Vol. 189, No. 3. 1992 December 30. 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 1591-1597
RATHEPATYXXTE HYALURONAN/GLYCOSAMINOGLYCAN BINDING PROTEIN3 EVIDENCE FOR DISTINCI’ DIVALENT CATIONINDEPENDENT AND DIVALENT CATION-DEPENDENT AClWHlES Stephen J. Frost, Grete M . Kindberg, Janet A. Oka and Paul H. Weigel Department of Human Biological Chemistry and Genetics The University of Texas Medical Branch, Galveston, Texas 77555-0647 Received
November
9,
1992
We have previously shown (Biochemistry, a 10425, 1990) that hepatocytes contain intracellular specific binding sites for hyaluronan (HA). Although HA-binding activity is not dependent on divalent cations, it is increased in the presence of Cae2. Here we report that a novel photoaffmity HA derivative (ASD-HA) crosslinks specifically to different proteins in permeable cells in the presence or absenceof Caf2. With Cat2 present, two proteins of approximately 24 kD and 43 kD were labeled. Additionally, a broad zone of specific crosslinking was observed in the region of 40-100 kD. However, in the presence of the chelator EGTA this zone was absent and the 24 and 43 kD proteins were also not crosslinked to the HA photoaffinity derivative. In the absenceof Ca+2,only a 54 kD protein was specifically labeled. The results indicate that different intracellular hepatocyte proteins are responsible for the Cat2-independent and the Cat2-dependent binding of HA. 0 1992 Academic Press,Inc.
Hyaluronic acid, or hyaluronan (HA), is a non-sulfated glycosaminoglycan. HA is a ubiquitous component of the vertebrate extracellular matrix and can noncovalently interact with extracellular matrix molecules such as proteoglycans,link protein and hyaluronectin (l3). Furthermore, HA can bind to the surface of SV3T3 fibroblasts (4) hepatocytes (5-7) liver sinusoidal endothelial cells (6,8), alveolar macrophages(9) and lymphoid cells (10-12). HA, other glycosaminoglycansand proteoglycanshave also been reported inside cells in the cytoplasm and nucleus (13-15). This latter unexpected finding raises the possibility that glycosaminoglycans,or related molecules, may also have an intracellular function. Although hepatocyteshave some surface HA-binding sites, a large number of specific HA-binding sites (> lo6 cell) are intracellular (7,16). These binding sites, which bind HA
1591
All
0006-29 I X/92 $4.00 Copyright 0 1992 by Academic Press, Inc. rights of reproduction in my form rrser\&
Vol. 189, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUhlICATIOtIS
with a & -2 x 10s7Mare not nuclear; they are recovered in crude m icrosomes and are either associatedwith membranes or cytoskeleton. This binding activity is specific for HA and several other related glycosaminoglycans,including chondroitin sulfate and heparin (7). Other anionic polysaccharides,RNA and DNA do not bind to these HA/glycosaminoglycanbinding sites. This HA-binding activity in permeable hepatocytes is stimulated 2- to 6-fold by 10 m M concentrations of several divalent cations (16). The greatest stimulation of HA binding was demonstrated with Cat2 and Mn+2. Here, we provide evidence that the specific Ca+2-stimulatedHA-binding activity in hepatocytesis different than the Ca+2-independent HA-binding activity.
