Abstract

Frutalin (FTL) is a multiple-binding lectin belonging to the jacalin-related lectin (JRL) family and derived from Artocarpus incisa (breadfruit) seeds. This lectin specifically recognizes and binds α-d-galactose. FTL has been successfully used in immunobiological research for the recognition of cancer-associated oligosaccharides. However, the molecular bases by which FTL promotes these specific activities remain poorly understood. Here, we report the whole 3D structure of FTL for the first time, as determined by X-ray crystallography. The obtained crystals diffracted to 1.81 Å (Apo-frutalin) and 1.65 Å (frutalin–d-Gal complex) of resolution. The lectin exhibits post-translational cleavage yielding an α- (133 amino acids) and β-chain (20 amino acids), presenting a homotetramer when in solution, with a typical JRL β-prism. The β-prism was composed of three 4-stranded β-sheets forming three antiparallel Greek key motifs. The carbohydrate-binding site (CBS) involved the N-terminus of the α-chain and was formed by four key residues: Gly25, Tyr146, Trp147 and Asp149. Together, these results were used in molecular dynamics simulations in aqueous solutions to shed light on the molecular basis of FTL-ligand binding. The simulations suggest that Thr-Ser-Ser-Asn (TSSN) peptide excision reduces the rigidity of the FTL CBS, increasing the number of interactions with ligands and resulting in multiple-binding sites and anomeric recognition of α-d-galactose sugar moieties. Our findings provide a new perspective to further elucidate the versatility of FTL in many biological activities.

Introduction

Plant lectins are proteins that are capable of interacting specifically and reversibly with glycans without altering their covalent structure. These proteins can mediate a variety of biological processes when in contact with glycoconjugates on the cell surface. This carbohydrate recognition ability also confers inflammatory and anti-inflammatory properties to plant lectins, as well as immunostimulatory activities [1]. Indeed, plant lectins are generally considered to be a very heterogeneous group of proteins, given that comparative biochemical studies indicate distinct biochemical/physicochemical properties, molecular structure, carbohydrate-binding specificity and biological activities, even among homologous molecules [2].

The genus Artocarpus (Moraceae) is a group among forest plants comprising more than 1400 species widely used in traditional medicines. Thus, prompting scientific interest in the secondary metabolites produced by this genus, which possess useful biological activities. Among these species, jackfruit (Artocarpus integrifolia) and breadfruit (Artocarpus altilis, also known as Artocarpus incisa) are well-known sources of lectins [3]. Jacalin is the predominant protein in jackfruit crude seed extract [4,5]. Since 1998, our group has surveyed A. incisa seeds and found lectins with similar characteristics to those in jackfruit seeds. The most abundant lectin in breadfruit is frutalin (FTL), first identified by Moreira et al. [6].

FTL belongs to the jacalin-related lectins (JRLs) with a preference for α-d-galactose moieties. At pH >8.0, FTL appears to be in a tetramer form composed of four protomers and with an apparent molecular mass of 60 kDa in native electrophoresis. FTL presents bands of 12 and 15 kDa upon SDS–PAGE, which correspond to a glycosylated fraction and a slightly or non-glycosylated fraction [7]. FTL is expressed in different isoforms, which mainly reflect differences in its post-translational glycosylation pattern [8]. Monteiro-Moreira et al. observed several masses ∼16 kDa in the deconvoluted spectra, which is consistent with the presence of these isoforms.

FTL has become an attractive and versatile biotechnological tool based on its ability to specifically recognize glycoconjugates. FTL has a multitude of activities, such as in the identification of prostate cancer tissues [9], as a tool for pivotal cancer biomarkers [10], neutrophil activation [11], gastroprotection in ethanol-induced lesions [12], and as an inhibitor of orofacial nociception in acute and chronic pain [13]. Therefore, based on such attractive properties, we aimed to elucidate the structure–activity relationship of FTL, to provide an understanding of its protein–carbohydrate interactions, which could help to identify and/or improve its therapeutic value.

Experimental

Isolation and purification of FTL from Artocarpus incisa seeds

A. incisa seeds were collected in Maranguape, Ceará, Brazil, and FTL extraction was performed according to previous studies [6]. Purity was confirmed by SDS–PAGE, and functional activity was assessed by a routine hemagglutination assay to measure the minimal concentration for agglutination (MCA). FTL-induced hemagglutination was determined as described previously [14]. After 1 h incubation at 37°C, duplicate wells were assessed to determine the MCA, i.e. the lowest lectin concentration that gave visible agglutination. PBS blanks were used as control [15]. After tryptic digestion, FTL was also submitted to mass spectrometry using a SYNAPT HDMS mass spectrometry (Waters, Manchester, U.K.) spectrometer, coupled to a nanoUPLC-ESI system. The sample was diluted to 1 mg/ml in 0.1% (v/v) formic acid. One microliter of the sample was used to perform reverse-phase chromatography using a gradient from 3% to 70% (v/v) acetonitrile for 30 min with 0.1% formic acid (v/v) at a flow rate of 300 nl/min in a BEH C4 column. For intact mass analysis, an aliquot was diluted in 0.1% (v/v) formic acid and applied directly to the spectrometer without being submitted to reverse-phase chromatography. The acquired MS data were processed using a technique that prioritized the maximal entropy (MaxEnt) to obtain a deconvoluted spectrum [16].

Crystallization, data collection, and processing and refinement of FTL structure

Crystallization and data collection of Apo-frutalin has been previously described [8]. The molecular replacement was performed with the use of PHASER [17], using jacalin co-ordinates (PDB ID: 3P8S) as the initial model (98% identity with FTL). Co-crystallization experiments were performed using FTL crystals incubated with 5 mM d-galactose (Sigma–Aldrich®, purchased as a mixture of α/β-d-galactose) and 10% PEG 20 000, 50% (v/v) PEG MME 550, 0.1 M bicine/Trizma base (pH 8.5). Data collection was performed at 100 K up to 1.6 5 Å using a Rigaku MicroMax 007 HF equipped with RAXIS IV++ IP, to 1.81 Å (Apo-frutalin) and 1.65 Å resolution (FTL–d-galactose complex). The diffraction data were indexed, integrated and scaled using the XDS package [18]. We used jacalin co-ordinates for molecular replacement, which shares 98% sequence identity with FTL (PDB ID: 3P8S); both structures were refined using Phenix [19] and Coot [20]. R and Rfree were monitored to evaluate the validity of the refinement protocol, and the stereochemistry of the model was assessed using Molprobity [21]. The co-ordinates and structure factors have been deposited in the RCSB Protein Data Bank with the accession codes 4wog and 5bn6 for the Apo and FTL–d-galactose complex, respectively. The representative figures for both models were generated, and the two structures were validated through the validation server of the Protein Data Bank.

Molecular docking studies

AutoDock 4.2 was used to perform the molecular docking analysis [22]. The grid maps of 40 Å x 40 Å x 40 Å centered on the CBS of FTL (PDB ID: 4WOG) and α-d-Gal-(1 → 6)-d-Man) and β-d-Gal-(1 → 2)-α-d-Xyl-(1 → 6)-d-Glu residues mimicking, respectively, d-galactose found in galactomannan and xyloglucan polysaccharides, were calculated with AutoGrid. These molecules were modeled with Chem3D software and had their energies minimized to maintain the most stable structure. The saccharide substrates exhibited all of the torsional bonds with free rotation, while the protein was held rigid. The 10 best structures were analyzed and ranked according to the predicted binding affinity (expressed in kcal mol−1). Three-dimensional images of the interactions between ligands and proteins were depicted with the aid of Pymol [21].

