- Email: [email protected]
A solid-phase immunoassay for the binding of cartilage proteoglycan to hyaluronic acid
ANALYTICAL
BIOCHEMISTRY
160,462-467
A Solid-Phase
(1987)
Immunoassay for the Binding of Cartilage Proteoglycan to Hyaluronic Acid HAROLDD. KEISER
Department of Medicine, Albert Einstein CoNege of Medicine, Bronx, New York 10461 Received August 1, 1986 A solid-phase assayfor detecting the binding of cartilage proteoglycan (PG) to hyaluronic acid (HA) is described. In the assay,HA is immobilized on protamine-treated microtiter wells, the wells are incubated with PG monomer and antibody to PG monomer, and then an ELISA system is used to detect binding of the PG to HA. The specificity of the assayis indicated by the failure to detect PG binding to chondroitin sulfate or albumin-coated microtiter wells, the absence of binding with tryptic fragments of PG monomer other than the HA-binding segment, the loss of binding after reduction and alkylation of PG monomer, and the inhibition ofbinding by preincubation of PG monomer with small amounts of HA. In contrast to the HA-PG interaction in solution, hyaluronidase digestion of HA does not affect its ability to inhibit the reaction of PG monomer with immobilized HA. The microtiter well-based assayappears to be a rapid, simple, and potentially versatile method for studying interactions with HA. o 1987 Academic Press, Inc. KEY WORDS:
hyaluronic acid binding; cartilage; proteoglycan; solid-phase assay; immunoassay; hyaluronic acid.
Glycoproteins and proteoglycans from a variety of mammalian tissues and cells have been found to bind specifically to hyaluronic acid (HA)’ (l-4). Most of our knowledge of this interaction is derived from studies of the proteoglycan (PG) which is present in the extracellular matrix of hyaline cartilage as high-molecular-weight aggregates of PG monomers with hyaluronic acid and link proteins (5). The noncovalent interaction between HA and PG monomer most often has been demonstrated by showing that the addition of small amounts of HA to a solution of PG monomer produces a large increase in the apparent hydrodynamic size of the PG upon Sepharose gel chromatography under associative conditions (1). We describe here a more rapid, simple, and versatile method for demonstrating this interaction, using HA’ Abbreviations used: HA, hyaluronic acid; PG, proteoglycan; BSA, bovine serum albumin; PBS, phosphate-buffered saline. 0003-2697187 $3.00 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.
462
coated microtiter wells, antibodies to PG monomer, and an ELISA detection system. METHODS
Proteoglycan fractions and antibodies. Bovine nasal cartilage was extracted with 4 M guanidine HCl, and PG aggregate and PG monomer fractions were prepared from the extract by associative and dissociative density gradient centrifugation using the method of Hascall and Sajdera (6). Tryptic fragment subfractions comprising the keratan sulfaterich, chondroitin sulfate-bearing, and HAbinding segments of the PG monomer were prepared from PG aggregates by dissociative density gradient centrifugation, digestion with chondroitinase ABC, Sepharose 6B column chromatography, and dissociative CLSepharose 6B column chromatography, as described in detail previously (7,8). PG monomer was reduced and alkylated by incubation at a concentration of 2 mg/ml in
SOLID-PHASE
ASSAY FOR HYALURONIC
4 M guanidine-HCI, 0.05 M Tris-HCI buffer, pH 7.4, and 5 mM dithiothreitol for 4 h at 37°C. Iodoacetamide was added to a concentration of 15 tnM and the solution was incubated for 20 h at room temperature, then dialyzed, and lyophilized (9). HA from human umbilical cord, grade 1, was purchased from Sigma Chemical Co. (St. Louis, MO). The HA was digested by incubation under toluene for 24 h at 37’C at a concentration of 1 mg/ml either in 0.01 M sodium acetate buffer, pH 5.0, with and without testicular hyaluronidase (Leo, Helsinborg, Sweden), 25 pg/ml, or in 0.01 M Tris-acetate buffer, pH 7.3, with and without Streptomyces hyaluronidase (Sigma), 100 units/ml. Chondroitin sulfate was prepared by incubating PG monomer fractions at 5 mg/ml with 0.4 M NaOH at room temperature overnight, followed by neutralization with 0.1 M HCl, dialysis against 0.01 M sodium phosphate buffer, pH 7.0, elution from DEAE-cellulose with KC1 concentration 0.5 to 1 M, dialysis against distilled water, and lyophilization. Antiserum BN4 1A was prepared by immunization of a rabbit with a chondroitinase ABC digest of PG monomer; although raised in a different rabbit, its properties are similar to those of previously described antiserum BN4 1B ( 10). The preparation and properties of monoclonal antibodies F1.2 and LC8.13 have been described previously (11). The reactivity of these antibodies with PG subfractions is summarized in Table 1. HA-Binding assay. Wells of Falcon No. 39 12 microtiter plates (Becton Dickinson and Co., Oxnard, CA) were incubated with 100 ~1 of protamine sulfate (Sigma, grade X), 0.0 1% in distilled water, for 90 min at room temperature and then washed three times with distilled water. One hundred microliters of HA, 100 pg/ml in distilled water, chondroitin sulfate, 100 rcg/ml, or 2% bovine serum albumin (BSA), was added to the wells and the plates were stored overnight, or until later use, at 4°C. Before use, wells were
463
ACID BINDING TABLE 1
REACTIVITYOFANTIBODIESTOCARTILAGE PROTE~GLYCANWITHPROTEOGLYCAN SUBFRACTIONS~
Antibodies PG subfraction
BN41A
LC8.13
F1.2
PG Monomer KS-Rich fragments CS-Bearing fragments HA-Binding fragments Link protein PGS/R + A
+++ ++ + +++ +
++ +++ + ++
++ +++ ++
a Abbreviations: PG, proteoglycan; KS, keratan sulfate; CS, chondroitin sulfate; HA, hyaluronic acid; PGS/R + A, reduced and alkylated proteoglycan monomer.