METHODS Materials. Digitonin was from Kodak, Collagenase (type I), Percoll, HA from human umbilical cord, calf thymus DNA, and Nonidet P-40 were from Sigma. SASD [sulfosuccinimidyl-2@-azidosalicylamido)ethyl-l,3’-dithiopropionate] and 1,3,4,6, tetrachloro3a,6cr diphenylglycouril (iodogen) were from Pierce. tzsI-NaI (lo-20 mCi/pg iodine) was from Amersham. Bovine serum albumin (BSA) was from Armour Biochemicals. Male Sprague Dawley rats (200 g) were obtained from Harlan Breeding Laboratories, Houston, TX. Preparation of m I-HA and ‘%ASD-HA. HA was further purified and converted to a hexylamine derivative at the terminal reducing sugar as described previously (17), with m inor modifications (7). The hydroxyphenylpropionyl derivative of the HA-hexylamine was prepared and iodinated (17) to give specific activities of 2-4 x lo6 dpm/ug HA. The lBI-HA was stable for at least 20 days at 4°C. SASD was handled in dim light or a red safety light. ASD-HA was prepared as recently described (18) by incubating 1 m l of HA-hexylamine (5 mg/ml) with 3.2 m l of SASD (2 mg/ml) in 0.1 M sodium carbonate, pH 9.0. The solution was incubated at room temperature for 30 m in in a foil-covered tube, and subsequently dialyzed in the dark against 150 m M NaCl, 20 m M sodium phosphate, pH 8.0 for 36 hr with three buffer changes at 4’C. Aliquots (500 ~1) were then stored at -7O’C. ASD-HA was iodinated (17) and stored at 4 oC in foil-covered tubes. Media and buffers. Medium l/BSA is a modified Eagle’s medium (Grand Island Biological Co.) supplemented with 2.4 g/L of 4-(2-hydroxyethyl)-l-piperazine-ethanesulfonic acid (HEPES), 0.22 g/L NaHCO,, and 0.1% (w/v) BSA, pH 7.4. Buffer 1 contains 143 m M NaCl, 6.8 m M KC1 and 10 m M HEPES, pH 7.4. Buffer l/BSA is Buffer 1 with 0.1% (w/v) BSA. BIClO is Buffer 1 with 10 m M CaCI,. Buffer l/EGTA is buffer 1 with 2 m M EGTA. Isolated rat hepatoqtes. Hepatocytes were prepared from male Sprague-Dawleyrats by a collagenaseperfusion procedure (19). Before use, the cells were first incubated at 37°C for 1 hr in Medium l/BSA to allow recovery from the isolation procedure. The cells were then chilled, washed, and loaded onto a discontinuous Percoll gradient to separate nonparenchymal cells and dead hepatocytes from viable hepatocytes (20). Hepatocytes (98% viability, 99% purity) at the bottom of the gradient were washed with Buffer 1, stored on ice and used within 1 hr. 1592
Vol.
189,
No.
3, 1992
BIOCHEMICAL
AND
BlOPHYStCAL
RESEARCH
COMMUNtCATtONS
mI-HA binding to pemeabilized hepatocytes. Cells (2-5 x 106cells/ml) were permeabilized with 0.055% digitonin in Buffer l/BSA for 20 min at 4”C, washed and then incubated with ‘251-HAin the absence(total binding) or presence (nonspecific binding) of 2 lOO-fold excess of nonradiolabeled HA in 12 x 75 mm plastic tubes at 4°C for 1 hr with gentle agitation every 5 min. The cells were washed three times by centrifugation and ‘ZI-radioactivity was determined using a Packard Multiprias 2 gamma spectrometer. Specific binding was calculated by subtracting the nonspecific binding from the total binding. Crosslinking of mI-A!XHIA. Hepatocytes (4 x 106/ml) were permeabilized with 0.055% digitonin in Buffer l/BSA plus 133 mM PMSF at 4”C, washed three times and resuspended at 1 x 10’cells/ml in either Buffer l/EGTA (minus Cat’) or BIClO (plus Ca+2). The cells were incubated for 20 min with (nonspecific binding) or without (total binding) 900 hg/ml of nonradiolabeled HA. ‘2SI-ASD-HA was added (12 pg/ml) and the cells were incubated for 45 min at 4°C in the dark. The cells were washed three times with either Buffer l/EGTA or BIClO, resuspendedand aliquots (lo6 cells in 75 ~1) were transferred to wells of a 96 well-plate. Photolysis was for 2 min with a short wave (254 nm), 4 watt UV lamp (Model UVG-11 from Ultra-violet Products, Inc., San Gabriel, CA.) placed on top of the plate. Laemmli (21) sample buffer (40 ~1 of 4X) containing 5% A-mercaptoethanol was added to each well and the final volume was adjusted to 150 ~1. Samples (75 ~1; 5 x 16 cells) were analyzed by SDS-PAGE on a 5-15% gradient polyacrylamide gel. The gel was dried and autoradiography was done at -70°C using preflashed X-Omat AR film (Kodak). General. Protein was measured using the method of Bradford (22) with BSA as a standard. Cell number was estimated by determining DNA as described by Labarca and Paigen (23) with calf thymus DNA as a standard. Data points represents the mean of triplicates and the error bars represent the sample standard deviation. RESULTS AND DISCWSION Effect of CM& on “I-HA
binding to permeabilized hepatocytes
Specific HA binding to permeabilized hepatocytesat 4’C is increased by Ca+2and other divalent cations such as Mg+2 and Mnc2 (16). A -lo-fold
increase in HA binding was
observed at different concentrations of lZI-HA when binding was assessedwith and without 10 mM Ca+2 (not shown). At a constant HA concentration, a linear increase in specific HA binding was observed with increasing concentrations of CaCl, (Fig. 1). The increase was 80 fmol of HA bound/lo6 cells per 1 mM increase in CaCl, concentration. It is important to note that in the absence of any free divalent cations (i.e. plus EGTA) there is still a significant amount of specific HA binding: in this experiment 200 fmol/106 cells or - 1.2 x ld intracellular sites/cell. The stimulator-y effect of Cat2 on HA binding is reversible. If hepatocytes are allowed to bind lZI-HA with 10 mM CaCl,, and then diluted to 2 mM CaCl, there is no significant increase in ‘=I-HA binding compared to that at 2 mM CaCl, alone. 1593
Vol.
