A proposed lectin‐mediated mechanism to explain the in Vivo antihyperglycemic activity of γ‐conglutin from Lupinus albus seeds

Abstract Experiments conducted in vitro and in vivo, as well as clinical trials for hypoglycemic therapeutics, support the hypoglycemic properties of the lectin γ‐conglutin, a Lupinus seed storage protein, by a mechanism not yet been clarified. Structural studies established that binding of γ‐conglutin, in native and denatured form, to insulin occurs by a strong binding that resists rupture when 0.4 M NaCl and 0.4 M galactose are present, suggesting that strong electrostatic interactions are involved. Studies on binding of γ‐conglutin in native and denatured form to HepG2 membrane glycosylated receptors were conducted, which reveal that only the native form of γ‐conglutin with lectin activity is capable of binding to these receptors. Glycosylated insulin receptors were detected on purified HepG2 cell membranes and characterized by 1D and 2D analyses. Preclinical assays with male mice (CD‐1) indicated that native and denatured γ‐conglutins display antihyperglycemic effect, decreasing glucose in blood comparable after 120 min to that exhibited by the animal group treated with metformin, used to treat T2D and used as a positive control. Measurement of organ injury/functional biomarkers (hepatic, pancreatic, renal, and lipid profile) was comparable to that of metformin treatment or even better in terms of safety endpoints (pancreatic and hepatic biomarkers).

ity but with no consensus about of its mechanism of action. In in vitro models, it exhibits hypoglycemic activity via glucose reuptake increase by HepG2 cells, with internalization and multiple phosphorylation (Capraro et al., 2013;Lovati et al., 2011), insulin binding , increased glucose transport GLUT4 translocation in C2C12 myoblastic cells (Terruzzi et al., 2011), and also lowered glucose levels in in vivo hyperglycemic rat models  and animal models with impaired glucose metabolism (insulin resistance and type 2 diabetes model) (González-Santiago et al., 2017).
There are many gaps in the knowledge of the hypoglycemic mechanism exhibited by γ-conglutin in type 2 diabetes mellitus.
Some studies reported that γ-conglutin binds to insulin in an electrostatic-dependent manner , whereas others suggested an insulin-mimetic action in mouse myoblasts (Terruzzi et al., 2011). Also, the relationship among insulin, insulin receptors, and glucose transporters GLUT2 was studied (Sandoval-Muñíz et al., 2018) and led the authors to conclude that γ-conglutin upregulates Slc2a2 gene expression in liver and normalizes GLUT2 protein content in the pancreas of streptozotocin-induced diabetic rats. In in vitro studies, in which γ-conglutin hydrolysates were tested in mature 3T3-L1 adipocytes, they significantly increased GLUT4 translocation to the plasma membrane compared with untreated control cells, meaning that increased glucose uptake was stimulated by L. mutabilis seed protein hydrolysates, thus indicating that glucose internalization into the cells could be mediated by insulin-dependent GLUT4 (Muñoz et al., 2018).
Diabetes is a chronic disease caused by metabolic disturbances that leads to an excess of glucose in blood and represents a serious worldwide epidemiologic problem. The aim of this work was to contribute to the improvement of knowledge on the γ-conglutin mechanism of action based on its lectin properties, which has not yet been described. Our results concerning the antihyperglycemic activity of γ-conglutin in in vitro (HepG2 cells) and in vivo (mice) models revealed new and promising data. The antihyperglycemic effect, as a result of structural modifications of γ-conglutin, and its lectin activity on insulin and insulin receptors binding to HepG2 cell membranes as molecular targets were explored. Also, the lectin activity effect on biomarkers of different organ functions was evaluated to complete the antihyperglycemic study. The results obtained suggest that γconglutin in native or denatured form could be a new pharmacologic tool in the management of diabetes and might have a role in the development of new therapies.

| Plants
Dry mature seeds from Lupinus albus cv amiga were kindly supplied by Dr. J.N. Martins (University of Lisbon, Lisbon, Portugal).

| HepG2 cells
HepG2 cell lines of hepatocellular carcinoma  were used in in vitro assays.

| Seed protein extraction
Seeds from Lupinus albus cv amiga were assayed.

| Isolation of albumins and total globulins based on solubility criteria
The seeds from Lupinus albus cv amiga were powdered, and the meal was defatted with n-hexane (34 ml/g of flour) for 4 hr with agitation and air-dried after decantation of the hexane. The albumin and globulin fractions were extracted according Ribeiro et al. (2014).
The albumins were extracted by stirring the powder for 4 hr at 4℃ in water (pH adjusted to 8.0) containing 10 mM CaCl 2 and 10 mM MgCl 2 (34 ml/g of dry mass). The insoluble proteins were removed by centrifugation at 30,000 g at 4℃ for 1 hr, and the albumin fraction was retained in the supernatant. For globulin extraction, the pellet was solubilized in 100 mM Tris-HCl buffer, pH 7.5, containing 10% (w/v) NaCl (34 ml/g of dry mass), and stirred for 4 hr at 4℃. The globulin-containing solution was centrifuged for 1 hr at 30,000 g, and the globulins remaining in the supernatant were subsequently precipitated by the addition of ammonium sulfate 80%. The precipitated globulins were centrifuged at 30,000 g for 20 min, suspended in 50 mM Tris-HCl buffer, pH 7.5, and desalted on PD-10 columns (GE Healthcare Life Sciences), previously equilibrated in the same buffer. All operations were performed at 4℃.