washed three times with 0.05% Tween 20 in 0.01 M sodium phosphate-buffered normal saline (Tween/PBS) and then blocked by incubation with 100 ~1 of 2% BSA in Tween/ PBS for 1 h at room temperature. The wells were washed three times with Tween/PBS, 100 ~1 of PG fraction in PBS was added, and the plate was again incubated for 1 h at room temperature. After three washes with Tween/PBS, 100 ~1 of a l/500 dilution of antiserum BN41A or a l/50 dilution of cell culture supematants containing monoclonal antibodies LC8.13 or F1.2 in BSA/Tween/ PBS was added and the plate was incubated for 1 h at room temperature. The wells were then washed three times with Tween/PBS, 100 ~1 of a l/500 dilution of peroxidase-conjugated goat anti-rabbit IgG or anti-mouse IgM (Cooper Biomedical Inc., Malvem, PA) was added, and the plate was again incubated at room temperature for 1 h. Following three final washes with Tween/PBS, 100 ~1 of a solution containing 0.5 ml of o-phenylenediamine, 10 mg/ml of methanol, and 0.05 ml of 3% hydrogen peroxide in 50 ml of distilled water was added and the plate was incubated at room temperature until the color fully developed. The reaction was stopped with a drop of 7 N sulfuric acid and the absorbance
464
HAROLD
D. KEISER
TABLE 2
the wells were coated with HA at a concentration of 100 pg/ml, binding could be deBINDING OF CARTILAGE PROTEOGLYCAN (PG) MONOMER AND TRYFVC FRAGMENT SUBFRACTIONS tected at levels of PG monomer as low as 3 OF PC MONOMER TO MICROTITER WELLS COATED pg/ml (Fig. 1). WITHHYALURONICACID(HA),BOVINESERUMALBUSubfractions containing partially purified MIN (BSA), ORCHONDROITIN SULFATE(CS)O tryptic fragments of PG monomer were A492 tested for their ability to bind to HA-coated wells. The fragments in the subfraction dePC Fraction HA BSA cs rived from the keratan sulfate-rich segment and the fragments in the Blank 0.271 0.270 0.247 of PG monomer PG monomer 1.079 0.263 0.275 subfraction derived from chondroitin sulKS-Rich segment 0.258 0.289 0.230 fate-bearing portions of PG monomer failed CS-Bearing segment 0.281 N.T. N.T. to bind to the HA-coated wells, whereas HA-Binding segment 1.737 0.252 0.198 binding was found with the subfraction “Binding was detected by ELISA using antiserum which contains fragments from the HABN41A as detailed under Materials and Methods. The binding segment of the PG monomer (TaPG antigens were used at a final concentration of 30 ble 2). pg/ml. KS: Keratan sulfate; N.T.: Not tested. Reduction and alkylation of PG monomer, which disrupts the sulfhydral bonds at 492 nm was read in an ELISA plate reader. that maintain the globular configuration of portion ( 12), completely Assays were carried out in triplicate and re- its HA-binding abolished reactivity with HA-coated wells sults, generally within five percent of each (Fig. 1). The reduced and alkylated PG other, were averaged. monomer retained approximately 30% of its For inhibition assays, 35 ~1 of PBS or the original reactivity with the rabbit antiserum appropriate dilution of inhibitor in PBS was to PG used in the HA-binding assay, as demixed with 3 15 microliters of a 20 /Ig/ml termined by competitive inhibition in a solution of PG monomer in PBS and incusolid-phase immunoassay (11) (data not bated at room temperature for 1 h. Ahquots shown), and the antiserum was used in the of 100 ~1 were added to each of three washed and blocked HA-coated wells and the assay HA-binding assay in great excess. was carried out using antiserum BN41A as described above. Percentage inhibition was calculated as ,oo _ A492 (PG + inhibitor) -A492 blank x100. A492 (PC + PBS) -A492 blank
1
IO t
OSC
RESULTS
The binding of PG monomer to HA immobilized on microtiter wells could be detected by sequential reactions with a rabbit antiserum to PG, peroxidase-conjugated antibodies to rabbit IgG, and peroxidase substrate (Table 2). The absorbance at 492 nm, presumably reflecting the amount of PG monomer bound, varied directly with the concentration of PG monomer used: when
2 2 06. 0402-
\
\ \ . ,PGS \
------,PGS/A.A __._ A----30 IO IO
'A.\ --__ 3 I
. 03
0'1
FIG. I. Binding of cartilage PG monomer fraction (PC%) and reduced and alkylated PG monomer fraction (PCS/R + A) to HA-coated microtiter wells. Binding was detected by ELISA using antiserum BN41A as described under Materials and Methods.
