Analysis of underivatised d -gluco-oligosaccharides (d.p. 2–20) by high-pressure liquid chromatography

Carbohydrate Research, 57 (1980) 165-l 73 Elstvier Scientific Publishing Company, Amsterdam

ANALYSIS OF UNDERIVATISED Z-20) BY HIGH-PRESSURE LIQULD

- Printed

in The Netherlands

(d-p.

D-GLUCO-OLIGOSACCHARIDES CHROMATOGRAPHY

CHARLES A. WHITE,PATRICI~H.CORRAN, National Imtihrte for Biologica (Great Britain) ASD

Srarrdards ami Control.

Holly Hill,

H~~t~~psrc~rr~l. Lo~rdo~rW 11’3 6Ra

JOHN F. KEXXEDY

Departtnetrt Britaitr) [Received

UuiverJit_v of Birutirtghai?l, P.O. Box 363, Birrnirrgharn BIS 2TT (Grcar

of Chetnistry, March

27th,

19SO: accepted

for publication, April 15th, 19SO)

ABSTRACT The (h.p.1.c.)

behaviour

of

oligosaccharidcs

has been investigated

different di- or poly-amines. hydrolysed

starch)

in high-pressure

by usin,m silica columns

chromatography modified

with

havin, = d-p. of up to 20 (from partially

Oligosaccharides

are well-resolved,

liquid

dynamically

wit!lout

derivatisation,

in under 40

using a simple isocratic system. The results agreed well with those obtained mated gel-permeation chromatography.

min by by auto-

INTRODUCTION Polymers of D-glucose, such as those present in corn syrups, are usually separated by low-pressure ion-exchange’ or by permeation chromatography (g.p_c_) using soft gels, such as cross-linked (e.g.,

Bio-Gel

P-series)3.

g_p.c. analyser+’ require

dextran (e-g_, Sephadex

The best separations

which

separate

G-series)’

are achieved

up to d.p.

oligosaccharides

analysis times of between S and 22 11. The use of small-bore columns and small particle-sizes

up to d-p. 11 to be resolved

in 4.5 h6 on Bio-Gel

or polyacrylamide

by using fully automated 15 (GI-G

IS),

but

permits oligosaccharides

P2 and up to d-p. 15 in 8 h’ on

Bio-Gel P4, but analysis times cannot be reduced further because of the compressibiiity of these packings at high operating pressures_ Silica of small particle size has been used for g.p.c. of dextrans of high molecular weight’ or oligosaccharides’ having d-p. up to 6, but intermediate fractionations have not been achieved_ In the absence of suitable rigid permeation-supports, alternative chromatographic methods for such analyses have been sought. Oligosaccharides up to d.p. 7 can be separated on ion-exchange resins’“-12 of small particle size in 30 min by elution with water or aqueous ethanol, but apparently the columns are unstable_ Oligosaccharides with d-p. higher than 7 cannot be separated, due to possible precipitation at high concentrations of ethanol. 0008-6215/80/0000-0000/S

02.25, @ 1980 -

Elsevier

Scientific

Publishing

Company

166

C. A. WHITE,

Separations by using

(up to d-p. satisfactorily analysed

by adsorption

ILBondpak

on chemically

carbohydrate13

7), and Partisil-1OPAC” in conjunction

bonded

phases

J. F. KENNEDY

have been achieved

(up to d-p. 5), aminopropyl-bonded

(up to d-p. 11). Refractometers

with isocratic

under these conditions_

P. H. CORRAN,

elution

Amino-bonded

and higher

oligomers

phases, however,

phaseI

can only be used cannot

be

tend to deteriorate

with prolonged use’ 6_ In an alternative approach, Wells and Lester’ ’ used an octadecyl-bonded phase to fractionate mixtures of peracetylated oligosaccharides. Gradient elution was required to separate oligosaccharides having d-p. > 11, but, with a convex gradient, those up to d-p. 30 could be separated in 75 min. Pure silica” has been used for adsorption chromatography of carbohydrates, but is not generally applicable because of the low solubility of oligosaccharides in the aqueous acetonitrile mixtures required. We have found that silica columns,

dynamically

modified

amines, provide a rapid, inexpensive method of separating but also oligosaccharides up to d.p. 20.