| γ-conglutin purification
A DEAE-Sepharose matrix was performed in batch to achieve the purification of γ-conglutin from the globulin fraction. The globulin fraction was mixed with DEAE matrix with a 50 mM Tris-HCl buffer, pH 7.0 (3.25 g of matrix/5 ml of volume), for 1 hr, at room temperature (25s). The matrix and the globulin fractions were mixed by stirring every 15 min, during 1 hr, to maximize the binding of the other protein contaminants, leaving γ-conglutin in the supernatant.
Finally, the matrix was regenerated with 50 mM Tris-HCl buffer, pH 7.0, containing 1 M NaCl. For albumin fraction, the same protocol was performed.

| Polypeptide and lectin characterization
2.3.1 | Polypeptide profile The discontinuous buffer system described by Laemmli (1970) was used for polyacrylamide gel electrophoresis (PAGE). Two types of electrophoresis were used, namely, nonreducing sodium dodecyl sulfate-PAGE (SDS-PAGE-NR) and reducing SDS-PAGE (SDS-PAGE-R), following the methodology previously described (Santos et al., 1997). Before electrophoresis, all protein samples were boiled for 3 min in the presence of sample buffer containing 2% (w/v) SDS (SDS-PAGE-NR) or 2% (w/v) SDS and 0.1 M β-mercaptoethanol (SDS-PAGE-R). Gels were stained with silver nitrate (Blum et al., 1987).

| Native gels
An electrophoretic run in native conditions was also performed.

| Tris-tricine 16.5% precast gels
The precast gels (Bio-Rad) were used in the in vitro assays performed with the human insulin. After the electrophoretic run, which was performed following the same methodology described by Santos et al. (Santos et al., 1997), the gel was placed on a different fixing solution in order to fix the proteins with a smaller molecular weight to the gel. The solution was composed of 50% (v/v) methanol, 10% (v/v) acetic acid, and 100 mM ammonium acetate.
The gel was placed on this solution for 1 hr, and a silver staining was performed afterward.

| Glycoprotein Detection
The glyosidic character of the polypeptide constituents of albumin and globulin fractions was carried out after electrophoretic run (SDS-PAGE-R), followed by its transfer and immobilization in nitrocellulose membrane, by the concanavalin A-Peroxidase method, proposed by Faye & Chrispeels. (1985).

Erythrocyte cells
Rabbit erythrocytes from total blood were treated according to Ribeiro and colleagues (Ribeiro et al., 2012). 5 ml of blood was washed three times in saline and incubated with trypsin for 1h at 37℃ at a final concentration of 1% (v/v). The 4% (v/v) suspension of the trypsinized erythrocytes was stored at 4℃ and used for the hemagglutination activity measurements.

Production of polyclonal antibodies
Polyclonal antibodies were produced in rabbits against γ-conglutin from Lupinus albus, according to Seabra et al. (1999). Samples containing the purified antigens (200 µg) were mixed with an equal volume of Freund's complete adjuvant (1 ml final volume) and injected subcutaneously into female New Zealand rabbits. To obtain a high titer, three booster injections were given every 2 weeks in complete Freund's adjuvant diluted 1:10 with incomplete adjuvant. After the third booster injection, total blood was taken from the heart 9 days after. Blood samples were allowed to clot, and the serum (with total Ab) was collected and stored frozen at −70℃.

Immunodetection
To confirm the purification of γ-conglutin, an immune detection was performed (Ribeiro et al., 2014). The obtained eluted DEAE matrix was placed on a SDS-PAGE-NR, and an electrophoretic run was performed with controls (albumin and globulin fractions) and blotted onto nitrocellulose membranes for immunodetection with polyclonal antibodies prepared against γ-conglutin.
After protein transfer onto nitrocellulose, the membranes were fixed as described above, then washed for 10 min in PBST (20 mM phosphate buffer, 140 mM NaCl, 20 mM KCl, and 0.05% v/v Tween-20, pH 7.4) prior to blocking for 1 hr in 1% (w/v) dry skimmed milk in PBST. Specific antibodies (antiγ-conglutin from lupinus) were used at 1:750 dilutions in PBST, followed by an incubation of 1 hr at room temperature with gentle shaking. Membranes were washed in PBST twice for 5 min prior to the addition of the secondary antibody specific for rabbit IgGs (A3812; Sigma-Aldrich), conjugated to alkaline phosphatase, and used at 1:7.000 dilution in PBST. Detection of the immune complex was achieved by hydrolysis of 5-bromo-4-chloro-3-indolyl phosphate (BCIP). This reaction was stopped by addition of water. The blot was rinsed twice for 30 s in water and allowed to dry at room temperature.

