Studies on Hemagglutinins from Maackia amurensis Seeds*

Two phytohemagglutinins (strongly hemagglutinating hemagglutinin and strongly mitogenic hemagglutinin which mitogenic activity against human peripheral lymphocytes, been purified from the seeds of Maackia amurensis by affinity chromatography on porcine thyroglobulin glycopeptides-Sepharose chromatography


Studies on Hemagglutinins
from Maackia amurensis Seeds* (Received for publication, October 20,1973) TSUTOIW KAWAGUCHI, Isaaru ~IATSU>IOTO, AND  From chemical analyses these hemagglutinins were found to be glycoproteins and only small differences in chemical composition were found between them. However, polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate revealed that MAH is a tetramer of subunits having an approximate molecular weight of 33,000 and MAM is a dimer of disulfide-containing subunits of an approximate molecular weight of 75,000. In hemagglutinating activity, MAH is much stronger than MAM, whereas MAM is a more potent mitogen than MAH.
Binding experiments with lnaIlabeled hemagglutinins indicate that normal human erythrocytes bind approximately 3.2 x lo6 molecules of MAH and 7.8 x 10" molecules of MAM with apparent association constants of 2.9 x lo7 M-I and 1.2 X lo7 M-l, respectively, and, furthermore, normal human lymphocytes bind approximately 1.1 x lo7 molecules of MAH and 2.4 X lo7 molecules of MAM with apparent association constants of 2.1 X 10' M-I and 1.1 X lo7 M-I, respectively. Inhibition assays using various sugars and glycoproteins as hapten inhibitors revealed that these hemagglutinins from M. amurensis seeds differ from each other in their specificities for sugars. 100 mg, was dissolved in 1 ml of 0.15 M NaCl-1 mM phosphate buffer (pH 7) and applied to a column (2 X 16 cm) equilibrated against the same buffer. Elution was carried out with the same buffer. Fractions of 4 ml were collected at 8 ml per hour at 4". Mitogenic activity was tested on each 10 ~1 of the fractions.
of sialic acid from PSMl and glycopeptide B was achieved. Highly purified samples of p-galactosidase, p-iv-acetylglucosaminidase, and or-mannusidase from Turbo corn&us were kindly provided by Seikagaku Kogyo Co. Ltd.
Each purified enzyme used in this study was found to be virtually devoid of other glycosidase activity.

Preparation of Porcine Thyroglobulin
Glycopeptides-Sepharose 4B Adsorbent-A mixture of glycopeptides prepared from porcine thyroglobulin by pronase digestion followed by gel filtration according to the method of Fukuda and Egami (3) was coupled to activated Sepharose 4B according to the method previously described (8). Purijcation of M. amurensis Hemagglutinins-Finely powdered M. amurensis seeds (100 g) were suspended in 1 liter of 0.15 M NaCl and allowed to stand overnight at 4" with stirring.
To the yellow clear supernatant obtained by centrifugation, solid (NH&S04 was added to give 50% and later 80 and 100% saturation.
The bulk of both hemagglutinating and mitogenic activities was detected in the fraction precipitating between 50 to 80% saturation. Further purification of this fraction was achieved by chromatography successively on affinity adsorbent prepared by coupling porcine thyroglobulin glycopeptides to activated Sepharose 4B, SE-Sephadex C-50 and Sepharose 6B as described in the legends of Figs. 2, 3, and 4, respectively.
Ultracentrifugation-Measurement of the sedimentation velocity of hemagglutinins was performed according to the band sedimentation method (9) in a Spinco model E ultracentrifuge equipped with an ultraviolet optical system at a speed of 56,100 rpm at 20" in 1 M NaCl.
Disc Electrophoresis-Disc electrophoresis in polyacrylamide gels was carried out in 7.5% gels in Tris-HCl buffer at pH 8.9 according to Ornstein (10) and Davis (11). Staining was performed with Amido black in 7% acetic acid, and destaining in an electric field with 7% acetic acid.