Molecular dynamics

The structure without ligands (PDB ID: 4WOG) was submitted to the program H++ using the webserver (http://biophysics.cs.vt.edu/H++) [23] to obtain pK estimates for residues and define which groups would be ionized in the molecular dynamics (MD) simulation at pH 9.0 (tetramer). Histidine residues were protonated at the delta positions of His68 and His95 on the four β-chains. The Asp, Glu, Lys, and Arg residues were ionized and the N- and C-terminal chains of the α- and β-chains resulting in a net charge of −4.

MD were analyzed using the GROMACS-5.1.2 package [24]. The system was defined by a 10 854 nm cubic box, and the crystallographic co-ordinates of the protein and α-d-galactose were used to generate the topology and place the geometric center of the protein in the box center. The water molecules that accompanied the crystallographic structure were excluded, and 38645 SPC water molecules were added. Additionally, four Na+ ions were used to neutralize the system. The system consisted of 122 107 atoms.

The GROMOS53A6 force field was used to model proteins and ions, and a modified version of the GROMOS53A6 was used to model sugar molecules [25]. A cutoff radius (rc = 1.4 nm) was used to determine the interactions between unrelated atoms for both van der Waals and electrostatic interactions, and the PME method [26] was used to treat long-range interactions. The bond distances of hydrogen atoms were constrained with LINCS [27] in the case of proteins and SETTLE [28] for water. The integration timestep was 2 fs during MD runs, at 1 atm and 298 K.

Root-mean-square deviations (RMSDs) from the starting structure for the alpha carbon (Cα) atoms from α- and β-chains were calculated during the 400 ns of the MD simulation. Contacts between the α- and β-chains were identified by the observation of percentage hydrogen bonds (HBs) between atoms of backbone (bb) and side-chain (sc) proteins formed during the MD (molecular dynamics) trajectory analysis. The geometric criterion for detecting HBs was used as proposed by IUPAC: , where the angle is >165° and the distance H……Y is <0.3 nm.

Results and discussion

In SDS–PAGE, the profile of an isolated FTL batch matched the typical jacalin-like mass pattern of two bands ∼12 and 15 kDa, corresponding to monomers with different glycosylation levels, as previously reported [6]. Evidence based on FTL mass deconvoluted spectra suggests that the protein is naturally expressed as a mixture of isoforms [29]. Using a hemagglutination assay, we found FTL to be in a biologically active form and able to agglutinate human erythrocytes, which was preliminarily required before submitting the lectin to crystallization assays (Supplementary Figure S1).

X-ray diffraction

The best solution for both Apo-frutalin and FTL–d-galactose complexes was obtained in space group I2 with four monomers per asymmetric unit. The data collection statistics are summarized in Table 1. The final models were refined to an Rfactor of 0.163 and an Rfree of 0.203 for Apo-frutalin, and to an Rfactor of 0.166 and an Rfree of 0.200 for the FTL–d-galactose complex. In both cases, the asymmetric unit contains for monomers, with 153 residues in each (Figure 1A,B). The residues (96.78%) are in the most favored regions, and 1.21% are in the allowed regions of the Ramachandran plot.

Predicted structural features of FTL.

Figure 1.
Predicted structural features of FTL.

(A) Sequence alignment of FTL (PDB ID: 5BN6) with jacalin (PDB ID: 1UGW) and CGB (PDB ID: 4AKB). Alignment was performed using Clustal Omega. The expanded carbohydrate-binding site for each protein is boxed in blue. (B) Tetrameric structure of FTL with the α- (magenta) and β-chains (red). Interactions between α- and β-chains can be observed in α1–β1–α2–β2 chains (set A) and α3–β3–α4–β4 chains (set B). (C) Ribbon diagram of β-prism, each part of the β-prism, is colored differently to facilitate the understanding of this structure (red, blue, and yellow). Topology diagram of FTL generated by PDBSUM. (D) The FTL structure is composed of three Greek keys and could be classified as N- (N- and C-termini) and C-types.

Figure 1.
Predicted structural features of FTL.

(A) Sequence alignment of FTL (PDB ID: 5BN6) with jacalin (PDB ID: 1UGW) and CGB (PDB ID: 4AKB). Alignment was performed using Clustal Omega. The expanded carbohydrate-binding site for each protein is boxed in blue. (B) Tetrameric structure of FTL with the α- (magenta) and β-chains (red). Interactions between α- and β-chains can be observed in α1–β1–α2–β2 chains (set A) and α3–β3–α4–β4 chains (set B). (C) Ribbon diagram of β-prism, each part of the β-prism, is colored differently to facilitate the understanding of this structure (red, blue, and yellow). Topology diagram of FTL generated by PDBSUM. (D) The FTL structure is composed of three Greek keys and could be classified as N- (N- and C-termini) and C-types.

Table 1
X-ray parameters for FTL structures

Values in parentheses are for the outer resolution shell.

 Apo-frutalin FTL–galactose complex 
PDB ID 4WOG 5BN6 
X-ray source Rigaku MicroMax 007HF Rigaku MicroMax 007HF 
Detector Rigaku Raxis IV++ Rigaku Raxis IV++ 
Cell parameters (Å)
a, b, c 
76.17, 74.36, 118.98 75.93, 74.60, 119.04 
Space group II
Resolution (Å) 26.74–1.81 (1.85–1.81) 19.65–1.65 (1.65–1.68) 
X-ray source Rigaku MicroMax 007 HF Rigaku MicroMax 007 HF 
λ (Å) 1.5418 1.5418 
Multiplicity 4.0 (3.7) 2.3 (1.5) 
Rmerge (%) 10.0 (53.3) 4.9 (44.6) 
Rpim (%) 4.9 (27.1) 3.9 (39.5) 
CC(1/2) 0. 996 (0.784) 0.998 (0.736) 
Completeness (%) 99.8 (97.0) 97.3 (75.6) 
Reflections 240 259 (12 493) 174 200 (4489) 
Unique reflections 59 769 (3386) 77 162 (2939) 
I/σ 10.3 (2.6) 13.7 (1.9) 
Reflections used for refinement 59 762 77 151 
Rfactor 0.163 0.166 
Rfree 0.203 0.200 
No. of protein atoms 4629 4626 
No. of ligand atoms 48 
B214.19 13.83 
Co-ordinate error (ML based) (Å) 0.21 0.15 
Phase error (°) 19.33 19.15 
Ramachandran plot 
 Favored (%) 96.59 96.59 
 Allowed (%) 3.41 3.41 
 Outliers (%) 0.00 
 All-atom clashscore 1.86 2.38 
 RMSD from ideal geometry 
 r.m.s. bond lengths (Å) 0.007 0.006 
 r.m.s. bond angles (°) 1.017 1.038 
 Apo-frutalin FTL–galactose complex 
PDB ID 4WOG 5BN6 
X-ray source Rigaku MicroMax 007HF Rigaku MicroMax 007HF 
Detector Rigaku Raxis IV++ Rigaku Raxis IV++ 
Cell parameters (Å)
a, b, c 
76.17, 74.36, 118.98 75.93, 74.60, 119.04 
Space group II
Resolution (Å) 26.74–1.81 (1.85–1.81) 19.65–1.65 (1.65–1.68) 
X-ray source Rigaku MicroMax 007 HF Rigaku MicroMax 007 HF 
λ (Å) 1.5418 1.5418 
Multiplicity 4.0 (3.7) 2.3 (1.5) 
Rmerge (%) 10.0 (53.3) 4.9 (44.6) 
Rpim (%) 4.9 (27.1) 3.9 (39.5) 
CC(1/2) 0. 996 (0.784) 0.998 (0.736) 
Completeness (%) 99.8 (97.0) 97.3 (75.6) 
Reflections 240 259 (12 493) 174 200 (4489) 
Unique reflections 59 769 (3386) 77 162 (2939) 
I/σ 10.3 (2.6) 13.7 (1.9) 
Reflections used for refinement 59 762 77 151 
Rfactor 0.163 0.166 
Rfree 0.203 0.200 
No. of protein atoms 4629 4626 
No. of ligand atoms 48 
B214.19 13.83 
Co-ordinate error (ML based) (Å) 0.21 0.15 
Phase error (°) 19.33 19.15 
Ramachandran plot 
 Favored (%) 96.59 96.59 
 Allowed (%) 3.41 3.41 
 Outliers (%) 0.00 
 All-atom clashscore 1.86 2.38 
 RMSD from ideal geometry 
 r.m.s. bond lengths (Å) 0.007 0.006 
 r.m.s. bond angles (°) 1.017 1.038 