| γ-conglutin binding assays to human insulin
Native γ-conglutin and denatured γ-conglutin (1,000 µg for each sample in 25 mM Tris-HCl, pH 7.2, containing 2 mM CaCl2 and 2 mM MgCl2) were incubated with 1,250 µg of human insulin solubilized in the some buffer, with a total reactional volume of about 3 ml. After incubation, in order to remove insulin unbound to native γ-conglutin and denatured γ-conglutin, was done one prior centrifugation at 12.096 × g for 10 min, followed by a cycle of three consecutive washes of the sediment, with 15 volumes of saline containing salts, by centrifugation at 12.096 g for 10 min at 4℃ (Beckman J2-21 641 m/E, Rotor JA 20 000). The supernatant was recovered, and the pellet, consisting of the insulin bound to gamma conglutin, was subsequently solubilized in 0.4 M of NaCl solution, pH 7.2. Aliquots of supernatant and pellet solubilized was measured for protein measurement, to detect whether the insulin bounded to native and denatured γ-conglutin.

| Galactose affinity assays
In parallel was evaluated the effect of galactose 0.4 M (in Tris-HCL 25 mM, pH 7.2) solution on the dissociation of the pellet formed by γ-conglutin and insulin, obtained after incubation. In order to clarify this possible effect, the pellet was solubilized in the galactose solution and incubated for 30 min at 25ºC with slight agitation, followed by a centrifugation at 12.096 × g for 10 min. If a pellet was visualized, it was solubilized in NaCl 0.4 M, pH 7.2, and the supernatant was passed through a PD-10 column previously equilibrated in the same solution to eliminate galactose. The solubilized pellet was treated in the same manner to eliminate galactose, which is performed with the hemagglutination assay.

| HepG2 cell culture
The HepG2 cells were seeded in plastic flasks and cultured in culture media with 50% of Dulbecco's modified Eagle's medium (DMEM) with low glucose content and 50% of Nutrient Mixture F-12 Ham and supplemented with 10% (v/v) fetal calf serum, 0.5% (v/v) of penicillin solution at 2 × 104 UI/ml, and 34 mM streptomycin (Oliveira et al., 2019). The cells were kept in a humidified atmosphere containing 5% CO2. The cells were supplied with fresh medium every second day, and trypsinized when the confluence was between 80% and 100%.

| Isolation of HepG2 cell membranes
HepG2 cells were grown as described previously, and the membranes were isolated by the method described by Vercoutter-Edouart et al (Vercoutter-Edouart et al., 2008). The cells preserved at −80℃ (26 × 10 6 cells) were thawed rapidly at 37℃, and the suspension of cells was washed with 10 volumes of HES buffer (20 mM HEPES, pH 7.4, and 250 mM sucrose), by centrifugation at 750 g for 10 min, at 21℃ (Beckman J2-21 m/E, Rotor JA 20,000). The supernatant was discarded, and the pellet was washed twice with HES buffer containing a protease inhibitor cocktail (without EDTA, Roche). Cell lysis was performed by cryolysis, in which cells were subjected serially to freezing and thawing cycles (4 times), during 30 min, at −20℃, and added from a 20-min sonication in ultrasound, followed by a centrifugation at 960 x g for 10 min at 4℃, with discard of the pellet. The final supernatant was ultracentrifuged at 126,000 x g for 45 min, 4℃ (Beckman J2-21 m/E, Rotor SW 32 Ti). The pellet containing the cell membranes was solubilized in 2 ml of physiological saline (0.9% NaCl), containing 2 mM of CaCl 2 and 2 mM of MgCl 2 , and was divided into aliquots containing 1 mg protein determined by the Bradford method (Bradford, 1976) and kept at -80℃, until use.

| γ-Conglutin affinity binding to HepG2 cell glycosylated receptors
A protocol according to Oliveira et al. (2019) was followed. The isolated HepG2 cell membranes were individually incubated with the native and the denatured γ-conglutin. Seven hundred (700 µg) of HepG2 protein membrane was solubilized in 3.5 ml of saline containing 2 mM CaCl 2 and 2 mM MgCl 2 and incubated during 45 min at 25℃, by gentle agitation, with 280 µg of native γ-conglutin and denatured γ-conglutin, both dissolved in 2 ml of saline containing 2 mM CaCl2 and 2 mM MgCl2. After incubation, in order to remove the protein unbound to membranes, was done one prior centrifugation at 12.096 g for 10 min, followed by a cycle of three consecutive washes of the pellet with 15 volumes of saline containing salts and centrifugation at 12.096 x g for 10 min at 4℃ (Beckman 684 J2-21 m/E, Rotor JA 20,000). The supernatant was discarded, and the obtained pellet, consisting of a complex formed between protein and membranes, was subsequently solubilized in saline solution (containing salts) and used, after protein measurement, to detect whether the native and denatured γ-conglutins were bounded to the membranes. The control was made by using saline instead of native and denatured γ-conglutins in the membrane incubation. The complex binding was visualized by 1D electrophoresis.

