Structural and Biophysical Characterization of C‐Glycosylic 1,2‐Thiodisaccharides Reveals Determinants of Selective Binding to Galectin‐7 and Galectin‐8N Authors: Anastasia S Tsagkarakou, Anastassia L Kantsadi, Vasiliki I Theodoridou, Nikolaos Veliotis, László Lázár, János József, László Juhász, George Kontopidis, Hakon Leffler, Ulf J Nilsson, László Somsák, Demetres D Leonidas Published: 2026 ABSTRACT AbstractAchieving isoform selectivity within the galectin family remains a central challenge in glycomimetic drug design due to the highly conserved architecture of their carbohydrate recognition domains. Here, we define the structural and thermodynamic basis of recognition of a series of C‐glycosylic 1,2‐thiodisaccharides targeting human galectin‐7 and the N‐terminal domain of galectin‐8 (galectin‐8N). Using an integrated approach combining fluorescence polarization, isothermal titration calorimetry, and high‐resolution X‐Ray crystallography, we establish a clear structure–activity relationship across the ligand series. Compound 17 emerges as the most potent galectin‐8N ligand (K d = 13 μM), outperforming thiodigalactoside, while compound 13 shows preferential binding to galectin‐7, demonstrating tunable isoform bias. Structural analysis reveals a conserved anchoring mechanism in which the β‐galactoside unit (Gly‐1) drives affinity through a rigid hydrogen‐bonding and π‐stacking network, whereas the second sugar (Gly‐2) modulates potency by adopting distinct orientations in galectin‐specific extended binding sites. Notably, ligand binding converges on conserved motifs while leaving nonconserved regions unexploited, highlighting clear opportunities for structure‐guided optimization. Collectively, this work establishes C‐glycosylic thiodisaccharides as a robust platform for selective galectin targeting and provides actionable design principles for next‐generation inhibitors. Introduction Galectins are a conserved family of β‐galactoside binding lectins that regulate key cellular processes, including cell adhesion, apoptosis, immune modulation, and tumor progression. Fifteen human galectins have been identified and are grouped into prototype, tandem‐repeat, and chimera types according to their carbohydrate recognition domain (CRD) architecture [1, 2]. Through their ability to engage specific glycan motifs on cell surface and extracellular matrix glycoproteins, galectins play central roles in cancer biology, inflammation, fibrosis, and immune regulation, making them attractive targets for therapeutic intervention [1, 2]. Recent studies emphasize the expanding network of galectin binding partners and signaling roles across immunity and cancer, reinforcing their value as tractable targets for chemical biology and drug discovery [3, 4]. Galectin‐7 is a prototype galectin expressed predominantly in stratified epithelia, where it contributes to epithelial homeostasis, differentiation, and stress responses. Increasing evidence links its dysregulation to tumor progression: Galectin‐7 overexpression correlates with poor prognosis in several epithelial cancers and has been implicated in metastasis, immune evasion, and apoptosis pathways [5, 9]. Furthermore, recent in vivo and ex vivo data show that galectin‐7 can reduce CD4+ (cluster of differentiation 4) T‐cell frequency via PD‐1 (programmed cell death protein 1) glycan‐dependent engagement, revealing direct cross talk with an immune checkpoint axis in cancer contexts [10]. Its emerging significance in oncology has stimulated efforts to develop selective inhibitors and molecular probes, including nanobodies and small‐molecule scaffolds that can modulate its biological activity [6, 11, 12]. Galectin‐8, a tandem‐repeat galectin, contains two CRDs connected by a flexible linker and participates in diverse processes such as cell adhesion, angiogenesis, autophagy initiation, and immune modulation [13, 17]. Notably, the N‐terminal CRD (galectin‐8N) displays higher affinity and distinct fine specificity toward sulfated and sialylated glycans compared to other galectins. Detailed structural, NMR, and biophysical studies have highlighted important differences between the N‐ and C‐terminal CRDs of galectin‐8 and established the molecular basis for their distinct glycan recognition profiles [18, 22]. Aberrant galectin‐8 activity has been associated with metastatic dissemination, chronic inflammation, and graft rejection [23, 24]. A recent comprehensive review has further highlighted the diverse functions of galectin‐8 in immune regulation and tumor biology, as well as current strategies for galectin‐8 inhibition and therapeutic targeting [25]. Structure‐guided design has yielded potent and selective monovalent ligands against galectin‐8N, underscoring the potential of targeting this domain for therapeutic applications [14, 15, 18]. Two recent structure‐guided studies demonstrated selective galectin‐8N ligands that exploit 8N‐specific pockets. First, 4‐halophenylphthalazinone–galactal hybrids achieve μM