration, differentiation, cellular trafficking, and cell signaling 4 . Galectins are also involved in several pathological processes such as inflammatory diseases, oncogenesis, cardiovascular disorders, and host-pathogen interactions [6][7][8][9][10] , making Galectins attractive therapeutic targets [11][12][13] . All these functions, at the surface of many different cell types, in all kind of environments, are achieved through their capability to oligomerize which leads to crosslinking of specific cell-surface glycoproteins or glycolipids into lattices 14 . Given the abundance of β-galactoside-derived glycoconjugates on cell surfaces, how these lectins select a subset of ligands among many available candidates to mediate a specific cellular function remains elusive.
The concept of Galectins as functioning in the extracellular compartment only through carbohydrate interactions has been recently challenged by the identification of nonglycosylated binding partners [15][16][17][18] . Whether these interactions participate to cell glycome decoding by Galectins is unknown. The first example of a carbohydrate-independent interaction in the extracellular compartment concerns the prototype Gal-1 homodimer and the pre-B cell receptor (pre-BCR) 19 . Gal-1 is an exo-type lectin i.e. it interacts specifically with terminal units of polysaccharides 20,21 . During B cell differentiation in the bone marrow, pre-BCRs are expressed at the surface of pre-B cells, which are present in a specialized cellular niche consisting of stromal cells secreting Gal-1. At this developmental stage, Gal-1 establishes interactions with glycoconjugates and the pre-BCRs at the contact zone between pre-B and stromal cells 19,22 . This Gal-1-dependent lattice drives pre-BCRs clustering, activation, and subsequent downstream signaling which has been implicated in cell survival, proliferation, and differentiation 23 . However, the molecular mechanism underlying Gal-1 functions at the pre-B cell surface remains unknown. Our previous study revealed that Gal-1 interacts with a non-glycosylated region of the pre-BCR, the λ5 unique region (λ5-UR), that docks onto a Gal-1 hydrophobic surface behind the CBS 24 . While binding at distance from the CBS, λ5-UR induces carbohydrate affinity changes as tested on glycan arrays 15 . These data suggested that λ5-UR binding to Gal-1 modulates Gal-1 binding activity for structurally related carbohydrates, decreasing affinity for branched and linear poly-N-acetylactosamine (poly-LacNAc) and increasing affinity for sulfated or α2,3 sialylated poly-LacNAc. Whether these affinity changes represent a mechanism to regulate Gal-1 interactions and function at the pre-B cell surface is still to be demonstrated on physiological cell surface ligands embedded in the plasma membrane. To date there are no published data analyzing at the structural level Galectin binding to native cell surface ligand.
Here, we addressed the challenge of studying Gal-1 binding to pre-B and stromal cell surfaces using solution-state "on-cell" NMR spectroscopy and investigated the effect of λ5-UR interaction on Gal-1 binding to its physiological ligands. We show that λ5-UR allows Gal-1 to select specific glycosylated receptors at the pre-B cell surface. We identified α2,3sialylated glycans as key targeted epitopes and demonstrated in vivo that this regulation can rescue pre-B cells from cell death. Defining the structural basis of this regulation highlighted a switch converting Gal-1 from an exo-to an endo-type lectin upon λ5-UR interaction.
Altogether, our study shows that Galectin/non-glycosylated protein interactions can act as regulators of Galectin functions and may be the missing piece of the puzzle to understand how Galectins acquire their target specificity at the cell surfaces.

NMR reveals the structural basis of Gal-1 binding to cell surface ligands
To investigate Gal-1 binding to glycoconjugates in their native membrane environment, we utilized stromal and pre-B cell lines (OP9 and Nalm6, respectively) to extract glycoconjugates-enriched membrane vesicles amenable for NMR studies. These cell lines were previously used to model the pre-B/stromal cell synapse and pre-BCR relocalization 24 .
