Isolation of a Hemagglutinating γ-conglutin from Seeds of the Andean Legume Lupinus mutabilis Sweet (Tarwi)

Lupinus mutabilis, known as tarwi, is one of the main legumes cultivated by the Andean populations and the only lupin species domesticated in the Andes (Atchison et al., 2016). Lupinus mutabilis has a wide weather adaptability and its seeds have superior grain quality and a high protein content (44%) in comparison to the other three domesticated lupin species (L. albus, L. angustifolius and L. luteus), which have a protein content between 34-42% (Gulisano et al., 2019; Gulisano et al., 2022). This legume is a key genetic resource that has many traditional applications such as antiparasitic, analgesic, anti-inflammatory and the treatment of diabetes (Jacobsen and Mujica, 2006). Unfortunately, L. mutabilis breeding is decreasing in South American countries due to replacement by other crops and the inconvenient need to treat L. mutabilis seeds before human consumption in order to reduce alkaloid content (Jacobsen and Mujica, 2006).

Lectins, or lectin-like proteins, are generally found in legume seeds, these proteins have multiple applications in biochemistry, molecular biology and medicine and their presence in L. mutabilis seeds could promote the use of this historically relegated plant. Lectins usually have hemagglutinating activity (HA) because they bind carbohydrates in the membrane of erythrocytes (Sharon and Lis, 2007). This activity has been demonstrated in the extract of mature seeds of L. albus, L. angustifolius and L. mutabilis (Falcon et al., 2000a). However, after the purification of the lectin from the extract of L. albus by affinity chromatography using galactose, the HA was loss. This lectin was reported to be a non-hemagglutinating protein with a circular dichroism spectrum similar to γ-conglutin (Falcon et al., 2000b).

The γ-conglutin is a basic globulin conserved between lupin species and it usually represents 4-5% of the seed protein content (Duranti et al., 2008). Duranti et al. (1995) reported a L. albus γ-conglutin with lectin-like activity and the ability to bind glycoproteins. Recently, an hemagglutinating γ-conglutin from L. albus was isolated (Gracio et al., 2021), but its activity was not constant in different purified batches, indicating that the methodology was not ideal. The hemagglutinating activity (Schoeneberger et al., 1982; Djabayan-Djibeyan et al., 2022) and the affinity for galactose of seed extracts from Lupinus mutabilis was reported (Falcon et al., 2000b), but the purification protocols were not suitable and caused the loss of the hemagglutinating capacity (Falcon et al., 2000b).

Thus, the objectives of the study were to purify and characterize a lectin (γ-conglutin) from an extract of Lupinus mutabilis seeds and to show that ¡-conglutin is responsible of the HA, evaluating this activity at each step of the process.

Lupinus mutabilis Sweet seeds were provided by the Germplasm Bank of the Grain and Oilseed Legumes Program of the Universidad Nacional Agraria La Molina. Seeds of three ecotypes were characterized by its composition and agronomic characteristics according to parameters stablished by Camarena et al., (2013). The ecotypes were cultivated at multiple locations of Peru: Cholo Fuerte seeds (AC TLM 14) were collected in the Oncoycancha locality (10°24'00"S77°23´46"W), Vicos seeds (AC TLM 47) in Vicos peasant community (9°19'39.5" S 77°33'8.1"W) and Patón Grande (AC TLM 12) in Otuzco (7°53'54"S 78°33'45"W).

Dry Lupinus mutabilis seeds were grinded and sifted. The obtained seed powder was defatted with n-Hexane by stirring for 2 h, this step was performed three times. The fat-free powder was dried and incubated overnight in saline extraction solution (0.15 M NaCl, 20 mM MnCl₂, 20 mM CaCl₂) at 10°C. The aqueous phase was separated by centrifugation (12000 x g, 4°C, 15 min) and proteins were precipitated with ammonium sulphate at 80% saturation. The suspension was centrifuged (21150 × g, 4°C, 15 min) and the pellet resuspended in saline extraction solution and dialyzed against the same solution overnight at 4°C (Falcon et al., 2000a, De Amat, 2016). Extracts were frozen (-20°C) and stored until its further use. This and the following procedures were performed in the Faculty of Biological Sciences of Universidad Nacional Mayor de San Marcos between 2015 and 2019.