binding and selectivity for 8N through double π‐stacking with R45, R59, and Y141 observed crystallographically [26]. Second, nanomolar 8N inhibitors were shown to bind via a noncanonical cation–π interaction, expanding the noncovalent design space for 8N selectivity [27]. Clinical interest in galectin inhibition has intensified, driven by ongoing trials of galectin antagonists in cancer, idiopathic pulmonary fibrosis, and nonalcoholic steatohepatitis [28, 30]. Current strategies encompass small molecules, glycomimetics, peptides, multivalent dendrimers, and antibody‐based approaches [28, 30, 32]. Recent immuno‐oncology overviews highlight combinations of galectin blockade with checkpoint inhibitors, chimeric antigen receptor T‐cell therapies, and vaccines, underscoring the need for isoform‐selective tools to avoid broad lectin effects [4]. Against this backdrop, achieving selectivity at galectin‐7 or ‐8N and other CRDs remains challenging, but now demonstrably tractable in case of galectin‐8N, via π‐stacking/cation–π strategies in the extended 8N pocket [26, 27]. Yet, achieving galectin selectivity remains a major challenge: The conserved architecture of the CRD limits the ability of conventional galactosides or thiogalactosides to discriminate among galectin family members. Consequently, identifying chemical modifications capable of exploiting subtle structural differences between related CRDs is essential for the development of next‐generation inhibitors. C‐Glycosylic 1,2‐thiodisaccharides represent a promising class of synthetic ligands that incorporate an extended sulfur linkage and increased rigidity, features previously shown to enhance binding to galectin‐3. For galectin‐3, C‐glycosylic 1,2‐thiodisaccharides leverage enthalpy‐driven affinity via water‐mediated hydrogen bonds with crystallographic support [33], motivating extension of this scaffold to other galectins and in particular galectin‐7 and galectin‐8N. Given the therapeutic appeal of these two galectins and the limited number of high‐affinity, selective ligands available, further exploration of this scaffold would be of interest. In this study, we investigate a series of C‐glycosylic 1,2‐thiodisaccharide derivatives and characterize their binding properties toward human galectin‐7 and the N‐terminal CRD of galectin‐8. Using fluorescence polarization (FP) assays, we assess their inhibitory potency, and through high‐resolution X‐ray crystallography, we elucidate the structural basis of ligand recognition. Comparative analysis across galectin‐7, galectin‐8N, and previously reported galectin‐3 complexes highlights conserved interactions in the primary β‐galactoside binding site, as well as nonconserved residues such as Q47, R59, R72, and I91 in galectin‐8N that may be exploited to achieve selectivity. Our structural analysis complements recent 8N‐selective ligand reports by mapping how Gly‐2 orientations in our series can engage Q47, R59, R72, and I91 through π‐stacking or cation–π enabling geometries, offering convergent design principles for 8N selectivity [26, 27]. Together, these findings provide a structure–activity framework for the rational optimization of thiodisaccharide scaffolds and advance efforts to develop potent, isoform‐selective inhibitors of galectins‐7 and ‐8N. Materials and Methods Chemical Synthesis The syntheses of the studied C‐glycosylic 1,2‐thiodisaccharide derivatives 13, 14, 15, and 17, and the non‐C‐glycosylic control molecule, were described in our earlier articles [33, 34], whereas TDG was purchased from Sigma‐Aldrich. Production and Purification of Recombinant Proteins Recombinant human galectin‐7 and the N‐terminal CRD of galectin‐8 (galectin‐8N) were produced in Escherichia coli BL21(DE3) cells using standard heterologous expression procedures. Chemically competent E. coli BL21(DE3) cells were transformed with 100 ng of the galectin‐7 pet22 expression vector, heat‐shocked at 42 °C for 45 s, and allowed to recover for 1 h at 37 °C in LB medium with shaking (210 rpm). Transformed cells were selected on LB agar plates containing 34 μg/mL ampicillin. A single colony was used to inoculate LB medium, and protein expression was induced at an OD600 of 0.6–0.7 by adding 0.5 mM IPTG, followed by overnight incubation at 18 °C with shaking (210 rpm). Cells were harvested by centrifugation (5000 g, 30 min, 4 °C), resuspended in Buffer A (MEPBS pH 7.4) and treated with DNase I (10 μg/mL). After clarification by centrifugation (18,000 g, 1 h, 4 °C), the supernatant was applied to a Lactosyl Sepharose 4‐CLB affinity column. Bound protein was eluted with 150 mM lactose. Pooled fractions were dialyzed overnight against 50 mM Tris–HCl (pH 8.0), 0.3 M NaCl, to remove lactose, and further purified by size exclusion chromatography on a Superdex 75 column. Protein purity (>95%) was confirmed by SDS‐PAGE. Galectin‐7 was concentrated to 20 mg/mL using a Centricon‐10 (Amicon) filters (MWCO 10 kDa) and stored at −80 °C. Typical yields were ~3 mg of purified protein