For reference, we characterized these membrane vesicle preparations by negative stain electron microscopy and observed on average 50 nm and 25 nm diameter vesicles extracted from stromal and pre-B cell lines, respectively (Supplementary Fig. 1a and 1b). NMR experiments carried out consisted in recording 1 H, 15 N HSQC spectra on 15 N-labeled Gal-1 alone and incubated with membrane vesicles (Fig. 1a). Upon vesicle addition, chemical shift deviations (CSDs) and line broadening were observed, indicative of complex formation with vesicle surface ligands ( Fig. 1b and Supplementary Fig. 1c and 1d). To provide evidence that glycans are mediating Gal-1 binding to cell vesicles, we released complex N-and O-glycans from the cell surface by enzymatic cleavage using PNGase F and O-glycosidase. After treatment, no variation nor intensity decrease were observed on Gal-1 spectra confirming that Gal-1 interaction to cell vesicles is glycan-dependent ( Fig. 1b and Supplementary Fig. 1e).
These results demonstrate not only the presence of functional glycosylated ligands in the isolated membrane vesicles but also the possibility to monitor Gal-1 interactions at atomic resolution in their physiological context.
Previously, the Gal-1 CBS has been defined as containing five subsites, A to E, where subsite C is accommodating the essential β-galactose unit, the other subsites interacting more specifically with the other units of the glycan 25 . When incubated with cell vesicles, CBS subsites C and D were mainly perturbed, thus confirming Gal-1 binding to cell surface βgalactosides (Fig. 1c, 1d and Supplementary Fig. 2b and 2c). However, significant differences were observed at the residue level depending on the cell vesicles added. Remarkably, additional perturbations in CBS subsites A and E were observed in the presence of stromal cell vesicles, indicating binding to β-galactoside ligands with extensions pointing towards these subsites ( Fig. 1d and Supplementary Fig. 2c). In addition to CBS perturbations, residues on the Gal-1 backside surface showed strong CSDs ( Fig. 1c and 1d). These variations outside the CBS could represent an additional interaction site for extended complex cellular glycans.
Several of these resonances experienced CSDs with both stromal and pre-B vesicles while some others are perturbed only with pre-B (A75, F77) or stromal (K28, C88, D102) vesicles, illustrating differences at the residue level depending on which cell vesicles were added.
Finally, residues at the dimer interface showed CSDs ( Fig. 1c and 1d) but also the strongest peak intensity decrease (Supplementary Fig. 2b and 2c), indicating that cell surface ligand binding to the CBS induces a long-range effect resulting in the stabilization of the dimer state.
Altogether, these data show that Gal-1 binding to cell surface glycosylated ligands impact not only the CBS but also other regions of the CRD (backside and dimer interface). Ultimately, while binding to glycoconjugates containing β-galactosides, significant differences are also observed at the residue level demonstrating that pre-B and stromal cells contain different sets of Gal-1 ligands. These results are in agreement with lectin microarrays data on pre-B and stromal cells showing cell specific glycomic signatures 15 .

λ5-UR regulates Gal-1 binding to pre-B cell ligands to target specific glycan epitopes
Within the pre-B cell niche, Gal-1 secreted by stromal cells contacts pre-B cells through binding to glycosylated receptors and to λ5-UR from the pre-BCR 19,22 . To determine whether λ5-UR can change Gal-1 binding properties to pre-B cell glycosylated ligands, we examined the effect of λ5-UR on the resulting 1 H, 15 N-HSQC spectra of Gal-1 in the presence of pre-B cell vesicles.
When added to Gal-1 bound to pre-B cell vesicles, λ5-UR induced CSDs in the λ5-UR binding site and within the CBS ( Fig. 2a and 2b, Supplementary Fig. 3a). Remarkably, several of these resonances were already affected by pre-B vesicles interaction but, upon λ5-UR binding, they shifted in the opposite direction towards their initial Gal-1 free resonance position (Fig. 2b). Calculated CSDs for these resonances decreased until complete cancelation at a 1:3 Gal-1:λ5-UR ratio ( Supplementary Fig. 3b). These observations indicate that pre-B cell ligands of Gal-1 are counter-selected in the presence of λ5-UR. Then, upon increasing λ5-UR addition, other CBS resonances located in CBS subsites B and C, and in loops surrounding the CBS started to shift are (Fig. 2b, 2c and Supplementary Fig. 3b).