Similar to jacalin, FTL is synthesized in vivo as an unusual preproprotein, which becomes two chains after co-translational and post-translational processing: an α-chain (133 amino acids) and a β-chain (20 amino acids). FTL also shows conserved consensus sequences, which suggests that three N-glycosylation sites may be present [6,8]. In contrast, mannose-specific lectins such as frutapin (also found in breadfruit seeds) consist of uncleaved protomers of ∼150 amino acid residues [15].

The FTL structure showed a typical symmetric β-prism fold, which is found in jacalin and other lectins [30]. This β-prism is composed of three 4-stranded β-sheets forming three antiparallel Greek key motifs, generating an approximate 3-fold symmetry (Figure 1C,D). FTL shares high structural similarity with jacalin, CGB, and frutapin, which superposed well and gave RMSDs ranging from 0.31 to 0.43 Å for the superposition of Cα atoms in these homologous structures (Supplementary Figure S2) [15,30,31].

Another major difference between the mannose-specific JRLs and galactose-specific JRLs is their biosynthesis, processing, and topogenesis. For example, jacalin is synthesized as a preproprotein, which undergoes a complex series of processing steps and is presumed to be located in the vacuolar compartment [32]. In contrast, the mature polypeptides of the mannose-specific JRLs correspond to the entire open reading frame of the respective lectin genes, presenting both α- and β-chains linked by a loop, and therefore are synthesized and located in the cytoplasm [32]. This alteration is believed to expose amino acids that are involved in carbohydrate recognition by FTL.

Indeed, leguminous lectins have considerable conservation in their primary, secondary, and tertiary structures. Comparisons of these sequences and structures demonstrate that differences in carbohydrate specificity appear to occur due to changes in carbohydrate-binding site-adjacent amino acids. The conformation of these loops is determined by the presence of calcium or transition metal ions in the protein structure, which helps CBS orientation and affinity for ligands. Although structurally analogous, with some reaching up to 90% similarity, these lectins present several distinct biological activities [33].

Carbohydrate-binding site and molecular docking studies

Although FTL recognizes a range of ligands, it has great affinity for α-d-galactose monosaccharides and complex carbohydrates that contain Galα1–3 glycans [6,9]. Galactose binding is dominated by hydrogen bonding with the sugar hydroxyl groups O1, O3, O5, and O6 (Figure 2A). Hydrogen bonding is the most dominant interaction in recognition of sugar molecules by the lectin CBS via carbonyl and hydroxyl groups (Figure 2B) of the backbone and side chains [34]. The crystal structures of the FTL–d-galactose complex also showed this pattern. The FTL–d-galactose complex does not present significant structural differences when compared with Apo-frutalin.

FTL carbohytdrate-binding site with D-galactose.

Figure 2.
FTL carbohytdrate-binding site with D-galactose.

(A) Composite omit map contoured at 1σ for a galactose molecule in the FTL structure. (B) Galactose-neighboring residues are labeled. Ligplus interaction drawing of galactose interactions in the FTL structure.

Figure 2.
FTL carbohytdrate-binding site with D-galactose.

(A) Composite omit map contoured at 1σ for a galactose molecule in the FTL structure. (B) Galactose-neighboring residues are labeled. Ligplus interaction drawing of galactose interactions in the FTL structure.

Overall, the FTL-binding site is similar to those in Moraceae lectins and consists of a domain close to the N-terminus of the α-chain consisting of four key residues: Gly25, Tyr146, Trp147, and Asp149 [3537]. In the three-dimensional structure of the jacalin–α-d-galactose complex (PDB ID: 1KU8), eight hydrogen bonds form directly between CBS amino acids and the hydroxyls, especially those of C3 and C6 positions in d-galactose. Similarly, in the galactose-binding lectin from champedak fruit (CGB)–Gal complex, FTL–galactose binding occurs via many hydrogen bonds between the O atoms on the sugar ring and with side-chain and main-chain N and O atoms on the α-chain (O3 and Gly1 N, O4 and Gly1 N and Asp125 OD1, O6 and Trp123 O, Trp123 N and Tyr122 N, O5 and Tyr122 N) through O3, O4, and O5 [31]. In contrast, the FTL–d-galactose complex displays 10 interactions through the C1 hydroxyl to residue Tyr146, C3 hydroxyl to residue Gly25, C4 hydroxyl to residues Gly25 and Asp149, and C6 hydroxyl to residues Tyr146, Trp147, and Asp149 (Figure 3A).

FTL CBS anchoring specific sugars.

Figure 3.
FTL CBS anchoring specific sugars.

(A) Close-up view of the FTL sugar-binding site in the d-galactose complex. Protein and carbohydrate molecules are depicted as ribbon and stick models, respectively. Amino acid residues, which interact with carbohydrates, are highlighted in blue. (B) Molecular docking experiments suggest that the FTL CBS can accommodate the α-d-Gal–d-Man disaccharide present in galactomannans, but not the β-d-Gal–Xyl–Glu trisaccharide present in xyloglucans (C), as evidenced through affinity chromatography in those cross-linked polysaccharide matrices.

Figure 3.
FTL CBS anchoring specific sugars.

(A) Close-up view of the FTL sugar-binding site in the d-galactose complex. Protein and carbohydrate molecules are depicted as ribbon and stick models, respectively. Amino acid residues, which interact with carbohydrates, are highlighted in blue. (B) Molecular docking experiments suggest that the FTL CBS can accommodate the α-d-Gal–d-Man disaccharide present in galactomannans, but not the β-d-Gal–Xyl–Glu trisaccharide present in xyloglucans (C), as evidenced through affinity chromatography in those cross-linked polysaccharide matrices.