with polyfunctional

not only simple sugars’ 6*1 g

EXPERIMENT_AL

Appm-atus. -The system used consisted of a Hewlett-Packard 10 IOB chromatograph with a Waters R401 refractometer for detection and a Watanabe Servocorder. Stainless-steel columns (100 x 5 mm i-d. and 200 x 8 mm id.) were packed with a

slurry of irregular silica (- 5 itrn diameter, H.S. Chromatography Packings) in 50 % (v/v) glycerol in methanol_ A pre-column (50 x 5 mm i.d.) containing LiChroprep Si60 (I 5-25 ltrn diameter, Merck) was fitted between the pump and septum injector_ Modzjkrs. I,%Diaminopropane, l,Pdiaminobutane, 1,3_diaminobenzene, i ,Idiaminopentane, 1,6_diaminohexane, 1,S-diaminooctane, tetraethylenepentamine, and 3-aminopropiononitrile (as fumarate) were commercial materials_ Polyamine modifier was obtained from H.S. Chromatography Packings, triethylenetetramine and di-(4aminodiphenyl)methane were gifts from Ciba-Geigy, and pentaethylenehexamine was a gift from Mr. B. B. Wheals.

General methods. volumes of acetonitrile

- The eluant was prepared and water. Percentages recorded

daily by mixing appropriate are for the water content. The

solution was filtered, 0.01 oA (v/v; w/v for solid samples which would not melt readily) of modifier was added, and the mixture was degassed by sonication. Columns were conditioned immediately before use with 300 ml of aqueous acetonitrile containing 0.1 % of the appropriate modifier. Modifiers were removed from the silica by washing with O.SO% aqueous phosphoric acid (-30 ml) and water (until acid free, - 100 ml). Removal of an amine was checked by confirming that a sample was not resolved when the aqueous acetonitrile contained no modifier. The silica was then equilibrated, as described above, with a different modifier. Starches were hydrolysed by chemical and/or enzymic methods, to provide various mixtures of oligosaccharides. on Bio-Gel P2. One malto-saccharide

All samples were analysed by automated g.p.c.4*5 sample” was a gift from Dr. M. Ohnishi.

MOI~IlXX

T3CC ICXl

for

tlClililS

Of

1.4 1.5 I .3 I ,I 0.7 I .2 I ,2 I2 I.2 0.4

chromntogrnpliy.

-

I .O I.0 0.9 0.8 0.6 0.8 0.9 0.9 0.9 0,3

%I

2.0 2.0 I .8 I ,5 0.9 I .7 I .7 I .7 I .7 0.4

~_.__._________ 3.5 3.8 3.2 2.G I .3 3,o 3.1 3.1 3.0 0.5

G7

487 5.0 5.0 4.7 0.6

IWII~OSC, G3

-----.

3.8 4.0 4.0 3,8 0,G

5,l 4,0 I,9

6,3

5.0 4.1 3.2 I.6

5,7

4.6

= u-glucose. G 2 ==

2,G 2,X 2.4 2,0 1.1 2,2 2.3 2.4 2.3 0.5

G6

_.-___-_I_-_--_____

No resolution No resolution

I.4 1.4 I.5 I.6

I .5 I.7 1.8 I.0

I.6

I.8

I.G 1.5

1.6

G?-+C3

I.7

GI-+G?

= n-~lucotrisaccllnri~i~,

5.9 6.2 6.2 5.8 -

6,4 4.8 2.2

7.9

7.2

(is

r/c.

I.6 I.9 I.9 2.0

--

I.9

I.9

I.5 I.7 I,7 I.8

1-G

I.7

I.6 I.6

G/I-G5 -

G3-+G4

G6+C7

I.4

I.4 I.5 I.5 I.7

I.2 I.5 I.5 I.5

1.3

I.6 I.3

1.3

-I.5

G5-+G6

Rc,soh!iorrbetII~~II oli~~o,srrrclm~i~l~~,s

ON Illi’lI:NTION AND RISOLUTION 01: OLlCiOSACCI-IAItIIIRi ON DYNAblIC~\LLY h101~11~11:I~ SILICA”

Tricthylcnclctraminc Tctracthylcncpcntamiw Pcntncthylcnchcxaminc Polyamine modifier 3-Aminopropiononitrilc I ,3-Diaminobcnzcnc Dian~inodiphcnylmcthanc

1,5-Diaminpentane l,6-Diaminohcxanc l,8-Diamino-oclanc

1,3-Diaminopropanc 1,4-Diaminobulnnc

lilFBC’r 01:

I_.