Consequently, the final binding pattern of Gal-1 to pre-B cell vesicles is different in the presence of λ5-UR indicating binding to new ligands (Fig. 1c vs. Fig. 2b). A control peptide, λ5-UR mutated on two essential residues for complex formation with Gal-1 (λ5-UR-L26A-W30A) 24 , has been used to verify that the effect observed is specific to λ5-UR binding to Gal-1 and not due to non-specific binding of λ5-UR to cell vesicle ligands. No CSDs were observed with the mutated peptide ( Fig. 2a and Supplementary Fig. 3c). Collectively, these experiments demonstrate glycan binding changes for Gal-1 at the pre-B cell surface upon λ5-UR interaction, decreasing affinity for some glycan epitopes and enhancing interaction with others.
Next, we set out to investigate what could be the λ5-UR effect on Gal-1 binding within the pre-B cell niche meaning when pre-B and stromal cell ligands are present. To mimic the pre-B cell niche, we mixed stromal and pre-B cell vesicles ( Fig. 2d and Supplementary Fig. 4).
Chemical shift variations observed in CBS subsites B and C were similar to the ones found when Gal-1 was titrated with λ5-UR in the presence of pre-B cell vesicles. These results indicate that, when pre-B and stromal cell ligands are present, λ5-UR regulates Gal-1 binding to target specific glycoconjugates on pre-B cells.

Dynamic allosteric coupling mediate λ5-UR-induced Gal-1 regulation
How λ5-UR binding on Gal-1 backside is transmitted to the CBS to target specific glycoconjugates is a fundamental question. A hypothesis would be an allosteric coupling throughout Gal-1 from the λ5-UR binding site to the CBS. Allostery involves coupling of ligand binding at one site with a conformational or dynamic change at a distant site, thereby affecting binding at that site. No structural changes within the Gal-1 CBS had been observed upon binding of λ5-UR 24 , hence we investigated the role of dynamics in the regulation of Gal-1 interactions mediated by λ5-UR. The backbone dynamics of Gal-1 were analyzed using nuclear spin relaxation parameters for Gal-1 free and bound to λ5-UR ( Supplementary Fig. 5).
The role of fast protein motions was determined by measuring changes in the order parameter S 2 . S 2 is a measure of the amplitude of internal motions on the picosecond-to-nanosecond timescale and can vary from S 2 = 1, for a bond vector with no internal motion, to S 2 = 0, for a bond vector that is rapidly sampling multiple orientations 26 . λ5-UR binding to Gal-1 caused a large number of residues to decrease their motions as evidenced by the corresponding increase in their S 2 values (Fig. 3a). More specifically, λ5-UR binding to Gal-1 resulted in line broadening and increased rigidity (increased S 2 ) for residues from the λ5-UR binding site but also for residues from the upper loop of the CBS subsite C and within subsite D and B (Fig.   3b). By contrast, increased flexibility (decreased S 2 ) was observed for the lower loop of the CBS subsite C. It should be noted that chemical shift analysis showed that λ5-UR binding to Gal-1 in the presence of pre-B cell vesicles (Fig. 2b) elicited changes in chemical shift for resonances precisely located within the same regions. These findings support the idea of an allostery-based regulation of Gal-1 binding activity triggered by λ5-UR interaction in order to select specific glycan epitopes on pre-B cells.

