COMPARISON OF SOLUTION BEHAVIOR OF POLYSACCHARIDES

Andrd M. Striegel 1 and Judy D. Timpa2"1University of New Orleans & zUSDA,

ARS, Southern Regional Research Center, P. O. Box 19687New Orleans, LA 70179

ABSTRACT

Polysaccharides function in key roles in textiles, biomass, foodstuffs, and plant cell walls, in addition to

pharmaceuticals and immunological interactions. Structural information is. badly needed, but complexity of the

polymers has been a limiting aspect. We investigated the solution behavior of polysaccharides representing typical

classes of natural products via GPC using refractive index and viscometric detectors. The solvent of choice for high-

molecular weight cellulose analysis, dimethylacetamide with lithium chloride, was used for direct dissolution without

extraction or derivafization. For more effective comparisons all of the compounds were dissolved in the same

soivant. Polysaccharides included a range of molecular weights, linear and branched polymers, as well as varying

linkages (i.e. celluloses, amyloses, amylopectins, dextrans, pullulans, and curdlan). In addition to molecular weights

and molecular weight distributions, calculated parameters defining the compounds' solution characteristics (intrln."_c

viscosity, radius of gyration, Mark-Houwink coefficients) were determined. Comparison between compounds with

similar molecular weights which differed either in their branching or their linkages were made. The calculated data

was correlated with theoretical results predicted in the titeratme and/or with known results, where such infonnmion

was available. Correlations of polymer linkage type and branching assessments with solution properties will be

presented.

'Corresponding author

567

6/INTRODUCTION

Gel permeation chromatography (GPC) when applied to the characterization of

polysaccharides offers a number of distinct advantages. Molecular weight averages and, more

importantly, molecular weight distributions can be calculated using this technique. Other

important solution parameters can also be obtained, such as intrinsic viscosity ([riD, radius of

gyration (Rg), polydispersity ratio (PD), and Mark-Houwink constants (a and log K), as well as

branching parameters. These last include branching frequency (_.) and branching number (Bn).

Physical properties are instrumental in the study of polysaccharides, as they are governedby the

molecular structure of the polymers. It is well known, for example, that a linear molecule

occupies a higher hydrodynamic volume than a branched molecule of equal molecular weight,

and thus solutions of the former will be moreviscous than those of the latter. In this fashion the

intrinsic viscosities and radii of gyration of polysaccharides can be directly related to chain

branching. Different properties or combinations of properties, such as chain length distribution,

branching of the molecular chain, type(s) of linkage(s) within the molecule, have a direct bearing

on the functional differences between polysaccharides. GPC provides an excellent means of

quantifying many important solution properties of polysaccharides.

Polysaccharides such as dextrans, pullulans, cellulose, are of great importance in the

medical and pharmaceuticalfields, as well as in the food, pulp and paper, and textile industries,

among others1 (Table I). They are essential to virtuallyall biological systems2. Optimizing the

functions these compounds perform is directly dependent upon knowledge of their structureand

molecular weight distribution(MWD), and the extent to which these affect polymer behavior.

This can only be achieved through a techniquesuch as GPC thatcan quantify these parameters

to establish effective comparisonsamong different polysaccharides.

While a number of polysaccharides are water-soluble and can thus be analyzed using

aqueous mobile phases, this property is by no means universal. Investigation of a number of

importantand ubiquitous polysaccharides such as cellulose and starch has been hinderedby their

568

poor solubility. Isolation and fractionation prior to solubilization, which detracts from

comparisons with the native or starting material, is often performed. Thus a solvent which

dissolves a wide range of carbohydrates (i.e. a variety of polysaccharide types over a wide

molecular weight range) is extremely important in GPC. In this study we have used N,N-

dimethylacetamide/lithium chloride (DMAc/LiC1), the solvent of choice for high-molecular

weight cellulose analysis, to dissolve all of the polysaccharides under consideration. The

appropriateness of choosing this solvent will be addressed.

In this work, GPC of polysaccharides, dissolved in DMAc/0.5% LiC1, has been used to

probe the dilute solution behavior of sugars representing different classes of polysaccharides. The

concept of universal calibration 3 has been applied (see Results and Discussion/Universal

Calibration). Variations in dilute solution properties have been related to both molecular weight

distribution and branching. Additionally, a comparison is made between number-average

molecular weight (1_) as determined with and without the use of a concentration (refractive

index or RI) detector. The latter average is generally referred to as the Goldwasser 1_ 4

EXPERIMENTAL

Materials. All polysaecharides analyzed are commercially available, and were used

without further purification. Samples included celluloses (J. T. Baker, Cat. Nos. 1525-1 and

