ANALYTICAL
BIOCHEMISTRY
190,
271-275
(1990)
Skeletal Keratan Sulfate Chain Molecular Weight Calibration by High-Performance Gel-Permeation Chromatography John
M. Dickenson,
Haydn
G. Morris,
Ian A. Nieduszynski,’
and Thomas
N. Huckerby
Division of Biological Sciences,Institute of Environmental and Biological Sciences,Lancaster University, Bailrigg, Lancaster LA1 4YQ, United Kingdom, and Department of Chemistry, Lancaster University, Bailrigg, Lancaster LA1 4YA, United Kingdom Received
April
4, 1990
A method has been developed for the molecular sizing of skeletal keratan sulfate chains using an HPLC gelpermeation chromatography system. Keratan sulfate chains and keratanase-derived oligosaccharides were prepared from the nucleus pulposus of bovine intervertebral disc (8year-old animals). A Bio-Gel TSK 30 XL column eluted in 0.2 M NaCl and at 30°C was calibrated with keratan sulfate oligosaccharides of known size as well as 3H-end-labeled keratan sulfate chains to yield the relationship log,&
= 4.588
is applicable
in the
acid residues remain unknown (10) and hence there is a need for detailed studies of both carbohydrate composition and molecular sizes. There have been several previous studies of the molecular sizes of glycosaminoglycans using the technique of gel-permeation chromatography, which include studies of chondroitin sulfates (11,12), heparins (13), hyaluronates (14), and keratan sulfates (1.516). In this study a rapid molecular size determination method is presented using HPLC gel-permeation chromatography. MATERIALS
- 2.128K,,
AND
METHODS
Materials which Q 1990
Academic
Press,
range
K,,
= O-15-0.65.
Inc.
Keratan sulfates (KS)2 have been classified according to their linkage to protein as KS-I for the N-linked chains derived from cornea and KS-II for the O-linked chains from skeletal tissues such as cartilage (1). A further type of keratan sulfate chain which is apparently O-linked from mannose to serine or threonine has been isolated from brain tissue (2). Analysis of skeletal keratan sulfate has assumed considerable significance recently with the advent of antiKS monoclonal antibodies (3-5) and clinical assays for KS in body fluids which are being used as monitors of, and markers for the disease, osteoarthritis (6-8). However, there is still uncertainty about the precise structures of skeletal keratan sulfates (for a review, see (9)) because the locations of fucose and a(2-6)-linked sialic
’ To whom correspondence and requests for reprints dressed. ’ Abbreviation used: KS, keratan sulfates. 0003.2697/90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
should
be ad-
Chemicals and enzymes used in this study are as previously described (10,17), except that lithium perchlorate (ACS grade) and piperazine (99%) were from Aldrich Chemical Co. (Gillingham, Dorset, UK), Ecostint A was from Mensura Technology Ltd. (Parbold, Wigan, Lanes, UK), and NaB3H, was from Amersham International plc (Bucks, UK). The Nucleosil5 SB column manufactured by Machery Nagel (Germany) was purchased from H.P.L.C. Technology (Cheshire, UK), the Mono-Q HR lO/lO column from Pharmacia, and the Bio-Gel TSK 30 XL column and Bio-Gel P-4 and P-10 from Bio-Rad Laboratories Ltd. (Watford, Herts, UK). The enzymes chondroitinase ABC (Proteus vulgaris, EC4.2.2.4) and keratanase (Pseudomonas species, EC3.2.1.103) were purchased from ICN Biomedicals (High Wycombe, Bucks, UK) and Sigma, respectively. Analytical
Methods
Assay procedures were performed as previously described (10,17). Radioactivity in samples was determined in a Packard Tri-Carb 300 scintillation counter, by taking lo-p1 aliquots and diluting with 2 ml of Ecostint A. 271
272
DICKENSON
High-Performance
Liquid
Chromatography
System
High-performance liquid chromatography was performed on a Bio-Rad series 700 HRLC titanium gradient and isocratic system (18) using uv and refractive index detectors. Preparation
of Keratan
Sulfate
Chains