The sodium dodecyl sulfate polyacrylamide gel electrophoresis was carried out according to Weber and Osborn (12). Twenty micrograms of protein were placed onto a 7.5% gel in 0.01 M phos- FIG. 3. SE-Sephadex C-50 chromatography of Fraction A. The concentrated Fraction A, containing 30 mg of protein, from the affinity column was dialyzed overnight against 50 mM phosphate buffer (pH 5) and applied to a column (2 X 30 cm) equilibrated against the same buffer. Elution was carried out with the same buffer and, after the first peak eluted out, gradient elution was performed with 100 ml of starting buffer in the mixing vessel and 100 ml of the same buffer containing NaCl (1 M) in the reservoir. Fractions of 4.4 ml were collected at 8 ml per hour at 4". Mitogenie activity was tested on each 10 pl of the fractions.
FIG. 4. The Fraction B2 (30 mg of protein; in A) and the Fraction Bl (20 mg of protein; in II) from the SE-Sephadex C-50 column were each concentrated by ultrafiltration, dialyzed against 0.15 M NaCl-50 mM phosphate buffer (pH 6) and applied to a column (1.4 X 58 cm) equilibrated against the same buffer. Elution was carried out with the same buffer and fractions of 2.8 ml were collected at a flow rate of 8 ml per hour at 4". Mitogenic activity was tested on each 10 ~1 of the fractions.
Elution volume of blue dextran was determined and is indicated by a vertical arrow. Disc gel electrophoreses of MAH (in A') and MAM (in B'). Electrophoresis was 1 hour at 5 ma per tube with 7.5% gels in Tris-HCl buffer, pH 8.9. phate buffer, pH 7.0, containing 0.1% sodium dodecyl sulfate. Experiments were performed both in the presence and absence of 1% /3-mercaptoethanol. Molecular Weight Estimation by Gel Filtration-A column of Bio-Gel P-200 (1 X 100 cm) was equilibrated with 0.15 M NaCl-5 mM phosphate buffer (pH 7). The relationship between elution volume and the logarithm of the molecular weights of various proteins was established on this column according to the pro-cedure of Andrews (13). The following proteins obtained from Mann Research Laboratories were used for this purpose (molecular weights shown in parentheses): sperm whaie myoglobin (17.800). beef Dancreas chvmotrvnsinogen A (25.000). ovalbumin (45$OOj; boviie serum albumin" 167,0&), and hhrn& r-globulin (160,000). Blue dextran (Pharmacia) was used for determining the void volume of the column.
Amino Acid Analysis-Samples were hydrolyzed with 6 N HCl in sealed tubes for 24,48, and 72 hours at 110". The amino acid content of the hydrolysates was determined on a Hitachi KLA-3D amino acid analyzer according to the method of Spackman et al. (14). Values for each amino acid were calculated as previously described (15). Tryptophan was determined on unhydrolyzed protein samples by the spectrophotometric method of Goodwin and Morton (16). Sugar Analysis-Amino sugar was determined according to the method of Belcher et al. (17). Hydrolysis for this assay was carried out with 4 N HCl for 8 hours at 100" in a sealed tube. For the identification and the determination of neutral sugars, gas-liquid chromatography was performed after reduction to the respective alditol followed by trifluoroacetylation according to the method of Matsui et al. (18), as described by Matsumoto and Osawa (19) and Akiyama and Osawa (20). Samples for this assay were hydrolyzed with 1 N HzSOd at 100" for 8 hours.
IO&nation of Hemagglutinins-Purified M. amurensis hemagglutinins were iodinated with 19 bv the chloramine-T method of kunter (21) using a 40-s exposure to the chloramine-T.
This procedure did not affect the hemagglutinating and mitogenic activities of the hemagglutinins.
Hemagglutination Assay-The titration and inhibition assays by use of human erythrocytes freshly obtained from a donor were carried out according to the methods previously described (15).