| Proteomic profile of HepG2 Cell membranes by 2D analysis
Two-dimensional (2D) electrophoresis (IEF/SDS-PAGE-R) of the cell membranes of HepG2 was carried out as follows, according to Oliveira et al. (2019). It was applied in the first dimension, isoelectric focusing (IEF), 1,400 µg of HepG2 membrane protein, separated using the IPGphor system (Amersham Pharmacia).
The second dimension (SDS-PAGE-R) was performed after IEF.

| Animals
Male CD-1 mice with 30-40 g body weight and 6-10 weeks old (Instituto de Higiene e Medicina Tropical) were used in the hyperglycemic study.
Animals were accommodated in polypropylene cages with free access to water and food and kept at a temperature of 18-23ºC and humidity of 40%-60% in a 12-hr light/dark cycle at the Animal Facility of the Faculty of Pharmacy (University of Lisbon).
Experiments were performed agreeing to the most recent rules and recommendations for the care and processing of laboratory animals, namely to the presently adopted European Commission regulations (Directive 2010/63/EU).
In addition, the studies were performed in agreement with the

| Administration
Animals were administered daily by gastric gavage during 7 days with the last administration being at the final day, 30 min prior to basal glycemic measurement and hyperglycemic induction.
Animals were randomly allocated to five experimental groups: Normoglycemic group-animals were treated orally with vehicle for 7 days, and at the last day, water was administered instead of glucose; Hyperglycemic group-animals were treated with vehicle for 7 days, and at the last day, a solution of glucose (65 mg/kg) was administered by gastric gavage; Native γ-conglutin group-animals were treated orally with γconglutin (60 mg/kg BW) for 7 days, and at the last day, a solution of glucose (65 mg/kg) was administered by gastric gavage; Denatured γ-conglutin group-animals were treated orally with denatured γ-conglutin (60 mg/Kg BW) for 7 days, and at the last day, a solution of glucose (65 mg/kg) was administered by gastric gavage; Metformin group-animals were treated orally with metformin (300 mg/kg BW) for 7 days, and at the last day, a solution of glucose (65 mg/kg) was administered by gastric gavage.

Experimental hyperglycemia
A basal measurement was performed at t0 immediately before administration of the glucose solution, and subsequent measurements of glycemia were performed at t30, t60, and t120 after glucose administration. Glycemic measurements were performed by collection of a blood drop by means of tail vein puncture and determined with a portable glucometer.

Blood collection
At the end of the experiment, animals were anesthetized by administration of a mixture of ketamine:xylazine (80 mg/kg:8 mg/kg) and blood was collected by cardiac puncture for measurement of pancreatic (insulin, amylase, and lipase), renal (creatinine and urea), hepatic (alanine aminotransferase and aspartate aminotransferase), and lipid (total cholesterol, HDL-cholesterol, LDL-cholesterol, and triglycerides) serum biomarkers. All these markers were evaluated in a clinical analysis laboratory certified under the ISO9001-2015 standard.

| General assays
Protein quantification was made by a modification of the Bradford method (Bradford, 1976), with bovine serum albumin used as the standard. SDS-PAGE, PAGE native, and 2D gels were stained by CBB R-250, CBB G-250 (Neuhoff et al., 1988), or silver stained (Blum et al., 1987).

| Statistical analysis
All results were expressed as mean ± S.E.M. of n observations, where n represents the number of animals studied. Results were compared using a two-factorial ANOVA test, followed by a Bonferroni's post hoc test using GraphPad Prism 5.0 software (GraphPad). A p value less than 0.05 was considered to be statistically significant. In order to confirm γ-conglutin lectin character, the hemagglutination activity was measured by a hemagglutination assay. For amounts of protein tested 15 and 30 µg, a hemagglutination unit (H.U.) of 0.14 µg ± 0.05 µg was obtained for purified γ-conglutin isolated from the globulin fraction, signifying that this value is the minimal protein concentration that produces erythrocyte hemagglutination. The controls used were saline as the negative control and concanavalin A (500 µg/ml) as the positive control.