Whole rabbit blood was washed with saline physiological solution and the erythrocytes pellet was incubated with 0.01% trypsin for 1 h at 37°C (Falcon et al., 2000a, De Amat, 2016). Trypsinized erythrocytes were washed three times with saline and resuspended in N-(2-Hydroxyethyl) piperazine N’-(2-ethanesulfonic Acid) (HEPES) 10 mM (pH 6.8, NaCl 0.15 M) buffer. For hemagglutination, 50 µL of samples of the extracts from the three ecotypes and 50 µL of erythrocytes were mixed in round bottom microplates and incubated at room temperature for 40 min.

Size exclusion chromatography was performed using HEPES 10 mM pH 6.8 (NaCl 0.15 M), which was the optimal buffer for purification. Initial tests were performed with Tris-HCl pH 8. Hemagglutinating fractions were submitted to anion exchange chromatography in DEAE-Sephadex (Life Sciences) in HEPES 10 mM pH 6.8 (and pH 7.2 in preliminary tests). Size exclusion peak I fractions were mixed and subjected to anion exchange chromatography in HEPES 10 mM pH 6.8. Size exclusion and anion exchange fractions were quantified by UV280 nm spectrophotometry and Bradford assay (Bradford 1976). HA was assessed following the methodology above described. 

Sodium dodecyl sulphate-polyacrylamide electro phoresis (SDS-PAGE) (Laemmli, 1970) was performed under reducing and non-reducing conditions in 10% polyacrylamide gels at 100 V for 2 h. Gels were stained with Coomassie G-250 and the molecular weight was calculated by relative migration analysed in ImageJ (Schneider et al., 2012).

SDS-page was performed in non-reducing conditions as above described. Gels were stained with Coomassie blue G-250 and the band of the isolated hemagglutinating protein was excised from the gel and sent to the Pasteur Institute of Montevideo (Uruguay) for the identification by Peptide Mass Fingerprinting (PMF). Excised band was digested with trypsin and subjected to Matrix-Assisted Laser Desorption Ionization Time of Flight Mass Spectrometry (4800 MALDI TOF/TOF) (Pappin et al., 1993). Obtained mass spectre was analysed in Mascot Software with the Swissprot database. 

To perform the hemagglutination inhibition test the following carbohydrates were used: glucose, lactose, galactose, mannose and melibiose were diluted in HEPES 10 mM pH 6.8 buffer. 25 µL of carbohydrates were incubated with 25 µL of 4 hemagglutinating units of the isolated protein for 30 min at room temperature. Following, 50 µL of rabbit erythrocytes were added to the wells. Carbohydrates final concentrations in the well were 50, 10, 2, 0.4 and 0.08 mM.

Determination of the optimal temperature for hemagglutination tests was performed using 10 mM HEPES, pH 6.8 as previously mentioned. The hemagglutination plates containing the purified protein and the rabbit erythrocytes were incubated in temperatures between 4 and 60°C (4, 10, 14, 20, 23, 30, 37, 50 and 60°C) for 1 h. After incubation, the hemagglutinating titers were compared.

To assess the thermostability, the isolated protein obtained by anion exchange was incubated at temperatures between -20 and 90°C (-20, 0, 10, 23, 30, 50, 70 and 90°C) for 30 min. The samples were incubated at room temperature for 20 min and their HA was assessed as previously mentioned.

The demonstration of the effect of a reducing and chelating agent on the hemagglutinating protein was carried up by incubating it with final concentrations of 1, 5, 10 and 20 mM of b-mercaptoethanol and ethylenediamine tetraacetic acid (EDTA) in HEPES 10 mM pH 6.8 for 1 h. After this step, hemagglutination was assessed as previously detailed.