per liter of culture. Galectin‐8N CRD was produced using an analogous protocol. Competent E. coli BL21(DE3) cells were transformed with 100 ng of the galectin‐8N expression plasmid (for residues 1–155; form 8 NM) and selected on LB agar plates containing 35 μg/mL kanamycin. Expression cultures were induced at OD600 of 0.6–07 with 1 mM IPTG and incubated for 3 h at 37 °C with shaking (210 rpm). Cells were collected by centrifugation (5000 g, 30 min, 4 °C), resuspended in Buffer A (MEPBS pH 7.4), and treated with DNAase I (10 μg/mL). After clarification (18,000 g, 1 h, 4 °C), the supernatant was loaded onto a Lactosyl Sepharose 4‐CLB affinity column, and galectin‐8N was eluted with 150 mM Lactose. Fractions were dialyzed overnight against PBS (pH 7.4) and concentrated to 20 mg/mL using a Centricon‐10 device. SDS‐PAGE analysis confirmed >95% purity. Typical yields were ~5 mg per liter of cell culture. Both recombinant proteins were flash‐frozen and stored at −80 °C. Crystallization, X‐Ray Data Collection, Structure Determination, and Refinement Crystals of human galectin‐7 crystals were obtained by the sitting‐drop vapor‐diffusion method. Protein at 10 mg/mL in 0.1 M Tris–HCl (pH 8.0), 0.1 M NaCl, was mixed in a 1:1 ratio with reservoir solution (0.1 M Tris–HCl (pH 8.0), 0.1 M NaCl, 15% (w/v) PEG 4000, 15% (v/v) glycerol) and incubated at 16 °C. Protein–ligand complexes with compounds 13, 14, 15, and TDG were prepared by two approaches: (a) Preformed galectin‐7 crystals were soaked for 48 h at 16 °C in crystallization solution supplemented with 2 mM ligand and (b) galectin‐7 (10 mg/mL) in 50 mM Tris–HCl (pH 8.0), 0.3 M NaCl, was mixed 1:1 with reservoir solution (0.1 M Tris–HCl, pH 8.0, 0.1 M NaCl, 15% (w/v) PEG 4000, 15% (v/v) glycerol) containing 5 mM ligand. Drops (2 µL) were equilibrated against 600 µL reservoir solution at 16 °C. Crystals appeared after ~3 days. Among the tested ligands, 13 was the only one observed bound in the final galectin‐7 structures. Galectin‐8N CRD crystals were grown using the sitting‐drop vapor‐diffusion method. Protein at 10 mg/mL in PBS (pH 7.4) was mixed 1:1 with reservoir solution containing 0.1 M Tris–sodium citrate (pH 5.6), 20% (v/v) 2‐propanol, and 20% (w/v) PEG 4000. Cocrystals with compounds 13, 14, 15, and TDG were obtained by incubating drops containing 1 mM ligand at 16 °C, with crystals forming within 24 h. A second crystal form of galectin‐8N was produced using SwissSci 96‐well plates and the Oryx Nano robot (Douglas Instruments). Drops of 300 nL were prepared in a 1:1 protein‐to‐reservoir ratio (protein at 5 mg/mL) with reservoir solutions containing 0.1 M Buffer System 1 (imidazole/MES, pH 6.5), 0.12 M mixed alcohols (1,6‐hexanediol, 1‐butanol, 1,2‐propanediol, 2‐propanol, 1,4‐butanediol, 1,3‐propanediol), and 22%–24% Precipitant Mix 4 (MPD/PEG1000/PEG3350). These crystals were subsequently used to generate the galectin‐8N–17 complex by soaking them for 48 h with 10 mM ligand. For data collection, crystals were cryoprotected by supplementing their respective crystallization solution with 30% (v/v) glycerol and then flash‐cooled in a nitrogen gas stream at 100 K. X‐Ray diffraction data were collected using synchrotron radiation at beamline P13 of the EMBL Hamburg outstation. Diffraction images were processed using the program XDS [35] for indexing, integration, and scaling. Intensities were merged using Aimless [36]. Structures were solved by molecular replacement using either human galectin‐7 (PDB 1BKZ) [5] for galectin‐7 complexes, or galectin‐8N CRD bound to lactose (PDB ID:5T7S) [16] for galectin‐8N complexes. Crystallographic refinement was performed by maximum‐likelihood methods in REFMAC [37] and a final round with PHENIX [38]. Ligand geometries and restrained dictionaries were generated using JLigand [39] and ligands were fitted into electron density maps by adjustment of torsion angle, followed by iterative refinement. Model validation was carried out through the PDB REDO server [40]. Hydrogen bonds and van der Waals interactions were analyzed using CONTACT from the CCP4 suite [41] with cutoffs 3.4 and 4.0 Å, respectively. Solvent accessible area calculations were performed with Naccess [42] and figures were prepared in PyMOL [43]. Crystallographic statistics for data collection and refinement statistics are shown in Table 1, and coordinates have been deposited in the RCSB Protein Data Bank (http://www.rcsb.org/pdb) under accession codes listed therein. TABLE 1Data collection and processing statistics for galectin‐7 and ‐8N CRD complexes. Values in parentheses are the outermost shell.InhibitorGalectin‐7 Galectin‐813, soak13, cocrystal13TDG141517Space group P4 1212 P212121 P212121 C 121 P212121 Unit cell dimensions, a, b, c, (Å), β (°)77.0, 77.0, 100.277.0, 77.0, 100.253.7, 61.8, 85.159.0, 39.2, 55.7, 107.561.4, 40.2, 69.7 115.153.8, 61.9, 84.754.1, 61.8, 85.0Resolution (Å)1.761.801.51.711.781.901.90Outer shell1.79–1.761.84–1.801.53–1.501.74–1.711.82–1.781.94–1.901.95–1.90Reflections