01
189, No. 3, 1992
CaC12 CONCENTRATION
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(mt.4)
02
HA CONCENTRATION
hM)
x 1O-2
Figure 1. Effect of CaClz concentration on ‘%HA binding to permeabilized rat hepatocytes. Permeabilized hepatocytes (5 x lo6 cells/ml) in Buffer 1 plus 0.1% BSA were
incubated with increasing concentrations of CaCl, and 6 pg/ml of ?-HA in the presence or absence of 600 pg/ml of nonradiolabeled HA for 1 hr at 4°C. The cells were washed and specifically bound ‘~5I- HA and DNA content were determined. The line determined by least squares linear regression analysis (cc = 0.996) had a slope of 80 fmoles of lEI-HA specifically bound/lo6 cells/mM Ca+*. specific HA binding to Figure 2. Effect of lzsI-HA concentration on Ca”-stimulated permeabilized rat hepatocytes. Permeabilized cells were incubated in Medium 1 (which
contains 1.8 m M CaCl,) with increasing concentrations of lZI-HA in the presence or absence of 20 PM nonradiolabeled HA and with (0) or without (0) an additional 10 m M CaCl, for 1 hr at 4°C. The cells were washed and specifically bound lEI-HA was determined. Since the HA concentration in Figure 1 is less than the I& of HA for its binding site
in permeabilized hepatocytes (16), the increase in HA binding in the presence of 10 m M CaCl, could be explained by either a decreasein the & of the HA binding interaction (te. greater affinity) or an increase in the relative number of HA-binding sites. To test this, the HA concentration was varied from 12.5 to 600 nM I-IA, which is well above the IQ. A -3fold increase in specific ‘251-HA binding was observed at all HA concentrations in the presence of 10 m M CaCl, as compared with 1.8 m M CaCl, (Fig. 2). Most of the previous studies have been performed with 1.8 m M CaCl, present (7,16). Although saturation occurred in the presenceof 1.8 m M Ca +‘, with 10 m M Cat2 HA binding was still increasing. Furthermore, in preliminary equilibrium binding studies (not shown) we observed no significant change in the & for ‘=I-HA binding in the presence or absenceof 10 m M CaCl, (Le. 3.4 x lo-‘M vs 3.6 x lo-‘M). The results are consistentwith the conclusion that divalent cations such as Cat2 cause an increase in the apparent number of binding sites rather than an increase in the affinity of HA binding. 1594
Vol.
189,
No.
3, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Pho&affinity crosslinking of mI-ASD-HA to the Ca+2-stimulated and the Ca+2-independentHA-binding activities We have recently developed the use of a unique photoaffinity crosslinking derivative of HA to identify specific HA-binding proteins or receptors, such as the sinusoidal liver endothelial cell HA receptor (18). The iodinatable arylazide derivative ASD-HA has a major advantage for these studies, After being photo-activated and crosslinked to a target protein(s), the HA can be released from this complex by reduction of the disulfide bond in the (S)ASD. This treatment, in effect, transfers only the lZ1-aryl group to the target protein, which can then be detected by autoradiography after SDS-PAGE (18). lsI-ASD-HA (-65%) as lZI-HA.
bound to permeable hepatocytes with the same level of specificity A different crosslinking pattern was observed in the absence of Ca+2
(i.e. in the presence of EGTA) compared to when 10 mM CaCl, was included during the binding and photolysis of lZI-ASD-HA to permeabilized hepatocytes. In the presence of EGTA only one band at -54 kD was specifically crosslinked to the lZI-ASD-HA
(Fig. 3).