1528-1); pullulans (Pfanstiehl, Cat. Nos. 12474 and 12476); dextran T10, T40, T70, TS00,

T2000, (Pharmacia, Code Nos. 17-0250-01, 17-0270-01, 17-0280-01, 17-0320-01, 17-0330-01,

respectively); dextran Low Fraction, High Fraction (J. T. Baker, Cat. Nos. G200-5, G202-05,

respectively); dextran Industrial Grade (Sigma, No. D-5501); arabinogalactan (Atomergic

Chemetals, 6030); corn amylose (Sigma, No. A-7043); potato amylose (Sigma, No. A-0512);

corn amylopectin (Sigma, No. A-7780); potato amylopeetin (Sigma, No. A-8515). Materials were

N,N-dimethylacetamide (Burdick & Jackson), dried with molecular sieves (Baker, Activated Type

3A); lithium chloride (Baker), oven-dried and stored in desiccator; I0 mL ReactaVials (Pierce);

heating block (Pierce); Teflon magnetic stirbars (2.5 cm); Baker 10 extraction apparatus;

disposable Teflon filters (Millipore, Millex-SR and TYPE FI-I, 0.5 l.tm); 10 em 3 glass syringes

569

(BD); 4 mL WISP vials with Teflon septa; 50 mL volumetric flasks. Polystyrene standards were

from Toyo Soda Manufacturing (Types F-288, F-20, F-80, F-10, F-128, F-4, F-40, F-2, A-5000,

F-I, with nominal molecular weights of 2.89 xl0 6, 1.9 xlO s, 7.1 xl(Y, 1.02 xlO s, 1.26 xl&,

4.39 xl04, 3.55 xl0 s, 1.96 xl04, 6.2 xl&, 1.03 xl(Y, respectively).

Sample Preparation. Samples were dissolved generally following the procedure reported

for cotton s. 30 mg of polysaccharide was added to 5 mL DMAc in 10 mL ReactaVials with a

conical magnetic stirrer in a heating block. The temperature was raised to 150"C and maintained

with stirring for 1 hour. The mixture was allowed to cool to 100"C. 0.250 g dried LiC1 was

added. The vials were shaken by hand and returned to the heating block, where the mixture was

maintained with stirring at 100*C for 1 hour. The temperature of the block was lowered to 50"C

and samples were stirred at this temperature overnight. The solutions were quantitatively

transferred to 50 mL volumetric flasks and diluted to volume with DMAc. They were then

filtered through a solvent-resistant Teflon disposable filter. An extraction apparatus was employed

with 10 cm 3 glass syringes fitted onto filters with 4 mL glass vials held in the small volumetric

holder. The final concentration of each polysaccharide was 0.6 mg/mL in DMAc/0.5% LiC1.

Chromatography. The mobile-phase/solvent for GPC was DMAc/0.5% LiCI prepared

by raising the temperature of 1 L of DMAc to 101YCand then adding 5 g of dried LiC1. The salt

was stirred until it dissolved, and the solvent was filtered through a Teflon filter with a glass

filter apparatus 5. Filtered samples were analyzed on a GPC system consisting of an automated

sampler (Waters WISP) with an HPLC pump (Waters Model 590), pulse dampener (Viscotek),

viscometer detector (Viscotek Model 100), and refractive index detector (Waters Model 410),

at a flow rate of 1.0 mL/min. The detectors were connected in series with the refractive index

detector last due to back-pressure considerations. The columns configuration consisted of three

101aMixed-B columns (Burdick & Jackson/Polymer Laboratories) preceded by a guard column

(Burdick & Jackson/Polymer Laboratories). The system was operated at 80"C, with temperature

controlled by a column heater (Waters column temperature system). Injection volumes were 100

and 150 laL with a run time of 34 rain per sample. Data acquisition and calculations were

performed using the software package TriSEC GPC 6. Universal calibration was determined with

570

polystyrene standards dissolved directly in DMAc/0.5% LiC1. The universal calibration curve

was linear as a logarithmic function of the product of the intrinsic viscosity times molecular

weight versus retention volume using a third-order fit. Data were obtained from two dissolutions

per sample with two GPC runs per dissolution.

RESULTS AND DISCUSSION

Solution Properties of Polysaccharides

In analyzing dilute solutions of the polysaccharides in T_ble I dissolved in DMAc/LiCI

via GPC, the following parameters were determined: weight-average molecular weight (/_),

number-average molecular weight (_7I,),Goldwasser _I,, molecular weight polydispersity ratio

(PD = _ (Table II), weight-average intrinsic viscosity ([rl]_), weight-average radius of

gyration (Rgw),Mark-Houwink constants (a and log K) (Table liT), and branching number (Bn)

and branching frequency (X).Branching calculations were made according to the theory of Zimm

and Stockmayer via TriSEC software6 and correlated well with calculated parameters. Small

deviations in branching results are likely due to underlying assumptions in the theory involving

long-chain branching and theta (0) solvents7"8.