measured343,068 (19766)320,446 (19459)197,255 (2240)84,703 (3722)100,536 (5662)128,727 (8233)110,825 (4455)Unique reflections30,613 (1700)55,975 (3285)42,942 (1230)12,509 (573)14,804 (832)22,926 (1444)21,211 (1076) R merge 0.136 (1.014)0.117 (0.717)0.164 (0.599)0.214 (1.306)0.077 (0.803)0.069 (1.100)0.037 (0.318)Completeness (%)100 (100)100 (100)93.3 (55.8)94.2 (84.1)99.0 (99.0)99.9 (100)92.0 (74.0) I/σ(I)10.9 (2.3)8.5 (2.0)5.9 (0.9)6.0 (2.2)11.5 (2.3)11.6 (1.6)21.4 (3.7)Multiplicity11.2 (11.6)5.7 (5.9)4.6 (1.8)6.8 (6.5)6.8 (6.8)5.6 (5.7)5.2 (4.1)CC1/2 0.998 (0.834)0.996 (0.784)0.977 (0.517)0.973 (0.503)0.998 (0.797)0.998 (0.542)0.999 (0.900)B Wilson (Å2)23.828.923.142.140.143.231.5Reflections used for refinement30,54555,55242,88812,48314,80022,87021,172No of water molecules28658617266808893No of ligand atoms521045223255044 R (%)16.8 (20.1)16.8 (24.0)19.2 (26.8)24.1 (30.0)18.5 (22.7)19.6 (26.4)18.8 (23.6) R free (%)21.1 (25.5)21.4 (31.7)22.4 (32.7)29.3 (32.1)22.1 (28.4)24.5 (36.7)24.9 (28.9)r.m.s.d. inBond lengths (Å)0.0100.0180.0060.0040.0050.0050.006Bond angles (o)1.41.70.90.70.80.80.8Average B (Å2)Protein atoms20.8/23.920.0/23.4/20.1/23.435.0/34.240.239.942.6/41.938.5/39.3Water molecules38.436.540.940.848.842.740.7Ligand atoms17.8/16.015.7/15.1/15.4/15.237.1/42.437.949.244.8/48.741.0/48.3PDB entry31LN31OK31ML31NV31NP31NW31OF Data collection and processing statistics for galectin‐7 and ‐8N CRD complexes. Values in parentheses are the outermost shell. Inhibitory Potency Measurements FP measurements were performed on a PHERAstar FS plate reader using PHERAstar Mars v2.10 R3 software (BMG LABTECH, Offenburg, Germany). Competitive displacement assays were carried out using a fluorescein‐labeled galactoside probe and purified recombinant galectin‐7 or galectin‐8N. Assays were conducted at room temperature in MEPBS buffer (pH 7.4). Polarization values were recorded at each inhibitor's concentration. Gal‐8N was tested at 0.4 μM with 100 nM of Lacto‐N‐neotetraose probe, and Gal‐7 was tested at 5 μM with 20 nM of another Lacto‐N‐neotetraose probe with slightly different linker between glycan and the fluorescein moiety [44, 45], and data were treated as described previously [33]. All FP‐derived parameters for compounds are shown in Table 2. Isothermal titration calorimetry (ITC) experiments were conducted at 25 °C using a Nano‐ITC (TA Instruments) with a cell volume of 1.038 mL and a 250‐µL syringe under continuous stirring at 170 rpm. The sample cell contained galectin‐7 or galectin‐8N, while ligands were prepared in the same buffer as the proteins and loaded into the syringe. Each titration consisted of 16 injections of 15 µL, spaced 150 s apart to allow complete baseline recovery between injections. More specific experimental details are in the legends of Figures S5–S9. Due to sample limitations, each ITC experiment consisted of a single titration with 16 injections. The data were analyzed individually using a single‐site binding model (Figures S5–S9; Table S5). No replicate titrations or global fitting procedures were applied. Heats of dilution were subtracted using the instrument's software. Binding isotherms were fitted using NanoAnalyze (TA Instruments) with a single‐site binding model as previously described [33]. TABLE 2The K d (μM) values of antagonists for galectins‐7, ‐8N, and ‐3. All K d values were determined by FP and in parentheses are those determined by ITC.Ligand nameChemical structure and crystallographic numberingGalectin‐7 Galectin‐8NGalectin‐3TDG 160 ± 18 [46] (101 ± 7) 615 ± [46] (75 ± 11) 49 ± 11 [46] (44 ± 5 [33]) General denotation of the two glycosyl units in 13 –17. 13 30 ± 337 ± 11 16 ± 1 (18 ± 4) [33] 14 88 ± 1337 ± 3 8 ± 1 (9 ± 1) [33] 15 138 ± 16 (26 ± 9) 28 ± 2 13.0 ± 1.0 (14 ± 2) [33] 16 (>500)2800 ± 4703476 ± 3 17 310 ± 37 (132 ± 28) 147 ± 24 (13 ± 1) 35 ± 4 (21 ± 3) [33] The K d (μM) values of antagonists for galectins‐7, ‐8N, and ‐3. All K d values were determined by FP and in parentheses are those determined by ITC. 615 ± [46] (75 ± 11) 49 ± 11 [46] (44 ± 5 [33]) General denotation of the two glycosyl units in 13 –17. 16 ± 1 (18 ± 4) [33] 8 ± 1 (9 ± 1) [33] 138 ± 16 (26 ± 9) 13.0 ± 1.0 (14 ± 2) [33] 147 ± 24 (13 ± 1) 35 ± 4 (21 ± 3) [33] Results and Discussion Binding Affinity Analysis Binding affinities (Table 2) were quantified using FP and ITC as described previously for related thiodisaccharides [33]. FP measurements provided a comparative overview of inhibitory potency across the ligand series, whereas ITC experiments yielded direct thermodynamic parameters for selected compounds. In the analysis below, FP‐derived values are used for systematic comparison across all ligands, while ITC values are included where available to validate binding trends and highlight thermodynamic behavior. For galectin‐8N, FP measurements indicate a gradual improvement in affinity across the series, with 15 (28 µM) and 14 (37 µM) showing enhanced potency compared to TDG (61 µM), while 13 displays a similar