Jacalin is among the most thoroughly studied lectins. Jeyaprakash et al. [38] identified three components of the jacalin sugar-binding site, a primary binding site and two secondary sites, named A and B. In this postulation, the primary site is responsible for binding galactose. The secondary site can bind any α-linked sugar moiety, but cannot tolerate any β-substitutions [39]. In addition, jacalin and CGB bind carbohydrates at one primary and two secondary binding sites [31,40]. The large number of interactions is consistent with FTL's affinity for galactose. The FTL agglutination of erythrocytes is due to its multi-subunit proteins, which contain multiple carbohydrate-binding sites that enable them to agglutinate cells [41]. These carbohydrate-binding domains are identical or very similar and bind a wide range of ligands. A comparison of the FTL CBS to other Moraceae lectins suggests that the site closely resembles that of jacalin. There are equivalent residues in both lectins (such as Asp149 in FTL and Asp125 in jacalin), and both form two hydrogen bonds: one with the C3 hydroxyl and another with the C4 hydroxyl. Based on these interactions, we suggest that FTL has an affinity for d-galactose epimers such as d-mannose, as the C4 hydroxyl in this sugar is equatorial instead of axial, and the interaction would also be allowed by Asp149.

Interactions between derivatives of Gal β-(1,3) Gal-α-OMe and jacalin have been structurally and thermodynamically characterized [39]. It is now known that distortion of the ligand occurs as a strategy for modulating affinity. β-Substituted methyl derivatives of disaccharides can also bind to jacalin without changing the pattern of interactions in the complexes involving the corresponding α-substituted derivatives. This is achieved through distortions of the ligand molecule at the anomeric carbon and the glycosidic linkage in addition to a small lateral shift. The higher internal energy caused by the distortion reduces the affinity of β-substituted b-(1,3)-linked disaccharides to jacalin when compared with the α-substituted variants [36,39,42]. Gal β-(1,3) Gal and its derivatives preferentially bind to jacalin with the reducing sugar at the primary binding site, although binding with the non-reducing Gal at the primary site is possible. α-Methyl substitution further strengthens binding in the first arrangement. In contrast, β-substitution weakens the binding due to ligand distortion. The β-substituted disaccharides continue to bind with the reducing Gal at the primary binding site, thus, indicating that the affinity reduction is not strong enough to overcome the intrinsic propensity of Gal β-(1,3) Gal to bind to jacalin with the reducing Gal at the primary site. This propensity is believed to be an important determinant in the biologically relevant interactions between jacalin and oligosaccharides [39].

Interestingly, native FTL was previously isolated in an affinity chromatography step using an Adenanthera pavonina cross-linked galactomannan [6]. In this type of polysaccharide, galactose is naturally α-linked (1-6-d-galactopyranose) to the β-1-4-d-mannopyranose backbone. On the other hand, in xyloglucan from Tamarindus indica seeds, these galactose branches are bound by β-(1-2) to xylosyl residues attached to the main glucan backbone. When using xyloglucan cross-linked matrices, we found FTL to always be easily removable from the column with PBS buffer during the washing steps. This result suggests that FTL has a lower or no affinity for β-galactose residues, implying an anomeric recognition. In looking for further evidence of this anomeric recognition, docking experiments were performed to check whether the FTL carbohydrate-binding site could accommodate the α-d-Gal-(1 → 6)-d-Man disaccharide and β-d-Gal-(1 → 2)-α-d-Xyl-(1 → 6)-d-Glu trisaccharide, which mimic the FTL interaction in galactomannan and xyloglucan cross-linked matrices, respectively. FTL interactions with those saccharides yielded 10 energetic clusters. Figure 3B,C shows the pivotal interface between the FTL–galactomannan residue (−4.9 kcal mol−1) and FTL–xyloglucan residue (−6.5 kcal mol−1), respectively. These results are consistent with affinity chromatography data in which xyloglucan matrices were inefficient at retaining FTL. In contrast, accommodation of the α-d-Gal-(1 → 6)-d-Man disaccharide with Ala17 and Val19 residues was observed in FTL–galactomannan docking, in addition to those involved in the CBS. Thus, FTL has anomeric carbohydrate recognition.

In addition, the tetrapeptide-linker ‘T-S-N-N’ is not a structural component of mature FTL, as it is excised during lectin processing to separate FTL β- and α-chains, giving rise to new N- and C-terminal sequences and reducing CBS rigidity, which results in the multiple-binding abilities of FTL, jacalin, and CGB [37]. The specificity of KM+ (Artocarpin), the mannose-binding lectin in jackfruit seeds, is attributed to the increased structural rigidity caused by the presence of the binding peptide GGPGGNGW. This peptide linker is not eliminated by post-translational processes, and is rich in glycine residues, thereby yielding extremely strong bonds between the two chains [43]. Therefore, we investigated MD to determine how this peptide linker may affect carbohydrate recognition by FTL.

Molecular dynamics

Simulations of the FTL structure with four α- and β-chains revealed an increasing RMSD with a remarkable peak ∼110 ns and 0.5 nm (Figure 4). This observed behavior indicates structural relaxation, which might occur through hydration after 110 ns and therefore reach equilibrium. Structural properties were then analyzed considering the trajectory at the same time point (equilibrium position) for all α- and β-chains (as shown in RMSD profiles by MD time in Supplementary Figure S3A,B). Table 2 shows the mean RMSD for the interaction of tertiary FTL structures (α- and β-chains), β-chain moieties (RMSD-M and RMSD-E), and the sum of the four α- and β-chains. Higher averages were observed for single β-chains when compared with single α-chains, implying that β-chains are more flexible.

RMSD FTL during MD simulation.

Figure 4.
RMSD FTL during MD simulation.

RMSD from the α-carbon positions after finding maximal overlap between α-carbons of the initial structure and structures collected during the MD simulation in all four FTL α- and β-chains, with its peak within 110 ns and 0.5 nm.

Figure 4.
RMSD FTL during MD simulation.

RMSD from the α-carbon positions after finding maximal overlap between α-carbons of the initial structure and structures collected during the MD simulation in all four FTL α- and β-chains, with its peak within 110 ns and 0.5 nm.

Table 2
The mean RMSD ± SD obtained in the last 300 ns of the MD simulation

The RMSD-M includes the superimposing step and calculating the RMSD between the Cα of residues in the 7–17 range. In the RMSD-E, Cα atoms between residues 7 and 17 are used for the superimposing step, and the deviation calculation is between the atoms of residues near the N- and C-termini of the β-chain from 1 to 6 and 18 to 19.