Proteoglycan monomers were extracted from the nucleus pulposus of bovine intervertebral discs (g-year-old animals) in 4 M guanidine hydrochloride in the presence of proteolytic inhibitors, dialyzed into associative conditions, and subjected to associative followed by dissociative cesium chloride density gradient centrifugation (as previously described in (19)). The AlDl fraction was dialyzed against 0.1 M Tris acetate, pH 7.3, and then digested with chondroitinase ABC (0.5 unit/300 mg proteoglycan) followed by diphenylcarbamyl chloridetreated trypsin (2 mg/l g proteoglycan). The digest was partially freeze-dried and chromatographed on a column of Sepharose CL-6B (152 X 3.2 cm) eluting with 0.5 M sodium acetate110 mM EDTA, pH 6.8. This produced the peptido-keratan sulfate fragments derived from the keratan sulfate-rich region of the proteoglycan as previously described (17). Keratan sulfate chains were subsequently prepared by alkaline borohydride reduction (20), followed by chondroitinase ABC digestion and chromatography on a column of Sephadex G-50 (82 X 1.5 cm), eluted with 0.2 M NH&O,. The reduced keratan sulfate chains were recovered by lyophilization after extensive dialysis against 1.0 M NaCl and then water, and purified further to remove any O-linked oligosaccharides on a Pharmacia Mono-Q HR 10110 column eluting with a linear gradient of O-O.5 M lithium perchlorate/lO mM piperazine, pH 5.0. The eluate was pooled into four fractions of increasing charge density, Ql-Q4. Preparation
of Keratan
Sulfate
Oligosaccharides
Fraction Q3 was dissolved in 0.2 M sodium acetate, pH 7.4, and digested for 24 h with keratanase (1 unitI2.8 mg KS) at 37°C. The digest was then chromatographed in 0.2 M NH&O, on a Bio-Gel P-10 column (82 X 1 cm), eluted at 1.86 ml/h. The eluate was analyzed for hexose, and the fractions were pooled as indicated (Fig. la). Dextran blue and Cl- were used as V, (void volume) and V, (total volume) markers, respectively. Calibration
of the Bio-Gel P-10 Column
The Bio-Gel P-10 column had previously been calibrated with 3H-labeled reduced cornea1 keratan sulfate oligosaccharides (21) kindly donated by Dr. E. Hounsell. There were insufficient quantities of these 3H-labeled oligosaccharides to allow any further column calibrations; therefore, a range of secondary keratan sulfate oligosaccharide standards were prepared. The oligosac-
ET
AL.
charides generated from bovine intervertebral disc keratan sulfate, fractions B-F in Fig. la, were shown (after comparison of the Kav values with those for the cornea1 standards and of the ‘H NMR spectra of C and E (not shown) with the cornea1 oligosaccharides characterized in (22)) to be the nonasulfated decasaccharide (KSlO), heptasulfated octasaccharide (KS8), pentasulfated hexasaccharide (KS6), trisulfated tetrasaccharide (KS4), and monosulfated disaccharide (KS2), respectively. Calibration of Nucleosil5 SB Oligosaccharides KS8 and KS10 were then used to calibrate the Nucleosil5 SB ion-exchange column (their elution positions are shown in Fig. lb), which has a greater resolving power than Bio-Gel P-10 and is capable of resolving oligosaccharides which are larger than KSlO. The chromatogram on Nucleosil5 SB of a keratanase digest of bovine intervertebral disc keratan sulfate is shown in Fig. lb. The column was eluted with a linear gradient of O-O.5 M lithium perchlorate/lO mM piperazine, pH 5.0. The first major peak in the Nucleosil 5 SB profile (Fig. lb) originates from a contaminant within the piperazine. The remainder of the profile, assuming that the oligosaccharides produced by keratanase form a homologous series, shows separation of keratan sulfate oligosaccharides up to KS18. The trisulfated tetrasaccharide (KS4) elution position (see Fig. lb) was confirmed by reducing this oligosaccharide with alkaline borohydride, followed by a complete structural analysis by high-field ‘H NMR spectroscopy (23). ‘H NMR