SOLID-PHASE
;-’*,I I 1,000
ASSAY FOR HYALURONIC
\ , 100
IO mwnpxmlml
1 I
\ _ 01
Ftc. 2. Inhibition by HA and chondroitin sulfate (CS) of the binding of PG monomer to HA-coated microtiter wells. The inhibition assay was carried out as described under Materials and Methods.
Microtiter plate wells were coated with chondroitin sulfate or BSA and the binding assay was carried out in the same manner as described for HA-coated wells. Binding of PG monomer or its substructures to chondroitin sulfate or BSA-coated wells was not detected (Table 2). The specificity of binding to HA-coated microtiter wells was further tested by competitive inhibition studies. Preincubation with HA at concentrations as low as several tenths of a nanogram per milliliter blocked the binding of PG monomer to HA-coated wells (Fig. 2). Depolymerization of HA by digestion with testicular hyaluronidase or Streptomyces hyaluronidase did not appreciably affect its ability to inhibit the binding of PG monomer to HA-coated wells when tested at several concentrations between 3 and 0.03 pg/ml, even though some of the testicular hyaluronidase digest and most of the Streptomyces hyaluronidase digest eluted from a Sephadex G-50 superfine column at the relative elution rate characteristic of HA octasaccharides and smaller fragments ( 13). Inhibition of PG monomer binding was also found after preincubation with chondroitin sulfate; however a lOOO-fold higher concentration of chondroitin sulfate than of HA was required for equivalent inhibition (Fig. 2). Antibodies specific for PG substructures can also be used to demonstrate the binding of PG monomer to HA-coated wells. With the appropriate peroxidase-conjugated anti-
465
ACID BINDING
mouse Ig as second antibody, monoclonal antibody LC8.13, which is directed to an epitope on keratan sulfate (1 I), and monoclonal antibody F1.2, which appears to be directed to a conformation-dependent core protein epitope in the keratan sulfate-rich segment of PG monomer (1 I), detected the binding of PG monomer to HA in the same manner as the rabbit antiserum to intact PG core (Table 3). Monoclonal antibody LC8.13 also detected the binding of the subfraction containing the HA-binding tryptic fragment of PG monomer to HA-coated wells, confirming that some keratan sulfate is present on this fragment (14); however, the epitope with which monoclonal F1.2 reacts appeared to be absent from the HA-binding fragment (Table 3). DISCUSSION
Hardingham and Muir first demonstrated that PG monomer forms aggregates by binding to HA on the basis of the large increase in the hydrodynamic size of disaggregated PG produced by the addition of small amounts of HA (1). A variety of other techniques, inTABLE 3
BINDINGOFCARTILAGEPG MONOMERANDTRYPTICFRAGMENT SUBFRACTIONS OFPG MONOMER TO HYALURONIC ACID-COATED MICROTITER WELU AS DETECTEDBYANTISERUMTOPGMONOMER(BN~IA) ANDBYMONOCLONALANTIBODIESTOKERATANSULFATE(LC~.~~)ANDKERATANSULFATE-RICH PGSEGMENT
(F1.2)y A492 PG Fraction
BN41A
LC8.13
F1.2
Blank PG monomer KS-Rich segment CS-Bearing segment HA-Binding segment
0.27 1 1.039 0.339 0.389 1.549
0.155 1.290 0.164 0.152 0.642
0.149 0.997 0.114 0.124 0.147
a The assay was performed as described under Materials and Methods using PG fractions at a final concentration of 30 fig/ml. The abbreviations are as indicated in Table 1.
466
HAROLD
eluding viscosometry (l), analytic ultracentrifugation (15), affinity chromatography ( 16), equilibrium dialysis ( 17), ultrafiltration (18), and laser light scattering ( 19) have also been used to detect this interaction. Nevertheless, CL-Sepharose 2B gel chromatography in the presence and absence of HA is the method used in most recent studies of the binding of components of cartilage and other mammalian tissues to HA in pathologic and physiologic states (3,20-22). Gel chromatography is conceptually straight-forward and easy to do compared with the alternative techniques previously used for demonstrating the PG-HA interaction, but it can be cumbersome and timeconsuming in that it entails two column runs followed by multiple carbazole or radioactivity determinations on column eluates and is only roughly quantitative. In comparison, the solid-phase assay described here is more rapid and even simpler to perform, requires relatively little material, can easily be made quantitative relative to an appropriate standard, and can be used to test multiple samples simultaneously. Although the PG monomer and antiserum to PG monomer used in developing the assay were produced in this laboratory, similar reagents, not tested by us, are commercially available (Miles Scientific, Naperville, IL). Under the conditions described here, the solid-phase assay for HA binding appears to be sensitive and fairly specific. The binding of PG monomer to HA could be detected with concentrations of PG monomer as low as several micrograms per milliliter. Even greater sensitivity might be achieved by altering the amount of HA coating or by using different antisera or different amounts of antiserum. The specificity of the solid-phase assay was tested in several ways. Partially purifted subfractions of tryptic fragments from structurally and functionally distinct segments of the PG monomer were tested for their reactivity with HA. Although the rabbit antiserum used reacts with all the PG fragment subfrac-
D. KEISER