Even though the dextrans are generally described as being branched glueans, it has been

reported that lower molecular weight (MW) dextrans possess a near-linear StruCtUre 9. Previously

reported results from our group support this conclusion and extend the range of near-linearity to

-< -2.5 × I05. The importance of this fact will become evident in the discussion to follow.

A Mark-Houwink equation for near-linear dextrans dissolved in DMAc/LiC1 was derived7, with

the values a = 1.3 and K = 1.7 × 107. A plot of the MWD of several dextran standards with

varying _ is given in Figure 1. It can be seen that the lower MW dextrans (T10, T70) have

fairly narrow MWD and polydispersity ratios -I-1.5, while the higher MW dextran (T2000) has

a broad MWD and large PD. Intrinsic viscosity was found to be more correlated to radius of

gyration (r = 0.958) than to lVI,,(r = 0.667), which is the ease for other classes of compounds 23.

571

As mentioned above, a linear molecule will occupy a larger hydrodynamic volume than

a branched molecule of equal molecular weight. The latter's higher coil density will be reflected

in lower values for its [rl] and Rg (as well as higher values for _. and Bn), and deviations from

linearity will be observed in the dependence of log [rl]wvs. log _7L,1°'1_.Archetypal behavior can

be observed when comparing arabinogalactan with dextran LF. The first is a highly branched

galactan consisting of two main fractions. Approximately 10% of the molecular weight fraction

is located below MW = 20,000 and comprises the smaller of the two main fractions, with 1_

~ 14,000. The larger fraction comprises the other approximately 90% of the molecular weight

fraction, with I_ ~ 90,000. The overall 1_ of this polysaccharide is 8.4 x 104. Dextran LF is

a near-linear dextran (see definition above) with an l_w equivalent to that of arabinogalaetan

(Table II). Even though the values for _ for these two saccharides are nearly identical, large

differences in their solution behavior earl be noted by comparing such parameters as [rl],, (0.06

dL/g for arabinogalactan vs. 0.38 for dextran LF) and Rg,, (5.2 nm for arabinogalactan vs. 10.0

for dextran LF) (Figure 2 and Table 11). The constant a, which has proven to be a good indicator

of branching for the dextrans, is similarly much smaller for arabinogalactan (-0.1) than for the

dextran (1.1) (Table 1/I). This wide variability in solution behavior can be explained by the

extensive branching of the arabinogalaetan (X = 0.61, Bn ~ 300). Much smaller values would

be expected for the near-linear dextran for the branching argument to hold, and indeed _ = 0.08

and Bn ~ 9 for dextran LF.

A further example of the influence of branching of the main polysaccharide chain on

solution behavior is given by contrasting potato amylose with dextran T500. The former is a

linear gluean while the latter is branched (Table I). Their NI_ values are fairly close (< 15%

difference), although the MWD of dextran T500 is slightly higher (Figure 3). Significant

differences may be observed in the high-MW portion of the dextran, where the majority of the

branching is believed to occur in dextrans 7. Here its [rl]w and Rg_ are lower than those of the

amylose.

When comparing the two major components of starch, amylose and amylopectin, to each

other, the second had a higher M_, regardless of the source (corn or potato). While the branched

572

amylopectins did have a lower [_],,. than the amyloses (Table III), it is not prudent to compare

solution parameters for these two classes of polysaccharides due to the large differences in their

1_,, values and MWD (Table II and Figures 4 and 5). This set of examples showcases the ability

of the GPC technique to provide multiple profiles of a number of important solution parameters.

Cellulose and pullulan are both linear molecules with a wide variety of applications (Table

I). The broad MWD and low MW ranges observed for the cellulose powders analyzed (Table II

and Figure 6) are in direct contrast to the high MW secondary wall component of native cotton

fiber5. Cellulose 5 was used as a linear polymer reference for the branching calculations. Even

though pulhlan is a linear carbohydrate, a small degree of branching is observed (Figure 7).

While it is impossible from this study to conclude with certainty the reason for the branching,

we postulate what we believe to be a likely cause. The structure of pullulan consists of repeating

units of maltotriose (a-D-(l_4) linkages)joined by tx-D-(I---_6) linkages in a step-wise manner.