spectra for KS6 and KS8 (not shown) showed identity with corresponding cornea1 keratan sulfate oligosaccharides characterized in (22), which confirmed both their structures and elution positions. The elution position of the oligosaccharide calibrant D confirmed its assignment as KSlO. The elution positions of KS12, KS14, KS16, and KS18 could then be inferred. NMR Spectroscopy Oligosaccharide samples were initially dissolved in 0.5 ml of 2H20 (99.8%), buffered to pH 7 with phosphate, and then filtered by centrifugation in a microfilterfuge tube (0.45 pm pore size). The internal reference 3-trimethylsilyl[2H,]propionic acid was then added, and the samples were exchanged three times with 99.8% and then once with 100% 2H20 to minimize the residual HOD signal before dissolution in 100% 2H20 (0.5 ml). Proton NMR spectra were determined at 500.14 MHz using a Bruker AM 500 spectrometer with 5-mm variable temperature probes at 60°C. Radiolabeling of Keratan Sulfate Chains Radiolabeled bovine intervertebral disc keratan sulfate chains were prepared by alkaline borohydride reduction (20) of the peptido-keratan sulfates. These (60 mg) were dissolved in 3 ml of 0.05 M NaOH and NaB3H,
KERATAN
a 0.07
g
0.04-
5: $
0.030.02
MOLECULAR
WEIGHT
CALIBRATION
BY
273
CHROMATOGRAPHY
vo
0.060.05
SULFATE
I -
-
0.01 0.00
j 0
10
20
30
40 Fraction
50
60
70
80
90
0
10
20
30
number
40
50
Fraction
60
70
m1---7--80 90
1
number
FIG. 1.
(a) Bio-Gel P-10 gel-chromatography profile of the keratanase digest of bovine intervertebral disc keratan sulfate chains (fraction Q3). The column (82 X 1 cm) was eluted at a flow rate of 1.86 ml/h with 0.2 M NH&O, and fractions (0.62 ml) were pooled as indicated. (b) High-performance anion-exchange chromatography profile of a keratanase digest of bovine intervertebral disc keratan sulfate chains. The Nucleosil 5 SB column eluent was monitored by uv detection at 206 nm. The gradient program was as follows: 10 min of buffer A (10 mM piperazine, pH 5.0) and then 230 min of O-100% buffer B (0.5 M lithium perchlorate, 10 mM piperazine, pH 5.0). The elution positions ofthe two calibrants (B and C) and the keratan sulfate oligosaccharides up to KS18 are shown.
(100 mCi, 5 Ci/mmol dissolved in 200 ~1 of 0.05 M NaOH) was added. After 1 h a further 3 ml of 0.05 M NaOH was added and sufficient NaBH, to produce a 1 M solution. The reaction was terminated by the dropwise addition of glacial acetic acid and the mixture dialyzed against frequent changes of 0.1 M Tris acetate, pH 7.3. The population of reduced glycosaminoglycan chains was then digested with chondroitinase ABC (0.5 unit/ 300 mg) and chromatographed on a column of Sephadex G-50 (82 X 1.5 cm), eluted with 0.2 M NH,HCO,. The radiolabeled keratan sulfate chains were recovered by lyophilization after extensive dialysis against 1.0 M NaCl and then water. The tritiated keratan sulfate chains had a specific activity of 20 &i/mg of KS. Calibration of the TSK 30 XL Column Oligosaccharides KS6, KS8, KS12, KS16, and KS18 obtained from Nucleosil 5 SB were desalted on a BioGel P-2 column (11.2 X 0.85 cm), eluted with water, and lyophilized. They were then chromatographed in 0.2 M
NaCl on a TSK 30 XL gel-permeation column (300 X 7.8 mm), eluting at 0.5 ml/min at a constant 30°C. The eluate was monitored with a Bio-Rad 1755 refractive index detector and 0.25-ml fractions were collected. The V,, and V, were determined by using Dextran Blue and NaCl, respectively. A proportion (1 mg) of the 3H-labeled keratan sulfate chains was also chromatographed under identical conditions. Fractions (0.25 ml) were collected and monitored for radioactivity. RESULTS
The oligosaccharides KS6, KS8, KS12, KS16, and KS18 were chromatographed in 0.2 M NaCl on TSK 30 XL and their elution positions and K., values were recorded. A plot of log,, M, against K,, for the oligosaccha120,
T
4.00
3.75 ; ”
3.50
? 20 3.25
FIG. 2. (KS18
24
26
28
30 Fraction
FIG. 3.