I.4 I.G I.5 1.5

I.1

I.7

I.3

-

G7-+GS

$ Fj G1

;

x

5

a P

F ‘e

16s

C. A. WHITE, P. H. CORRAN, J. F. KENNEDY

TABLE

II

EFFECT

OF

FLOW

RATE

OK

THE

RESOLUTION

OF

OLIGOS.ACCHARlDES

Flow rate (ntl!tttin) 02

BY

DYSAXIICALLY

MODIFIED

SILICA’

0.1

1.0

1.5

1.6 1.9 2.0 1.8 1.7

1.3 1.5 1.6 1.5 1.3 1.3

Resohiott

Gl-+G2 G1-+G3 G3+G4 G4-tG5 G5+G6 G6-+G7 G7-+G8

1.6 2.3 2.4 1.1 1.9 1.6 1.5

1.7 1.1 2.1

1.8

1.5 1.5

Efficiency (N)

2200

1700

1400

1.8

1.s 1.7

“See test for details of chromatography. TABLE

1.2 1000

“For key, see Table 1.

III

EFFECT

OF

MODIFIED

XVATER

COXTEbT

OF

ELUAXI-

OS

THE

RETEBTION

OF

0LICOSACCHARIDF.S

BY

DYNAMICALLY

SILICA”

IVahtercottfettt (“,L v/v) of eluattr 30 45

SO

5i

0.52

G3 G4 G5 G6 G7 GS

0.93 1.33 1.90 2.57 3.47 4.40 5.60 7.07

1.10 1.41 1.82 2.31 2.85 3.41 4.13

0.60 0.76 0.95 1.15 1.3s 1.60 1.85 2.10

0.45 0.53 0.61 0.71 0.79 0.89 0.98

Flow rate

2.0

1.5

1.5

1.0

Efficiency (N)

2800

3600

5700

6500

Oligosacclrarideb

Gl G2

“See text for details of chromatography. Effect

of different

“For key, see Table I.

tttorlifiet-s. -

Using

the

lOO-mm

column

and

45%

aqueous

acetonitriie, the effect of various modifiers on the elution pattern of one sample of hydrolysed starch was studied. Phase-capacity ratios (k’) are given in Table I. Where possible, the resolution between successive peaks (R,) was calculated by-using peak widths at half height for successive peaks and the relationship R, = J5.54/2

x (t” -

t’)/(wv; + w;),

H.P.L.C.

ANALYSIS

OF D-GLUCO-OLIGOSACCHARIDES

169

where t’ and t” are the elution times and w’+ and IV; the widths at half height, respectively, for successive peaks. Effect offrow rate. - The effect of different flow-rates on the resolution was studied by using the lOO-mm column and 45 “/oaqueous acetonitrile containing 0.01 x, of theoretical of polyamine modifier (Table II). The column efficiency (number

plates, N) was measured with respect to D-ghrcose. Flow rates of 2.0 ml.min-L and above gave inadequate resolution. Effect of ehrant composition. - The effect of varying the water content of the aqueous acetonitrile containing 0.01 0A of 1,4-diaminobutane was studied by using the 200-mm column. For lower concentrations of water, the flow rate was increased to avoid excessive times of analysis. Water contents below 40’% were not employed

because of insolubility of the oligosaccharides, and concentrations above 55’%, gave no resolution (Table IfI). The concentration of 1,4-diaminobutane (between 0.005 and 0.02 “/,) in 40 “/, aqueous acetonitrile had no significant effect on the retention. Ana!ytical sepat-ations. - Partial hydrolysates of starch were analysed by using the 200-mm column and SO’% aqueous acetonitrile containing 0.01 “/0of either 1,4-

Fig_ I_ lkctionation of o-gluco-oligosaccharides presentin asample ofhydrolysed starch. Chromatographic conditions: column (200 mm x S mm i-d.) eluted at 3.0 ml/min with 50% aqueous acetonitriic containing O.Ol”/b of (a) polyamine modifier or (h) 1,4-diaminobutane; the upper trace in (a> shows the signal-to-noise ratio at twice the normal sensitivity_

170

C.

A.

30-

WHITE,

P. H. CORRAN,

J. F. KENNEDY

(al

2 ;

20.

s 8 z! 10.