ID RMSD (nm) RMSD-M RMSD-E 
β1 0.26 ± 0.05 0.07 ± 0.02 0.32 ± 0.11 
β2 0.16 ± 0.05 0.08 ± 0.02 0.39 ± 0.08 
β3 0.41 ± 0.04 0.07 ± 0.02 0.38 ± 0.14 
β4 0.28 ± 0.04 0.09 ± 0.02 0.60 ± 0.09 
β1 +  β2 + β3 + β4 0.37 ± 0.02   
α1 0.15 ± 0.01   
α2 0.16 ± 0.01   
α3 0.16 ± 0.03   
α4 0.14 ± 0.02   
α1 + α2 + α3 + α4 0,36 ± 0.03   
All β- and α-chains 0.35 ± 0.05   
ID RMSD (nm) RMSD-M RMSD-E 
β1 0.26 ± 0.05 0.07 ± 0.02 0.32 ± 0.11 
β2 0.16 ± 0.05 0.08 ± 0.02 0.39 ± 0.08 
β3 0.41 ± 0.04 0.07 ± 0.02 0.38 ± 0.14 
β4 0.28 ± 0.04 0.09 ± 0.02 0.60 ± 0.09 
β1 +  β2 + β3 + β4 0.37 ± 0.02   
α1 0.15 ± 0.01   
α2 0.16 ± 0.01   
α3 0.16 ± 0.03   
α4 0.14 ± 0.02   
α1 + α2 + α3 + α4 0,36 ± 0.03   
All β- and α-chains 0.35 ± 0.05   

The RMSDs for α-chains varied between 0.14 and 0.16 nm, whereas β-chain RMSDs varied between 0.16 and 0.41 nm, suggesting that the positions of the α-chains have shorter distanced than the beginning of the β-chains until reaching the equilibrium position of the quaternary structure. Similarly, the RMSDs were higher for the β-chains (range 0.04–0.05 nm), suggesting that after reaching the equilibrium position, β-chain positions vary more than α (range 0.01–0.03). These variations can be viewed and compared between the RMSD profiles in Supplementary Figure S3A,B. Therefore, even though β-chains move more than α, they are maintained in the structure, and the chain set stabilizes the octameric structure of FTL. Taking into account the relative positions of the four α- and β-chains in the quaternary structure, the RMSDs of (α1 + α2 + α3 + α4) 0.36 ± 0.03 nm and (β1 + β2 + β3 + β4) 0.37 ± 0.02 nm presented similar variations in the means and SD, suggesting that positions of the eight chains supporting the quaternary structure are preserved in equilibrium (Table 2). The RMSD-E and standard deviation were much larger than for the overall side-chain (RMSD). The high RMSD-E values suggest that the residues placed in these regions do not establish effective interchain contacts with the α-chains, which remain free to move. However, the RMSD-M for the 7–17 region happens to be smaller than the RMSD and RMSD-E, which indicates increased rigidity in this regions.

Figure 5 shows the distribution of the intermolecular interaction potential (IIP) between the residues, building up the four β-chains and water molecules within a 0.5 nm cutoff radius. It is evident that regions from 1 to 6 and 18 to 19 are hydrated (energy less than −5 kcal mol−1), which is in contrast with those from 7 to 17. The IIP results combined with the RMSD-E values found in the simulations allow us to conclude that the β-chain region from 7 to 17 residues increases the rigidity of the referred chain by interacting with the α- and β-chain residues. These interactions reflect the intermolecular HBs between groups of α- and β-chains, as can be seen in Table 3. We expected to find interactions among these α- and β-chains by the breakdown in FTL structure, which are mostly by α1–α2 and α3–α4 pairs. However, the majority of pivotal HBs were negligible. β-Chains appear to be useful in maintaining the structure, and they occur close to the amino acids from 10 to 18 in α-chains and interact through HB between the backbone amino and carbonyl groups (C=O……H–N) of both α-chains (Table 3). Figure 6 shows a representation of these connections formed by HBs between pairs of α- and β-chains. The β1 and β2 chains connect the α1 and α2 chains, while the β3 and β4 connect the α1 and α2 chains via intermolecular HBs.

PII between FTL residues and water molecules.

Figure 5.
PII between FTL residues and water molecules.

The distribution of IIP between the residues of the four β-chains and the water molecules within a 0.5 nm cutoff radius.

Figure 5.
PII between FTL residues and water molecules.

The distribution of IIP between the residues of the four β-chains and the water molecules within a 0.5 nm cutoff radius.

Interface between 03B1; and 03B2; chains of FTL connected by hydrogen bonds.

Figure 6.
Interface between 03B1; and 03B2; chains of FTL connected by hydrogen bonds.

Carbonyl and amine group HBs (C=O……H–N) between α1, β1-chains from monomer 1 and the α2-chain of monomer 2 are depicted in the tetrameric FTL structure (magenta). The interactions occur on two faces of the lectin and involve both α- and β-chains from monomers 1 and 2, resulting in an α2–β2–α1 cluster. Likewise, the connections are prone to occur in α3–β3–α4 and α4–β4–α3 interchain clusters.

Figure 6.
Interface between 03B1; and 03B2; chains of FTL connected by hydrogen bonds.

Carbonyl and amine group HBs (C=O……H–N) between α1, β1-chains from monomer 1 and the α2-chain of monomer 2 are depicted in the tetrameric FTL structure (magenta). The interactions occur on two faces of the lectin and involve both α- and β-chains from monomers 1 and 2, resulting in an α2–β2–α1 cluster. Likewise, the connections are prone to occur in α3–β3–α4 and α4–β4–α3 interchain clusters.

Table 3
Percentage of HBs observed (% OBS) between amino backbone (N–H) and carbonyl groups (C=O) for α- and β-chain residues

The residue name and position is represented by Res in the α1–4 chains, which make intermolecular HBs with residues of the β1–4 chains. Only the HBs between the amino backbone (NH) and carbonyl (C=O) groups with % OBS > 25 are shown.

Chain α1 α2 α3 α4 
Residue N–H C=O N–H C=O N–H C=O N–H C=O 
Gly5    Ser156* β3 (47)     
Lys6     Thr10 β2 (45)    
Ser7  Thr34 β4 (95)     Ser156* β1 (23)  
Gln8  Leu157CT β4 (39) Leu157CT β3 (96)    Leu157CT β1 (31)  
Val10 Leu155 β1 (59) Leu155 β1 (100) Leu155 β2 (60) Leu155 β2 (100)  Leu155 β3 (99) Leu155 β4 (96) Leu155 β4 (99) 
Ile11 Asn134
β2 (89) 
Glu133 β2 (55)
Asn134 β2 (93) 
Asn134 β1 (85) Glu133 β1 (74)
Asn134 β1 (86)