Binding and crosslinking of the HA derivative to this 54 kD protein were competed > 90% by nonderivatized HA. The labeling pattern with Ca” present was more complex, with diffuse labeling in the region between 40-100 kD (Fig. 4). That this material was dramatically decreasedin the presence of excessnonderivatized HA, suggestsit represents specific binding sites. Surprisingly, the 54 kD band seen without Ca+’ was not obviously labeled in the presence of Cat2 (lanes 1 and 2, Fig. 4). Better resolution of labeled bands was obtained in the region between 14-40 kD. In this region only a 43 kD and a 24 kD protein were consistently labeled specifically; that is, they were competed by nonlabeled HA. These two proteins specifically labeled in the presence of Cat2 were not labeled in the absenceof Ca+2 (lanes 1 and 3, Fig. 4). The different specific crosslinking pattern with ‘=IASD-I-IA provides further evidence that the increased HA binding to hepatocytes in the presence of Ca+’ is to different protein(s) than in the absenceof Cat’. Verification of this
1595
Vol.
189,
No.
3, 1992
03
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
04
Figure 3. Crosslinking of ‘251-ASD-HA to pemeabilized hepatocytes in the presence Digitonin-treated hepatocytes were incubated at 4°C with the photoaffinity of EGTA
derivative ‘“I-ASD-HA with 2 mM EGTA in the presence (lanes 1 and 2) or absence (lanes 3 and 4) of excess nonradioactive HA. UV photolysis, SDS-PAGE analysis and autoradiography were as described in Methods. The duplicate lanes are independent replicates. The migration of standard proteins (M, x 10”) and the top (T) of the resolving gel are indicated on the right. The arrow points to the 54 kD band that was specifically competed with nonradiolabeled HA. Figure 4. Effect of CaC& on the photoaffinity crosslinking of ‘%ASD-HA to permeable hepatocytes. Permeable hepatocytes were incubated with ‘%ASD-HA in the
presence of 2 mM EGTA (lane 1) or 10 mM CaCI, (lanes 2-4) with (lane 4) or without (lanes l-3) excess nonradioactive, underivatized HA. Cells were incubated, photolyzed and analyzed as described in Methods. Molecular weight markers (M, x 10m3)and the top (T) of the resolving gel are indicated on the left. The open arrow shows the 54 kD band specifically labeled in the absence of Cat2 (lane 1). Lane 2 is a shorter exposure of the same sample in lane 3 to show the near absence of this band when Ca+* is present. The solid arrows indicate bands (24 and 43 kD) that are specifically crosslinked to HA in the presence of Ca+2.
conclusion will require purification and elucidation of the structures of the two major proteins covalently crosslinked to the photoaffinity ASD-HA derivative. ACKNOWLEDGIbWNlS We thank Shirley Chapman for preparation of hepatocytes and Lisa Raney for help preparing the manuscript. This work was supported by National Institutes of Health grant GM 35978. REFERENCES 1.