0,

L

I I 20

40 Elution

Inject

2b

time

40

Inject

Elution

time

60

80

60

80

(min)

(mid

of D-gluco-oligosaccharides present in the maltosaccharidezOhaving average conditions: column (200 mm ?: 8 mm i-d.) eluted at 1.O ml/min with SOo/:, aqueous acetonitrile containing O.Oi~/, of (a) polyamine modifier or (6) l+diaminobutane. Fig. 2. Fractionation

d.p_ 17. Chromatographic

diaminobutane

or polyamine

modifier_ Injections

were usually of 25 ~tl of a solution

(up to 100 mg/ml) of the hydrolysate in water. Percentage compositions were calculated by expressing the area under each peak in terms of the total area, using an interpolated baseline (Figs. 1 and 2). DISCUSSION

in agreement

with previous work16-’ ‘, the use of various di- and poly-amines

to modify silica dynamically provided a useful system for separation of sugars. The retention data given in Tables I and II show that, at higher concentrations of water than used by Wheals and Whitelg, there was little difference between the various polyamines used in this study. Fig. 1 shows the separation obtained by using polyamine modifier and 1,4_diaminobutane. The improved separation obtained with the diamine can be seen, notably in the range Gl-G8 where troughs between the peaks are

deeper_ This effect is more marked

with higher oligomers.

Using

1,4-diamino-

H.P.L.C.

ANALYSIS

Fig. 3. Relationship conditions as in Fig.

77

171

OF D-GLUCO-OLIGOSACCHARIDES

between lo.

iogtu

k’

and

d.p.

of

o-gluco-oligosaccharides.

Chromatographic

2.8 -

.-E c

: ‘;

E

2.5-

8-L 0 $

2.2.

1.9

0

5

10

15

20

25

D-p.

Fig. 4. Relationship between loglo [retention graphic conditions as in Fig. la.

(Fig. saccharide”

butane

time] and d.p. of o-gluco-oligosaccharides.

Chromato-

2), good separation up to G20 was achieved with a sample of malto(average d-p. 17) chosen because of its high content of oligomers

above G8. The effect of these modifiers

is complex,

as can be seen from plots of log,,

k’

against d-p. (Fig. 3). A straight-line relationship might be expected, but the higher oligomers were eluted more rapidiy than would be predicted_ This may be because the dynamically modified packing has some permeation properties analogous to amino-bonded phases’, or because interaction between the amine present in the mobile phase and oligosaccharides (analogous to the interaction between poly-

172

C. A. WHITE,

P. H_ CORRAN,

J. F. KENNEDY

saccharides and polyaromatic amines”) reduces the interaction with the packing. The overall effect is to permit analysis of higher oligomers without loss of resolution of the lower ohsomers. The relationship between elution volumes (or retention time) specific and d-p. was logarithmic (Fig. 4) and provides a method of identifying components in an incompIete series. LModifiers can be removed from the system with orthophosphoric acid and the system regenerated without significant loss of performance_ Whilst the acid was passing through the column, the pH of the eluate rose from neutrality to 10 before dropping to t7. This effect seems to be associated with displacement of bound amine from the silica. The pre-column is unable to protect the analytical column from these pH changes. Repeated treatment with acid causes an increase in column backpressure (ultimately column b!ockage), a drop of > 50 % in efficiency, and decreased column life, although phase-capacity ratios are unaffected. For routine use, it is not necessary to remove the modifier, and storage of the column overnight in eluant had no adverse eflects. Columns stored for longer periods were washed with water (50 ml) and methanol (50 ml), and then stored in methanol. Re-equilibration at high concentrations of modifier, as described above, was necessary before re-use. Resolution of the oli,oosaccharides can be increased by lowering the flow rate (Table II)_ Flow rates of l-2 mljmin with the ZOO-mm column provide a satisfactory compromise between time of analysis and resolution_ Lowering the water content of the eluant increased the resolution (Table III). However, the proportion of water was maintained as high as possible in order to avoid the risk of precipitation of higher oligomers. Variation in modifier content did not affect the separation_ Variation in TABLE

I\’

CO>IPARISON

OF

HYVROLYSED

STARCHES

RESULTS

Gl

G2 G3 G4

G5 GG G7 GS G9 GIO Gil G12

Ol%TAISEV

UY

H.P.L.C.

3.0

AUTO~IATEV

2.6 10.5 14.3 9s 9-s 33.0 16.3 4.9 2.9 2.1 2.0 1.8

11.7 15.4 9.6 9.4 22.0 16.1 5.4 2.7 i.6 1.2 0.8 __--_-

uFor key, see Table I.