tions (lo), only the subfraction containing fragments from the HA-binding segment was found to react with the HA-coated wells. Reduction and alkylation of PG monomer, to disrupt sulfhydral bonds which maintain the structural conformation required for HA binding (6,12), resulted in the total loss of reactivity with HA-coated wells. Finally, binding of PG monomer to microtiter wells coated with chondroitin sulfate, another highly anionic glycosaminoglycan, or with BSA, an anionic protein, was not detected by the ELISA immunoassay. The specificity of the solid-phase assay was also tested by competitive inhibition. Preincubation with several tenths of a microgram per milliliter of HA blocked the reactivity of PG monomer with the HA-coated wells. Similar levels of inhibition were obtained by preincubation with chondroitin sulfate, but a lOOO-fold higher concentration was required. In addition, hyaluronidase digestion of HA did not alter its ability to inhibit the interaction of PG monomer with HA-coated wells, even though some of the HA fragments produced by testicular hyaluronidase and most of those produced by Streptomyces hyaluronidase were smaller than the HA decasaccharide required for strong inhibition of the interaction of PG monomer with HA in solution (23). The fine specificity of the HA-PG interaction may thus be somewhat different when the HA is immobilized. With the feasibility of studying interactions with HA-coated microtiter wells established, components of the system can be altered as required for additional types of studies. One variation yielding information with respect to PG structure is described here. Two different mouse monoclonal antibodies reactive with the keratan sulfate-rich segment of PG monomer were used in the assay in place of the polyclonal rabbit antiserum to PG. The monoclonal antibody reactive with keratan sulfate itself was able to detect the interaction of the tryptic fragment containing the HA-binding segment of the PG core protein with HA-coated wells, consistent
SOLID-PHASE
ASSAY FOR HYALURONIC
with the previous observation that this fragment contains some keratan sulfate ( 14). No reactivity was found with the monoclonal antibody reactive with a keratan sulfate-associated core protein epitope, indicating that the HA-binding fragment does not include this amino acid sequence or that it is not present in the appropriate conformation. Many other variations on the solid-phase assay for HA binding appear feasible. Proteoglycans or glycoproteins from other types of cartilage or from noncartilage sources can be tested directly for their ability to bind to HA by the ELBA technique described here, if antisera to them are available. In the absence of specific antisera, reactivity with HA may be determined on the basis of the ability of these proteoglycans or glycoproteins to compete with PG monomer and reduce the amount of its binding to HA-coated wells as detected by antiserum to PG. If the proteoglycans or glycoproteins are radioactively labeled, HA binding can be demonstrated directly, without the ELISA detection system, by allowing them to react with HA-coated wells and determining the number of counts per minute bound. These and other potential uses of the solid-phase assay for HA binding can readily be explored because of the ease and convenience of the microtiter well-based methodology. ACKNOWLEDGMENTS I am grateful to Mrs. Fannie Keyserman for her excellent technical assistance. This work was supported by Grant AM 257 11 from the National Institutes of Health.
REFERENCES 1. Hardingham, T. E., and Muir, H. (1972) Biochim. Biophys. Actn 279,401-405.
ACID BINDING
467
2. Delpech, B., and Halavent, C. (198 1) J. Neurochem. 36,855-859. 3. Wagner, W. D., Rowe, H. A., and Connor, J. R. (1983) J. Biol. Chem. 258, 11136-l 1142. 4. Turley, E. A., and Torrance, J. ( 1984) Exp. Cell Rex 161, 17-28.
5. Hascall, V. C. (1977) J. Supramol. Struct. 7, 101-120. 6. Hascall, V. C., and Sajdera, S. W. (1969) J. Biol. Chem. 2442384-2396. 7. Keiser, H. D., Adlersberg, J. B., and Steinman, H. M. (1982) Biochem. J. 203,683-689. 8. Keiser, H. D., and Keyserman, F. Z. (1985) Connect. Tissue Rex 13, 157- 167. 9. Heinegard, D. (1977) J. Biol. Chem. 252, 1980- 1989. 10. Keiser, H. D. (1982) Biochem. J. 203,691-698. 11. Keiser, H. D., and Diamond, B. A. (1986) Connect. Tissue Rex, in press. 12. Hardingham, T. E., Ewins, R. J. F., and Muir, H. (1976) Biochem. J. 157, 127-143. 13. Christner, J. E., Brown, M. L., and Dziewiatkowski, D. D. (1979) J. Biol. Chem. 254,4624-4630. 14. Bonnet, F., Dunham, D. G., and Hardingham, T. E. (1985) Biochem. J. 228,77-85. 15. Gregory, J. (1973) Biochem. J. 133,383-386. 16. Christner, J. E., Brown, M. L., and Dziewiatkowski, D. D. (1978) Anal. Biochem. 90,22-32. 17. Nieduszynski, I. A., Sheehan, J. K., Phelps, C. F., Hardingham, T. E., and Muir, H. (1980) Biothem. J. 185, 107-l 14. 18. Cleland, R. L. (198 1) Arch. Biochem. Biophys. 210, 565-572. 19. Reihanian, H., Jamieson, A. M., Tang, L. H., and Rosenberg, L. (1979) Biopolymers 18, 1727-1747. 20. Plaas, A. H. K., and Sandy, J. D. (1984) Biochem. J. 220,337-340. 21. Roughley, P. J., White, R. J., Poole, A. R., and Mot-t, J. S. (1984) Biochem. J. 221,637-644. 22. Cole, T. C., Ghosh, P., and Taylor, T. K. F. (1986) Biochim. Biophys. Acta 880,209-2 19. 23. Hardingham, T. E., and Muir, H. (1973) Biochem. J. 135.905-908.