 
Asn134 β4 (98) Glu133 β4 (52)
Asn134 β4 (89) 
Asn134 β3 (96) Glu133 β3 (78)
Asn134 β3 (86) 
Val12 Met153 β1 (99) Met153 β1 (99) Met153 β2 (99) Met153 β2 (99) Met153 β3 (99) Met153 β3 (99) Met153 β4 (99) Met153 β4 (99) 
Gly13 Pro131 β2 (87)  Pro 131 β1 (86)  Pro131 β4 (89)  Pro131 β3 (83) Phe 151 β4 (96) 
Trp15 Phe151 β1 (98) Phe151 β1 (99) Phe151 β2 (98) Phe151 β2 (99) Phe151 β3 (98) Phe151 β3 (99) Phe151 β4 (98)  
Ala17 Asp149 β1 (73)  Asp149 β2 (88)  Asp149 β3 (90)  Asp149 β4 (67)  
Chain α1 α2 α3 α4 
Residue N–H C=O N–H C=O N–H C=O N–H C=O 
Gly5    Ser156* β3 (47)     
Lys6     Thr10 β2 (45)    
Ser7  Thr34 β4 (95)     Ser156* β1 (23)  
Gln8  Leu157CT β4 (39) Leu157CT β3 (96)    Leu157CT β1 (31)  
Val10 Leu155 β1 (59) Leu155 β1 (100) Leu155 β2 (60) Leu155 β2 (100)  Leu155 β3 (99) Leu155 β4 (96) Leu155 β4 (99) 
Ile11 Asn134
β2 (89) 
Glu133 β2 (55)
Asn134 β2 (93) 
Asn134 β1 (85) Glu133 β1 (74)
Asn134 β1 (86)

 
Asn134 β4 (98) Glu133 β4 (52)
Asn134 β4 (89) 
Asn134 β3 (96) Glu133 β3 (78)
Asn134 β3 (86) 
Val12 Met153 β1 (99) Met153 β1 (99) Met153 β2 (99) Met153 β2 (99) Met153 β3 (99) Met153 β3 (99) Met153 β4 (99) Met153 β4 (99) 
Gly13 Pro131 β2 (87)  Pro 131 β1 (86)  Pro131 β4 (89)  Pro131 β3 (83) Phe 151 β4 (96) 
Trp15 Phe151 β1 (98) Phe151 β1 (99) Phe151 β2 (98) Phe151 β2 (99) Phe151 β3 (98) Phe151 β3 (99) Phe151 β4 (98)  
Ala17 Asp149 β1 (73)  Asp149 β2 (88)  Asp149 β3 (90)  Asp149 β4 (67)  
*

HB is formed between the backbone and α-chain side-chain groups. Blank locations represent the absence of HB.

The amino and carbonyl groups from 10, 12, 15, and 17 β-chain residues form HBs with amino and carbonyl groups of one α-chain, while residues 11 and 13 form HBs with groups of another. Thus, the HBs establish β1–α1, β2–α2, β3–α3, and β4–α4 pairs; while for residues 11 and 13, the pairs are β1–α2, β2–α1, β3–α4, and β4–α3. Indeed, the α-chain segment from 10 to 19 residues (VIVGPWGAQ), with hydrophobic side chains, helps to maintain these HBs by keeping the water molecules away from the amino and carbonyl groups.

The stability of the entire FTL structure can be understood when we consider interaction sets between α- and β-chains, which occur as α1–β1–α2–β2 (set A) and α3–β3–α4–β4 chains (set B) (Figure 1B). Sets A and B form an interface maintained through interactions between the C-terminal regions of the β-chains (residues 1–8) and complementary regions of the α-chains in set B. These interactions result in the formation of HBs between α1–β4, α2–β3, α3–β2, and α4–β1 (mainly by the carboxylic acid of the C-terminal Leu157 in α-chains and amine of Gln8 in β-chains) (Table 3).

Additionally, Asp149 in the α-chain is stabilized in the carbohydrate-binding site by HBs between the amine and carbonyl groups of Ala17 in the β-chain (Figure 1D and Table 3). The occurrence of HB is greater than 67% in all α–β pair interactions. Furthermore, Ala17 and Val19 promote a hydrophobic environment by enhancing interchain HBs and preventing the α-d-Gal-(1 → 6)-d-Man disaccharide hydration, which helps its accommodation into the FTL CBS. Meanwhile, Ala17 appears to play a structural role, helping to stabilize Asp149, with Val19 oscillating in its position, as shown by the RMSD-E (Table 2). The MD results validate the functional importance of the β-chains in maintaining the FTL structure. In addition, Ala17 in the β-chains helps to stabilize the Asp149 position in the α-chains through HB formation, which corroborates Asp149 functioning as the key residue for interactions and carbohydrate-binding. Indeed, Asp149 allows entry of α-d-Gal-(1 → 6)-d-Man disaccharide to the CBS site and its subsequent stabilization.

In regard to this carbohydrate recognition, FTL has been evaluated in several biomedical applications and found to establish hydrogen bonds with the glycosylated fraction of the TRPV1 ion channel, inhibit the orofacial pain mechanism [13], and specifically recognize the NMDA receptor in its glycosylated fraction, interfering in a cellular mechanism and producing antidepressant-like NMDA receptor-mediated activity [44].

Conclusion

Over the years, the therapeutic relevance of FTL has been demonstrated in many biomedical mechanisms. However, little was known about FTL's structure. Taken together, the results in this work provide a new perspective for further elucidation of the functional properties of this lectin. While FTL is conserved across species, which explains the high similarity to other lectins in the Moraceae family, some plants have developed unique and specialized mechanisms to deal with their complex intracellular signaling pathways, mostly triggered by the intricate relationships between the carbohydrate-binding specificity of lectins. Thus, a better understanding of FTL binding to sugar moieties provides insights into its functionality and sheds light on the biological activities of this lectin.

Abbreviations

     
  • CBS

    carbohydrate-binding site

  •  
  • CGB

    champedak galactose-binding lectin

  •  
  • FTL

    frutalin

  •  
  • HB

    hydrogen bond

  •  
  • IIP

    intermolecular interaction potential

  •  
  • JRL

    jacalin-related lectins

  •  
  • KM+

    artocarpin

  •  
  • MCA

    minimal concentration for agglutination

  •  
  • MD

    molecular dynamics

  •  
  • PME

    particle mesh ewald

  •  
  • RMSD

    root-mean-square deviation

  •  
  • UPLC-ESI

    ultra performance liquid chromatography with electrospray ionization

Author Contribution

A.E.V.N., T.B.G., and F.D.S. were involved in all aspects of purifying the protein, while H.M.P. and F.B.M.B.M determined the X-ray structure. M.R.L. performed the MD experiments. A.C.O.M.M. and R.A.M. obtained funding for the study and provided overall supervision. A.E.V.N and F.D.S. wrote the initial drafts of the paper and all authors contributed to the final manuscript.

Funding

This work was supported by National Council for Scientific and Technological Development (CNPq), Fundação Cearense de Amparo á Pesquisa (FUNCAP), and Agency for Financing Studies and Projects (FINEP).

