Biochimica et Biophysica Acta 1674 (2004) 282 – 290 http://www.elsevier.com/locate/bba
Specificity of Amaranthus leucocarpus syn. hypocondriacus lectin for O-glycopeptides Pedro Herna´ndeza,b, Daniel Tetaertc, Ge´rard Vergotend, Henri Debrayd, Maria del Carmen Jimenezb, Georgina Alvareze, Concepcio´n Agundise, Pierre Degandc, Edgar Zentenoe,* a Departamento de Bioquı´mica, Instituto Nacional de Enfermedades Respiratorias, Tlalpan, D.F., 14080, Mexico Centro de Investigaciones Medicas-Biolo´gicas Facultad de Medicina, Universidad Auto´noma del Estado de Oaxaca, Benito Jua´rez, 68020 Oaxaca, Mexico c Institut National de la Sante´ et la Recherche Me´dicale U 560, Les mucines e´pithe´liales, du ge`ne a` la fonction, 59045 Lille, France d Unite´ Mixte de Recherche 8576, Glycobiologie Structurale et Fonctionnelle Universite´ de Sciences et Technologies de Lille, 59655 Villeneuve d’Ascq, France e Laboratorio de Inmunologı´a, Departamento de Bioquı´mica, Facultad de Medicina, UNAM, P.O. Box 70159, 04510 Mexico b
Received 21 April 2004; received in revised form 2 July 2004; accepted 20 July 2004 Available online 21 August 2004
Abstract Amaranthus leucocarpus syn. hypochondriacus lectin (ALL) has been shown to be specific for N-acetyl-d-galactosamine (GalNAc). In this work, we determined a value of 1.010 2 M for the association constant of ALL for GalNAc, calculated using fluorescence spectroscopy assays. Using neoglycopeptides obtained by in vitro O-glycosylation, we determined the main features of O-glycopeptides recognized by ALL using molecular dynamics simulations, capillary electrophoresis, and ELISA. Neo-glycopeptides were obtained by in vitro O-glycosylation reaction using microsomal preparations of murine thymocytes, human gastric fundus and colonic mucosa. ELISA assays were performed with peroxidase-labeled murine monoclonal IgG2, n light chain (5D4) antibodies against ALL. Among the in vitro neoglycopeptides, only those of TTSAPTTS containing GalNAc at Thr in #2 and #6 reacted with ALL. Neither the TTSAPTTS glycopeptide, containing a unique GalNAc residue at Thr in #2, nor others (with more than two GalNAc residues) interacted with the lectin. Computational docking assays of the lower energy conformers for interactions between glycopeptides and lectins confirmed that ALL recognized GalNAc residues when they are spaced out in glycan structures, whereas GalNAc residues arranged in clusters prevented interaction with the lectin, indicating that ALL is specific for a special GalNAc-containing motif found in different O-glycoproteins. D 2004 Elsevier B.V. All rights reserved. Keywords: Amaranthus leucocarpus; Plant lectin; T- and Tn-specific lectin; O-glycan
1. Introduction Lectins are important tools for oligosaccharide characterization as well as for isolation of cellular populations [1,2]. In particular, lectins with specificity for O-linked glycans (containing Galh1–3 GalNAca1-O-Ser/Thr and GalNAca1-O-Ser/Thr) have been widely used in the fractionation of glycoproteins and cellular sub-populations [3,4]. Lectins from the Amaranthaceae family have been identified in A. caudatus [5], A. cruentus [6], and A. leu* Corresponding author. Tel.: +52 5 623 2175; fax: +52 5 616 2419. E-mail address:
[email protected] (E. Zenteno). 0304-4165/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbagen.2004.07.008
cocarpus syn. hypocondriacus [7]. All these lectins interact specifically with GalNAc [5–7] and have been proven useful tools to recognize O-linked glycans in different tissues, tumors, and cells [8–10]. Amaranthus leucocarpus syn. hypocondriacus is a Mexican plant representative of the Amaranthaceae family from the seeds of which a lectin inducing immunosuppression in animals can be isolated [11]. This lectin from A. leucocarpus (ALL) recognizes non-stimulated murine peritoneal macrophages and naive CD4+ murine and human lymphocytes [12]. ALL agglutinates human erythrocytes with the M phenotype [8], and has been shown to interact with O-linked glycoproteins in some neurodegenerative processes
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
[13,14]. In recent works, it has been suggested that ALL may recognize a particular conformation of O-glycan determinants arranged in clusters on the peptidic backbone of a specific cellular receptor, conferring the lectin its particular biological properties [15,16]. In this work, we compared the specificity of ALL toward different Oglycoproteins, which in previous works showed capacity to inhibit the hemagglutinating activity of ALL [15], and with neoglycopeptides obtained by in vitro O-glycosylation using different microsomal preparations. The ELISA study and capillary electrophoresis, in addition to computational docking analysis, allowed us to determine the main features of the specific receptor for ALL.