BIOCHEMISTRY
160,462-467
A Solid-Phase
(1987)
Immunoassay for the Binding of Cartilage Proteoglycan to Hyaluronic Acid HAROLDD. KEISER
Department of Medicine, Albert Einstein CoNege of Medicine, Bronx, New York 10461 Received August 1, 1986 A solid-phase assayfor detecting the binding of cartilage proteoglycan (PG) to hyaluronic acid (HA) is described. In the assay,HA is immobilized on protamine-treated microtiter wells, the wells are incubated with PG monomer and antibody to PG monomer, and then an ELISA system is used to detect binding of the PG to HA. The specificity of the assayis indicated by the failure to detect PG binding to chondroitin sulfate or albumin-coated microtiter wells, the absence of binding with tryptic fragments of PG monomer other than the HA-binding segment, the loss of binding after reduction and alkylation of PG monomer, and the inhibition ofbinding by preincubation of PG monomer with small amounts of HA. In contrast to the HA-PG interaction in solution, hyaluronidase digestion of HA does not affect its ability to inhibit the reaction of PG monomer with immobilized HA. The microtiter well-based assayappears to be a rapid, simple, and potentially versatile method for studying interactions with HA. o 1987 Academic Press, Inc. KEY WORDS:
hyaluronic acid binding; cartilage; proteoglycan; solid-phase assay; immunoassay; hyaluronic acid.
Glycoproteins and proteoglycans from a variety of mammalian tissues and cells have been found to bind specifically to hyaluronic acid (HA)’ (l-4). Most of our knowledge of this interaction is derived from studies of the proteoglycan (PG) which is present in the extracellular matrix of hyaline cartilage as high-molecular-weight aggregates of PG monomers with hyaluronic acid and link proteins (5). The noncovalent interaction between HA and PG monomer most often has been demonstrated by showing that the addition of small amounts of HA to a solution of PG monomer produces a large increase in the apparent hydrodynamic size of the PG upon Sepharose gel chromatography under associative conditions (1). We describe here a more rapid, simple, and versatile method for demonstrating this interaction, using HA’ Abbreviations used: HA, hyaluronic acid; PG, proteoglycan; BSA, bovine serum albumin; PBS, phosphate-buffered saline. 0003-2697187 $3.00 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.
462
coated microtiter wells, antibodies to PG monomer, and an ELISA detection system. METHODS
Proteoglycan fractions and antibodies. Bovine nasal cartilage was extracted with 4 M guanidine HCl, and PG aggregate and PG monomer fractions were prepared from the extract by associative and dissociative density gradient centrifugation using the method of Hascall and Sajdera (6). Tryptic fragment subfractions comprising the keratan sulfaterich, chondroitin sulfate-bearing, and HAbinding segments of the PG monomer were prepared from PG aggregates by dissociative density gradient centrifugation, digestion with chondroitinase ABC, Sepharose 6B column chromatography, and dissociative CLSepharose 6B column chromatography, as described in detail previously (7,8). PG monomer was reduced and alkylated by incubation at a concentration of 2 mg/ml in
SOLID-PHASE
ASSAY FOR HYALURONIC
4 M guanidine-HCI, 0.05 M Tris-HCI buffer, pH 7.4, and 5 mM dithiothreitol for 4 h at 37°C. Iodoacetamide was added to a concentration of 15 tnM and the solution was incubated for 20 h at room temperature, then dialyzed, and lyophilized (9). HA from human umbilical cord, grade 1, was purchased from Sigma Chemical Co. (St. Louis, MO). The HA was digested by incubation under toluene for 24 h at 37’C at a concentration of 1 mg/ml either in 0.01 M sodium acetate buffer, pH 5.0, with and without testicular hyaluronidase (Leo, Helsinborg, Sweden), 25 pg/ml, or in 0.01 M Tris-acetate buffer, pH 7.3, with and without Streptomyces hyaluronidase (Sigma), 100 units/ml. Chondroitin sulfate was prepared by incubating PG monomer fractions at 5 mg/ml with 0.4 M NaOH at room temperature overnight, followed by neutralization with 0.1 M HCl, dialysis against 0.01 M sodium phosphate buffer, pH 7.0, elution from DEAE-cellulose with KC1 concentration 0.5 to 1 M, dialysis against distilled water, and lyophilization. Antiserum BN4 1A was prepared by immunization of a rabbit with a chondroitinase ABC digest of PG monomer; although raised in a different rabbit, its properties are similar to those of previously described antiserum BN4 1B ( 10). The preparation and properties of monoclonal antibodies F1.2 and LC8.13 have been described previously (11). The reactivity of these antibodies with PG subfractions is summarized in Table 1. HA-Binding assay. Wells of Falcon No. 39 12 microtiter plates (Becton Dickinson and Co., Oxnard, CA) were incubated with 100 ~1 of protamine sulfate (Sigma, grade X), 0.0 1% in distilled water, for 90 min at room temperature and then washed three times with distilled water. One hundred microliters of HA, 100 pg/ml in distilled water, chondroitin sulfate, 100 rcg/ml, or 2% bovine serum albumin (BSA), was added to the wells and the plates were stored overnight, or until later use, at 4°C. Before use, wells were
463
ACID BINDING TABLE 1
REACTIVITYOFANTIBODIESTOCARTILAGE PROTE~GLYCANWITHPROTEOGLYCAN SUBFRACTIONS~
Antibodies PG subfraction
BN41A
LC8.13
F1.2
PG Monomer KS-Rich fragments CS-Bearing fragments HA-Binding fragments Link protein PGS/R + A
+++ ++ + +++ +
++ +++ + ++
++ +++ ++
a Abbreviations: PG, proteoglycan; KS, keratan sulfate; CS, chondroitin sulfate; HA, hyaluronic acid; PGS/R + A, reduced and alkylated proteoglycan monomer.