3.00
22
T 0.40
i 0.45
0.50
0.55 K 0”
0.60
0.65
A plot of log,,, M. vs Kav for the oligosaccharide KS16, KS12, KS8, and KS6) on TSK 30 XL.
0.70
standards
32
34
36
38
40
42
44
number
TSK 30 XL high-performance gel chromatography profile of radiolabeled intervertebral disc keratan sulfate chains. The column (300 X 7.8 mm) was eluted at a flow rate of 0.5 ml/min with 0.2 M NaCl and fractions (0.25 ml) were collected. The eluent was monitored by refractive index (--) and the fractions for radioactivity (O0). The keratan sulfate chains were collected in 10 fractions (27-36) as indicated.
274
DICKENSON TABLE
The Refractive
&
27 28 29 30 31 32 33 34 35 36
6,904 20,592 52,741 89,427 105,096 95,416 63,235 27,417 10,436 4,993
RI/dpm
b&o (RI/dpm)
792 4,613 14,511 23,211 41,315 52,878 50,717 27,228 14,628 7,911
8.717 4.46 3.634 3.85 2.543 1.804 1.247 1.007 0.713 0.6311
0.94 0.65 0.56 0.59 0.41 0.26 0.10 0.0032 -0.15 -0.20
the microvolt time
ride standards is shown in Fig. 2. Similarly, 3H-labeled KS chains from bovine intervertebral discs were also chromatographed in 0.2 M NaCl on the TSK 30 XL column and the elution was monitored by refractive index and by scintillation counting (Fig. 3). The refractive index (represented in microvolts) and radioactivity data obtained are shown in Table 1. If the 3H label was incorporated solely in the galactosaminitol ends and the refractive index was a monitor of mass, then the ratio of refractive index response to radiolabel would be a monitor of relative molecular mass. These assumptions appear to be justified because a plot of log,, of these relative molecular masses against KaV (24) yields a straight line, of which the slope almost exactly matches that obtained from the calibration run which uses KS oligosaccharides of known structure and molecular mass (Fig. 4). 5
5 4
5
3 /
0.1
1
3H @pm)
u Fraction size, 0.25 ml. b Refractive index (RI) response was determined by measuring each fraction. c The K,, for each fraction was calculated by noting the elution
0
AL.
Index and Radioactivity Data for the TSK 30 XL High-Performance Gel Chromatography of Radiolabelled Intervertebral Disc Keratan Sulfate Chains b
Fraction” number
ET
0.2
0.3
0.4
0.5
0.6
0.7
(
K 0"
FIG. 4. A plot of log,, M, vs and oligosaccharide standards sulfate chain fractions (which from the TSK 30 XL column; the oligosaccharide standards
Kav for reduced keratan sulfate chains on TSK 30 XL. The reduced keratan from left to right are fractions 27-36 see Fig. 3) are represented as (A) and as (0).
output
at the midpoint
0.129 0.187 0.245 0.303 0.362 0.420 0.478 0.536 0.594 0.652
from
the refractive
24,100 12,400 10,000 10,800 7,100 5,000 3,500 2,800 2,000 1,700
index
detector
Profile
4.382 4.092 4.002 4.032 3.852 3.702 3.542 3.445 3.292 3.242
at the midpoint
of
of each fraction.