affinity (37 µM). Compound 17 appears weaker by FP (147 µM); however, ITC measurements identify 17 as the most potent ligand in the series, with a K d of 13 µM, followed by TDG (75 µM). In contrast, compound 16 exhibits very weak binding (3 mM by FP), and 13 showed no detectable binding under the ITC conditions employed. Taken together, these results establish a consistent potency trend at galectin‐8N in ITC (17 > TDG), while FP highlights more modest differences across 13–15, reflecting method‐dependent sensitivity. For galectin‐7, FP data show that 13 is the most potent ligand (K d = 30 µM), followed by 14 (88 µM), 15 (138 µM), and 17 (310 µM), whereas 16 does not inhibit up to 0.5 mM. ITC measurements, available for selected compounds, reveal a stronger binding of 15 (K d = 28 µM) and 17 (K d = 107 µM) compared to their FP‐derived values, indicating some quantitative differences between the two methods while preserving the overall ranking of potencies. Collectively, these data highlight moderate affinities for galectin‐7 and greater variability across the ligand series compared to galectin‐8N. Compounds 14 and 15, which differ only in Gly‐2 C4 configuration, show a modest but reproducible difference in affinity for galectin‐8N, with FP‐derived K d values of 37 and 28 µM, respectively, consistent with the ITC value obtained for 15 (28 µM). These differences correlate with distinct hydrogen‐bonding patterns observed in the crystal structures, where 15 engages the binding site through its amino substituent, while 14 primarily interacts via hydroxyl groups with the conserved R69–E89 pair. The relatively small difference in affinity between these epimers indicates that Gly‐2 stereochemistry fine‐tunes binding rather than acting as the dominant determinant of interaction strength. In contrast, compound 16, which alters the C4 configuration of the Gly‐1 unit, displays very weak or no binding, demonstrating that modifications at Gly‐1 are detrimental to interaction with the conserved primary binding site. Collectively, these results show that the Gly‐1 unit serves as the primary determinant of binding across the series, consistent with its extensive network of conserved hydrogen‐bond and π‐stacking interactions within the CRD. In contrast, the Gly‐2 moiety modulates affinity through additional contacts that vary between galectins. These findings highlight the functional importance of the C‐glycosylic linkage at Gly‐2 and support the potential for further optimization of this scaffold to enhance galectin selectivity. Structural Analysis Human galectin‐7 consists of 136 residues forming the characteristic galectin β‐sandwich fold composed of two antiparallel β‐sheets of six (S1−S6) and five (F1‐F5) strands. The concave face of strands S1−S6 constitutes the canonical carbohydrate‐binding groove, which is subdivided into subsites A−E (Figure 1) [47, 48]. Subsite C, formed by residues H49, N51, R53, N62, and W69, is highly conserved among galectins and provides the primary recognition site for β‐galactosides. Subsites A, B, D, and E accommodate adjacent sugars and being more variable contribute to different specificities among galectins [48]. The concave surface of galectin‐7 CRD. Subsites A–E are labeled. Crystal structures of the galectin‐7–13 complexes were determined at high resolution (Table 1). Compound 13 binds within the canonical carbohydrate‐binding groove along strands S4–S6, without inducing significant conformational changes in the protein structure. Although galectin‐7 crystallizes as a dimer, the presence of 13 promotes the formation of a tetrameric assembly in the crystal lattice (Table 1; Figure S1). The native dimer interface is formed by interactions between β‐strands F1–F5, consistent with previous structural reports [5]. In both dimeric and tetrameric forms, 13 is bound in all available CRD sites (Figures 2 and S1). The β‐galactoside Gly‐1 moiety of 13 occupies subsites C and D in a conserved orientation across all monomers. It engages an extensive hydrogen‐bonding network typical of galectin–galactose interactions: O4 interacts with H49, N51, and R53; O6 interacts with R53 and E72; and O6′ forms hydrogen‐bond interactions with N62 and the main chain of G72. W69 provides the expected π‐stacking interaction with the Gly‐1 ring (Table S1). In contrast, the orientation of the Gly‐2 moiety varies among monomers, revealing multiple discrete and well‐defined binding modes within the crystal lattice (Figure 2). The observation of distinct conformations with interpretable electron density indicates that these states are sufficiently stable rather than arising from unresolved dynamic disorder. At the same time, the presence of multiple binding modes suggests an inherent plasticity of the extended binding surface, although the exact distribution of conformations is likely influenced, at least in part, by crystal packing interactions. Consistent with this interpretation, the refined temperature factors of the ligand are comparable