λ5-UR regulates binding of Gal-1 by selecting α2,3-sialylated glycan motifs found on pre-B cell surface
Our previous glycome exploration of pre-B and stromal cells highlighted a glycomic signature specific to pre-B cells corresponding to biantennary N-glycans, sulfated-and α2,3-sialylated β-galactoside glycans 15 . Importantly, the latter showed increased binding to Gal-1 on glycan arrays in the presence of λ5-UR 15 . In addition, sialylated glycans have been shown to be involved in immunological processes 27,28 . We therefore hypothesized that α2,3 sialylated glycans could be one of the specific glycosidic epitopes targeted by Gal-1 at the pre-B cell surface. To test this hypothesis on live cells, pre-B cells were incubated with the MAL II lectin which specifically recognize α2,3-sialylated glycans and is thus expected to antagonize Gal-1 binding. Remarkably, when incubated with MAL II, cells showed a strong loss in viability ( Fig. 4a) indicating that interaction of MAL II with the broad range of α2,3-sialylated receptors at the pre-B cell surface triggers a signaling pathway leading to cell death (Fig. 4b).
Neither Gal-1 nor λ5-UR reversed the induction of cell death by MAL II, thus illustrating their individual inability to perturb MAL II interactions with α2,3-sialylated receptors ( Fig. 4a and 4c). However, when Gal-1 is combined to λ5-UR, the MAL II-induced cell death was significantly inhibited (Fig. 4a). These observations demonstrate that λ5-UR binding to Gal-1 provokes increased affinity for α2,3-sialylated receptors containing β-galactosides, perturbing MAL II network interactions and precluding cell death signaling (Fig. 4d). Of note, in the absence of MAL-II, neither Gal-1, λ5-UR or a combination of both are able to induce strong pre-B cell death as seen with MAL II (Fig. 4a). Altogether, these results demonstrate that λ5-UR mediates Gal-1 specific targeting of ligands containing α2,3-sialyl moieties at the pre-B cell surface and can promote pre-B cell survival. The structural basis of such ligand binding selection remains to be unraveled.

λ5-UR induces increased Gal-1 contacts with α2,3 sialylated di-LacNAc, converting Gal-1 from exo-to endo-type lectin
To further investigate at atomic resolution the effect of λ5-UR on the Gal-1/α2,3 sialylated glycan interaction, we synthesized using a chemo-enzymatic based methodology the α2,3 sialyl di-LacNAc pentasaccharide called hereafter SdiLN ( Supplementary Fig. 6a). Since galactose moieties are essential to Gal-1 binding, we also synthesized the pentasaccharide fully 13 C-labeled on the galactose moieties ( Supplementary Fig. 6). First, we tested its interaction with Gal-1 using saturation transfer difference (STD) NMR spectroscopy. This technique is based on selective irradiation of protein protons and subsequent detection of magnetization transfer to small ligands 29,30 . In the presence of Gal-1, magnetization transfer was observed indicating intermolecular contacts with the pentasaccharide (Fig. 5a). The STD signal mainly arises from interaction with the terminal galactose (GalA) as expected for an exo-type lectin ( Fig. 5a and Supplementary Fig. 7a). In addition, 1 H, 13 C-HSQC spectra recorded before and after addition of Gal-1 to the 13 C-labeled SdiLN showed severe peak loss for GalA which is therefore the core binding unit to Gal-1 ( Fig. 5b and 5c). These data are consistent with the Gal-1 classification as an exo-type lectin 31,32 . Remarkably, while only galactose moieties were 13 C-labeled, peaks corresponding to resonances from the terminal Nacetyl glucosamine (GlcNAc) and sialic acid (NeuAc) were visible on the HSQC spectrum in the presence of Gal-1 (Fig. 5b). This observation suggests that binding to the protein changes the dynamic of these groups resulting in a higher flexibility which, combined to the HSQC experiment sensitivity, make them visible on the spectrum despite the weak 13 C natural abundance. The reverse experiment, where Gal-1 is 15 N-labeled and the SdiLN is unlabeled has also been performed and showed CSDs and decreased peak intensities mainly for Gal-1 resonances located in CBS subsite C where the core binding galactose GalA of the SdiLN is likely interacting (Fig. 5d and Supplementary Fig. 8). Surprisingly, in the presence of λ5-UR, STD signal contributions became equivalent for GalA and GalB demonstrating that Gal-1 interaction is not exclusive to the terminal galactose when bound to λ5-UR ( Fig. 5e and Supplementary Fig. 7b). As well, 1 H, 13 C-HSQC spectrum of SdiLN showed loss of GalA and GalB resonances upon addition of λ5-UR confirming interaction with both terminal and internal galactose ( Fig. 5f and 5g). Moreover, NeuAc and GlcNAc resonances diseappeared indicating increased contacts with these glycan moieties. On the Gal-1 side, addition of λ5-UR to the Gal-1/SdiLN complex not only intensified the initial perturbations but also propagated the effect towards CBS subsites A to E illustrating increased intermolecular contacts throughout the CBS with the pentasaccharide (Fig. 5h and Supplementary Fig. 8).