2. Materials and methods 2.1. Lectin isolation A. leucocarpus syn. hypochondriacus seeds were obtained in Tulyehualco (Mexico), and the lectin was purified by affinity chromatography using a column containing stroma from human O-desialylated erythrocytes entrapped in Sephadex G-25, as described previously [7]. 2.2. Fluorescence spectroscopy The lectin from A. leucocarpus (0.1 mg/ml) in phosphate-buffered saline (PBS: 0.01 M sodium phosphate, 0.15 M sodium chloride, pH 7.2) was placed in a 1-cm pathlength-cell and the titration curve with GalNAc was obtained at 26 8C. The fluorescence emission spectrum was recorded to obtain the maximum peak of fluorescence of ALL (data not shown) using a Perkin Elmer spectrofluorometer (LS-5B) with 285 nm excitation wavelength and 385 nm emission wavelength. The excitation and emission slits were set at 5 and 2.5 nm, respectively. The association constant was determined using the Quasi-Newton algorithm by quenching of fluorescence at 390 nm. 2.3. Microsomal preparation Thymocytes were obtained from surgically removed thymuses of 28-day-old male BALB/c mice and minced in DMEM with a 20-gauge needle in a Petri dish, passed through a nylon mesh and the harvested cells were washed thrice in DMEM. Thymocytes were suspended in 2 ml of PBS and counted with a hemocytometer. Their viability was assessed by the trypan blue exclusion test, and they were adjusted to 1106 cells/ml in PBS. Afterwards, the thymocyte microsomal suspensions were prepared by ice-sonication of the cell pellets and stored at 70 8C in a 0.2 M NaCl/0.25 M sucrose solution until used for GalNAc transferase assays. Protein concentration was determined according to the method of Peterson [17].
283
Human stomach and colon samples were from healthy tissues of cancer patients, who had undergone resection. Microsomes from these tissues were prepared through Potter–Elvehjem tissue homogenization in ice-cold 0.25 M sucrose and 0.2 M NaCl, and centrifuged at 10,000g for 10 min followed by ultracentrifugation at 140,000g for 20 min to collect the microsomal pellet. The pellet was suspended in NaCl-sucrose solution and protein concentration was determined on a Hitachi 717 spectrophotometer (Boehringer) using the Biotrol kit protein assay reagent (Biotrol, Chennevieres les Louvres, France). 2.4. GalNAc transferase assays Different synthetic peptides were used as exogenous substrates: the octapeptide NH 2 -TTSAPTTS-COOH deduced from a tandem repeat sequence of MUC 5AC of human stomach mucin, NH2-GTTPSPVP-COOH tandem repeat sequence of fundic human gastric tissue, and NH2GTTPSPVPTTSTTSAP-COOH tandem repeat sequence of human colon [18–20]. Purity (more than 95%) of the peptides was assessed by HPLC analysis and capillary electrophoresis. Their mass was verified by electrospray MS and their amino acid sequences were also verified with an Applied Biosystems gas-phase sequencer 477A [21]. Assays were performed with microsomal suspensions in a total volume of 40 Al, containing the following components at final concentrations: 5 mM (10 Al) peptide acceptor; 125 mM (10 Al) MES buffer (pH 7.0), containing 0.2% (v/v) Triton X-100, 12.5 mM MnCl2; 1 mM PMSF, 1 mM AEBSF, 1 mM E64; 1 mM (5 Al) AMP, UDP-[3H]-GalNAc (0.3106 dpm) in 1 mM (5 Al) cold UDP-GalNAc, and 12 Ag (10 Al) microsomal preparation. After addition of the microsomal preparation, the samples were incubated for 3 h at 37 8C. The reaction was stopped by adding 8 volumes (320 Al) of 20 mM sodium borate-1 mM ethylenediamine tetra-acetic acid (pH 9.1). The products of the in vitro Oglycosylation reaction were separated from the excess UDPGalNAc using an AG1-X8 anion exchange resin column by elution with water (31 ml), and [3H]GalNAc incorporation was counted in a Beckman LS6000 SC liquid scintillation spectrometer. Before capillary electrophoresis, the aqueous eluates from the AG1-X8 exchanger were desalted on SepPak C18 reversed-phase cartridges activated with 10 ml methanol followed by 10 ml deionized water containing counter-ion trifluoroacetic acid (TFA, 0.1%). Polar materials were washed with 10 ml water/TFA, whereas the glycopeptide fractions were obtained by elution with 10 ml of 25% acetonitrile in water/0.1% TFA [19]. 2.5. Capillary electrophoresis Zone capillary electrophoresis was performed on a P/ ACE Model 5000 system (Beckman). UV absorbance was monitored at 200 nm. The fused silica capillary had a 57 cm75 Am i.d. and was fitted into a cartridge with a
284
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
modified mandrel to improve cooling. Migrations were run at 23 8C using a 2N formic acid buffer and the voltage across the capillary was maintained at 20 kV. Injections were performed under pressure (duration 6 s, approximately 10–20 pmol) and after each separation the capillary was flushed with 25% acetonitrile in PBS, pH 7.2, for 2 min [21,22]. Interaction of the A. leucocarpus lectin with neoglycopeptides was achieved by using 4 Ag of ALL-FITC with 16 Ag of peptides in PBS, pH 7.2, and after incubation for 1 h at room temperature, 5 ng of the sample were injected into the P/ACE Model 5000 system. Migrations were run at 23 8C using PBS, pH 7.2, during 30 min and the voltage across the capillary was maintained at 20 kV, and absorbance was monitored at 560 nm [21,22].
