GEL FILTRATION STUDIES ON SOLUBLE POLYSACCHARIDES FROM SEEDS OF THE WINGED

Absirack The apparent molecular weight of three soluble polysaccharide fractions of winged bean seeds (Psophocarpru tetragonolobus [L.]) were investigated using gel filtration chromatography. Winged bean polysaccharides tend to be aggregated in water solution (molecular weight of 1 million plus), but dissociate in salt solution into a mixture containing components with an essentially continuous distribution of molecular weight (1 million to 1000). The degree of aggregation and disaggregation, and the range of molecular weight in solution, depends to some extent on the method of isolation and the method of analysis.


Introduction
As a part of an investigation aimed at characterizing winged bean (Psopl~oca~us tetragoi~olobits [L.]) seed polysaccharides, soluble and insoluble polysaccharides were isolated by different methods and their compositions were studied.' Analysis of the soluble fractions revealed a high proportion of galactose and small amounts of other neutral sugars, such as arabinose, xylose and glucose, and of the acidic sugar, galacturonic acid.
All fractions contained protein and ash.The present communication reports the results of investigations on the size distributions of the soluble polysaccharides.

Experimental isolation of Polysaccharides
Seed material and extraction procedures used have previously been described.'In this study, soluble polysaccharides were isolated by three methods: (i) the phenolacetic acidwater method (PAW), (ii), the trichloroaceticacid method (TCA) and (iii) the enzymatic method (Enz).The PAW method yielded two fractions, PAW-S fraction I and PAW-S fraction 11, while the other two methods yielded one fraction each, TCA-S and ENZ-S respectively.

Determination ,of Molecular Weight by Gel Filtration (GF)
The molecular weights (MW) of the soluble polysaccharides PAW-S, TCA-S and ENZ-S were determined by GF on '~ bondagel" high performance liquid chromatography columns (Waters Associates, Milford, MA).Columns with a variety of fractionation ranges were tested.An "E-125" column, with a reported fractionation range of 600-b0,000 and an "E-Linear" column with 1500-2,000,000 fractionation range yielded the best separations.In prelimizary experiments, 20 p l aliquots of a 5 mgml aqueous soluble polysaccharide solution were eluted with water at 0.3 mumin and the MW was estimated using dextran standards.Samples of the same concentration were also prepared in and eluted with 0.05M, 0.1M and 0.2M Na2S04, respectively, at 0.3 mvmin.Dextrans dissolved in Na2S04 were used as standards.The high purity dextrans were obtained from Pharmacia Fine Chemicals AB (Uppsala, Sweden) and had molecular weights of 10,000, 40,000, 70,000, 500,000, and 2 million.Plots of retention time against log molecular weight yielded a typical linear calibration curve within the fractionation range of the column and also indicated the void volume.Galactose dissolved in 0.05 M Na2S04 was used to establish the total permeation volume of the columns.Preparative scale GF of ENZ-S was also performed on Ultrogel AcA 44 (LKB Instruments, Rockville, MD.) with a reported fractionation range of 180-140,000.A 100 ml volume of gel was mixed with 40 ml of 0.05M Na2S04 and packed into a LKB 2137 chromatography column of 45 cm length and cross-sectional area of 2 cm2.An aliquot (1 ml) of a 10 mglml sample was loaded and then eluted with 0.05 M Na2S04 at 0.3 mumin.Fractions were collected and analysed for sugar by the phenoi-sulphuric acid method2 and uronic acid by the method of Ahmed and ~a b a v i t c h .~ In one case, the eluent was also monitored witn a refractive index detector.Some of the Cactions collected from the preparative columns were later analysed on the "E-Linear" column to obtain additional information from MW distribution.