In addition, pulhlan contains -6.6% ofa maltotetraose subunit 2'. It should not be concluded that

the maltotetraose units are arranged in such a fashion as to cause short-chain branching, as Catley

and Whelan showed that the tetramer is linked exclusively through its ends in pullulan 12.We

propose that it is possible that the combination of different repeating units, linkage types, and the

step-wise arrangement of the molecule all contribute to the amount of branching observed.

Choice of Solvent

A number of polymer properties are dependent upon the solvent used. Macromolecular

chains or coils tend to form secondary valence bonds within the molecules themselves. If a poor

solvent is used, undissolved associated segments will be found within the coil, and consequently

the coil will be tighter and its viscosity lower. A molecule dissolved in a better solvent will have

fewer secondary valence links within its coil, as molecules of the dissolved material will form

secondary valence bonds with molecules of the solvent. The coil density will thus be lower and

the viscosity higher. Polymers such as cellulose, which possess strong secondary valence links

in the form of hydrogen bonds will be dissolved only if the solvent attaches itself to the chain

with a strength equal to or greater than that of the hydrogen bonds. A logical conclusion to be

573

drawn is that any solvent capable of dissolving cellulose will be by definition a good solvent for

this type of molecule. In such a solvent coil densities will be low and, accordingly, viscosities

will be high s'l°.

At the theta (0) temperature s the chemical potential of the chain-solvent reaction in the

volume element becomes zero. Accordingly the 0 temperature is also known as the ideal

temperature, and usually lies in the range 0-40°C. If the exponent a = 0.5 in a solvent at the

0 temperature, the solvent is known as a 0 solvent. Good solvents exert a buttressing effect on

the molecular backbone; the coil will be stiffened and expanded, and a will not be 0.5 but a

much higher value which can reach 1 or 2 for bulky and/or maximally extended chains 1°.

The DMAc/LiC1 solvent system has been employed to dissolve cellulose 13._4.Our

laboratory has developed procedures for the dissolution and characterization of high MW

cellulose chains in cotton fiber without the need for prior extraction, derivatization, or

fractionation 5. In this study branching was related to decreases in the value of a (Table 111).

Intrinsic viscosity provides an index of random coil dimensions and chain conformation through

the Mark-Houwink equation _. Values for [rl],, <- -1.0 were observed with the exception of

arabinogalactan, which has been addressed above. All carbohydrates were dissolved without

difficulty. Comparison of observed solution behavior to that characteristic of polymers in dilute

solutions leads to the conclusion that DMAc/LiC1 is a good solvent for polysaccharides. A major

advantage gained by employing this solvent system is that it can be used both for dissolving the

starting material as well as the processed or derivatized products t3-ts'_6.Therefore the solvent used

for dissolution is the same as the mobile phase being utilized in the GPC analysis, and the

polymer properties will not vary as a function of solvent.

Another parameter of importance in GPC analysis of polysaccharides is temperature. It

will exert an effect on the secondary valence bonds, and consequently on parameters such as [rl].

Accordingly, it is extremely important that the temperature be kept constant over the course of

the experiment. To this effect a column heater was used to control the temperature and maintain

it at 80°C. In addition, the elevated temperature assists in decreasing the viscosity of the mobile

574

l ]phase.

Goldwasser 1_7I.

The number-average molecular weight, 1_,, provides a way of measuring the Maverage"

chain length in a polymer sample. It has been correlated with a number of polymer physical

properties such as fiber and film strength, polymer solubility, fiber tenacity, solution viscosity,

etc 17.Goldwasser has proposed a method for the determination of 1_ using an online differential

viscometer as the sole detector 4. This method eliminates the need to utilize Mark-Houwink

constants or to determine the weight fraction of the eluting sample with a concentration (e.g.

refractive index) detector. He derived (I) to calculate the number-average molecular weight.

1_n = {3_bC_V_}/{4/_VsE(rl.p,/Vhi) } (1)

Where t_ = Flory constant

C_ = Injected polymer concentration

VI = Injection volume

Vs = Slice volume

q,pi = Specific viscosity of eluting species i

Vhi = Hydrodynamic volume of eluting polymer species i

Goldwasser obtained good agreement between the number-averages determined by his

method and those expected using Mark-Houwink constants for several different polymers and

polymer mixtures. Other research _oups who applied this concept to the characterization of

copolymers by GPC found that Goldwasser lfl. values sometimes correlated well with averages

obtained by other methods (e.g., light scattering), but sometimes showed large deviations which

remained unexplained is.