Unbound conjugate was removed by washing six times with PBS containing 0.1% Tween-20 followed by two washes with PBS alone. The presence of ALL was revealed by incubating the wells with 100 Al j-phenylenediamine hydrochloride and 100 Al hydrogen peroxide in 0.1 M citric acid at room temperature for 30 min. The reaction was stopped by adding 100 Al of 6 M HCl to each well and the absorbance was read at A 490 on a Dynatech MR 5000 microplate reader. Control assays were performed using hen ovomucoid, an N-glycosylated glycoprotein that has shown no capacity to inhibit the hemagglutinating activity of ALL [5–7].
2.6. Hybridoma fusions and screening
The 3-D structure of ALL was determined using the SWISS-MODEL program version 3.5 (http://us.expasy.org/ sprot/). This program searches for identity between amino acid sequences; the sequence of Amaranthus caudatus (ACA [PDB ID 1JLX]) showed the highest identity with ALL. The 3-D structures were analyzed as PDB files using ViewerLite version 4.2 (www.accelrys.com).
Balb/c female mice (20–30 g) were immunized intraperitoneally four times with 50 Ag of ALL in complete Freund’s adjuvant. Spleen cells were harvested and fused to the myeloma cell line P3-X63Ag8, according to the method described by Kohler and Milstein [23] and modified by Gefter et al. [24], using polyethylene glycol to promote hybridization of mouse myeloma cells. To determine production and secretion of ALL reactive monoclonal antibodies (mAbs), supernatants of HAT selection growthpositive clones were analyzed through ELISA using ALL and supernatant of P3-X63Ag8 cell culture as negative control. Positive hybridomas were expanded and subcloned by limiting dilution to ensure stability of mAb production. The antibodies were purified by fractionation with 27% solid ammonium sulfate and ion exchange chromatography on DEAE-Sephadex (Pharmacia) and finally labeled with horseradish peroxidase, according to the method of Avrameas and Ternynck [25]. 2.7. Specificity of A. leucocarpus lectin determined by ELISA The specificity of ALL was determined by comparing the capacity of the lectin to interact with glycopeptides and glycoproteins with well-known glycan composition. Samples (200 Ag) in 200 Al carbonate buffer (100 mM, pH 9) were incubated in sterile flat bottom microwell plates (Nunc Delta, Roskilde, Denmark) for 12 h at 4 8C. The plates were then washed with PBS and the wells covered with 300 Al of 5% skimmed milk in PBS and incubated for 12 h at 4 8C. After removing the milk, the wells were incubated with 100 Ag ALL in 200 Al PBS at 37 8C for 3 h and then incubated overnight at 4 8C. The excess lectin was washed with PBS and the wells were incubated with 100 Al of mAbs (5D4) against the lectin (at a dilution of 1:400) for 1 h at 37 8C, and then incubated overnight at 4 8C. Excess mAb was discarded by washing five times with PBS and the wells were incubated with 100 Al of antimurine IgG labeled with peroxidase for 1 h at 37 8C.
2.8. Molecular modeling of ALL
2.9. Molecular modeling of glycopeptides The minimization and random searching tools within SYBIL 2002 (version SYBIL, Tripos, St. Louis, MO) allowed to determine the structures of the low energy conformers of the glycopeptides TT[GalNAc]SAPTTS with linkage at Thr position #2; TTSAPTT[GalNAc]S at Thr position #6; TT[GalNAc]SAP TT [GalNAc]S at Thr positions #2 and # 6; LST[GalNAc]T[GalNAc]EVAM with GalNAc at Thr positions #3 and #4. 2.10. Docking of glycopeptides in ALL molecule For the five energy conformers of glycopeptides, a surface docking with ALL was investigated by means of The Global Range Molecular Matching (GRAMM 2002 version) program. This technique located the area of global minimal intermolecular energy for the structure. The generic mode and grip step were used, 1000 matches were stored, from which 1000, 10, and 5 were displayed. For the lower energy glycopeptide/ALL complexes, a dynamic energybased conformation was performed. Data from the docking experiments were analyzed as PDB files using ViewerLite version 4.2 (www.accelrys.com). 2.11. Analytical methods Protein concentration was determined by the method of Lowry modified by Peterson [17], using bovine serum albumin as standard. Carbohydrate content and composition of the lectin, its isoforms, and of all glycoproteins used in this study were determined with heptafluorobutyrate derivatives of O-methyl-glycosides, obtained after
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
methanolysis with 0.5 N methanol–HCl for 24 h at 100 8C, by gas-chromatography with a capillary column (250.32 mm) of 5% Silicone OV 210 (Applied Science Lab., Buffalo, NY), in a Varian 2100 gas chromatograph (Orsay, France). The carrier gas was helium at a pressure of 0.6 bar, and the oven temperature was programmed from 100 to 240 8C at 2 8C per min [26]. l-lysine (Sigma Chem., St. Louis, MO) was used as internal standard. Identification of each sugar residue was confirmed by mass spectrometry using the Finnigan Automass II mass spectrometer, the HFB derivatives of the O-methyl glycosides of monosaccharides presented relatively high masses: 978 for hexoses, 977 for hexosamines, and 1275 for sialic acid [26]. The exact structure of the neo-synthesized Oglycopeptides was established by MALDI-MS MS/MS as described in Ref. [19].