Evered, D. and Whelan, J., eds., (1989) The Bioloe
Chichester, UK 1596
of Hyaluronan. John Wiley,
Vol. 189, No. 3, 1992
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Hascall, V.C. and Heinegard, D. (1974) J. Biol. Chem. 249, 4242-4249. Delpech, B. and Halavant, C. (1981) J. Neurochem. 3, 855-859. Underhill, C.B. and Toole, B.P. (1980) J. Biol. Chem. 255, 4544-4549. Truppe, W., Basner, R., Von Figura, K. and Kresse, H. (1977) Biochem. Biophys. Res. Commun. a, 713-719. Smedsrod, B., Pertoft, H., Eriksson, S., Fraser, J.R.E. and Laurent, T.C. (1984) Biochem. J. 223.617-626. Frost, S.J., McGary, C.T., Raja, R.H. and Weigel, P.H. (1988) Biochim. Biophys. Acta 946, 66-74. Raja, R.H., McGary, CT. and Weigel, P.H. (1988) J. Biol. Chem. a, 16661-16668. Green, S.J., Tarone, G. and Underhill, C.B. (1988) Exp. Cell Res. m , 224-232. Lesley, J., Schulte, R. and Hyman, R. (1990) Exp. Cell Res. 187, 224-233. M iyake, K., Underhill, C.B., Lesley, J. and Kincade, P.W. (1990) J. Exp. Med. 172, 69-75. Aruffo, A., Stamenkovic, I., Melnick, M ., Underhill, C.B. and Seed,B. (1990) Cell f& 1303-1313. Ripellino, J.A., Margolis, R.U., and Margolis, R.K. (1988) J. Cell Biol. 105, 845-855. Fedarko, N.S. and Conrad, H.E. (1986) J. Cell Biol. _142,587-599. Fromme, H.G., Buddecke, E., Von Figura, K. and Kresse, H. (1976) Exp. Cell Res. m , 445-449. Frost, S.J., Raja, R.H. and Weigel, P.H. (1990) Biochemistry a, 10425-10432. Raja, R., LeBoeuf, R., Stone, G. and Weigel, P.H. (1984) Anal. Biochem. 139, 168177. Yannariello-Brown, J., Frost, S.J. and Weigel, P.H. (1992) J. Biol. Chem. Z,2045 l20456. Clarke, B.L., Oka, J.A. and Weigel, P.H. (1987) J. Biol. Chem. a, 17384-17392. Dalet, C., Fehlmann, M . and Debey, P. (1982) Anal. Biochem. 122, 119-123. Laemmli, U.K. (1970) Nature 227, 680-685. Bradford, M .M. (1976) Anal. Biochem. 2, 248-254. Labarca, C. and Paigen, K. (1980) Anal. Biochem. l& 344-352.
1597
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 1591-1597
RATHEPATYXXTE HYALURONAN/GLYCOSAMINOGLYCAN BINDING PROTEIN3 EVIDENCE FOR DISTINCI’ DIVALENT CATIONINDEPENDENT AND DIVALENT CATION-DEPENDENT AClWHlES Stephen J. Frost, Grete M . Kindberg, Janet A. Oka and Paul H. Weigel Department of Human Biological Chemistry and Genetics The University of Texas Medical Branch, Galveston, Texas 77555-0647 Received
November
9,
1992
We have previously shown (Biochemistry, a 10425, 1990) that hepatocytes contain intracellular specific binding sites for hyaluronan (HA). Although HA-binding activity is not dependent on divalent cations, it is increased in the presence of Cae2. Here we report that a novel photoaffmity HA derivative (ASD-HA) crosslinks specifically to different proteins in permeable cells in the presence or absenceof Caf2. With Cat2 present, two proteins of approximately 24 kD and 43 kD were labeled. Additionally, a broad zone of specific crosslinking was observed in the region of 40-100 kD. However, in the presence of the chelator EGTA this zone was absent and the 24 and 43 kD proteins were also not crosslinked to the HA photoaffinity derivative. In the absenceof Ca+2,only a 54 kD protein was specifically labeled. The results indicate that different intracellular hepatocyte proteins are responsible for the Cat2-independent and the Cat2-dependent binding of HA. 0 1992 Academic Press,Inc.
Hyaluronic acid, or hyaluronan (HA), is a non-sulfated glycosaminoglycan. HA is a ubiquitous component of the vertebrate extracellular matrix and can noncovalently interact with extracellular matrix molecules such as proteoglycans,link protein and hyaluronectin (l3). Furthermore, HA can bind to the surface of SV3T3 fibroblasts (4) hepatocytes (5-7) liver sinusoidal endothelial cells (6,8), alveolar macrophages(9) and lymphoid cells (10-12). HA, other glycosaminoglycansand proteoglycanshave also been reported inside cells in the cytoplasm and nucleus (13-15). This latter unexpected finding raises the possibility that glycosaminoglycans,or related molecules, may also have an intracellular function. Although hepatocyteshave some surface HA-binding sites, a large number of specific HA-binding sites (> lo6 cell) are intracellular (7,16). These binding sites, which bind HA
1591
All
0006-29 I X/92 $4.00 Copyright 0 1992 by Academic Press, Inc. rights of reproduction in my form rrser\&
Vol. 189, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUhlICATIOtIS
with a & -2 x 10s7Mare not nuclear; they are recovered in crude m icrosomes and are either associatedwith membranes or cytoskeleton. This binding activity is specific for HA and several other related glycosaminoglycans,including chondroitin sulfate and heparin (7). Other anionic polysaccharides,RNA and DNA do not bind to these HA/glycosaminoglycanbinding sites. This HA-binding activity in permeable hepatocytes is stimulated 2- to 6-fold by 10 m M concentrations of several divalent cations (16). The greatest stimulation of HA binding was demonstrated with Cat2 and Mn+2. Here, we provide evidence that the specific Ca+2-stimulatedHA-binding activity in hepatocytesis different than the Ca+2-independent HA-binding activity.