Comparison of the l_n obtained by both methods can be made from those results shown

in Table II. 1_ values obtained with and without the concentration detector follow the same

575

pattern detected by Gores and Kilz, sometimes showing good agreement with each other and

sometimes not. The overall correlation between the two was poor (r2 = 0.641). A different

picture appears when looking at the molecular weight polydispersity ratio, lVI,JlVl,.This ratio is

a measure of the breadth of the polymer molecular weight distribution. Results derived from

using the 1Vl,obtained with the refractive index detector were comparable to those obtained using

the average derived from Goldwasser's method (r2 = 0.999). While no logical explanation has

been found for the sometimes large deviations in 1VI,values from method to method, we agree

with other researchers in that calculated values are dependent upon baseline settings and baseline

noise tS. It is likely that these factors become even more important at the lower molecular weight

end of the chromatogram, which is more heavily influential upon the number-average molecular

weight.

Universal Calibration

Universal calibration as applied to GPC takes into account the behavior of the polymer

in solution, thereby providing accurate MW values 3. Linear calibration curves are obtained

employing this method, which describes the relationship between hydrodynamic volume (log MW

x [rl]) and retention volume. It avoids the problem of limited availability of polymer standards

having identical chemical structure and narrow MWD at discrete levels of molecular weight over

the range of MW of the unknown polymers. It was thus possible to use polystyrene (PS)

standards of sharp MWD, known MW and covering a wide range of MW valuess. This

methodology has been applied in our laboratory for monitoring cell wall polymers during cotton

fiber development _9'-'°, and was subsequently extended to quantify the effects of extrusion

processing on starches from corn and wheat tS'2t'22.PS standards were also employed in the

calculations for the samples described in this paper. Universal calibration curves were obmlned

with the carbohydrates of known MW studied here (data not shown). Although the data appeared

to agree with the PS calibration, the carbohydrate standards do not extend over as large a MW

range as the calibration curve obtained with the polystyrene stalldards 7.

576

CONCLUSIONS

GPC with differential viscometer and refractive index detectors is an effective means of

characterizing important solution parameters of polysaccharides dissolved in DMAc/LiC1, over

the range of the MWD. Applying the concept of universal calibration based on polystyrene

standards afforded 1_ values comparable to those supplied with the carbohydrates. This curve

extended over a greater MW range than curves obtained using the carbohydrates. The parameters

determined were related to branched versus linear composition of the polymers within the limits

of the determinations, lVlnvalues determined with the refractive index as a detector and without

(Goldwasser's approach) show wide variability in their correlations, sometimes being closely

correlated and sometimes not. While no suitable theory has emerged to explain this behavior,

molecular weight polydispersity ratios calculated using both number averages do show good

correlation. Finally, DMAe/LiC1 seems to behave as a thermodynamieaUy good solvent for

polysaccharides, one in which dilute solution behavior can be readily studied.

ACKNOWI _EDGEMENTS

We thank Dr. Richard Cole (University of New Orleans) for his valuable advice and partial

sponsorship of this work under USDA, ARS Cooperative Agreement 58-6435-2-110. We also

thank Drs. Margaret Clarke and Les Edye (Sugar Processing Research Institute) for the gift of

the dextrans, Drs. Bruce Wasserman and Mary Politz (Rutgers University) for the amyloses and

amylopectins, and Dr. Barbara Triplett (SRRC) for helpful discussions.

NOTES

Names of companies or commercial products are given solely for the purpose of providing

specific information; their mention does not imply recommendation or endorsementby the Unites

States Department of Agriculture over others not mentioned.

577

SAFETY CONSIDERATIONS

N,N-dimethylacetamide is an exceptional contact hazard that may be harmful if inhaled or

absorbed through skin and may be fatal to embryonic life in pregnant females (Baker Chemical

C. N,N-dimethylacetamide, Material Safety Data Sheet, 1985, D5784-01; pp. 1-4).

REFERENCES

1. R.L. Whistler and J.N. BeMiller (Eds.), Industrial Gums, 3rd ed., Academic Press, San

Diego (1993).

2. A. Varld, Glycobiol., 3 (1993) 97-130.

3. A. Grubisic, P. Rempp, H.A. Benoit, Polym. Lett., 5 (1967) 753-759.

4. J.M. Goldwasser, in T. Provder (Ed.), Chromatography ofPolymers -Charcterization by

SEC and FFF, ACS Symposium Series 521, American Chemical Society, Washington,

D.C. (1993) 243-251.

5. J.D. Timpa, J. Agr. Food Chem., 39 (1991) 270-275.

6. TriSEC GPC Software, Release V 2.61, (c) 1994, Viscotek Corp.

7. A.M. Striegel and J.D. Timpa, Carbohydr. Res. (in press).

8. P.J. Flory, Principles of Polyn_r Chemistry, ComeU University Press, Ithaca, N.Y.

(1953) 610-612, 622-626.