3. Results 3.1. Fluorescence spectroscopy The association constant of A. leucocarpus syn. hypocondriacus lectin (ALL) for GalNAc was considered as 1.0010 2 M (Fig. 1). Other monosaccharides, such as GlcNAc, Gal, Glc and Man, were used as controls, but showed no interaction with ALL. 3.2. GalNAc transferase assays The first step of the O-glycosylation reaction was evaluated by measuring GalNAc-transferase activities toward different synthetic peptides determined with crude microsomal homogenates prepared from thymocytes and from human gastric fundus and human colon. After elimination of the excess UDP-GalNAc, GalNAc transferase activities were measured by the transfer of radiolabeled
Fig. 1. Fluorescence titration of A. leucocarpus syn. hypocondriacus lectin (ALL) with GalNAc. Fluorescence intensity of ALL was recorded at 26 8C and wavelength fixed at 385 nm. The algorithm used to obtain the association constant was Quasi-Newton.
285
Table 1 Incorporation rate of GalNAc in peptides after in vitro reaction Peptide TTSAPTTS GTTPSPVP GTTPSPVPTTSTTSAP
Microsomal preparations Thymocytesa
Fundic gastrica
Colona
84,000F100 ND ND
17,800F5000 4,070F500 51,300F460
19,900F6400 4070F30 5300F2660
ND: not determined. a Reported as nM of GalNAc/h/mg of microsomes after 3 h of incubation at 37 8C.
[3H]GalNAc to the peptide substrates. Incorporation rates for the tested peptides were determined and the values ranged from 4000 to 40,000 nmol GalNAc/h/mg of microsomal proteins (Table 1). The V max values were found four times higher with thymocyte preparations than with colonic or gastric preparations. 3.3. Capillary electrophoresis To visualize the differences in O-glycosylation reactions with all peptides and the different microsomal preparations, samples were analyzed by electrophoretic techniques. Capillary electrophoresis allowed us to identify O-glycosylated products, according to their relative electrophoretic mobilities, and to compare them with previously wellcharacterized glycopeptide standards. Incubation without UDP-GalNAc always revealed a specific fraction, identified at 19.6 min, whereas in the presence of UDP-GalNAc, additional fractions were identified at 21.4 and 22.2 min, corresponding to glycopeptides containing one or two Olinked GalNAc residues, respectively (Fig. 2). Only the peptide TTSTTSAP was glycosylated with microsomal preparations from thymocytes with two threonine positions (Thr #2 and #6) occupied by GalNAc residues, as confirmed by MALDI-MS technique. TTSTTSAP was able to interact
Fig. 2. Electrophoregram of the TTSAPTTS peptide O-glycosylated by murine thymocyte microsomes. Capillary electrophoresis was performed on a P/ACE Model 5000 system. UV absorbance was monitored at 200 nm. Migrations were run at 23 8C using a 2N formic acid buffer and the voltage across the capillary was maintained at 20 kV, in a 30-min program. Roman numbers indicate the position of mono- and di-glycosylated peptides.
286
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
tion of antibodies against ALL by indirect ELISA. Hybridoma cells were subcloned and one of them was chosen for further characterization and designated as DG4. Characterization of the antibody (named 5D4) produced by DG4 cells indicated that it belongs to IgG2 with n light chains. Specificity of mAbs was verified in the lectin purification procedures by affinity chromatography of ALL samples, and the collected fractions were assayed for the presence of the lectin: only the protein fraction eluted from the affinity chromatography matrix was positively identified by the 5D4 antibody in the ELISA assays.
4. ELISA Fig. 3. Electrophoregram of the interaction of A. leucocarpus syn. hypocondriacus lectin with the TTSAPTTS O-glycopeptide. The assay was performed using 4 Ag of ALL-FITC with 16 Ag peptides in PBS, pH 7.2, and incubated during 1 h at 25F3 8C; 5 ng of the sample was injected into a P/ACE Model 5000 system. Migration rates of the lectinglycopeptide (dotted line) were compared with free lectin (solid line). Assays were performed at 23 8C using PBS, pH 7.2, at 20 kV, in a 30-min program. Absorbance was monitored at A 560.
with ALL, as identified by the delay in migration time, whereas the same peptide that was O-glycosylated with gastric fundus and human colon microsomal preparations did not interact with the lectin (Fig. 3; Table 2). Peptides GTTPSPVP and GTTPSPVPTTSTTSAP, known to be Oglycosylated with different microsomal preparations at multiple positions [27,28], did not interact with the lectin regardless of the substituted Thr positions (not shown).
The ELISA method to evaluate the interaction of ALL with glycoproteins with O-glycosidically linked glycans and with different O-neo-glycopeptides implied utilization of mAbs against the purified ALL. Standardization of ELISA assays was achieved with the purified lectin. The sensitivity of the assay allowed us to detect concentrations as low as 3 ng/ml (1 fmol of lectin). Our results indicate that the glycoproteins OSM, BSM, glycophorin AM, fetuin, and, to a lesser extent, glycophorin AN and human IgA were recognized by ALL (Table 3). In the presence of neoglycopeptides, our results confirmed that only the peptide TTSTTSAP, which was glycosylated at Thr positions #2 and #6, was recognized by ALL. Neoglycopeptides containing only one or several GalNAc residues were not recognized by the lectin. Control assays with hen ovomucoid showed ELISA values of 0.02 at A 490.