Saline extraction and hemagglutinating protein isolation from seeds
 
Collected Lupinus mutabilis seeds ecotypes presented the phenotype described in literature (Camarena et al., 2013) (Fig 1). Saline extraction from the seeds of the three ecotypes yielded yellow-coloured extracts with a pH between 5-6 and similar amounts of protein quantified by Bradford (100-200 mg protein per 100 g of seed). Extracts from the three ecotypes showed HA on rabbit erythrocytes, with Patón Grande ecotype (PG) presenting the highest hemagglutinating  titer (rank of 16-32) in comparison to Cholo Fuerte and Vicos ecotypes (rank of 8-16). Then, PG ecotype was chosen for the subsequent procedures. Falcon et al., (2000a) reported a hemagglutinating titer of 512 for L. mutabilis crude, which was higher than the overall titer from our ecotypes. However, not Falcon et al., (2000a) or other studies have compared HA between L. mutabilis ecotypes.



The size exclusion profiles for PG showed four peaks (Fig 2A). Proteins from peak I retained their HA when 10 mM HEPES buffer pH 6.8 (0.15 M NaCl) was used, while proteins from peaks II, III and IV did not show this activity (Fig 2A). Villamarin (2016) also found 4 peaks in the chromatographic profile using 2.5 mM phosphate buffer pH 7.6. In this study, the peak IV presented the highest hemagglutinating titer with human group O erythrocytes and the activity was probably caused by other lectins. Again, L. mutabilis ecotypes were not considered in this study. In our study, the first peak of the anion exchange chromatography showed HA (titer of 8), while no activity was found in the peaks with lower absorbance (Fig 2B).

Chromatographic profiles were similar when size exclusion was initially performed using Tris-HCl buffer pH 8 in the initial tests; however, the HA was completely lost (data not shown). When anion exchange was carried out at pH 7.2, the activity was reduced to 50%, with hemagglutinating titers of 4. Therefore, considering the two chromatographic methods applied, it is shown that the isolated lectin has better HA at lower pH (pH 6.8). The pH sensitivity contrasted with the high pH stability observed in lectin-like proteins or lectins from other legumes. For example, lectin from Parkia biglobosa showed optimal HA on rabbit erythrocytes between pH 6-8 and decreased at pH 9-10 (Silva et al., 2012). Similarly, the Dioclea lasiophylla mannose-binding lectin has optimal agglutinating activity on rabbit erythrocyte at pH 8 (Pinto-Junior et al., 2013).


 
Electrophoretic analysis
 
The extract contained multiple proteins (Fig 3A) that were lost during the purification procedure. The lectin isolated in HEPES pH 6.8 buffer had a molecular weight of 46 kDa in non-reducing conditions (Fig 3D). Under reducing condition, it was separated into two bands of 31 and 16 kDa (disulphide-bond subunits) (Fig 3E). Similar results were obtained by Santos et al., (1997) when analysing the g-conglutin of L. mutabilis (Potosí and Inti cultivars), who identified a 42-43 kDa monomer under non-reducing conditions and two 30 and 18 kDa subunits under reducing conditions. Nadal et al., (2011) found that the g-conglutin from L. albus is formed by a 43 kDa monomer under non-reducing conditions and two subunits of 30 and 17 kDa under reducing conditions. It is important to mention that these authors did not perform hemagglutination tests. Thus, our purified lectin had the expected molecular weight of lupins g-conglutin and differed from the small lectin (17 kDa) isolated from L. mutabilis extracts under different purification conditions (Villamarin, 2016).


 
γ-conglutin identification
 
The γ-conglutin isolated in the first peak of the anion exchange chromatography at pH 6.8 presented hemagglutinating titers of 8 (50 mg/mL) quantified by Bradford, equivalent to 0.31 mg of protein per hemagglutinating unit (HU) (2.5 mg in 50 mg/L per well/titer of 8), while Gracio et al., (2021) found lower required amounts of protein per HU (0.14 mg) for L. albus.