9. K. Gekko, Makromol. Chem., 148 (1971) 229-238.

10. B. Vollmert, Polymer Chemistry, Springer-Verlag, New York (1973) 513-529.

11. L. Vilenchik and R. Ayotte, J. App. Polym. Sci.: App. Polym. Syrup. 51 (1992) 73-90.

12. B.J. Cafley and W.J. Whelan, Arch. Biochem. Biophys., 143 (1971) 138-142.

13. C.L. MeCromick, P.A. CaUais and B.H. Hutchinson, Macromol., 18 (1985) 2394-2401.

14. A.F. Turbak, A. E1-Kafrawy, F.W. Snyder and A.B. Auerbach, U.S. Patent 4,302,252

0981).

15. M.L. Politz, J.D. Timpa, A.R. White and B.P. Wasserman, Carbohydr. Polym. 24

(1994) 91-99.

578

16. T.R. Dawsey and C.L. McCormick, JMS-Rev. Macromol. Chem.Phys. , C30 (1990) 405-

440.

17. W.W. Yau, J.J. Kirkland,and D.D. Bly, Modern Si2v-Exclusion &'quidChromatography,

John Wiley & Sons, New York (1979).

18. F. Gores and P. Kilz, in T. Provder (Ed.), Chromatography of Polymers -Characteriza-

tion by SBC and FFF, ACS Symposium Series 521, American Chemical Society,

Washington, D.C. (1993) 122-148.

19. J.D. Timpa, in T. Provder (Ed.), Chromatography of Polymers -Characterization by SBC

and FFF, ACS Symposium Series 521, American Chemical Society, Washington D.C.

(1993) 302-313.

20. J.D. Timpa and B.A. Triplett, Planta, 189 (1993) 101-108.

21. B.P. Wassermanand J.D. Timpa, Starch/Starl_, 43 (1991) 389-392.

22. M.L. Politz, J.D. Timpa and B.P. Wasserman, Cereal Chem. (in press).

23. M.L. Fishman, D.T. Gillespie and B. Levaj, in T. Provder (Ed.), Chromatography of

Pol3_ners -Characterization by SEC and FFF, ACS Symposium Series 521, American

Chemical Society, Washington, D.C. (1993) 314-325.

24. Y.Tsujisaka and M. Mitsuhashi,in I. 447-460.

579

Table I

Structure and applications of selected polysaccharides.

Polysaccharide Branching and Linkage Applications

Amylopeetin branched (l---_4)-a-D-glucan Food product.

Amylose linear (1---_4)-a-D_lucan Food product.

Arabinogalactan branched (l---_3)-I_-D-galactan In printing, mining,pharmaceutical, andfood industries.

Cellulose linear (1--+4)-[_-D-glucan In textile, pulp andpaper, and plasticsindustries.

Dextran branched (l_6)-_-D-glucan In the medical field.In the cosmetics and

photography indus-tries. Chromato-

graphy standards.

Pullulan linear (l---_6)-ot-D-glucan In the food, printing,plastics, cosmetics,medical, and pharma-ceutical industries.Chromatographystandards.

580

Table !1

Calculated molecular weight averages of polysaccharides dissolved in DMAc/LiCI.

Sample ID l_w l_w" l_. ° 1_.° PD(I) b PD(2y(Supplied) (with RI) (Goldwasser)

Amylopectin (Potato) N/A d 1.2 × 106 5.8 x 104 7.1 x 104 21.4 18.0Amylopectin (Corn) N/A 2.1 x 107 6:1 x 104 3.3 x 104 349.0 689.0

Amylose (Potato) N/A 4.9 x I0"_ 5.0 x 104 4.3 x 104 9.8 I 1.4Amylose (Corn) N/A 6.2 x lO-_ 8.6 x 104 7.3 x 104 7.2 8.7Arabinogalactan 8.0 x 104 8.4 x 104 3.2 x 104 3.6 x 104 2.6 2.3Cellulose 4 1.8 x 10"_ 1.8 x 10"_ 2.7 x 104 2.7 x 104 6.8 6.8Cellulose 5 3.2 x 10_ 3.3 x I05 4.2 x 104 4.6 x 104 7.9 7.2

Dextran TIO l.O × 104 1.9 x IO4 1.7 x 104 1.8 x 104 I.I 1.0Dextran T40 4.4 x IO4 4.7 x IO' 3.6 x IO' 3.4 x 104 1.3 1.3Dextran T70 7.0 x IO4 7.6 x 104 4.8 x IO' 3.9 x 104 1.6 2.0