washed three times with 0.05% Tween 20 in 0.01 M sodium phosphate-buffered normal saline (Tween/PBS) and then blocked by incubation with 100 ~1 of 2% BSA in Tween/ PBS for 1 h at room temperature. The wells were washed three times with Tween/PBS, 100 ~1 of PG fraction in PBS was added, and the plate was again incubated for 1 h at room temperature. After three washes with Tween/PBS, 100 ~1 of a l/500 dilution of antiserum BN41A or a l/50 dilution of cell culture supematants containing monoclonal antibodies LC8.13 or F1.2 in BSA/Tween/ PBS was added and the plate was incubated for 1 h at room temperature. The wells were then washed three times with Tween/PBS, 100 ~1 of a l/500 dilution of peroxidase-conjugated goat anti-rabbit IgG or anti-mouse IgM (Cooper Biomedical Inc., Malvem, PA) was added, and the plate was again incubated at room temperature for 1 h. Following three final washes with Tween/PBS, 100 ~1 of a solution containing 0.5 ml of o-phenylenediamine, 10 mg/ml of methanol, and 0.05 ml of 3% hydrogen peroxide in 50 ml of distilled water was added and the plate was incubated at room temperature until the color fully developed. The reaction was stopped with a drop of 7 N sulfuric acid and the absorbance
464
HAROLD
D. KEISER
TABLE 2
the wells were coated with HA at a concentration of 100 pg/ml, binding could be deBINDING OF CARTILAGE PROTEOGLYCAN (PG) MONOMER AND TRYFVC FRAGMENT SUBFRACTIONS tected at levels of PG monomer as low as 3 OF PC MONOMER TO MICROTITER WELLS COATED pg/ml (Fig. 1). WITHHYALURONICACID(HA),BOVINESERUMALBUSubfractions containing partially purified MIN (BSA), ORCHONDROITIN SULFATE(CS)O tryptic fragments of PG monomer were A492 tested for their ability to bind to HA-coated wells. The fragments in the subfraction dePC Fraction HA BSA cs rived from the keratan sulfate-rich segment and the fragments in the Blank 0.271 0.270 0.247 of PG monomer PG monomer 1.079 0.263 0.275 subfraction derived from chondroitin sulKS-Rich segment 0.258 0.289 0.230 fate-bearing portions of PG monomer failed CS-Bearing segment 0.281 N.T. N.T. to bind to the HA-coated wells, whereas HA-Binding segment 1.737 0.252 0.198 binding was found with the subfraction “Binding was detected by ELISA using antiserum which contains fragments from the HABN41A as detailed under Materials and Methods. The binding segment of the PG monomer (TaPG antigens were used at a final concentration of 30 ble 2). pg/ml. KS: Keratan sulfate; N.T.: Not tested. Reduction and alkylation of PG monomer, which disrupts the sulfhydral bonds at 492 nm was read in an ELISA plate reader. that maintain the globular configuration of portion ( 12), completely Assays were carried out in triplicate and re- its HA-binding abolished reactivity with HA-coated wells sults, generally within five percent of each (Fig. 1). The reduced and alkylated PG other, were averaged. monomer retained approximately 30% of its For inhibition assays, 35 ~1 of PBS or the original reactivity with the rabbit antiserum appropriate dilution of inhibitor in PBS was to PG used in the HA-binding assay, as demixed with 3 15 microliters of a 20 /Ig/ml termined by competitive inhibition in a solution of PG monomer in PBS and incusolid-phase immunoassay (11) (data not bated at room temperature for 1 h. Ahquots shown), and the antiserum was used in the of 100 ~1 were added to each of three washed and blocked HA-coated wells and the assay HA-binding assay in great excess. was carried out using antiserum BN41A as described above. Percentage inhibition was calculated as ,oo _ A492 (PG + inhibitor) -A492 blank x100. A492 (PC + PBS) -A492 blank
1
IO t
OSC
RESULTS
The binding of PG monomer to HA immobilized on microtiter wells could be detected by sequential reactions with a rabbit antiserum to PG, peroxidase-conjugated antibodies to rabbit IgG, and peroxidase substrate (Table 2). The absorbance at 492 nm, presumably reflecting the amount of PG monomer bound, varied directly with the concentration of PG monomer used: when
2 2 06. 0402-
\
\ \ . ,PGS \
------,PGS/A.A __._ A----30 IO IO
'A.\ --__ 3 I
. 03
0'1
FIG. I. Binding of cartilage PG monomer fraction (PC%) and reduced and alkylated PG monomer fraction (PCS/R + A) to HA-coated microtiter wells. Binding was detected by ELISA using antiserum BN41A as described under Materials and Methods.