Mascot analysis of the mass spectrometry result (MS/MS spectra) for the purified hemagglutinating protein identified the γ-conglutin from Lupinus angustifolius (Gene ID: 109345795) (p<0.05) by the peptide RQLEENLV VFDLAKS. The number of hits is limited by the lack of genomic information for L. mutabilis, but it was another evidence that, in conjunction with the results obtained in the electrophoretic analysis, showed that the purified lectin from L. mutabilis PG ecotype was a g-conglutin lectin (LMC). The methodology used for its purification allowed to preserve its HA, unlike the isolation of γ-conglutin from L. albus, which lost its HA while maintaining its affinity for galactose (Falcon et al., 2000b) and glycosylated polypeptides (Duranti et al., 1995). Gracio et al., (2021) isolated γ-conglutin from L. albus at pH 7.0 preserving its HA; although, it was not constant in different purification batches. These findings indicate that pH is an important factor in maintaining the hemagglutinating capacity of lupins g-conglutin.
 
Inhibition of hemagglutination with carbohydrates
 
The HA of LMC was inhibited by 2 mM galactose and 50 mM melibiose (disaccharide containing galactose and glucose) (Fig 4A). None of the other assessed carbohydrates (glucose, mannose, lactose) inhibited hemagglutination. Rabbit erythrocytes have a high abundance of galactose terminal residues (Yamakawa, 2005), which explain the hemagglutination by LMC. Therefore, the isolated protein has specificity for galactose, as previously found for L. mutabilis extracts (Falcon et al., 2000a). Villamarin (2016) reported the inhibition of hemagglutinating activity (IHA) of human erythrocytes for mannose and glucose; besides, the IHA was negative for galactose, differing from our results and possibly due to the extraction method used by the author, that yielded other hemagglutinating lectin.
 
Optimal hemagglutination temperature and thermostability
 
LMC showed optimal HA at temperatures between 15 and 20°C (Fig 4B). At 60°C, HA was completely depleted while lower temperatures did not inhibit hemagglutination. In the case of legume lectins, hemagglutinating activities are diverse. For example, the optimal HA of the Sophora japonica lectin is near 0°C, while the Canavalia ensiformis lectin (concanavalin A) has optimal activity at 35°C (Gilboa-Garber and Sudakevitz 1999).

HA was totally preserved after short term incubation (30 min) at temperatures between -20°C and 50°C, was dramatically reduced after incubation at 70ºC and lost after incubation at 90°C (Fig 4C). A similar thermostability was found for the 17 kDa lectin isolated from L. mutabilis (Villamarin, 2016), although temperatures below 37°C were not assessed. Other legume lectins, such as the P. biglobosa lectin, retain their HA up to 50°C (Silva et al., 2013) while the D. lasiophylla and black turtle bean lectins retain their HA up to 70°C (Pinto-Junior et al., 2013; He et al., 2014) and those of P. vulgaris up to 80°C (Sharma et al., 2009).


 
Hemagglutinating activity of LMC is cation-independent and resistant to reducing conditions
 
LMC was not inhibited by  b-mercaptoethanol or EDTA at the concentrations tested (1-20 mM). Legume lectins HA can be cation dependent or independent, as it has been seen in the lectins from C. ensiformis and Vataiera guianensis, respectively (Sharon and Lis 2007; Silva et al., 2012). In cation-dependent lectins, such as those from P. vulgaris var. Jade, the addition of EDTA at concentrations as low as 10 mM can almost completely deplete the HA (Cheung et al., 2013). Unlike our findings, Villamarin (2016) reported that their L. mutabilis lectin lost its HA on human erythrocytes (A and AB type) after incubation with EDTA (25 mM). To our knowledge, there are no reports of HA under reducing conditions; however, the fact that LMC hemagglutinates under these conditions suggests that only one of the g-conglutin subunits exerts the HA.

In this study, we successfully isolated and characterized a g-conglutin from the seeds of Lupinus mutabilis Sweet ecotype Patón Grande. The isolated g-conglutin has a molecular weight of 46 kDa, is formed by two subunits and has affinity for galactose. The used methodology allowed it to preserve its hemagglutinating activity, verified at each step of the isolation process. We consider that this methodology is an important contribution for the purification of g-conglutin from L. mutabilis and can be scaled up to a semi-industrial level.
 

We thank Professor Patricia Woll for the guide in chromatography and Douglas Ricardo for the advice about protein identification. This work was supported by Universidad Nacional Mayor de San Marcos [Rectorate 2015] and [PIBA RR N°03912-R-15].

None