3.4. Monoclonal antibodies Hybridomas were produced by fusion of mouse myeloma cells and spleen cells from mice immunized with ALL. Hybridoma culture supernatants were screened for produc-
Table 2 Time (min) of migrationa of A. leucocarpus syn. hypocondriacus lectin in the presence of different neoglycopeptides Peptide
Table 3 Interaction of A. leucocarpus syn. hypocondriacus lectin with glycoproteins and neoglycopeptides as determined by ELISA method Glycoprotein or neoglycopeptide
Sequence
A 490
Ovine mucin Bovine mucin Fetuin**
ST*T*GST*Sa S*EST*T*QLPa AGPT*SAAGb VAS*VVV GPT*PSA APS*AVP TPST*PST*Pa LS*TT*EVAMa SS*TT*GVAMa TT*SAPT*TSc TT*SAPTTSd GT*T*PSPVPd GT*T*PSPVPT*T*ST*T*SAPd
0.5 0.7 0.9
Microsomal preparations b
TTSAPTTS TTSAPTTSc LSTTEVAMd SSTTGVAMe GTTPSPVPf GTTPSPVPTTSTTSAPg
Thymocytes
Gastric fundus
Colon
15 13 13 13 13 13
13 13 13 13 13 13
13 13 13 13 13 13
a Determined by capillary electrophoresis in a 30-min run program. Positive interaction of the lectin was determined by inhibition of the migration time. b Neoglycopeptides with two GalNAc residues. c Neoglycopeptides with one GalNAc residue. d Neoglycopeptides with one GalNAc residue. e Neoglycopeptides with one GalNAc residue. f Neoglycopeptides with one GalNAc residue. g Neoglycopeptides with six GalNAc residues.
IgA AN glycophorin AM glycophorin Neoglycopeptides
0.4 0.1 0.8 0.5 0.1 0.1 0.1
*Indicates the putative glycosylation site. Glycopeptide concentration was adjusted to 100 nM and incubated with 10 Ag lectin, and revealed with anti-ALL HPR labeled mAbs. Control assays with non-lectin-inhibitor N-glycosylated hen ovomucoid reported values of 0.02. a Based on netOglyc program [35]. b As proposed by Jones [36]. c The peptide was glycosylated by thymocytes. d The peptide was glycosylated by fundic gastric microsomes.
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
4.1. Molecular modeling of ALL and docking with glycopeptides The ALL polypeptide chain is folded in two similar domains, each one with the h-trefoil domain, and showed high identity with the 3-D structure reported for A. caudatus lectin (Fig. 4). Docking of glycopeptides into the ALL’s 3-D revealed that the major interactions were the hydrogen bonds between GalNAc residues and the polypeptide chain of glycopeptides with the amino acids of the carbohydrate recognition domain (CRD). The best ALL interaction was observed with TTSAPTTS glycopeptides possessing GalNAc in Thr #2 and #6, whereas docking of glycopeptides with only one or adjacent GalNAc residues showed poor interactions. The interaction of ALL with TTSAPTTS was stabilized by four hydrogen bonds: one hydrogen bond (2.50 2) between the oxygen of Thr #1 and the carboxylic group of Gln #27, one hydrogen bond (2.57 2) between the oxygen of Thr #2 and the carboxylic group of Gln #26, one hydrogen bond (2.59 2) between the oxygen of Ser #3 and
Fig. 4. Molecular comparison between (A) A. caudatus lectin (ACA) and (B) A. leucocarpus syn. hypocondriacus lectin (ALL). ACA structures were obtained from PDB (1JLX) and ALL structure was obtained using the SWISS-MODEL program.
287
the phenolic group of Tyr #28, and one hydrogen bond (2.66 2) between the oxygen of Ser #81 and the N-Acetyl group of GalNAc in Thr #6 (Fig. 5).