Results and Discussion
isolated soluble fractions were analyseti by SF to estimate the MW of the polysaccharides and to determine if discrete polysaccharides with a particular structure could be isolated.
1.l preliminary studies witE a w?ter mobile phase, all polysacchariies were eluted :n the void volume indicating that the MW of 811 polysaccharide components were in excess of 1,000,000.When samples were dissolved in and eluted with 0.05 M Na2S04, there was still a high MW peak at or near the void volumes but additional peaks corresponding to intermediate or low MW components appeared on the chromatograms.These results suggested that ionisable groups, probably uroniues, and possibly also the associated protein caused the polysaccharides to be excluded from the gel pores by electrostatic repulsion, when samples were eluted with water.Addition of salt "saturated" the charges and minimized exclusion.A second possibility is that the added salt reduced aggregation of polysaccharidc molecules.I n any case, elution with 0.05 Na2S04 was adequate to bring about the maximum cfCcct (elutions with 0.05M, 0.1M and 0.2 M Na2S04 yielded essentially identical clution patterns).All subsequent G F analyses were therefore.performed on samplcs dissolved in and eluted with a 0.05 M NalS04 mobile phase.
Figures 1 and 2a and 2b show GF separations of the various solublc polysaccharide fractions on "E-125" and "E-linear" columns, respectively.Thc approximate mean MW versus dextran standards is shown for each peak.The elution profiles on these two columns for PAW-S fraction 11, TCA-S and ENZ-S fractions were similar.On the low MW(E-125) column, these three fractions yielded two broad, ill-defined peaks, one at the void volume (MW 40,000 or greater) and onc indicating a mean,MW in the range of 1500 (Figure 1).The profiles did not come to baseline between the peaks, indicating the presence of intermediate molecules.On the high MW (E-Linear) column (Figures 2a and 2b) the predominant peak for these three samples was at low MW, the mean MW being below 1200 for ENZ-S, about 3000 for PAW-S fraction I1 and about 5000 for TCA-S.Fraction I1 and TCA-S also showed evidence (shoulder peaks) for the presence of components with lower MW, in the range of 1600 to 2300.All three samples contained some material estimated at 1.1 to 1.8 million, as measured from the retention time of the early peak.All three samples also yielded one or more smaller peaks corresponding to components of intermediate MW.The ENZ-S fraction differed somewhat from the other two, in containing a higher proportion of high MW material, and less of intermediate and low MW materials.
PAW-S fraction I yielded noticeably different patterns.The G F profile on "E-125" was one broad, ill-defined "peak" indicating the presence of a range of MW from 40,000 + down to about 600 (Figure 1).The profile on "E-Linear" also indicated an essentially continuous spectrum of MW, but with one broad peak corresponding to MW of about 10,500 (Figure 2a).Comparison of PA.W-s fractions I and I1 profiles on both columns indicates fraction I1 to contain considerably more low MW material.Fraction I yielded a precipitate with 1.5 volumes of alcohol and fraction I1 with 5 volumes; therefore fraction I1 would be expected to contain material of lower MW.Further study of graded alcohol precipitation might be useful as it appears to be effective in separating high and low MW fractions.
Based on the separations obtained on the high performance "E-125" and "E-Linear" columns, preparative scale GF of polysaccharide fraction ENZ-S was performed on Ultrogel ACA 44. Figure 3 shows the MW profile obtained in the run monitored by refractive index.Polysaccharide fractions were collected as indicated in Figure 3 and analysed for total sugars and uronic,acids.Considerable difficulty (poor replicates, obviously too high values) was experienced with these analyses, due to interference of galactose with uronic acid analyses.However, the results did confirm the presence of both free sugars and uronic acids in fractions 2-6 (Figure 3).In a preliminary separation of the 5 ml fractions under the same conditions, (but not monitored by a refractive index detector) a plot of either sugars or uronic acid content yielded an elution profile similar to Figure 3.In this run, calculated recovery of the sugars added to the column was about 90%.The uronic acid recovery was about 89%, and again indicated the presence of uronic acid in both high and low MW peaks.The presence of sugar and uronic acid in all the peaks indicates aggregation, rather than exclusion from the gel pores by the charged groups of the uronic acid.
In an attempt to obtain additional information of the MW and purity of the high and low MW components, a sample each of fractions 2 and 6 (Figure 3) were reanalysed on the E-Linedr column.Fraction 2 yielded two peaks, the elution pattern being similar to that for the original ENZ-S (Figure 2b) indicating that additional disaggregation occured sometime during the collection and reanalysis.Fraction 6 yielded only a low MW peak, indicating no change in aggregation.
Overall, the GF results suggest that the soluble WB polysaccharides tend lo 5e a aggregated in aqueous solution.When dissolved in water, the aggregates are of high MW (I million plus).In 0.05 M Na2.504, a proportion of the polysaccharide is apparently dissociated into components of lower MW ranging from about 1 million down to components with MW in the oligosaccharide range (DP 3 to DP 10).The degree of disaggregation and particular range of MW obtained is quite variable and apparently depends in part on the method of isolation and in part on the method used to dissolve and analyse the sample.Thus each soluble sample yielded a distinct];/ different value of MW on the E-Linear column (Figures 2a and 2b) even though all samples were treated identically.The additional disaggregation observed in the sample collected from the preparative column and reanalysed, indicates that particular experimental manipulations will affect the extent of aggregation.In a study of water soluble and chelator soluble apple pectins, OJBeirne and Van ~u r e n ~ reported that the pectins occured in solutions as an essentially continuous distribution of high MW species in the range of 1 million to 12.5 million.Since the nectins had been isolated by "non-degradative" methods, and since the distributio~ was not affected by addition of upto 1 M salt, the authors concluded that these high MW species (probably aggregates) reflected an intrinsic property of these pectins.
The G F data for the WB polysaccharide fractions, especially PAW-S fraction I and I1 and TCA-S, show an essentially continuous distribution of MW, over the range of about 1000.This distribution, plus the observation that the WB polysaccharides occur as species with MW above 1 million in water, suggests strong aggregation as an intrinsic property of the WB polysaccharide:' However, unlike the apple pectins, the WB polysaccharides exhibit considerable disaggregation when dissolved in salt solution.
Finally it should be emphasized that the crude soluble W B fractions analysed via GF contained considerable (7-12%) protein.The protein, whether present as impurity-or as integral part of the-polysaccharides, probably affected the GF behaviour.Future studies of MW distribution, aggregation and related phenomena will require further purification of the polysaccharides.

Figure 1 :
Figure 1: GF chromntograms of WB soluble polysacchnrides on E-125 column.Arrows indicate approximate MW based on calibration curves prepared with dextrans on the same column under the same conditions.

Figure 2b :
Figure 2b: CF chromatograms of WB soluble polysaccharides on an E-Linear column.MW was as indicaled in Figure 1.

Figure 3 : 1 .
Figure 3: GF of ENZ-S (WB) on Ultrogel AcA 44 column MW was determined as indicaled in Figure 1.Fractions were collected as indicated by the horizonlal arrows.