| γ-Conglutin denaturation
In order to denature γ-conglutin, the purified protein was exposed to a denaturing "cocktail" associated with deglycosylation conditions. In an attempt to understand the consequences of incubating γ-conglutin with this "cocktail," an electrophoretic analysis under native conditions was performed, with native γ-conglutin as the control. F I G U R E 1 γ-Conglutin purification (a) Supernatants containing γ-conglutin (Cγ) from DEAE-Sepharose, after different contact times with matrix and globulin fraction. Polypeptide profile by a SDS-PAGE-NR, 17.5% (m/v) acrylamide revealed by silver staining. It was applied 7 µg of total globulin fraction (Glb) as control and γ-conglutin (Cγ) and 3 µg of molecular markers (M). (b) Immunodetection in a nitrocellulose membrane using antiγ-cong. (lupinus) as primary antibody and anti-rabbit IgG as secondary antibody. It was applied 15 µg of globulins (Glb) and albumins (Alb) as control, 10 µg of purified γ-conglutin supernatant (Cγ), and 12 µl of molecular weight (M)

| Binding capacities to human insulin
To study the binding capacities to human insulin, both native and denatured γ-conglutins were incubated with insulin, having ensured previously the solubilization of all molecules (Materials and Methods section). For each γ-conglutin (native with lectin activity and denatured), two incubations were performed with insulin and the resulting pellet was collected for each one. The first incubation of native γ-conglutin incubation with insulin results in a pellet, which was  As the glycomic profile is very important to characterize lectin glycoreceptors, the study of the glycoprotein content in the HepG2 cell membranes was done. The purified HepG2 cell membrane polypeptides were separated by SDS-PAGE-R, which was used to transfer the polypeptide profile of these membranes to a nitrocellulose membrane. In this assay, the positive control used was mucin, a well-known glycoprotein. As shown in Figure 4c, the HepG2 cell membrane polypeptide profile is poor in glycoproteins. Only a few glycosylated bands were revealed, and the most representative ones exhibit high molecular weights. The insulin receptors are marked in Figure 4c with an arrow highlighting the polypeptide bands of 95 and 130 kDa.

| Binding of native and denatured γ-conglutin to HepG2 membrane receptors
The study of the lectin character on the binding capacity of γconglutin (with and without lectin activity) to HepG2 cells was evaluated. An incubation of HepG2 cell membranes with three forms of γ-conglutin was analyzed: native γ-conglutin without lectin activity (batch 1), native γ-conglutin with lectin activity (batch 2), and denatured γ-conglutin (see Table 1). The protein profiles were analyzed using purified HepG2 cell membranes and native γ-conglutin as controls. Figure 5a shows that after incubation of

| In vivo evaluation of the hypoglycemic effect of γ-Conglutin
The antihyperglycemic effect of γ-conglutin was evaluated in male mice (CD-1) by an in vivo protocol described in the Materials and Methods section. The data showed that native γ-conglutin (Cγ-Nat)

Pancreatic functional biomarkers
The results obtained for pancreatic biomarkers when γconglutin is administered to mice are evidenced in Figure 7. and there is no significant difference among them.

Liver injury biomarkers
The results for the liver biomarkers when γ-conglutin is administered to mice are shown in Figure 8. The biomarkers evaluated were alanine aminotransferase (ALT) (Figure 8a) and aspartate aminotransferase (AST) (Figure 8b). For the ALT, the results showed that the Cγ-DT group presented values similar to the normoglycemic control group while the Cγ-Nat group exhibited 10% lower values and metformin group 10% higher values.
The AST value obtained for the groups treated with native (Cγ-Nat) and denatured γ-conglutin (CyDT) was lower than that of the normoglycemic group control, although not significantly, and much lower than the group treated with metformin, about 50%, and 40% lower, respectively. The hyperglycemic control group and the metformin group revealed similar AST levels, being the highest among all groups.

Renal functional biomarkers
The renal functional biomarkers evaluated in this assay were creatinine and urea, and the results are displayed in Figure 9a

Lipid profile biomarkers
The lipid profile biomarkers studied were total cholesterol, HDL, LDL, and triglycerides (TG), and the results for all these biomarkers are exhibited in Figure 10a,b,c,d, respectively. The results for total

F I G U R E 6
In vivo evaluation of the hypoglycemic effect of native γ-conglutin and denatured γ-conglutin. Five mouse groups were tested: hyperglycemic group control, normoglycemic group control, group exposure to native γ-conglutin, group exposure to denatured γ-conglutin, and group exposure to metformin. It was tested 60 mg/kg BW CD-1 mouse of the native γ-conglutin and denatured γ-conglutin and 300 mg/kg BW for the metformin group *P<0,001 vs Hyperglycemic cholesterol levels revealed that this biomarker value for the normoglycemic group control is similar to the groups treated with the native and denatured γ-conglutin (slightly lower). The metformin group and the hyperglycemic group control showed very similar values of the total cholesterol content. Regarding β-ME (Figure 10b), the results showed that, for this biomarker, the same values were obtained for the normoglycemic and hyperglycemic group controls and also for the group treated with metformin. The groups treated with denatured and native γ-conglutin revealed higher values of HDL, but the difference between these values is not significant in comparison with the other groups evaluated.
The group with the lowest LDL levels is the group treated with denatured γ-conglutin (230 mg/dl), although this result is not significantly different than the one obtained for the normoglycemic group control, having increased when compared to the hyperglycemic group. The groups treated with native γ-conglutin (Cγ-Nat) and metformin exhibited very similar values of LDL. This value is lower than that obtained for the hyperglycemic group control and higher than that obtained for the normoglycemic group control (300 mg/ dl). The hyperglycemic group control has the highest value (550 mg/ dl) of LDL values.
Regarding TG, results revealed no difference between normoglycemic, hyperglycemic, and CyDT group values (≈230 mg/dl). The group treated with native (Cγ-Nat) exhibited values higher (≈290 mg/ dl) than the control groups, but the difference between these is not significant. The group treated with metformin revealed the lowest value of TG content (≈200 mg/dl).