SOLID-PHASE
;-’*,I I 1,000
ASSAY FOR HYALURONIC
\ , 100
IO mwnpxmlml
1 I
\ _ 01
Ftc. 2. Inhibition by HA and chondroitin sulfate (CS) of the binding of PG monomer to HA-coated microtiter wells. The inhibition assay was carried out as described under Materials and Methods.
Microtiter plate wells were coated with chondroitin sulfate or BSA and the binding assay was carried out in the same manner as described for HA-coated wells. Binding of PG monomer or its substructures to chondroitin sulfate or BSA-coated wells was not detected (Table 2). The specificity of binding to HA-coated microtiter wells was further tested by competitive inhibition studies. Preincubation with HA at concentrations as low as several tenths of a nanogram per milliliter blocked the binding of PG monomer to HA-coated wells (Fig. 2). Depolymerization of HA by digestion with testicular hyaluronidase or Streptomyces hyaluronidase did not appreciably affect its ability to inhibit the binding of PG monomer to HA-coated wells when tested at several concentrations between 3 and 0.03 pg/ml, even though some of the testicular hyaluronidase digest and most of the Streptomyces hyaluronidase digest eluted from a Sephadex G-50 superfine column at the relative elution rate characteristic of HA octasaccharides and smaller fragments ( 13). Inhibition of PG monomer binding was also found after preincubation with chondroitin sulfate; however a lOOO-fold higher concentration of chondroitin sulfate than of HA was required for equivalent inhibition (Fig. 2). Antibodies specific for PG substructures can also be used to demonstrate the binding of PG monomer to HA-coated wells. With the appropriate peroxidase-conjugated anti-
465
ACID BINDING
mouse Ig as second antibody, monoclonal antibody LC8.13, which is directed to an epitope on keratan sulfate (1 I), and monoclonal antibody F1.2, which appears to be directed to a conformation-dependent core protein epitope in the keratan sulfate-rich segment of PG monomer (1 I), detected the binding of PG monomer to HA in the same manner as the rabbit antiserum to intact PG core (Table 3). Monoclonal antibody LC8.13 also detected the binding of the subfraction containing the HA-binding tryptic fragment of PG monomer to HA-coated wells, confirming that some keratan sulfate is present on this fragment (14); however, the epitope with which monoclonal F1.2 reacts appeared to be absent from the HA-binding fragment (Table 3). DISCUSSION
Hardingham and Muir first demonstrated that PG monomer forms aggregates by binding to HA on the basis of the large increase in the hydrodynamic size of disaggregated PG produced by the addition of small amounts of HA (1). A variety of other techniques, inTABLE 3
BINDINGOFCARTILAGEPG MONOMERANDTRYPTICFRAGMENT SUBFRACTIONS OFPG MONOMER TO HYALURONIC ACID-COATED MICROTITER WELU AS DETECTEDBYANTISERUMTOPGMONOMER(BN~IA) ANDBYMONOCLONALANTIBODIESTOKERATANSULFATE(LC~.~~)ANDKERATANSULFATE-RICH PGSEGMENT
(F1.2)y A492 PG Fraction
BN41A
LC8.13
F1.2
Blank PG monomer KS-Rich segment CS-Bearing segment HA-Binding segment
0.27 1 1.039 0.339 0.389 1.549
0.155 1.290 0.164 0.152 0.642
0.149 0.997 0.114 0.124 0.147
a The assay was performed as described under Materials and Methods using PG fractions at a final concentration of 30 fig/ml. The abbreviations are as indicated in Table 1.
466
HAROLD
eluding viscosometry (l), analytic ultracentrifugation (15), affinity chromatography ( 16), equilibrium dialysis ( 17), ultrafiltration (18), and laser light scattering ( 19) have also been used to detect this interaction. Nevertheless, CL-Sepharose 2B gel chromatography in the presence and absence of HA is the method used in most recent studies of the binding of components of cartilage and other mammalian tissues to HA in pathologic and physiologic states (3,20-22). Gel chromatography is conceptually straight-forward and easy to do compared with the alternative techniques previously used for demonstrating the PG-HA interaction, but it can be cumbersome and timeconsuming in that it entails two column runs followed by multiple carbazole or radioactivity determinations on column eluates and is only roughly quantitative. In comparison, the solid-phase assay described here is more rapid and even simpler to perform, requires relatively little material, can easily be made quantitative relative to an appropriate standard, and can be used to test multiple samples simultaneously. Although the PG monomer and antiserum to PG monomer used in developing the assay were produced in this laboratory, similar reagents, not tested by us, are commercially available (Miles Scientific, Naperville, IL). Under the conditions described here, the solid-phase assay for HA binding appears to be sensitive and fairly specific. The binding of PG monomer to HA could be detected with concentrations of PG monomer as low as several micrograms per milliliter. Even greater sensitivity might be achieved by altering the amount of HA coating or by using different antisera or different amounts of antiserum. The specificity of the solid-phase assay was tested in several ways. Partially purifted subfractions of tryptic fragments from structurally and functionally distinct segments of the PG monomer were tested for their reactivity with HA. Although the rabbit antiserum used reacts with all the PG fragment subfrac-
D. KEISER