The use of both keratanase-derived oligosaccharides and reduced KS chains permits calibration of a wide range of K., values for the column TSK 30 XL, eluted in 0.2 M NaCI, and yields the formula log,,,M,
= 4.588 - 2.128K,,
which is applicable in the range K,, = 0.15-0.65. Protein standards were chromatographed under identical conditions and yielded the K,, values in brackets, bovine serum albumin (0.316), ovalbumin (0.377), carbonic anhydrase (0.478) and trypsin inhibitor (0.574). DISCUSSION Clearly the calibration of this TSK 30 XL column is based upon the calculated molecular weights of the keratanase-resistant oligosaccharides, which will be highly sulfated because the enzyme cleaves unsubstituted (unsulfated) galactose residues. Thus, the accuracy of the calibration will be greater for highly sulfated keratan sulfates. Similarly, low molecular weight keratan sulfates will have their structures (see Fig. 5) dominated by chain ends, the potentially sialic acid-containing linkage (19), and capping regions (unpublished results) and these have a lower homology with the predominantly sulfated poly-N-acetyllactosamine sequences used for the calibration. However, the major advantage of this method is that it provides relative molecular mass determinations in 30 min and requires very little sample (e.g., less than 0.1 mg). Furthermore, different molecular size parameters such as M, and M,, can be measured either by using different methods of detection or by calculation using the above formula. Detection procedures that give a
KERATAN
SULFATE
MOLECULAR
WEIGHT
6-SO,
CALIBRATION
BY
275
CHROMATOGRAPHY
6-SO,
6--SO,
NeuAca2-3Gal~1-4GlcNAc~‘-3Gal~‘-4[GlcNAc~’-3Gal]~~‘-3GlcNAc’-~GalNAc-oi
3 I
NeuAcaZ-3Gal’ M, 1000 \
A
--/
Generalized
structure
of a skeletal
keratan
sulfate
measure of the number of chain ends such as radioactivity and, in the case of skeletal keratan sulfate chains, sialic acid analysis yield mole or number average molecular weights (M,) whereas refractive index and uv detection produce weight average molecular weights (M,). The hexosamine ratio (GlcNAc/GalNAc or GlcNAc/ GalNAc-01) method of determining keratan sulfate relative molecular size is very susceptible to error when chondroitin sulfate contamination is present. Electrophoresis using polyacrylamide gels has been employed as a method of separating oligosaccharides of hyaluronate, dermatan sulfate, chondroitin sulfate (25), and heparin (26). Molecular weight determinations using this method are comparatively time-consuming and require the inclusion of oligosaccharide standards in each experiment. The bovine intervertebral disc keratan sulfate chains used in this study range in molecular weight from 1750 (for fraction 36) to 24,100 (for fraction 27) with a weight average, M,, of 6450, whereas Hopwood and Robinson (15) reported molecular weights of around 20,000. The M, was calculated to be 4550 which yields a MJM, ratio of 1.42 (a monodisperse population would have a M,l M,, ratio of 1). Accurate molecular weight determinations of keratan sulfate chains will be of considerable value in both structural studies that are aimed at proposing a general model for KS-II structure, and in the elucidation of the epitope requirements of the various anti-KS monoclonal antibodies. ACKNOWLEDGMENTS The authors thank neal keratan sulfate Council for support,
Dr. E. F. Hounsell for kindly donating the coroligosaccharides, the Arthritis & Rheumatism and the SERC for a studentship (to J.M.D.).
B. A., Lieberman,
R., and Meyer,
K. (1967)
J. Biol.
C/rem.