| D ISCUSS I ON
Diabetes is a chronic disease that affects nearly 420 million of people in all world (Laemmli, 1970), being one of the leading causes of death (World Health organization). Type 2 diabetes mellitus (T2D) is characterized by insulin resistance and the inability of the β cell to compensate for it (Zatterale et al., 2020).
In the last years, plant-based research has been improving via increased ethnobotanical and ethnopharmacological knowledge, more precisely in legume seeds consumed on a daily basis. This allowed for the establishment of a relationship between chronic intake and exhibition of nutraceutical and functional behavior, which resulted in health benefits (Arnoldi et al., 2015;Guzmán et al., 2019;Scarafoni et al., 2007). Lupinus is a leguminous plant whose seeds have been consumed as a functional food and revealed pharmacologic activity with hypoglycemic effect, evidenced by different metabolic mechanisms, such as enhanced glucose reuptake and reduction in gluconeogenesis in vitro Muñoz et al., 2018). This is of importance as it can be used in the managements of diabetes through modulation of different targets (insulin, glycosylate insulin receptors, and glucose transporters). Studies with animals and humans indicate that moderate consumption of seeds of the genus Lupinus or their derived and toxic alkaloids has positive effects on hyperglycemia and glucose homeostasis (Wink, 2005. γ-Conglutin is a seed protein exclusive to seeds of Lupinus species with documented pharmacologic activity regarding the hypoglycemic activity in both in vitro and in vivo models and human studies.
The first aim of this work was to understand the influence of the lectin character of γ-conglutin on its hypoglycemic activity. Until now, no study has been reported based on this property in order to understand whether γ-conglutin hypoglycemic activity remains effective when its lectin activity is absent. This fact is highly relevant as it can influence the behavior of γ-conglutin binding to target molecules such as insulin, membrane insulin receptors, or glucose transporters, among others. Native γ-conglutin may have specificity for glycan components in these target molecules. Insulin is the only molecule among the others studied in this work that is not glycosylated, meaning that γ-conglutin lectin activity is not directly involved in insulin binding.
The binding studies carried out between human insulin (MW 6-10 kDa) and: (a) purified native γ-conglutin with lectin activity (H.U. = 0.14 ± 0.05 µg); (b) γ-conglutin without lectin activity (Table 1), and (c) denatured γ-conglutin, which revealed that the presence of lectin activity is not required for γ-conglutin binding to insulin (Figure 3a-P1), as also its native structure is not demanding (Figure 3b-P1), meaning that the insulin binding capacity of γ- From Figure 3b, it is observed that the aggregates in denatured γconglutin (Figure 3b-P1 and PG) could be due to the structural pH changes where the incubation medium has a pH near 7.5 resulting in oligomeric forms of γ-conglutin of 100, 250, and 480 kDa, as also hexameric forms (Capraro et al., 2010).
It has been described that γ-conglutin binds to insulin via electrostatic interactions, by an unknown mechanism, which required its native conformation to established this bound . peptide of insulin β-chain (Sparrow et al., 2007), suggesting that γconglutin may show specificity to these domains as it is galactosespecific (Ribeiro et al., 2014). Others authors (Klein et al., 2009) reported that the accumulation of INS-1 Ser1011 GlcNAc in HepG2