to those of surrounding residues (Table 1), supporting well‐defined positioning rather than highly dynamic disorder. Depending on the monomer, Gly‐2 forms additional hydrogen bonds with R53, G70, E72, and R74 (Table S1 and Figures 2, S1) as well as numerous van der Waals contacts. Overall, compound 13 establishes extensive van der Waals contacts within the CRD (Table S3), consistent with a stable and well‐defined ligand placement. However, the occurrence of multiple resolved binding modes suggests that these conformations are energetically accessible within the binding site and may be differentially stabilized by crystal packing interactions, rather than reflecting dynamic averaging. Binding of 13 to galectin‐7 CRD dimer molecules A (left) and B (right) (PDB IDs 31LN and 31OK). Hydrogen‐bond interactions are represented as dashed lines and water molecules as blue spheres. Galectin‐8N CRD adopts the canonical galectin β‐sandwich fold with the carbohydrate‐binding site positioned along strands S4–S6. The primary recognition site consists of residues R45, H65, N67, R69, N79, W86, and E89, most of which are conserved across the galectin family (Figure 3). Notably, galectin‐8N also possesses several nonconserved residues at the extended binding interface, including Q47 (S3), R59 (S3‐S4 loop), R72 (S2), and I91 (S6). These residues contribute to selective recognition of sulfated and sialylated glycans and represent key determinants of galectin selectivity that have been successfully targeted by structure‐guided design of synthetic inhibitors [19, 21, 47]. Binding of 13 (A), 14 (B), 15 (C), 17 (D), and TDG (E) to galectin‐8N CRD (PDB IDs: 31ML, 31NP, 31NW, 31OF, and 31NV, respectively). Hydrogen‐bond interactions are represented as dashed lines and water molecules as blue spheres. High‐resolution structures were obtained for galectin‐8N in complex with 13, 14, 15, 17, and TDG (Table 1). All ligands bind at the canonical site along S4–S6 without inducing major conformational changes in the protein. As observed in galectin‐7, the Gly‐1 β‐galactoside ring occupies the primary binding pocket in an essentially identical manner across all complexes. Gly‐1 forms conserved hydrogen bonds with R45, H65, N67, R69, N79, and E89, and engages in π‐stacking against W86 (Table S2). The Gly‐2 moieties of 13, 14, 15, 17, and TDG adopt distinct well‐defined orientations with clear electron density and predominantly contact the conserved R69–E89 pair via different donors/acceptors. For example, 13 forms four hydrogen bonds via O7 and O8; 14 and 17 use O3′ to engage R69 and E89; 15 forms hydrogen bonds via its amino group; and TDG uses O2′ for similar interactions. Across our galectin‐8N complexes, direct productive contacts with nonconserved residues (Q47, R59, R72, and I91) were not observed, so we view these positions as design opportunities rather than demonstrated contributors to the present potency order (17 > 15 > 14 > TDG). The total number of hydrogen bonds per ligand ranges from 12 to 15 (Table S2). These subtle differences in Gly‐2 positioning are visually evident in structural overlays (Figures 3 and S4). A clear SAR emerges from these data. Gly‐1—the β‐galactoside moiety, acts as the primary anchoring element across all galectins examined. Every active ligand preserves a strict Gly‐1 geometry that maintains the conserved hydrogen‐bond network and π‐stacking interaction with residues such as H65/H49, N67/N51, R69/R53, N79/N62, and W86/W69 (galectin‐8N/‐7 numbering). Correspondingly, alteration of Gly‐1 C4 configuration in compound 16 abolishes binding to both galectin‐7 and galectin‐8N, highlighting that Gly‐1 is responsible for baseline affinity and is largely nonnegotiable from a design perspective. Our results agree with what has been discussed in detail previously [47], and how additional moieties on either side of the Gal can enhance affinity very much even with relatively small interactions. In contrast, Gly‐2 serves as the selectivity‐modulating region. Its conformation and substituent orientation vary significantly between galectins, and these changes correlate with the potency differences observed. At galectin‐8N, 14 and 15 differ only in Gly‐2 C4 configuration and exhibit a small affinity difference (37 vs. 28 µM). This correlates with slightly different hydrogen‐bonding patterns where 15 engages the R69/E89 region via its amido group in a water‐mediated interaction, while 14 interacts through hydroxyl groups, including water‐mediated contacts to E89. Given the modest affinity difference, these interactions likely reflect subtle variations in binding geometry rather than a dominant energetic contribution. The structural basis for these differences is evident in the crystal structures of the galectin‐8N complexes, which show that all ligands converge on the conserved R69–E89 “clamp,” while none directly contact the nonconserved residues Q47, R59, R72, or I91. This explains why the series achieves measurable affinity but only moderate