METHODS Materials. Digitonin was from Kodak, Collagenase (type I), Percoll, HA from human umbilical cord, calf thymus DNA, and Nonidet P-40 were from Sigma. SASD [sulfosuccinimidyl-2@-azidosalicylamido)ethyl-l,3’-dithiopropionate] and 1,3,4,6, tetrachloro3a,6cr diphenylglycouril (iodogen) were from Pierce. tzsI-NaI (lo-20 mCi/pg iodine) was from Amersham. Bovine serum albumin (BSA) was from Armour Biochemicals. Male Sprague Dawley rats (200 g) were obtained from Harlan Breeding Laboratories, Houston, TX. Preparation of m I-HA and ‘%ASD-HA. HA was further purified and converted to a hexylamine derivative at the terminal reducing sugar as described previously (17), with m inor modifications (7). The hydroxyphenylpropionyl derivative of the HA-hexylamine was prepared and iodinated (17) to give specific activities of 2-4 x lo6 dpm/ug HA. The lBI-HA was stable for at least 20 days at 4°C. SASD was handled in dim light or a red safety light. ASD-HA was prepared as recently described (18) by incubating 1 m l of HA-hexylamine (5 mg/ml) with 3.2 m l of SASD (2 mg/ml) in 0.1 M sodium carbonate, pH 9.0. The solution was incubated at room temperature for 30 m in in a foil-covered tube, and subsequently dialyzed in the dark against 150 m M NaCl, 20 m M sodium phosphate, pH 8.0 for 36 hr with three buffer changes at 4’C. Aliquots (500 ~1) were then stored at -7O’C. ASD-HA was iodinated (17) and stored at 4 oC in foil-covered tubes. Media and buffers. Medium l/BSA is a modified Eagle’s medium (Grand Island Biological Co.) supplemented with 2.4 g/L of 4-(2-hydroxyethyl)-l-piperazine-ethanesulfonic acid (HEPES), 0.22 g/L NaHCO,, and 0.1% (w/v) BSA, pH 7.4. Buffer 1 contains 143 m M NaCl, 6.8 m M KC1 and 10 m M HEPES, pH 7.4. Buffer l/BSA is Buffer 1 with 0.1% (w/v) BSA. BIClO is Buffer 1 with 10 m M CaCI,. Buffer l/EGTA is buffer 1 with 2 m M EGTA. Isolated rat hepatoqtes. Hepatocytes were prepared from male Sprague-Dawleyrats by a collagenaseperfusion procedure (19). Before use, the cells were first incubated at 37°C for 1 hr in Medium l/BSA to allow recovery from the isolation procedure. The cells were then chilled, washed, and loaded onto a discontinuous Percoll gradient to separate nonparenchymal cells and dead hepatocytes from viable hepatocytes (20). Hepatocytes (98% viability, 99% purity) at the bottom of the gradient were washed with Buffer 1, stored on ice and used within 1 hr. 1592
Vol.
189,
No.