λ5-UR alone (without the SdiLN) is not able to prompt such Gal-1 signal intensity decrease which is consistent with the binding of a 24 amino acid peptide in fast exchange regime 13 .
Thus, λ5-UR interaction modifies Gal-1 binding to α2,3 sialylated glycan by allowing additional contacts with the pentasaccharide including with the internal galactose unit. While recent studies have elegantly rubber-stamped the selectivity of Gal-1 towards terminal galactose 21 , here we establish that λ5-UR binding to Gal-1 induces a selectivity switch converting Gal-1 from an exclusive exo-type lectin to an endo-type lectin interacting with both terminal and internal galactose units (Fig. 5i). This switch is therefore at the basis of Gal-1 target specificity at the pre-B cell surface.

Discussion
Over the past decades, many cellular activities have been ascribed to Galectins 4 . Such a wide range of functions in so many cell types and environments have conducted many laboratories to develop methodologies to understand the varied and intricate roles played by these ubiquitous lectins. While many of these experiments developed provide evidence and enable visualization for Galectin-glycoprotein lattices on cell surfaces 33,34 , further studies were needed to understand the Galectins structural features regulating lattice assembly/disassembly at the cellular level. Here, we directly examined for the first time the structural properties of Gal-1 binding to native membranes by combining solution-state NMR spectroscopy to the preparation of pre-B and stromal membrane vesicles. Importantly, we could demonstrate that Gal-1 binding to its physiological cellular ligands involves long-range effects throughout the CRD beyond the canonical CBS limits. While multivalency of Gal-1 and its glycoconjugate ligands is essential for the formation of supramolecular lattices 35 , the additional interaction surface revealed for extended complex sugar ( Fig. 1c and 1d) as well as the conformational exchange observed (Supplementary Fig. 2b and 2c) for residues at the dimer interface could be a key mechanism allowing increased overall stability and biological functionality of the assembled lattice on the cell surface.
Remarkably, our on-cell NMR approach allowed us to address how Gal-1 is able to decode the glycome of pre-B cells. Indeed, given the abundance of β-galactosides on the cell in the form of glycoproteins and surface glycolipids, one would expect that Galectin/glycoconjugates lattice translates into a large heterogeneous complex on the surface of cells. But conversely, this lectin is able to cross-link a single species of glycoproteins to form a uniform lattice 34,36,37 . Here, we demonstrate that Gal-1 is able to achieve target specificity at the pre-B cell surface through a direct interaction with a non-glycosylated protein domain of the pre-BCR, λ5-UR. This interaction allows Gal-1 to interact with α2,3 sialylated glycoconjugates. Sialic acids are ubiquitously and abundantly found on the surface of all human cells as the terminating sugar in glycolipids (gangliosides) and glycoproteins (complex N-glycans and mucin type O-linked glycans). Moreover, sialic acids have been shown to play essential roles in immunological processes and in particular during B cell receptor (BCR) signaling 28,35,[38][39][40] . In this context, the α2,6 sialylated CD22 receptor, which is a lectin recognizing α2,6 sialylated glycoreceptors, is recruited close to the BCR and inhibits its downstream signaling (Fig. 6a). The objective of this control being to prevent inadvertent activation to weak signals that could be considered a form of self-recognition and/or unwanted B cell responses under the appropriate circumstances 40 . Our results suggest that Gal-1 bound to λ5-UR would help recruitment of α2,3 sialylated receptors at the pre-B synapse. In addition, although our experimental setup using MAL II lectin is not a physiological condition, we showed that clustering of a broad range of α2,3 sialylated receptors lead to pre-B cell death and that Gal-1/ λ5-UR complex is able to disrupt the clustered sialylated receptors thus promoting pre-B cell survival (Fig. 3). Similarly to the CD22 dependent regulation of BCR signaling, it is tempting to