Acknowledgements

We acknowledge the Physics Institute of São Carlos (University of São Paulo, USP) for assistance with crystal testing and data collection. We also thank the Fundação Edson Queiroz for providing infrastructure at the University of Fortaleza (UNIFOR) and CAPES (Coordination for the Improvement of Higher Education) for financial support.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Lagarda-Diaz
,
I.
,
Guzman-Partida
,
A.
and
Vazquez-Moreno
,
L.
(
2017
)
Legume lectins: proteins with diverse applications
.
Int. J. Mol. Sci.
18
,
1242
2
Jiang
,
S.-Y.
,
Ma
,
Z.
and
Ramachandran
,
S
. (
2010
)
Evolutionary history and stress regulation of the lectin superfamily in higher plants
.
BMC Evol. Biol.
10
,
79
PMID:
[PubMed]
3
Jagtap
,
U.B.
and
Bapat
,
V.A
. (
2010
)
Artocarpus: a review of its traditional uses, phytochemistry and pharmacology
.
J. Ethnopharmacol.
129
,
142
166
4
de Azevedo Moreira
,
R.
and
Ainouz
,
I.L
. (
1981
)
Lectins from seeds of jack fruit (Artocarpus integrifolia L.): isolation and purification of two isolectins from the albumin fraction
.
Biol. Plant.
23
,
186
192
5
Kabir
,
S.
,
Aebersold
,
R.
and
Daar
,
A.S
. (
1993
)
Identification of a novel 4 kDa immunoglobulin-A-binding peptide obtained by the limited proteolysis of jacalin
.
Biochim. Biophys. Acta
1161
,
194
200
PMID:
[PubMed]
6
Moreira
,
R.A.
,
Castelo-Branco
,
C.C.
,
Monteiro
,
A.C.
,
Tavares
,
R.O.
and
Beltramini
,
L.M.
(
1998
)
Isolation and partial characterization of a lectin from Artocarpus incisa L. seeds
.
Phytochemistry
47
,
1183
1188
7
Oliveira
,
C.
,
Felix
,
W.
,
Moreira
,
R.A.
,
Teixeira
,
J.A.
and
Domingues
,
L
. (
2008
)
Expression of frutalin, an α-d-galactose-binding jacalin-related lectin, in the yeast Pichia pastoris
.
Protein Expr. Purif.
60
,
188
193
8
Monteiro-Moreira
,
A.C.O.
,
D'Muniz Pereira
,
H.
,
Vieira-Neto
,
A.E.
,
Moreno
,
F.B.M.B.
,
Lobo
,
M.D.P.
,
Sousa
,
F.D
et al.  et al.  (
2015
)
Crystallization and preliminary X-ray diffraction studies of frutalin, an α-d-galactose-specific lectin from Artocarpus incisa seeds.
Acta Crystallogr. Sect. F Struct. Biol. Commun.
F71
,
1282
1285
PMID:
[PubMed]
9
Oliveira
,
C.
,
Teixeira
,
J.A.
,
Schmitt
,
F.
and
Domingues
,
L
. (
2009
)
A comparative study of recombinant and native frutalin binding to human prostate tissues
.
BMC Biotechnol.
9
,
78
10
Lobo
,
M.D.
,
Moreno
,
F.B.
,
Souza
,
G.H.
,
Verde
,
S.M.
,
Moreira
,
R.A.
and
Monteiro-Moreira
,
A.C.
(
2017
)
Label-free proteome analysis of plasma from patients with breast cancer: stage-specific protein expression
.
Front. Oncol.
7
,
14
PMID:
[PubMed]
11
Brando-Lima
,
A.C.
,
Saldanha-Gama
,
R.F.
,
Henriques
,
M.D.G.M.O.
,
Monteiro-Moreira
,
A.C.O.
,
Moreira
,
R.A.
and
Barja-Fidalgo
,
C
. (
2005
)
Frutalin, a galactose-binding lectin, induces chemotaxis and rearrangement of actin cytoskeleton in human neutrophils: involvement of tyrosine kinase and phosphoinositide 3-kinase
.
Toxicol. Appl. Pharmacol.
208
,
145
154
12
De Vasconcellos Abdon
,
A.P.
,
Coelho De Souza
,
G.
,
Noronha Coelho De Souza
,
L.
,
Prado Vasconcelos
,
R.
,
Araújo Castro
,
C.
,
Moreira Guedes
,
M.
et al.  (
2012
)
Gastroprotective potential of frutalin, a d-galactose binding lectin, against ethanol-induced gastric lesions
.
Fitoterapia
83
,
604
608
13
Damasceno
,
M.B.M.V.
,
De Melo Júnior
,
J.D.M.A.
,
Santos
,
S.A.A.R.
,
Melo
,
L.T.M.
,
Leite
,
L.H.I.
,
Vieira-Neto
,
A.E.
et al.  (
2016
)
Frutalin reduces acute and neuropathic nociceptive behaviours in rodent models of orofacial pain
.
Chem. Biol. Interact.
256
,
9
15
14
Moreira
,
R.A.
and
Perrone
,
J.C
. (
1977
)
Purification and partial characterization of a lectin from Phaseolus vulgaris
.
Plant Physiol.
59
,
783
787
15
de Sousa
,
F.D.
,
da Silva
,
B.B.
,
Furtado
,
G.P.
,
Carneiro
,
I.S.
,
Lobo
,
M.D.P.
,
Guan
,
Y.
et al.  (
2017
)
Frutapin, a lectin from Artocarpus incisa (breadfruit): cloning, expression and molecular insights
.
Biosci. Rep.
37
,
BSR20170969
16
Ferrige
,
A.G.
,
Seddon
,
M.J.
,
Jarvis
,
S.
,
Skilling
,
J.
and
Aplin
,
R
. (
1991
)
Maximum entropy deconvolution in electrospray mass spectrometry
.
Rapid Commun. Mass Spectrom.
5
,
374
377
17
McCoy
,
A.J.
,
Grosse-Kunstleve
,
R.W.
,
Adams
,
P.D.
,
Winn
,
M.D.
,
Storoni
,
L.C.
and
Read
,
R.J
. (
2007
)
Phaser crystallographic software
.
J. Appl. Crystallogr.
40
(Pt 4)
,
658
674
PMID:
[PubMed]
18
Kabsch
,
W
. (
2010
)
Xds.
Acta Crystallogr. Sect. D Biol. Crystallogr.
66
,
125
132
19
Adams
,
P.D.
,
Afonine
,
P.V.
,
Bunkóczi
,
G.
,
Chen
,
V.B.
,
Davis
,
I.W.
,
Echols
,
N.
 et al. . (
2010
)
PHENIX: a comprehensive Python-based system for macromolecular structure solution
.
Acta Crystallogr. Sect. D Biol. Crystallogr.
66
,
213
221
20
Emsley
,
P.
and
Cowtan
,
K
. (
2004
)
Coot : model-building tools for molecular graphics research papers
.
Acta Crystallogr. D Biol. Crystallogr.
60
(Pt 12 Pt 1),
2126
2132
PMID:15572765
21
Chen
,
V.B.
,
Bryan
,
W.
,
Iii
,
A.
,
Headd
,
J.J.
,
Keedy
,
D.A.
,
Robert
,
M.
et al.   (
2010
)
Molprobity : all-atom structure validation for macromolecular crystallography research papers
.
Acta Crystallogr. D Biol. Crystallogr.
66
(Pt 1)
,
12
21
PMID:
[PubMed]
22
Trott
,
O.
and
Olson
,
A.J
. (
2011
)
NIH public access
.
J. Comput. Chem.
31
,
455
461
PMID:
[PubMed]
23
Gordon
,
J.C.
,
Myers
,
J.B.