liver cells and MC3T3-E1 preosteoblasts upon inhibition of O-
GlcNAcase indicates that O-GlcNAcylation of endogenously expressed INS-1 is a dynamic process that occurs at normal glucose concentrations (5 mM).
In this work, HepG2 cell membranes purified by an optimized methodology (Oliveira et al., 2019;Vercoutter-Edouart et al., 2008) were analyzed by 1D and 2D (IEF/SDS-PAGE-R) electrophoresis for proteomic profiling (Figure 4a,b) analysis, followed by glycodetection of polypeptide profile of these membranes (Figure 4c), for INS-1 detection as identified in Figure 4a cell membranes, or/and by an intrinsic binding mechanism with internalization (Ribeiro et al., 2018;Fu et al., 2011) via mitochondria. In this study, only the extrinsic mechanism is evidenced, although the intrinsic mechanism can also occur. The internalization of γ-conglutin has already been described as the multiple phosphorylation suffered by this molecule (Capraro et al., 2013), suggesting that both mechanisms participate in hypoglycemic activity.
From the point of view of glucose transporters, insulin binding to cell membrane receptor initiates a series of cascading phosphorylation/dephosphorization reactions, which results in cellular metabolic changes, necessary after binding of this hormone to cells to increase glucose absorption by, as well as expression, recruitment and translocation of glucose transporters such as GLUT2 (Ohtsubo et al., 2013) and GLUT4 (Saltiel & Kahn, 2001). It has been reported that cell stimulation (rat myoblasts) with γ-conglutin led to decreased blood glucose and to the recruitment and translocation of GLUT4 (the primary effects of binding insulin to its membrane receptor) (Terruzzi et al., 2011).
The glucose transporters are also glycosylated molecules with a single N-linked oligosaccharide (Mueckler & Thorens, 2013). Concerning GLUT2 in hepatic tissue, it is the most expressed in beta cells, basolateral surface of kidney, hepatocytes, and small intestine epithelia (Orci et al., 1990;Thorens et al., 1990), with expression being regulated by sugars and hormones (Leturque et al., 2009). Patients with diabetes mellitus and with prolonged hepatitis C virus (HCV) infection have been related to viral-induced reduction in hepatocyte expression of GLUT2 (Kasai et al., 2009), which exhibit an N-glycan moiety for which GnT-IV, a glycosyltransferase is required. GLUT2-mediated glucose sensing is essential for maintaining normal glucose-stimulated insulin secretion in pancreatic beta cells. The N-glycan structure acts as a ligand for galectins to form the glycan-galectin lattice that maintains the stable cell surface expression of GLUT2, and cellular glucose transport activity (Ohtsubo et al., 2013). Knowing that galectins are lectins and are β-galactoside-binding proteins sharing homology in the amino acid residue sequence of their carbohydrate-recognition domain, and that several glycan ligand recognition in GLUT exists for galectins, as well as disaccharide Galβ1-4GlcNAc and other substrates with substitution at O-2 and O-3 of galactose residue, as well as core fragments ("right" from GlcNAc), that provides significant increase in affinity (Rapoport & Bovin, 2015), we could hypothesize that γ-conglutin may establish a competition binding mechanism with galectins in order to bind to N-glycan structure of glucose transporters that exhibit galactose domains, resulting in an antihyperglycemic effect.
The in vivo assays provided plausible information by doseresponse about different molecular targets involved in the hypoglycemic effect observed in animal models, after γ-conglutin exposure.
Previous works in hyperglycemic animal models reported that γconglutin decreased plasma and serum glucose levels (González-Santiago et al., 2017;Lovati et al., 2011;Magni et al., 2004;Sandoval-Muñíz et al., 2018;Terruzzi et al., 2011), with exhibition of some molecular targets as it was evidenced by Vargas-Guerrero et al. (2014), when STZ animal models exposed with γ-conglutin showed benefic effects due to reductions in glucose, increments in serum insulin, and increases in Ins-1 gene expression and beta cell insulin content compared with the STZ control group. A recent study (Sandoval-Muñíz et al., 2018) described that after γ-conglutin administration to streptozotocin-induced rats, a slight increase in Slc2a2 and Pdx-1 mRNA levels in pancreas and up-regulated Slc2a2 expression in the liver were observed, but with no effect on hepatic glucokinase (GCK) expression, and with normalized GLUT2 protein content in pancreas of the of streptozotocin-induced rats (Sandoval-Muñíz et al., 2018). Also, modification of the gene expressions of enzymes G6pc and Fbp1 (fructose-bisphosphatase 1 gene), involved in glucose hepatic production in vivo, and also Pck1 (phosphoenolpyruvate carboxykinase 1 gene) was studied in two experimental animal models of impaired glucose metabolism (insulin resistance IR and STZ models), where, after γ-conglutin treatment, G6pc (glucose-6-phosphatase gene) expression was decreased in the IRγ-conglutin and STZCγ groups. Post-treatment, Fbp1 and Pck1 expressions were reduced in the IRγ-conglutin group but increased in STZγ-conglutin animals. The authors suggested that γ-conglutin is involved in reducing hepatic glucose production, mainly through G6pc inhibition in impaired glucose metabolism disorders (González-Santiago et al., 2017).
In this work, the animal groups were established to test the antihyperglycemic activity of different structural forms of γ-conglutin, in order to produce reliable and robust results, in agreement with the 3Rs policy. The results obtained in this hyperglycemic male mouse model aimed to evaluate serum glucose levels by the use of glucose oral tolerance test (GOTT) and the measurement of organ injury/ functional biomarkers.