3, 1992
BIOCHEMICAL
AND
BlOPHYStCAL
RESEARCH
COMMUNtCATtONS
mI-HA binding to pemeabilized hepatocytes. Cells (2-5 x 106cells/ml) were permeabilized with 0.055% digitonin in Buffer l/BSA for 20 min at 4”C, washed and then incubated with ‘251-HAin the absence(total binding) or presence (nonspecific binding) of 2 lOO-fold excess of nonradiolabeled HA in 12 x 75 mm plastic tubes at 4°C for 1 hr with gentle agitation every 5 min. The cells were washed three times by centrifugation and ‘ZI-radioactivity was determined using a Packard Multiprias 2 gamma spectrometer. Specific binding was calculated by subtracting the nonspecific binding from the total binding. Crosslinking of mI-A!XHIA. Hepatocytes (4 x 106/ml) were permeabilized with 0.055% digitonin in Buffer l/BSA plus 133 mM PMSF at 4”C, washed three times and resuspended at 1 x 10’cells/ml in either Buffer l/EGTA (minus Cat’) or BIClO (plus Ca+2). The cells were incubated for 20 min with (nonspecific binding) or without (total binding) 900 hg/ml of nonradiolabeled HA. ‘2SI-ASD-HA was added (12 pg/ml) and the cells were incubated for 45 min at 4°C in the dark. The cells were washed three times with either Buffer l/EGTA or BIClO, resuspendedand aliquots (lo6 cells in 75 ~1) were transferred to wells of a 96 well-plate. Photolysis was for 2 min with a short wave (254 nm), 4 watt UV lamp (Model UVG-11 from Ultra-violet Products, Inc., San Gabriel, CA.) placed on top of the plate. Laemmli (21) sample buffer (40 ~1 of 4X) containing 5% A-mercaptoethanol was added to each well and the final volume was adjusted to 150 ~1. Samples (75 ~1; 5 x 16 cells) were analyzed by SDS-PAGE on a 5-15% gradient polyacrylamide gel. The gel was dried and autoradiography was done at -70°C using preflashed X-Omat AR film (Kodak). General. Protein was measured using the method of Bradford (22) with BSA as a standard. Cell number was estimated by determining DNA as described by Labarca and Paigen (23) with calf thymus DNA as a standard. Data points represents the mean of triplicates and the error bars represent the sample standard deviation. RESULTS AND DISCWSION Effect of CM& on “I-HA
binding to permeabilized hepatocytes
Specific HA binding to permeabilized hepatocytesat 4’C is increased by Ca+2and other divalent cations such as Mg+2 and Mnc2 (16). A -lo-fold
increase in HA binding was
observed at different concentrations of lZI-HA when binding was assessedwith and without 10 mM Ca+2 (not shown). At a constant HA concentration, a linear increase in specific HA binding was observed with increasing concentrations of CaCl, (Fig. 1). The increase was 80 fmol of HA bound/lo6 cells per 1 mM increase in CaCl, concentration. It is important to note that in the absence of any free divalent cations (i.e. plus EGTA) there is still a significant amount of specific HA binding: in this experiment 200 fmol/106 cells or - 1.2 x ld intracellular sites/cell. The stimulator-y effect of Cat2 on HA binding is reversible. If hepatocytes are allowed to bind lZI-HA with 10 mM CaCl,, and then diluted to 2 mM CaCl, there is no significant increase in ‘=I-HA binding compared to that at 2 mM CaCl, alone. 1593
Vol.
01
189, No. 3, 1992
CaC12 CONCENTRATION
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(mt.4)
02
HA CONCENTRATION
hM)
x 1O-2
Figure 1. Effect of CaClz concentration on ‘%HA binding to permeabilized rat hepatocytes. Permeabilized hepatocytes (5 x lo6 cells/ml) in Buffer 1 plus 0.1% BSA were
incubated with increasing concentrations of CaCl, and 6 pg/ml of ?-HA in the presence or absence of 600 pg/ml of nonradiolabeled HA for 1 hr at 4°C. The cells were washed and specifically bound ‘~5I- HA and DNA content were determined. The line determined by least squares linear regression analysis (cc = 0.996) had a slope of 80 fmoles of lEI-HA specifically bound/lo6 cells/mM Ca+*. specific HA binding to Figure 2. Effect of lzsI-HA concentration on Ca”-stimulated permeabilized rat hepatocytes. Permeabilized cells were incubated in Medium 1 (which
contains 1.8 m M CaCl,) with increasing concentrations of lZI-HA in the presence or absence of 20 PM nonradiolabeled HA and with (0) or without (0) an additional 10 m M CaCl, for 1 hr at 4°C. The cells were washed and specifically bound lEI-HA was determined. Since the HA concentration in Figure 1 is less than the I& of HA for its binding site
in permeabilized hepatocytes (16), the increase in HA binding in the presence of 10 m M CaCl, could be explained by either a decreasein the & of the HA binding interaction (te. greater affinity) or an increase in the relative number of HA-binding sites. To test this, the HA concentration was varied from 12.5 to 600 nM I-IA, which is well above the IQ. A -3fold increase in specific ‘251-HA binding was observed at all HA concentrations in the presence of 10 m M CaCl, as compared with 1.8 m M CaCl, (Fig. 2). Most of the previous studies have been performed with 1.8 m M CaCl, present (7,16). Although saturation occurred in the presenceof 1.8 m M Ca +‘, with 10 m M Cat2 HA binding was still increasing. Furthermore, in preliminary equilibrium binding studies (not shown) we observed no significant change in the & for ‘=I-HA binding in the presence or absenceof 10 m M CaCl, (Le. 3.4 x lo-‘M vs 3.6 x lo-‘M). The results are consistentwith the conclusion that divalent cations such as Cat2 cause an increase in the apparent number of binding sites rather than an increase in the affinity of HA binding. 1594
Vol.