,
Folta
,
T.
,
Shoja
,
V.
,
Heath
,
L.S.
and
Onufriev
,
A
. (
2005
)
H++: a server for estimating pKas and adding missing hydrogens to macromolecules
.
Nucleic Acids Res.
33
,
W368
W371
24
Abraham
,
M.J.
,
Murtola
,
T.
,
Schulz
,
R.
,
Páll
,
S.
,
Smith
,
J.C.
,
Hess
,
B.
et al.  (
2015
)
Gromacs: high performance molecular simulations through multi-level parallelism from laptops to supercomputers
.
SoftwareX
1–2
,
19
25
25
Pol-Fachin
,
L.
,
Rusu
,
V.H.
,
Verli
,
H.
and
Lins
,
R.D
. (
2012
)
GROMOS 53A6GLYC, an improved GROMOS force field for hexopyranose-based carbohydrates
.
J. Chem. Theory Comput.
8
,
4681
4690
26
Darden
,
T.
,
York
,
D.
and
Pedersen
,
L
. (
1993
)
Particle mesh Ewald: an N⋅log(N) method for Ewald sums in large systems
.
J. Chem. Phys.
98
,
10089
27
Hess
,
B.
,
Bekker
,
H.
,
Berendsen
,
H.J.C.
and
Fraaije
,
J.G.E.M
. (
1997
)
LINCS: a linear constraint solver for molecular simulations
.
J. Comput. Chem.
18
,
1463
1472
28
Miyamoto
,
S.
and
Kollman
,
P.A
. (
1992
)
Settle: an analytical version of the SHAKE and RATTLE algorithm for rigid water models
.
J. Comput. Chem.
13
,
952
962
29
De Oliveira Monteiro-Moreira
,
A.C.
,
D'Muniz Pereira
,
H.
,
Vieira Neto
,
A.E.
,
Mendes Batista Moreno
,
F.B.
,
Duarte Pinto Lobo
,
M.
,
de Sousa
,
F.D.
et al.  (
2015
)
Crystallization and preliminary X-ray diffraction studies of frutalin, an α-d-galactose-specific lectin from Artocarpus incisa seeds
.
Acta Crystallogr. Sect. Struct. Biol. Commun.
71
(Pt 10)
,
1282
1285
30
Pratap
,
J.V.
,
Jeyaprakash
,
A.A.
,
Rani
,
P.G.
,
Sekar
,
K.
,
Surolia
,
A.
and
Vijayan
,
M
. (
2002
)
Crystal structures of artocarpin, a Moraceae lectin with mannose specificity, and its complex with methyl-alpha-d-mannose: implications to the generation of carbohydrate specificity
.
J. Mol. Biol.
317
,
237
247
31
Gabrielsen
,
M.
,
Abdul-Rahman
,
P.S.
,
Othman
,
S.
,
Hashim
,
O.H.
and
Cogdell
,
R.J
. (
2014
)
Structures and binding specificity of galactose- and mannose-binding lectins from champedak: differences from jackfruit lectins
.
Acta Crystallogr. Sect. F Struct. Biol. Commun.
70
,
709
716
PMID:
[PubMed]
32
Peumans
,
W.J.
,
Hause
,
B.
and
Van Damme
,
E.J.M.
(
2000
)
The galactose-binding and mannose-binding jacalin-related lectins are located in different sub-cellular compartments
.
FEBS Lett.
477
,
186
192
33
Gadelha
,
C.A.A.
,
Moreno
,
F.B.M.B.
,
Santi-Gadelha
,
T.
,
Cajazeiras
,
J.B.
,
Rocha
,
B.A.M.
,
Assreuy
,
A.M.S.
 et al. . (
2005
)
Native crystal structure of a nitric oxide-releasing lectin from the seeds of Canavalia maritima
.
J. Struct. Biol.
152
,
185
194
34
Thirumalai
,
D.
,
Reddy
,
G.
and
Straub
,
J.E
. (
2012
)
Role of water in protein aggregation and amyloid polymorphism
.
Acc. Chem. Res.
45
,
83
92
35
Pinedo
,
M.
,
Orts
,
F.
,
Carvalho
,
A.O.
,
Regente
,
M.
,
Soares
,
J.R.
,
Gomes
,
V.M.
et al.  (
2015
)
Molecular characterization of Helja, an extracellular jacalin-related protein from Helianthus annuus: insights into the relationship of this protein with unconventionally secreted lectins
.
J. Plant Physiol.
183
,
144
153
36
Abhinav
,
K.V.
,
Sharma
,
K.
,
Surolia
,
A.
and
Vijayan
,
M
. (
2016
)
Effect of linkage on the location of reducing and nonreducing sugars bound to jacalin
.
IUBMB Life
68
,
971
979
37
Gabrielsen
,
M.
,
Abdul-Rahman
,
P.S.
,
Othman
,
S.
,
Hashim
,
O.H.
and
Cogdell
,
R.J
. (
2014
)
Structures and binding specificity of galactose- and mannose-binding lectins from champedak: differences from jackfruit lectins
.
Acta Crystallogr. Sect. F Struct. Biol. Commun.
70
,
709
716
38
Jeyaprakash
,
A.A.
,
Katiyar
,
S.
,
Swaminathan
,
C.P.
,
Sekar
,
K.
,
Surolia
,
A.
and
Vijayan
,
M
. (
2003
)
Structural basis of the carbohydrate specificities of jacalin: an X-ray and modeling study
.
J. Mol. Biol.
332
,
217
228
39
Abhinav
,
K.V.
,
Sharma
,
K.
,
Surolia
,
A.
and
Vijayan
,
M
. (
2017
)
Distortion of the ligand molecule as a strategy for modulating binding affinity: further studies involving complexes of jacalin with β-substituted disaccharides
.
IUBMB Life
69
,
72
78
40
Jeyaprakash
,
A.A.
,
Srivastav
,
A.
,
Surolia
,
A.
and
Vijayan
,
M
. (
2004
)
Structural basis for the carbohydrate specificities of artocarpin: variation in the length of a loop as a strategy for generating ligand specificity
.
J. Mol. Biol.
338
,
757
770
41
Varki
,
A
. (
2017
)
Biological roles of glycans
.
Glycobiology
27
,
3
49
42
Abhinav
,
K.V.
,
Sharma
,
K.
,
Swaminathan
,
C.P.
,
Surolia
,
A.
and
Vijayan
,
M
. (
2015
)
Jacalin-carbohydrate interactions: distortion of the ligand molecule as a determinant of affinity
.
Acta Crystallogr. Sect. D Biol. Crystallogr.
D71
,
324
331
43
Rosa
,
J.C.
,
De Oliveira
,
P.S.
,
Garratt
,
R.
,
Beltramini
,
L.
,
Resing
,
K.
,
Roque-Barreira
,
M.C.
et al.  (
1999
)
KM+, a mannose-binding lectin from Artocarpus integrifolia: amino acid sequence, predicted tertiary structure, carbohydrate recognition, and analysis of the beta-prism fold
.
Protein Sci.
8
,
13
24
44
Araújo
,
J.R.C.
,
Júnior
,
J.M.A.M.
,
Damasceno
,
M.B.M.V.
,
Santos
,
S.A.A.R.
,
Vieira-Neto
,
A.E.
,
Lobo
,
M.D.P.
et al.  (
2018
)
Neuropharmacological characterization of frutalin in mice: evidence of an antidepressant-like effect mediated by the NMDA receptor/NO/cGMP pathway
.
Int. J. Biol. Macromol.
112
,
548
554

Supplementary data