Lymphocyte Cultures for Mitogenic Assay-Human peripheral lymphocytes were cultured by the method previously described (22). The mitogenic activity of the hemagglutinin was expressed as a mitogenic dose (microgram per ml) to give 10,000 cpm of [6-3Hlthymidine incorporation against 3 X 105 cells. Inhibition assays for [6-sH]thymidine incorporation into the lymphocytes wi t.h sugars and glycopeptides were performed as described previously (23).

Preparation
of Purified Lymphocytes and Erythrocytes for Hemagglutinin-binding Studies-Human group 0 venous blood was withdrawn into syringes previously treated with heparin.
The heparinized blood was transferred to glass cylinders, and the erythrocytes were allowed to sediment by gravity.
The leukocyte-rich plasma was withdrawn, and layered over the same volume of the following solution in a glass centrifuge tube: 9.1 g of Ficoll (Pharmacia Fine Chemical, Sweden) and 25 ml of Urografin solution (600/,, Schering, Germany) in 125 ml of twice-distilled water. The centrifugation was performed at 400 g for 20 min. The white fluffy interface layer between the serum phase and the Urografin-Ficoll phase was removed by aspiration.
The average proportion of lymphocytes in this layer was higher than 98yo and, the viability was found to be 98 to 100% by the Trypan blue exclusion test. The cells were then washed with 0.25yo bovine serum albumin in 0.15 M NaCl-5 mM phosphate buffer (pH 7) and presented for the binding studies.
The erythrocyte layer, after removal of leukocyte-rich plasma and buffy coat, was washed three times by centrifugation with 0.25yo bovine serum albumin in 0.15 M NaCl-5 mM phosphate buffer (pH 7) each time carefully removing the top layer of cells. The erythrocytes thus obtained were found to be free from leukocytes or cell debris.

Binding
Studies-Binding reactions were carried out in siliconecoated tubes presoaked overnight with 0.25% bovine serum albumin solution in 0.15 M NaCl-5 mM phosphate buffer (pH 7). The reaction mixtures contained 4 X lo6 erythrocytes or 1 X lo6 lymphocytes, 0.4 to 8pg of '261-labeled hemagglutinin in a final volume of 0.3 ml of 0.25yo bovine serum albumin solution in 0.15 M NaCI-5 mM phosphate buffer (pH 7). After 90 min of incubation at room temperature with occasional mixing, the cells were washed twice with 3 ml of cold 0.259$ bovine serum albumin solution in 0.15 M NaCl-5 mM phosphate buffer (pH 7), and the amount of bound lz61 was determined in an Alloka Autogamma counter. When the radioactivity of the supernatant washings was checked, the dissociation of lzaI during the washing process was found to be negli-gible. Appropriate corrections were made for nonspecific binding to the tube.

Ammonium
Sulfate Fraclionation- Table  I summarizes data pertaining to the purification of hemagglutinins from M. amurensis seeds. From Table I, it can be seen that most of the hemagglutinating and mitogenic activities were precipitated by (NH&SO4 fractionation at 50 to 80% saturation (crude hemagglutinin).

Afinity
Chromatography on Porcine Thyroglobulin Glycopeptides-Sepharose @-The crude hemagglutinin was applied to a porcine thyroglobulin glycopeptides-Sepharose 4B affinity column in 0.15 M NaCl-1 mM phosphate buffer (pH 7.0) and elution was performed by the same buffer as shown in Fig. 2. After the bulk of protein was eluted in the void volume of the column, strong hemagglutinating activity appeared first, and strong mitogenie activity was recovered in the more retarded fractions. Thus, most of the inert proteins were effectively removed and, furthermore, the presence of two hemagglutinins differing in biological activity was ascertained by this chromatograhy.
These overlapping active fractions were combined (Fraction A) and subjected to SE-Sephadex column chromatography to clearly separate these two hemagglutinins.
In another experiment, the crude hemagglutinin was applied to a column of Sepharose 4B having the same size as the column in the case of the glycopeptides-Sepharose 4B adsorbent.