Dextran LFt 6.0 - 9.0 x 104 8.4 x lO' 6.4 x 104 2.5 x 104 1.3 3.4Dextran HF g 2.0 - 3.0 x !0 "_ 2.5 x IOs 5.6 x 104 3.8 x 104 4.4 6.6Dextran T500 5.0 x 10"_ 5.5 x 10"_ 6.8 x 104 4.5 x 104 7.9 12.7Dextran T2000 2.0 x 106 !.9 x I0_ 5.4 x 104 5.4 x !04 36.2 36.3 o_Dextran 1Gh 5.0- 40.0 x 106 5.1 x 106 4.8 x 104 4.8 x 104 104.9 108.8

Pullulan I 1.0 x l0 "_ l.l x l0 s 3.5 x l04 3.4 x 104 3.0 3.2Pullulan 3 3.0 x l0 "_ 3.1 x l0 _ 3.4 × 104 3.3 x l04 9.3 9.6

"Values calculated using GPC. bPD(I) = M,,/I_, (with RI). 'PD(2) = l_wlM, (Goldwasser). JN/A = Value not available.Walues from ref 13. rLF = Low Fraction. SHF -- High Fraction. biG = Industrial Grade.

Table III

I Calculated solution parameters for polysaccharides in DMAc/LiC1.

Sample [D [qk Rg,, Mark-Houwink

(dL/g) (tam) a . log K

Amylopectin (Potato) 0.59 25.7 0.4 -2.7

Amylopectin (Corn) 0.40 33.0 0.2 -1.9Amylose (Potato) 0.83 17.7 0.9 -5.1Amylose (Corn) 0.65 19.2 0.7 -4.0Arabinogalactan 0.06 5.2 -0.1 -0.9Cellulose 4 1.04 14.0 1.0 -5.1Cellulose 5 0.71 16.2 0.7 -4.0Dextran TI0 0.15 4.5 1.1 -5.5

Dextran T40 0.29 7.3 1.8 -8.8Dextran T70 0.40 9.4 1.4 -7.4Dextran LF_ 0.38 10.0 1.1 -5.8Dextran HF ° 0.64 16.2 0.5 -3.0

Dextran TS00 0.84 21.0 0.4 -2.3Dextran T2000 0.78 26.1 0.4 -2.5Dextran IGc 0.14 21.1 0.4 -3.6Pullulan 1 0.69 12.3 1.1 -5.6Pullulan 3 0.89 14.5 0.8 -4.3

*LF = Low Fraction. bHF = High Fraction. ClG = I.ndn_trial Grade.

582

rLEGENDS FOR FIGURES

Figure 1. Relationship between molecular weight distributions (MWD) of dextrans TI0,

T70, and T2000.

Figure 2. Solution parameters of arabinogalactan. (A) Molecular weight distribution

(MWD). (B) log-log plot of intrinsic viscosity ([rl]) vs.

molecular weight (MW). (C) Plot of branching frequency CA)vs.

log molecular weight. (D) log-log plot of radius of gyration (Rg)

vs. molecular weight (MW).

Figure 3. Solution parameters of arabinogalactan. (A) Molecular weight

distribution (MWD). (B) log-log plot of intrinsic viscosity

([rl]) vs. molecular weight (MW). (C) Plot of branching frequency

CA)vs. log molecular weight (MW). (D) log-log plot of radius of

gyration (Rg) vs. molecular weight (MW).

Figure 4. Solution parameters of corn amylose. (A) Molecular weight

distribution (MWD). (B) log-log plot of intrinsic viscosity

([rid vs. molecular weight (MW). (C) Plot of branching frequency

CA)vs. molecular weight. (D) log-log plot of radius of gyration

(Rg) vs. molecular weight (MW).

Figure 5. Solution parameters of corn amylopectin. (A) Molecular weight

distribution (MWD). (B) log-log plot of intrinsic viscosity

([rid vs. molecular weight (MW). (C) Plot of branching frequency

CA)vs. molecular weight. (D) log-log plot of radius of gyration

(Rg) vs. molecular weight.

Figure 6. Solution parameters of cellulose 4. (A) Molecular weight

distribution (MWD). (B) log-log plot of intrinsic viscosity

583

([rl]) vs. molecular weight (MW). (C) Plot of branching frequency (_.) vs.

molecular weight. (D) log-log plot of radius of gyration (Rg) vs. molecularweight.

Figure 7. Solution parameters of pullulan 1. (A) Molecular weight distribution (MWD).

(B) log-log plot of intrinsic viscosity ([q]) vs. molecular weight (MW). (C) Plot

of branching frequency ('A)vs. molecular weight. (D) log-log plot of radius of

gyration (Rg) vs. molecular weight.