5. Discussion Plant lectins can be classified in seven families of structurally and evolutionary related proteins; within a given lectin family most, but not necessarily all, members are built up of protomers with a similar primary structure and an overall 3-D folding. The overall structure of native lectins is not only determined by the structure of the protomers but also depends on the degree of oligomerization and, in some cases, on the post-translational processing of the lectin precursors. In general, lectin families are fairly homogeneous concerning the overall specificity of the individual lectins, which illustrates that the 3-D structure of the binding site has been conserved during evolution [29,30]. Lectins from the Amaranthaceae family have been demonstrated in A. caudatus [5], A. cruentus [6], and A. leucocarpus syn. hypochondriacus [7]. These lectins have been considered as important tools to characterize O-glycans [9]. The lectin from A. leucocarpus syn. hypocondriacus (ALL) agglutinates preferentially human erythrocytes with glycophorin AM rather than those with glycophorin AN [8]: both phenotypes reveal similar rates of O-glycosylation and show differences at their NH2-terminal [31]. ALL recognizes antigenic T (Galh1,3GalNAca1, O-Ser/Thr) and Tn (GalNAca1 O-Ser/Thr) structures [8,15]. In this work, we identified the fine specificity of ALL for glycoproteins, peptides, and neoglycopeptides. Neoglycopeptides were prepared by in vitro reaction using microsomes from mouse thymocytes [18], human gastric fundus and colon mucosa [19,20]. Bovine [32] and ovine submaxillary mucin [33], human IgA [34], glycophorin AM, as well as AN [31], are rich in T- and Tn-antigens, although ovine mucin possesses Tn-antigens in a higher proportion. Hen ovomucoid, which contains oligomannosidic N-glycosidically linked glycans [35], was used as control in ELISA assays, and showed no interaction with ALL confirming the specificity of ALL for O-glycosidically linked glycoproteins. The interactions between ALL and these structures showed to be heterogeneous, very probably due to the position of GalNAc residues on the peptidic backbone, as determined by prediction analysis through the NetO-glyc program [36]. In fetuin, the O-linked glycans represent 30% of the oligosaccharidic structure, and they are confined to the C-terminal part of the molecule [37,38], but fetuin was the glycoprotein mostly recognized by ALL. Glycophorin AM was better recognized than glycophorin AN, and differences in their interaction might be due to ALL interacting with GalNAc. Furthermore, the 3-D presentation of the glycan or the possible participation of OH from serine residues in the amino-terminal of the AM-glycophorin has also been implied [8,15]. From the results obtained with neoglycopeptides and capillary
288
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
Fig. 5. Molecular docking of the interaction of A. leucocarpus lectin with the glycopeptide TT*SAPT*TS.
electrophoresis, it can be inferred that the glycopeptide TTSAPTTS, with two GalNAc residues at Thr #2 and Thr #6 in its sequence, was able to modify the migration time of ALL. It is interesting to point out that the glycopeptide GTTPSPVPTTSTTSAP with six incorporated GalNAc residues, but in cluster configuration, did not interact with ALL. In contrast to other lectins with similar sugar specificity, such as Salvia sclarea [39] or Vicia villosa lectins [40], GalNAc residues arranged in cluster were not recognized by ALL. Previous works reported that A. caudatus (ACA) showed a great identity with ALL [8,15,41]; but, despite this identity, ALL displayed slight differences in the recognition of oligosaccharidic sequences. Similarly to ACA, A. leucocarpus lectin can bind to the T- and Tnantigens, and substitutions in these antigens of their Gal-C3 and GalNAc-C6 by sialic acid residues are well tolerated by both lectins. The ability of lectins to bind carbohydrates depends on their 3-D structure [42,43] and on their capacity to detect subtle variations in the conformation of carbohydrate structures from the cell surfaces [43]. This ability could be explained by the variability in the size of the CRD and the variability in quaternary association [44,45]. In A. caudatus lectin, one hydrogen bond was involved between the N-acetyl group of GalNAc and the chain polypeptide amide of tyrosine in position #76 [41]; however, it might be possible that, in ALL, the same hydrogen bond interacts
with the oxygen of Ser at position #81. Presence of other hydrogen bonds between the polypeptide chain of the glycopeptide and the amino acid in the proximities of CRD suggests the existence of subsites for the recognition of glycopeptides. These subsites could be responsible for the specificity of ALL for T- and Tn-antigen structures arranged orderly in a receptor. On the contrary, a high density of GalNAc residues forming glycosylated clusters on the receptor interfered negatively with the active site of the lectin. Our results support further the hypothesis that the interaction of ALL with O-glycopeptides might be determined by steric interactions of the GalNAc residue with adjacent peptide residues. The O-linked glycans on specific peptide sequences present on membranes and cell surfaces play important biological roles [39,41], and these structures would be responsible for the 3-D organization of glycoproteins. Based on the aforementioned data, our results reinforce the notion that lectins from the Amaranthus genus, particularly from A. leucocarpus, represent a very useful tool for the study of mucin structures and their biological roles.
Acknowledgments This work was supported in part by CONACyT, PAED 202317, DGAPA-UNAM (PAPIIT-IN507907 and
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
IX232504) and Program ECOS Mexico–France (M97B05). We are indebted to Colette Richet and Dominique Demeyer for their skilful technical help.