AXV

G.P.C.

FOR

10.5 10.5 10.0 10.3 9.0 5.5 7.6 6.4 6.7 7.5 7.6 5.4

Cl-G12

IS

SASIPLES

s-9 9.7 9.7 9.5 9.3 s.9 9.1 5.2 s.0 7.2 6.2 5.6

OF

H.P.L.C.

ANALYSIS

OF D-GLUCO-OLIGOSACCHARIDES

173

retention times due to changes in how rate or eluant composition can be overcome by use of an internal standard, for example, erythritol (k’ 0.5; ~-glucose, k’ 0.7; resolution

1.5). Using the system described above, it is possible to detect IO jcg of u-glucose. This result is comparable with the detection limit o f40 nequiv. of acetylated hexose (- 8 jog of D-glucose) using u-v. detection16. However, the system described here has the advantage of faster analysis times, requires no derivatisation (pre- or postcolumn), and uses simpler equipment with isocratic elution. Moreover, the column is cheaper and probably more stable than amino-bonded phases”, and its higher efficiency allows the use of lower proportions of acetonitrile. thereby decreasing the risk of precipitation of higher oligosaccharides. The present method is capable of resolvin g more peaks than automated g.p.c.. but in a more restricted ~vay (polysaccharides of high molecular weight arc not detected). Consequently, direct comparison of results is not possible; over the range Cl-G 12, for which information is available for both methods, comparison of the results shows good agreement (Table IV). We consider that this procedure could be the basis of an automated system similar to that based on g.p.c_, but possessing the advantages discussed. REFERENCES 1 S. A. B;.R~;ER. B. W. HATI-,J. F. KENNEDY. AWL>P. J. Soxwts, Cir~Dol~_rc/~-. fkx.. 9 (1969) 317-334. 2 P. NORDLN, Ar-c/r. Ek~cIrem. Bioplzn.. 99 (1962) 101-104. 3 M. JOHX, G. TR~EL, AND H. DELLWTG, 1. C/wonrorogr.. 42 ( 1969) 476-4S4. 4 J. F. KENNEDY AND J. E. Fox Carbohyrlr. Acx, 54 (1977) 13-11. 5 J. F_ KENNEDY, J. E. Fos AND J. C. SKIRRO\V,.Srmdc, in press. 6 N. K. SABBXH AND 1. S. FAGERSON,J. CXronrafogr., 120 ( 1976) 55-63. 7 N. K. SABBAGH AND 1. S. FMXRSON. J. Chrotnarogr., S6 (1973) 154-159. S H. ESGELHARDT AXD D. MATHES, J. Chromatugr.. IS5 (1979) 30%319_ 9 P. J. KNUDSEN, P. B. ERIKSES, M. FENCER,ASD K. FLoRENIz, J. Clrromrrro~gr-.. I S7 ( 19SO) 373-379. 10 M. R. LADISCH, A. L. HUEBNER, AND G. T. TAO. J. C/wommgr., 147 (1975) 185-193. I I J. HAVLICEI; AND 0. S~~hmxsoN. Anal. Circw., 47 (1975) IS%%-lS57. 12 L. E. FITT, W. HASSLER. AND D. E. Jusr, J. C/mmart~.~r., IS7 (1950) 3SI-3W. 13 J. C. LINDEN AND C. L. LAWHEAD, J. C/zrornafogt-., 105 (1975) 125-133. 14 R. SCHWARZENBACH. J. Chromofogr., 117 (1976) 206-110. 15 F. M. RABEL, A. G. CAPUTO, AND E. T. Burrs, J. Chronmrogr., 126 (1976) 731-740. 16 K. AITZET~I~~LLER, J. C/zronmtogr.. I56 (1978) 345-358. 17 G. B. WELLS AND R. L. LESTER, Anal. Biodtenr., 97 (1979) 154-190. IS J. F. Rocc~ AXD A. ROUCHOUSSE, J. Chronr~fugr.. 117 (1976) 216-221. 19 B. B. WHEALS AND P. C. WHITE. J. C/wunrafugr.. 176 (1979) 411-426. 20 M. OHNISHI, Carbol~ydr. /(es., 61 ( 1979) 33.5-344. 31 J. F. KENNEDY. S. A. BARKER, AND C. A. WHI’TC. Cr.‘rbol~ulr. Rc.s_\., 3Y (1974) 13-23.