degrees of galectin selectivity. For galectin‐7, a different picture emerges. Although 13 binds moderately in solution (K d = 30 µM), it occupies all CRD sites in the tetrameric crystal lattice, adopting multiple Gly‐2 orientations (A–D monomers) as shown in Figures S1–S3. The observation of multiple, well‐resolved poses in different monomers indicates that the galectin‐7 extended site can accommodate alternative binding modes. While these conformations are stabilized within the crystal lattice, their coexistence suggests a permissive and structurally adaptable binding surface capable of accommodating multiple ligand orientations and van der Waals contact patterns (Table S3), without necessarily implying a defined energetic advantage. Correspondingly, affinity differences among 14, 15, and 17 at galectin‐7 arise predominantly from how well each Gly‐2 configuration complements this more permissive surface. Compound 15 (27.6 µM) achieves the best fit, followed by 14 (88 µM) and 17 (107 µM). Taken together, these analyses reveal a SAR governed by a dual mechanism, Gly‐1 controls absolute affinity via deep anchoring into the conserved CRD pocket, and Gly‐2 fine‐tunes potency and isoform selectivity by engaging distinct secondary regions whose conformational variability as captured by X‐Ray crystallography differs between galectin‐7 and galectin‐8N. These structural and biophysical insights provide a coherent framework for rational design: Future selectivity improvements must arise from tailoring Gly‐2 substituents to exploit the unique nonconserved microenvironments of galectin‐8N, while preserving the essential Gly‐1 geometry. Comparative Structural Analysis with the Binding to Galectin‐3 Comparison of the crystal structures of galectin‐3 [33], galectin‐7, and galectin‐8N bound to compounds 13, 14, 15, 17, and TDG reveals a shared overarching recognition mode, together with distinct isoform‐specific features that account for differences in affinity and selectivity. Across all three galectins, the Gly‐1 β‐galactoside ring occupies an almost identical position within the primary binding pocket, engaging a conserved network of hydrogen bonds and π‐stacking interactions. In each protein, Gly‐1 interacts with a structurally conserved set of residues located on β‐strands S4–S6 and the S4–S5 loop, including H158/H49/H65, N160/N51/N67, R162/R53/R69, N174/N62/N79, and E184/E72/E89 (in galectin‐3/‐7/‐8N, respectively). This highly conserved interaction geometry underscores the dominant role of Gly‐1 as the primary anchoring element of the thiodisaccharide scaffold and explains why modifications at Gly‐1 have dramatic effects on affinity. In contrast, the Gly‐2 moiety adopts protein‐specific orientations that differ markedly between galectin‐3, galectin‐7, and galectin‐8N (Figure 4). These variations arise from differences in the shape and residue composition of the extended binding site beyond subsite C, allowing Gly‐2 to act as the principal driver of isoform selectivity (Table S4). In galectin‐3, the Gly‐2 moiety of 13, 15, and TDG forms additional hydrogen bonds with R186 as well as conserved residues R162 and E184 [13, 33]. These interactions are enabled by the relatively open geometry of the extended site and contribute to the higher affinities observed for galectin‐3. In galectin‐7, distinct, well‐defined Gly‐2 positions are observed across the dimeric and tetrameric complexes. Hydroxyl groups O3′ and O4′ of 13 form hydrogen bonds with R53, E72, and R74 in several monomers, while other orientations position Gly‐2 for additional contacts with N62, G70, and G72. The presence of multiple crystallographically resolved binding modes (Figure S3) suggests that the extended binding site of galectin‐7 is more permissive and capable of accommodating multiple ligand conformations, which may partly explain the moderate binding affinities of the series. In galectin‐8N, all four ligands form hydrogen bonds between Gly‐2 substituents and the conserved pair R69 and E89, although the specific interacting atoms differ among compounds. Importantly, the extended binding site of galectin‐8N contains several nonconserved residues, in subsites B (Q47 and R59) and D (R72 and I91) that participate directly or indirectly in ligand recognition through hydrogen‐bond networks or water‐mediated interactions (Figure 3). These residues create a more enclosed and chemically distinct environment than in galectin‐3 or ‐7, accounting for the unique conformations of Gly‐2 in each galectin‐8N complex (Figure 5). These observations are consistent with previous structural and biophysical analyses, demonstrating that galectin‐8N possesses unique recognition properties compared with the C‐terminal CRD and other galectins, reflecting differences in the architecture of the extended carbohydrate‐binding surface [22]. Right: superposition of galectin‐3‐CRD‐13 (cyan), galectin‐7‐13 (Chain A) (green), galectin‐7‐13 (Chain B) (yellow), and galectin‐8N‐13 (pink). Left: the side chains