242,3373-3380. 2. Krusius,
T., Finne, J., Margolis, R. K., and Margolis, R. U. (1986) J. Biol. Chem. 261,8237-8242. 3. Caterson, B., Christner, J. E., and Baker, J. R. (1983) J. Biol. Chem. 258,8848-8854. 4. Zanetti, M., Ratcliffe, A., and Watt, F. M. (1985) J. Cell. Biol.
101,53-59.
chain,
Linkage region
showing
the approximate
5. Funderburgh, Biol. Chem.
weights
J. L., Caterson, 262, 11,634-11,640.
of the various
regions.
B., and Conrad,
G. W. (1987)
J.
6. Thonar, E. J.-M., Lenz, M. E., Klintworth, G. K., Caterson, B., Pachman, L. M., Glickman, P., Katz, R., Huff, J., and Kuettner, K. E. (1985) Arthritis Rheum. 28,1367-1376. 7. Thonar, E. J.-M., Pachman, L. M., Lenz, M. E., Hayford, Lynch, P., and Kuettner, K. E. (1988) J. Clin. Chem. Clin. them. 26,57-63. 8. Williams, tisRheum.
J. M., Downey, 3 1,557-560.
C., and Thonar,
E. J.-M.
(1988)
J., Bio-
Arthri-
9. Stuhlsatz, H. W., Keller, R., Becker, G., Oeben, M., Lennartz, L., Fischer, D. G., and Greiling, H. (1989) in Keratan Sulphate: Chemistry, Biology, Chemical Pathology (Greiling, H., and Scott, J. E., Eds.), pp. 1-11, The Biochemical Society, London. 10. Thornton, D. J., Morris, H. G., Cockin, G. H., Huckerby, T. N., Nieduszynski, I. A., Hardingham, T. E., and Ratcliffe, A. (1989) Biochem. J. 260,277-282. 11. Wasteson,
A. (1969)
12. Yamaguchi,
Biochim.
H. (1980)
Biophys.
J. Biochem.
13. Johnson, 127.
E. A., and Mulloy,
14. Laurent, 481-492.
U. B. G., and Granath,
15. Hopwood, 631-637.
J. J., and
Acta
Robinson,
177, 152-154.
87,969-977.
B. (1976)
16. Oeben, M., Keller, R., Stuhlsatz, Biochem. J. 248,85-93.
Carbohydr.
Res.
K. A. (1983)
Exp.
H. C. (1973)
Biochem.
51. 119-
Eye Res. 36,
H. W., and Greiling,
J.
135,
H. (1987)
17. Thornton, D. J., Morris, H. G., Cockin, G. H., Huckerby, T. N., and Nieduszynski, I. A. (1989) Glycoconjugate J. 6, 209-218. 18. Nieduszynski,
I. A., and Huckerby,
T. N., (1990)
Lab Pratt.
39,
73. 19. Dickenson, J. M., Huckerby, Biochem. J. 269,55-59.
20. Carlson, 21. Scudder, met,
22. Hounsell, 23. Huckerby, Magn.
D. M. (1968)
P., Tang, H., and Feizi,
(1986)
REFERENCES 1. Bray,
t
L----Main repeat sequence
Capping or chain terminating region FIG. 5.
M, 1000
M, (W”
T. N., and Nieduszynski,
J. Biol.
Chem.
I. A. (1990)
243,616-626.
P. W., Hounsell, E. F., Lawson, A. M., MehT. (1986) EUF. J. Biochem. 157, 365-373.
E. F., Feeney, EUF. J. Biochem.
J., Scudder,
P., Tang,
P. W., and Feizi,
T.
157, 375-384.
T. N., Dickenson, J. M., and Nieduszynski, Reson. Chem., in press.
I. A. (1990)
24. Laurent, T. C., and Killander, J. (1964) J. Chromatogr. 14, 317330. 25. Min, H., and Cowman, M. K., (1986) Anal. Biochem. 155, 275285. 26. Rice, K. G., Rottink, M. K., and Linhardt, R. J. (1987) Biochem. J. 244,515-522.