The hemagglutinating and mitogenie activities were, however, not separated from each other and recovered together as a single peak in the void volume of the column.
Chromatography on SE-Sephadex C-SO-Fraction A obtained by affinity chromatography was concentrated by ultrafiltration, dialyzed against 50 mM phosphate buffer (pH 5) and applied to a column of SE-Sephadex C-50 in the same buffer. After a large peak (Fraction Bl) was eluted, gradient elution was performed as shown in Fig. 3, and another protein peak (Fraction B2) was obtained.
The bulk of mitogenic activity was recovered in Fraction Bl, which had relatively weak hemagglutinating activ- ity, and very strong hemagglutinating activity was observed in Fraction 132, which had relatively weak mitogenic activity. These two fraction pools were further purified by passage over Sepharose 6B to remove trace contaminants. b d Chhromatography on Sepharose G&--Both Fractions Bl and B2 obtained by SE-Sephades chromatography were further purified by gel filtration on Sepharose 6B. As shown in Fig. 4, strong mitogenic and strong hemagglutinating activities were recovered in Fractions Cl and C2, respectively, each as a single peak. These fractions were designated MAM (strongly mitogenic) and MAH (strongly hemagglutinating).
The minimum hemagglutinating dose of MAH was 0.18 pg per ml against human erythrocytes regardless of their blood group types, whereas the mitogenic activity of this fraction against human peripheral lymphocytes was 5.3 pg per ml. On the other hand, the mitogenic activity of MAN against human peripheral lymphocytes was 1.2 c(g per ml, and its minimum hemagglutinating dose against human erythrocytes was found to be 1.5 pg per ml. These data are also listed in Table 1. Homogeneity and Molecular Weight Esfimation-Ultracentrifu-+ gation of both MAH and MAM yielded a single peak in each case during the whole of the run in the densitometer chart obtained from the experiment. The sedimentation coefficients (~20,~) of MAH and MAM calculated from the sedimentation velocity data were 7.1 and 7.4 S, respectively. The electrophoretic homogeneity of MRH or MAhI was confirmed by disc electrophoreses on polyacrylamide gel. A single band was observed in both instances at pH 8.9, as shown in Fig. 4 Polyacrylamide gel electrophoreses of XAII and MAM in the presence of sodium dodecyl sulfate without /%mercaptoethanol treatment gave a single discrete band in each case (Fig. 5). The molecular weights of these subunits were estimated to be 33,000 and 75,000, respectively.
ViThen MAH and MAM were subjected to electrophoresis in the presence of sodium dodecyl sulfate with P-mercaptoethanol treatment, however, both hemagglutinins gave a single discrete band (Fig. 5) corresponding to molecular weights of 33,000 and 35,000, respectively. These facts strongly suggest that MAH consists of four subunits of the approximate molecular weight of 33,000 and MAM consists of two disulfide-containing subunits of the approximate molecular weight of 75,000.
Chemical Composition-Results of amino acid and carbohydrate analyses of MAH and MAM are shown in Table II. The most notable feature of the amino acid composition of these proteins is the absence of cysteine in MAH and the high proportion of acidic and hydroxy amino acids in both hemagglutinins. Major carbohydrate constituents of both MAH and MAM were mannose and glucosamine.
The remainder of the carbohydrate was made up of smaller amounts of fucose, xylose, and arabinose in MAH, and fucose, xylose. arabinose, and galactose in MAM. Total carbohydrate contents of MAH and MAM were 9.5% and 8.7 To, respectively.   and concanavalin A, lymphocytes were treated with various quantities of the hemagglutinins.