References [1] I.J. Goldstein, R.C. Hughes, M. Monsigny, T. Osawa, N. Sharon, What should be called a lectin? Nature 28 (1980) 286. [2] H. Lis, N. Sharon, Lectins as molecules and as tools, Annu. Rev. Biochem. 55 (1986) 35 – 67. [3] G. Alvarez, R. Lascurain, A. Perez, P. Degand, L.F. Montano, S. Martinez-Cairo, E. Zenteno, Relevance of sialoglycoconjugates in murine thymocytes during maturation and selection in the thymus, Immunol. Invest. 28 (1999) 9 – 18. [4] N. Sharon, H. Lis, The Lectins Properties, Function and Application in Biology and Medicine, Academic Press, 1986, pp. 266 – 270. [5] S.J. Rinderle, I. Goldstein, K.L. Matta, R.M. Ratcliffe, Isolation and characterization of amaranthin, a lectin present in the seeds of Amaranthus caudatus, that recognizes the T- (or cryptic T)- antigen, J. Biol. Chem. 264 (1989) 16123 – 16131. [6] A. Calderon de la Barca, L. Vazquez-Moreno, Amaranthus cruentus lectin: purification, stability and some biochemical properties, J. Food Biochem. 12 (1988) 117 – 126. [7] E. Zenteno, J.L. Ochoa, Isolation and characterization of Amaranthus leucocarpus lectin, Phytochemistry 27 (1988) 313 – 317. [8] E. Zenteno, R. Lascurain, L.F. Montan˜o, L. Vazquez, H. Debray, J. Montreuil, Specificity of Amaranthus leucocarpus lectin, Glycoconj. J. 9 (1992) 204 – 208. [9] W.J. Peumans, E. Van Damme, Plant lectins: specific tools for the identification, isolation, and characterization of O-linked glycans, Crit. Rev. Biochem. Mol. Biol. 33 (1998) 209 – 258. [10] J.T. Gallagher, Carbohydrate-binding properties of lectins: a possible approach to lectin nomenclature and classification, Biosci. Rep. 4 (1984) 621 – 632. [11] E. Zenteno, J.L. Ochoa, C. Parra, L.F. Montan˜o, I. Rayon, G. Maldonado, B. Ruiz, R. Carvajal, Mitogenic immunosuppressive and phagocytic activity of Machaerocerus eruca and Amaranthus leucocarpus lectin, in: T.C. Bfg-Hansen, J. Breborowicz (Eds.), Lectins—Biology, Biochemistry, Clinical Biochemistry, vol. 4, Walter de Gruyter, Germany, 1985, pp. 537 – 546. [12] R. Lascurain, R. Cha´vez, P. Gorocica, A. Pe´rez, L.F. Montan˜o, E. Zenteno, Recognition of CD4+ mouse medullary thymocyte subpopulation by Amaranthus leucocarpus lectin, Immunology 83 (1994) 410 – 413. [13] J. Guevara, B. Espinosa, E. Zenteno, L. Va´zquez, J. Luna, G. Perry, R. Mena, Altered glycosylation pattern of proteins in Alzheimer disease, J. Neuropathol. Exp. Neurol. 57 (1998) 905 – 914. [14] B. Espinosa, J. Guevara, P. Hernandez, M.C. Slomianny, A. Guzman, S. Martinez-Cairo, E. Zenteno, Characterization of an O-glycosylated plaque-associated protein from Alzheimer disease brain, J. Neuropathol. Exp. Neurol. 62 (2003) 34 – 41. [15] P. Herna´ndez, M. Bacilio, F. Porras, S. Jua´rez, H. Debray, E. Zenteno, B. Ortiz, A comparative study on the purification of the Amaranthus leucocarpus syn. hypocondriacus lectin, Prep. Biochem. Biotechnol. 29 (1999) 219 – 234. [16] P. Herna´ndez, H. Debray, H. Jaekel, Y. Garfias, M.C. Jimenez, S. Martı´nez-Cairo, E. Zenteno, Chemical characterization of the lectin from Amaranthus leucocarpus syn. hypocondriacus by 2-D proteome analysis, Glycoconj. J. 18 (2001) 321 – 329. [17] G.L. Peterson, A simplification of the protein assay method of Lowry et al., which is more generally applicable, Anal. Biochem. 83 (1977) 346 – 356. [18] J.P. Audie´, A. Janin, N. Porchet, M.C. Copin, B. Gosselin, J.P. Aubert, Expression of human mucin genes in respiratory, digestive, and
[19]
[20]
[21]
[22]
[23] [24]
[25]
[26]
[27]
[28]
[29]
[30] [31]
[32]
[33]
[34]
[35]
289
reproductive tracts ascertained by in situ hybridization, J. Histochem. Cytochem. 41 (1993) 1479 – 1485. V. Guyonnet-Duperat, J.P. Audie´, V. Debailleul, A. Laine, M.P. Buisine, S. Galiegue-Zouitina, P. Pigny, P. Degand, J.P. Aubert, N. Porchet, Characterization of the human mucin gene MUC5AC: a consensus cysteine-rich domain for 11p15 mucin genes? Biochem. J. 305 (1995) 211 – 219. H. Krotkiewski, E. Lisowska, A.S. Angel, B. Nilsson, Structural analysis by fast-atom-bombardment mass spectrometry of the mixture of alditols derived from the O-linked oligosaccharides of murine glycophorins, Carbohydr. Res. 184 (1988) 27 – 38. B. Soudan, S. Hennebicq, D. Tetaert, A. Boersma, C. Richet, D. Demeyer, G. Briand, P. Degand, Capillary zone electrophoresis and MALDI-mass spectrometry for the monitoring of in vitro O-glycosylation of a threonine/serine-rich MUC5AC hexadecapeptide, J. Chromatogr., B, Biomed. Sci. Appl. 729 (1999) 65 – 74. D. Schmalzing, C.A. Piggee, F. Foret, E. Carrilho, B.L. Karger, Characterization and performance of a neutral hydrophilic coating for the capillary electrophoretic separation of biopolymers, J. Chromatogr., A 652 (1993) 149 – 159. G. Kohler, C. Milstein, Continuous cultures of fused cells secreting antibody of predefined specificity, Nature 256 (1975) 495 – 497. M.L. Gefter, D.H. Marulies, M.D. Scharff, A simple method for polyethylene glycol-promoted hybridisation of mouse myeloma cells, Somatic Cell Genet. 