Mice were administrated by oral gavage with a dose of 60 mg/ kg BW, selected according to previous reports , of both native γ-conglutin with lectin activity and denatured γconglutin. GOTT results showed that both conglutin forms exhibited antihyperglycemic effect although the denatured form originated better results than the native γ-conglutin (exhibiting lectin activity), evidenced 30 min after glucose administration, whereas glucose values after 120 min were the same for all hyperglycemic groups (untreated and treated). The antihyperglycemic effect of γ-conglutin was already reported; however, it is the first time that the primary structure of γ-conglutin, rather than its native form, is shown to be one key factor for its antihyperglycemic activity, leading to the interpretation that, after physiological digestion, its effectiveness remains. It is possible that both γ-conglutin forms exhibit antihyperglycemic activities, which are established by different mechanisms.
Native γ-conglutin has lectin activity, and in this state, it could bind to cell membrane-glycosylated receptors, producing a mimetic effect of insulin in glucose reuptake. Also, a mechanism similar to the galectins could be responsible for binding to glucose transporters stimulating insulin secretion in pancreatic beta cells. The denatured form of γ-conglutin could exhibit an antihyperglycemic activity by a mechanism that could be explained by binding to insulin, although this result needs to be further studied.
Several serum biomarkers with importance in the screening of T2D were evaluated to understand the behavior of Lupinus conglutins in this etiology, ruling out possible safety issues. For the pancreatic function evaluation, the biomarkers studied were insulin, amylase, and lipase. For amylase and lipase (Figure 7a,b) all animal groups exhibited similar results to those obtained for the normoglycemic group, with the exception for insulin values (Figure 7c), where the metformin group and both conglutins showed higher values (around 8.5 mU/l) than the hyperglycemic group (7m U/l), with this group showing a great increment compared with the normoglycemic group (1.5 mU/l). These markers reflect the normal state of the pancreas, and these results indicate the lack of toxicity to the molecules used in the assay, highlighting the increase in insulin production at a glucose overload, comparable to the result obtained for metformin.
Regarding biomarkers for hepatic function, the ALT and AST results are presented in Figure 8 (a,b). For both biomarkers, marginal changes, although significant, were observed in serum levels of the hyperglycemic group when compared to the normoglycemic group. In general, both conglutin forms were able to attenuate those changes, with results not statistically different from the normoglycemic group. The metformin group was the one that had higher variation of these enzymes when compared to the normoglycemic group, although the magnitude of the changes is small and does not necessarily indicate any toxic effect by rather a consequence of direct pharmacodynamics effects in the liver function.
The effect on renal function by measurement of creatinine and urea is presented in Figure 9a,b. Induction of hyperglycemia led to a reduction in creatinine levels, although given the magnitude of the effect, it can be assumed to be a transitory physiological compensation without specific relevance. Other than the Cγ-Nat group (in which no significant change was observed compared with the normoglycemic group), all other groups exhibited the same pattern of reduction. For urea, all groups exhibited no significant differences, as all groups originated results comparable to the normoglycemic group (between 34-38 U/L).
For total cholesterol (TC), HDL, LDL, and triglycerides (TG), regarding lipid profile evaluation, denatured γ-conglutin (Cγ-DT) was the experimental group that, in general, was associated with less variability in lipid profile, when compared to the normoglycemic group.
The result with the highest magnitude of change was associated with the increase in LDL-cholesterol levels in the hyperglycemic group, and increase that was inhibited by all experimental groups. Although these markers were performed solely with the aim of clarifying preliminary effects in several physiological parameters, specific studies in hypercholesterolemic models might enlighten some of the results obtained and complement already existing literature regarding the effect of Lupinus albus in dyslipidemia (Sirtori et al., 2004).

| CON CLUS IONS
This work reveal that in the modulation of diabetes exerted by γconglutin, several mechanisms are involved in order to regulate blood glucose levels. The display of lectin activity, associated with the native structure, as well as the activity shown by the primary structure of γ-conglutin, could be reflected in its antihyperglycemic activity. These capacities were involved in the binding of native γ-conglutin to HepG2 cell glucose receptors, including insulin receptors, as well as their binding to human insulin, but also with the ability of denatured γ-conglutin to bind to insulin.
The skills of γ-conglutin give a versatility in its action mechanism.
These skills were evidenced in in vivo experiments where the denatured form of γ-conglutin had better results than native form, at OGTT (oral glucose tolerance test) and in pancreatic, hepatic, and dyslipidemic biomarkers. Also, the native γ-conglutin reveals its antihyperglycemic effect, which seems to be necessary to exhibit its lectin activity effect.
In conclusion, native γ-conglutin with lectin activity and denatured γ-conglutin seem to have therapeutic potential, sometimes can rival with the institutional drug metformin.

ACK N OWLED G M ENT
Authors thank to "Dr Joaquim Chaves, Lab Análises Clínicas" for the support for publishing this research work.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest to publish the results.