of the well‐conserved residues of each galectin are labeled. Right: superposition of galectin‐8N‐CRD ‐13 (blue), ‐14 (yellow), ‐15 (pink), ‐17 (green), and TDG (salmon). Left: the side chains of the well‐conserved residues of each galectin are labeled. Collectively, this comparative analysis shows that whereas the conserved Gly‐1 binding mode provides a common anchoring framework across galectins, differences in the extended binding site—and specifically the identity and positioning of nonconserved residues, govern the orientation and interaction pattern of the Gly‐2 moiety. These distinctions define the structural basis for isoform‐specific ligand preferences and highlight the Gly‐2 position as the most promising site for further chemical optimization aimed at improving selectivity toward galectin‐7 or galectin‐8N. Conclusions This work integrates biophysical measurements and high‐resolution crystallography to define the molecular determinants of C‐glycosylic 1,2‐thiodisaccharide binding to galectin‐7 and the N‐terminal CRD of galectin‐8. Across all complexes, the Gly‐1 β‐galactoside consistently adopts a conserved anchoring mode, engaging the characteristic hydrogen‐bond network and π‐stacking interactions that underlie productive recognition. This strict requirement explains the severely reduced affinity of compound 16, confirming that even subtle distortions in Gly‐1 stereochemistry disrupt binding at all galectins examined. In contrast, Gly‐2 governs selectivity, sampling distinct orientations depending on the architecture and flexibility of each galectin's extended binding site. In galectin‐7, the extended binding surface appears permissive, as it is evident by the multiple Gly‐2 orientations observed for 13 across the A–D monomers of the crystallographic tetramer. While no structures are available for 14, 15, and 17, their moderate and variable affinities are consistent with a similarly adaptable binding environment. In galectin‐8N, however, the extended site is more constrained, all ligands converge on the conserved R69–E89 axis, and none directly engage the nonconserved residues Q47, R59, R72, or I91. These residues nonetheless define an enclosed, chemically distinct environment around the secondary pocket, offering clear opportunities for targeted chemical elaboration in future inhibitor design. Although the compounds reported here display micromolar affinities, they provide valuable chemical probes for investigating galectin‐dependent biology. Compound 13, which shows preferential binding to galectin‐7, and compound 17, the most potent galectin‐8N ligand identified by ITC, represent useful starting points for dissecting the distinct biological roles of these galectins in cellular systems. Galectin‐7 has been implicated in epithelial tumor progression, apoptosis, and immune regulation, whereas galectin‐8N contributes to processes, including cell adhesion, autophagy, angiogenesis, and immune modulation. The structural insights obtained in this work identify clear opportunities for enhancing affinity and selectivity through Gly‐2 optimization. Such next‐generation analogues could facilitate mechanistic studies in cells and ultimately support the development of galectin‐selective therapeutic agents for cancer, inflammatory disorders, and immune‐mediated diseases. Together, these findings establish C‐glycosylic 1,2‐thiodisaccharides as a robust, tunable scaffold for developing next‐generation galectin inhibitors. The deeply conserved Gly‐1 interactions provide a reliable affinity baseline, while the chemically pliable Gly‐2 substituents offer a strategic location for engineering isoform selectivity. Galectin‐8N presents promising nonconserved microenvironments that can be exploited through π‐stacking, cation–π, electrostatic, or hydrophobic design elements. These insights provide a blueprint for rational modification of Gly‐2, including incorporation of aromatic motifs, extended linkers, or negatively charged groups capable of reaching Q47, R59, R72, or I91 (Figure 6). By leveraging these principles, it should now be possible to engineer inhibitors with significantly enhanced potency and selectivity for galectin‐8N, thereby advancing the broader effort to therapeutically target galectin‐mediated pathways in cancer, immunity, and inflammation. Schematic summary of the structural determinants governing affinity and selectivity of C‐glycosylic 1,2‐thiodisaccharides. Gly‐1 (green) provides the conserved affinity driving anchor with conserved residues (green), whereas Gly‐2 (cyan) interacts with variable regions of the CRD and therefore represents the principal determinant of isoform selectivity. Nonconserved residues of galectin‐8N (Q47, R59, R72, I91; in red) constitute attractive targets for future ligand optimization. Funding This work was supported by the Hungarian National Research, Development and Innovation Office (2024‐1.2.3‐HU‐RIZONT‐2024‐00005). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supplementary Material