3 (1977) 213 – 236. S. Avrameas, T. Ternynck, Peroxidase labelled antibody and Fab conjugates with enhanced intracellular penetration, Immunochemistry 8 (1971) 1175 – 1179. J.P. Zanetta, P. Timmerman, Y. Leroy, Gas–liquid chromatography of the heptafluorobutirate derivatives of O-methyl-glycosides on capillary columns: a method for the quantitative determination of monosaccharide composition of glycoproteins and glycolipids, Glycobiology 9 (1999) 255 – 266. X. Czeszak, G. Ricart, D. Tetaert, J.C. Michalski, J. Lemoine, Identification of substituted sites on MUC5AC mucin motif peptides after enzymatic O-glycosylation combining h-elimination and fixedcharge derivatization, Rapid Commun. Mass Spectrom. 16 (2002) 27 – 34. D. Tetaert, K.G. Ten Hagen, C. Richet, A. Boersma, J. Gagnon, P. Degand, Glycopeptide: N-acetylgalactosaminyl transferase specificities for O-glycosylated sites in MUC5AC mucin motif peptides, Biochem. J. 357 (2001) 313 – 320. E.J. Van Damme, W.J. Peumans, A. Barre, P. Rouge, Plant lectins: a composite of several distinct families of structurally and evolutionary related proteins with diverse biological roles, Crit. Rev. Plant Sci. 17 (1998) 575 – 692. M. Rini, Lectin structure, Annu. Rev. Biophys. Biomol. Struct. 24 (1995) 551 – 577. M. Tomita, H. Furthmayer, V.T. Marchesi, Primary structure of human erythrocyte glycophorin A. Isolation and characterization of peptides and complete amino acid sequence, Biochemistry 17 (1978) 4756 – 4770. A.V. Savage, C.M. Donoghue, S.M. D’Arcy, C.A. Koeleman, D.H. van den Eijnden, Structure determination of five sialylated trisaccharides with core types 1, 3 or 5 isolated from bovine submaxillary mucin, Eur. J. Biochem. 192 (1990) 427 – 432. H.D. Hill Jr., J.A. Reynolds, R.L. Hill, Purification, composition, molecular weight, and subunit structure of ovine submaxillary mucin, J. Biol. Chem. 252 (1977) 3791 – 3798. A. Pierce-Cretel, M. Panblanco, G. Strecker, J. Montreuil, G. Spik, Heterogeneity of the glycans O-glycosidically linked to the hinge region of secretory immunoglobulins from human milk, Eur. J. Biochem. 114 (1981) 169 – 178. H. Egge, J. Peter-Katalinic, J. Paz-Parente, G. Strecker, J. Montreuil, B. Fournet, Carbohydrate structures of hen ovomucoid. A mass spectrometric analysis, FEBS Lett. 156 (1983) 357 – 362.
290
P. Herna´ndez et al. / Biochimica et Biophysica Acta 1674 (2004) 282–290
[36] D.T. Jones, Protein secondary structure prediction based on positionspecific scoring matrices, J. Mol. Biol. 292 (1999) 195 – 202. [37] K.D. Smith, A.M. Harbin, R.A. Carruthers, A.M. Lawson, E.F. Hounsell, Enzyme degradation, high performance liquid chromatography and liquid secondary ion mass spectrometry in the analysis of glycoproteins, Biomed. Chromatogr. 4 (1990) 261 – 266. [38] R.G. Spiro, V.D. Bhoyroo, Structure of the O-glycosidically linked carbohydrate units of fetuin, J. Biol. Chem. 249 (1974) 5704 – 5717. [39] A. Medeiros, S. Bianchi, J.J. Calvete, H. Balter, S. Bay, A. Robles, D. Cantacuzene, M. Nimtz, P.M. Alzari, E. Osinaga, Biochemical and functional characterization of the Tn-specific lectin from Salvia sclarea seeds, Eur. J. Biochem. 267 (2000) 1434 – 1440. [40] A. Babino, D. Tello, A. Rojas, S. Bay, E. Osinaga, P.M. Alzari, The crystal structure of a plant lectin in complex with the Tn antigen, FEBS Lett. 536 (2003) 106 – 110.
[41] T.R. Transue, A.K. Smith, H. Mo, I.J. Goldstein, M.A. Saper, Structure of benzyl T-antigen disaccharide bound to Amaranthus caudatus agglutinin, Nat. Struct. Biol. 4 (1997) 779 – 783. [42] I. Weiss, K. Drickramer, Structural basis of the lectin-carbohydrate recognition, Annu. Rev. Biochem. 65 (1996) 441 – 473. [43] Y. Yamashita, Y.S. Chung, R. Horie, R. Kanagi, M. Sowa, Alterations in gastric mucin with malignant transformation: novel pathway for mucin synthesis, J. Natl. Cancer Inst. 87 (1995) 441 – 446. [44] V. Sharma, A. Surolia, Analysis of carbohydrate recognition by legume lectin: size of the combining site loops and their primary specificity, J. Mol. Biol. 267 (1997) 433 – 445. [45] M. Vijayan, N. Chadra, Lectins, Curr. Opin. Struck. Biol. 9 (1999) 704 – 714.