584

Figure 1.

,oo Molecular Weight Distribution of Dextrans

T10

3.00- _ T70,_ f

0..! 2.00-"10

1.00-

2.O0- -O.50-

1.00- _ ..

• 150-

0.00 , -2,50! !3oo ,0o _0o e0o 7_o aoo 30o ,0o 5_o 6_o '• 7.00 8.0O

Log(MolecularWeight) Log (MolecularWeight) _o

,5.o Lambdavs. MW ofArabinogalactan ,.= Rg vs. MW of Arabinogalactan oo

12.0-

_) 1.00"

U.

.G

"_C 6.0-0.50-

rn

3.0-0.0 J 0.0O3.00 ,.0o 5.0o e._o _._o a.0o 3.00 ,._o 5._o e._o 7._o a.0o

Log(MolecularWeight) ..:Log(MolecularWeight)

_.)- • ....

IV vs. MW of DextranT500MWD of DextranT500 o.5o

; 1.50 I

0._1"

1.00"

• _ _ -0.I_"

0.50-

: .1.1111-J.

o.oo . 7:oo a.oo 3.oo ,.oo s._o 6.'0o 7:o0 8.oo3.00 4.00 5.00 6.00

Log (MolecularWeight) Log (MolecularWeight) t--

Lambda vs. MW of Dextran T500 zoo Rgvs. MW of Dextran T5002.50

2.00-

-_'O 1.50

x

_ 1.50- _,- _ 1.(_-

._= l._-r-

_ o._-IX)

0.50-f

• _ _ J o._o.oo , ' 4._o s._o 6._o 7._o 6.go3.o0 ,,.oo 5:oo sJoo 7.00 6.oo =.ooLog (MolecularWeight) " Log(MolecularWeight)

/

f j

|

O.OC -I.S0• !

3.o0 _oo 5.oo 6.oo 7.oo 8,oo 3.oo 4.oo 5._o e._o 7._o 8.oo

Log(MolecularWeight) Log(MolecularWeight)oo

,._o Lambda vs. MW of Corn Amylose Rg vs. MWof CornAmylose _

bx

.1 1.50-

U.

GI

IO.OO L. O.OO

i

3.00 ,.oo s._o e._o 7._o a.oo 3.oo ,._o s._o e._o 7._o aooLog (MolecularWeight) .Log(MolecularWeight)

• AbUA_; ,..),

MWD of CornAmylopectin IV vs. MW of Corn Amylopectin0.00 0.50

I

0.00 , , -250 ,3.00 ,.oo 5._o 6._o 7.00 a.oo 3.oo ,,.oo 5._o e_oo 7;00 6.00Log(MolecularWeight) Log (MolecularWeight) oo

Lambdavs. MW.of CornAmylopectin Rg vs. MW of CornAm_5.00

4.00-

¢..

O').cJ_ 2.00-e-

1.00-

0.00 _---- _ '3.0o ,ioo s_oo 67oo _;oo a.oo 3.oo ,._o s._o e._o 7._o aooLog (MolecularWeight) Log(MolecularWeight)

0.50-

1.00-

-0.50-

0.50--

-1.00-

0.00. , -1.503.oo ,.oo 5.oo e.oo 7._o a.oo 3.oo ,._o 5._o 6._o 7._o aoo

Log(MolecularWeight) Log(MolecularWeight)

o.oo Lambdavs. MW of Cellulose4 2.= Rg vs. MW of Cellulose4 o_

8.00- 2.00-

X

i e°°- _."_

_) 4.0o- 1.0o-,mr-¢.)C

O0

2.00- 0.50"

0.00- , 0.003oo _._o 5._o e_ ",_oo a.oo 3.0o 4Joo s:oo e._o _.oo 80oLog (MolecularWeight) _- Log(MolecularWeight)

MWD of Pullulan 1 IV vs. MW of Pullulan 11.50 1.50

/_ 0.50-1.00-

0.50-

0.00 _ -2.50300 ,'00 s._o e_oo 7._0 aoo 3.00 4_00 s_o e_o 7._0 aoo

Log(MolecularWeight) Log(MolecularWeight)

Lambda vs. MW of Pullulan1 Rg vs. MW of Pullulan1 o_ZOO 2.00

I

1.50- 1.50-

LL, 1.00- 1.00-

II1

0.50- 0.50-

,o.oo ..... o.ooi

_.oo __ s'oo e_oo 7.'oo a.oo 3.oo ,._o 5._o 6._o 7._ 8oo

Log (,MolecularWeight) ' Log(MolecularWeight)

m ............ i ,, ,, i i ii ........ -- -- _ __