propose that MAL II in our experiments would mimic a physiological receptor which would trigger specific cellular responses (Fig. 6b). The presence of Gal-1 bound to the pre-BCR would disrupt this clustering by recruiting specific α2,3 sialylated receptors at the pre-B synapse to activate proper signaling (Fig. 6b). Indeed, restricted receptor segregation into membrane microdomains would result in a homogenous platform suitable for efficient pre-BCR signaling leading to proliferation and survival. In line with this model of pre-BCR activation, it has been previously observed that Gal-1 binding resulted in the formation of large highly immobile pre-BCR aggregates at the pre-B cell surface 41 . Moreover, the clustering of specific glycoproteins and/or glycolipids could generate mechanical stress which translates into membrane curvature necessary for pre-BCR internalization following cell signaling. This model is strongly supported by previous evidence showing that another Galectin, the chimeric Galectin-3 (Gal-3), is able to mediate endocytic invaginations by helping the clustering of cargo proteins and glycosphingolipids 39, 40 . Overall, our study reveals that the Gal-1/pre-BCR interaction is part of the glycome decoding mechanism on the surface of pre-B cells and would aim to re-orchestrate receptors localization to ensure cell signaling regulation and proper B cell development.
At the structural level, upon λ5-UR interaction, dynamic fluctuations throughout Gal-1 occurs with an overall increased rigidity which locks Gal-1 into a conformation prone to interact with specific glycans and exclude others. This allostery-based mechanism of regulation for Gal-1/glycan interactions leads to specific selection of α2,3 sialylated glycoconjugates at the pre-B cell surface. Allostery based phenomena have been previously shown for Gal-1 binding to synthetic inhibitor molecules designed to target Gal-1 interactions [42][43][44] . In addition, studies on glycan recognition by Gal-1 showed that sugar binding affected residues far from the CBS and described significant changes in the dynamics of the protein thus evidencing Gal-1 conformational plasticity and allostery upon glycan interaction 45 . These data, together with our study, emphasize the existence of intramolecular communication pathways encoded within Gal-1 structure which link glycan binding to dimer interface but also to protein interaction. Rather than a simple modification of ligand targeting, these dynamic changes result in Gal-1 conversion from exo-to endo-type lectin. To the best of our knowledge, this result is the first example of such conversion dictating the glycan preference of a Galectin.
Gal-1 bound to λ5-UR is able to interact with terminal and internal galactose like the tandemrepeat Gal-8 21 . Tandem-repeat Galectins (Galectins with two CRDs linked by a flexible linker) are known to be more potent in triggering cellular responses 46 , therefore λ5-UR induced conversion might increase Gal-1 efficiency at the pre-B cell surface through the formation of higher-order multimers as seen for tandem-repeat Galectins 46 .
Beyond the pre-BCR case, our results raise a fundamental question: is there a universal mechanism of allosteric modulation of glycan binding that is conserved across all Galectins through binding to non-glycosylated protein partners? Indeed, there are increasing evidence for direct protein interactions with Galectins [16][17][18]47 . We also demonstrated that CXCL4 interacts with Gal-1 and control its glycan binding activities at the base of the immunoregulatory function of galectins 47 . The pre-BCR case and these latter examples all involve the three different Galectin subfamilies (prototype, chimera and tandem-repeat) 4 .
Further structural, cellular and glycomic studies of these new complexes are needed to demonstrate a universal allosteric mechanism of regulation for Galectins functions. Moreover, taking into account Galectins involvement in many crucial pathologies such as cancer 6,48 and inflammation 49 or infection by pathogens 7 , the discovery of this selectivity switch should also increase the potential to develop molecules modulating Gal-1 interactions and function in specific pathological situations.