tions (lo), only the subfraction containing fragments from the HA-binding segment was found to react with the HA-coated wells. Reduction and alkylation of PG monomer, to disrupt sulfhydral bonds which maintain the structural conformation required for HA binding (6,12), resulted in the total loss of reactivity with HA-coated wells. Finally, binding of PG monomer to microtiter wells coated with chondroitin sulfate, another highly anionic glycosaminoglycan, or with BSA, an anionic protein, was not detected by the ELISA immunoassay. The specificity of the solid-phase assay was also tested by competitive inhibition. Preincubation with several tenths of a microgram per milliliter of HA blocked the reactivity of PG monomer with the HA-coated wells. Similar levels of inhibition were obtained by preincubation with chondroitin sulfate, but a lOOO-fold higher concentration was required. In addition, hyaluronidase digestion of HA did not alter its ability to inhibit the interaction of PG monomer with HA-coated wells, even though some of the HA fragments produced by testicular hyaluronidase and most of those produced by Streptomyces hyaluronidase were smaller than the HA decasaccharide required for strong inhibition of the interaction of PG monomer with HA in solution (23). The fine specificity of the HA-PG interaction may thus be somewhat different when the HA is immobilized. With the feasibility of studying interactions with HA-coated microtiter wells established, components of the system can be altered as required for additional types of studies. One variation yielding information with respect to PG structure is described here. Two different mouse monoclonal antibodies reactive with the keratan sulfate-rich segment of PG monomer were used in the assay in place of the polyclonal rabbit antiserum to PG. The monoclonal antibody reactive with keratan sulfate itself was able to detect the interaction of the tryptic fragment containing the HA-binding segment of the PG core protein with HA-coated wells, consistent
SOLID-PHASE
ASSAY FOR HYALURONIC
with the previous observation that this fragment contains some keratan sulfate ( 14). No reactivity was found with the monoclonal antibody reactive with a keratan sulfate-associated core protein epitope, indicating that the HA-binding fragment does not include this amino acid sequence or that it is not present in the appropriate conformation. Many other variations on the solid-phase assay for HA binding appear feasible. Proteoglycans or glycoproteins from other types of cartilage or from noncartilage sources can be tested directly for their ability to bind to HA by the ELBA technique described here, if antisera to them are available. In the absence of specific antisera, reactivity with HA may be determined on the basis of the ability of these proteoglycans or glycoproteins to compete with PG monomer and reduce the amount of its binding to HA-coated wells as detected by antiserum to PG. If the proteoglycans or glycoproteins are radioactively labeled, HA binding can be demonstrated directly, without the ELISA detection system, by allowing them to react with HA-coated wells and determining the number of counts per minute bound. These and other potential uses of the solid-phase assay for HA binding can readily be explored because of the ease and convenience of the microtiter well-based methodology. ACKNOWLEDGMENTS I am grateful to Mrs. Fannie Keyserman for her excellent technical assistance. This work was supported by Grant AM 257 11 from the National Institutes of Health.
REFERENCES 1. Hardingham, T. E., and Muir, H. (1972) Biochim. Biophys. Actn 279,401-405.
ACID BINDING
467
2. Delpech, B., and Halavent, C. (198 1) J. Neurochem. 36,855-859. 3. Wagner, W. D., Rowe, H. A., and Connor, J. R. (1983) J. Biol. Chem. 258, 11136-l 1142. 4. Turley, E. A., and Torrance, J. ( 1984) Exp. Cell Rex 161, 17-28.
5. Hascall, V. C. (1977) J. Supramol. Struct. 7, 101-120. 6. Hascall, V. C., and Sajdera, S. W. (1969) J. Biol. Chem. 2442384-2396. 7. Keiser, H. D., Adlersberg, J. B., and Steinman, H. M. (1982) Biochem. J. 203,683-689. 8. Keiser, H. D., and Keyserman, F. Z. (1985) Connect. Tissue Rex 13, 157- 167. 9. Heinegard, D. (1977) J. Biol. Chem. 252, 1980- 1989. 10. Keiser, H. D. (1982) Biochem. J. 203,691-698. 11. Keiser, H. D., and Diamond, B. A. (1986) Connect. Tissue Rex, in press. 12. Hardingham, T. E., Ewins, R. J. F., and Muir, H. (1976) Biochem. J. 157, 127-143. 13. Christner, J. E., Brown, M. L., and Dziewiatkowski, D. D. (1979) J. Biol. Chem. 254,4624-4630. 14. Bonnet, F., Dunham, D. G., and Hardingham, T. E. (1985) Biochem. J. 228,77-85. 15. Gregory, J. (1973) Biochem. J. 133,383-386. 16. Christner, J. E., Brown, M. L., and Dziewiatkowski, D. D. (1978) Anal. Biochem. 90,22-32. 17. Nieduszynski, I. A., Sheehan, J. K., Phelps, C. F., Hardingham, T. E., and Muir, H. (1980) Biothem. J. 185, 107-l 14. 18. Cleland, R. L. (198 1) Arch. Biochem. Biophys. 210, 565-572. 19. Reihanian, H., Jamieson, A. M., Tang, L. H., and Rosenberg, L. (1979) Biopolymers 18, 1727-1747. 20. Plaas, A. H. K., and Sandy, J. D. (1984) Biochem. J. 220,337-340. 21. Roughley, P. J., White, R. J., Poole, A. R., and Mot-t, J. S. (1984) Biochem. J. 221,637-644. 22. Cole, T. C., Ghosh, P., and Taylor, T. K. F. (1986) Biochim. Biophys. Acta 880,209-2 19. 23. Hardingham, T. E., and Muir, H. (1973) Biochem. J. 135.905-908.