The results are shown in Fig. 6. Maximum incorporation doses per 3 x lo5 cells were: 2 pg per ml for MAM, 20 pg per ml for MAH, 3 pg per ml for P. uulgaris hemagglutinin and 7 pg per ml for concanavalin A. In order to demonstrate (  Procedures"). The data have N-Acetyl-n-galactosbeen plotted by the method of Steck and Wallach (24)  d Porcine thyroglobulin glycopeptides were tested at concentrations no higher than 1.3 or 5 rmoles per ml because of their limited availability.
human lymphocytes was carried out and the data were plotted according to the method of Steck and Wallach (24) (Fig. 7). From the data, the apparent association constants and the average number of molecules bound per cell were calculated assuming molecular weights of 130,000 for both MAH and MAM, as shown in Table III. These data demonstrate that MAH or MAM binds to either erythrocytes or lymphocytes with almost the same association constant, but the apparent association constants of MAH for these cells are twice those of MAM.
Of interest are the findings that the number of binding sites for strongly mitogenic MAM on lymphocytes is much larger than that on erythrocytes, and strongly hemagglutinating MAH binds more on erythrocytes than on lymphocytes.
Inhibition of Hemagglutination with Various Sugars and Glycoproteins-The results of inhibition tests of hemagglutination with sugars and glycoproteins on MAH and MAM are given in Table  IV (50) 87 (16) 17 (25) >100 (25) Among various glycopeptides and glycoprotcins tested, porcine submasillary muciu (PSM) was foulltl to tlemollstrate marked inhibitory activity only against ;\IAII, but tlcsialyzntion of the glycoprotein gave rise to a remarkable loss of inhibitory activity. On the other hantl, glgcopeptitle 13 (Fig. I) from porcine thyroglobulin exerted potent inhibitory activity oiily against lIA1lI. In this cast, removal of sialic acid tlitl not affwt the iiihibitory activity, but the sequential enzymic tlcgratlation with purifictl &qilactosidasc, P-iV-acet~lglucosamillidnse, ancl oc-mannositlase gratlually diminishctl the inhibitory activity against MAM, leaving apprrciablc inhibitory activity evcu after cu-mannosidase treatment.
.1moiig the simple sugars ant1 oli~osnccliaritles t&cd, only 1\T-acetyl-i)-galactosamitle and maltow have motlerate iirhibitory activity against both MA\11 ant1 AIMI. III spite of the fact that glycopcptide 13 from porciiw th~roglobulin is not illhibitory agaillst, 1IAH in the l~emagglr~tillatioli-illllibitioll assays, this glycopeptide and its srqucntial ciizymic tlrgratlatioii products are potrnt inhibitiors against the mitogcllic activity of X11-I. The fact that the glgcopeptitlc after cr-mnnirositlasc treatment still retains sigilificant inhibitory activity against both hemagglutinins intlicates that the core region of the glycopcptitle can interact with the licmagglutiiiiiis.
A recent survey in our laboratory 011 the Ilcmaggl~itilratillg and mitogenic activities of the extracts of matry plnirt scctls rcvealed that JI. a,wrensis seeds containctl stroilg hcmagglutiiiating and mitogenic activitifx2 These activities wcrc cffcctivcly scparatetl from inactive proteins by afiinity chromatography using a column prepared by coupling the, glycopcptitles obtainetl by pronasc digestion of porcine thyroglobulin to activated Sepharosc.
As shown ill 'l'ablc I, the Ilclnagglutillatirlg activity of 1I.M is stronger than that of MXM, but the mitogcnic activity of MAH is much weaker than that of MAM.
111. terestingly, a gootl correlation was found htvmw thcsc biological characteristics of these hemagglutinins ant1 the results of binding studies of the '*jI-labeled hemagglutinins to crythrocytes and lymphocytes.
The apparent association constant of MA11 to human erythrocytcs was about twice that of MhbI, aud the number of receptor sites on crythrocytrs for MA11 was four times greater than for M.&M.
In contrast, the number of receptor sites on lymphocytes for MXH was approximately half of that for XIAAI, which was in turil 30 times more abundant than on erythrocytcs.
More striking tliffercnces bctneen MAH and X11\1 can bc seen iu their specificities for sugars. Ill hemagglutination inhibition assays, 1'SAI was found to be a