189,
No.
3, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Pho&affinity crosslinking of mI-ASD-HA to the Ca+2-stimulated and the Ca+2-independentHA-binding activities We have recently developed the use of a unique photoaffinity crosslinking derivative of HA to identify specific HA-binding proteins or receptors, such as the sinusoidal liver endothelial cell HA receptor (18). The iodinatable arylazide derivative ASD-HA has a major advantage for these studies, After being photo-activated and crosslinked to a target protein(s), the HA can be released from this complex by reduction of the disulfide bond in the (S)ASD. This treatment, in effect, transfers only the lZ1-aryl group to the target protein, which can then be detected by autoradiography after SDS-PAGE (18). lsI-ASD-HA (-65%) as lZI-HA.
bound to permeable hepatocytes with the same level of specificity A different crosslinking pattern was observed in the absence of Ca+2
(i.e. in the presence of EGTA) compared to when 10 mM CaCl, was included during the binding and photolysis of lZI-ASD-HA to permeabilized hepatocytes. In the presence of EGTA only one band at -54 kD was specifically crosslinked to the lZI-ASD-HA
(Fig. 3).
Binding and crosslinking of the HA derivative to this 54 kD protein were competed > 90% by nonderivatized HA. The labeling pattern with Ca” present was more complex, with diffuse labeling in the region between 40-100 kD (Fig. 4). That this material was dramatically decreasedin the presence of excessnonderivatized HA, suggestsit represents specific binding sites. Surprisingly, the 54 kD band seen without Ca+’ was not obviously labeled in the presence of Cat2 (lanes 1 and 2, Fig. 4). Better resolution of labeled bands was obtained in the region between 14-40 kD. In this region only a 43 kD and a 24 kD protein were consistently labeled specifically; that is, they were competed by nonlabeled HA. These two proteins specifically labeled in the presence of Cat2 were not labeled in the absenceof Ca+2 (lanes 1 and 3, Fig. 4). The different specific crosslinking pattern with ‘=IASD-I-IA provides further evidence that the increased HA binding to hepatocytes in the presence of Ca+’ is to different protein(s) than in the absenceof Cat’. Verification of this
1595
Vol.
189,
No.
3, 1992
03
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
04
Figure 3. Crosslinking of ‘251-ASD-HA to pemeabilized hepatocytes in the presence Digitonin-treated hepatocytes were incubated at 4°C with the photoaffinity of EGTA
derivative ‘“I-ASD-HA with 2 mM EGTA in the presence (lanes 1 and 2) or absence (lanes 3 and 4) of excess nonradioactive HA. UV photolysis, SDS-PAGE analysis and autoradiography were as described in Methods. The duplicate lanes are independent replicates. The migration of standard proteins (M, x 10”) and the top (T) of the resolving gel are indicated on the right. The arrow points to the 54 kD band that was specifically competed with nonradiolabeled HA. Figure 4. Effect of CaC& on the photoaffinity crosslinking of ‘%ASD-HA to permeable hepatocytes. Permeable hepatocytes were incubated with ‘%ASD-HA in the
presence of 2 mM EGTA (lane 1) or 10 mM CaCI, (lanes 2-4) with (lane 4) or without (lanes l-3) excess nonradioactive, underivatized HA. Cells were incubated, photolyzed and analyzed as described in Methods. Molecular weight markers (M, x 10m3)and the top (T) of the resolving gel are indicated on the left. The open arrow shows the 54 kD band specifically labeled in the absence of Cat2 (lane 1). Lane 2 is a shorter exposure of the same sample in lane 3 to show the near absence of this band when Ca+* is present. The solid arrows indicate bands (24 and 43 kD) that are specifically crosslinked to HA in the presence of Ca+2.
conclusion will require purification and elucidation of the structures of the two major proteins covalently crosslinked to the photoaffinity ASD-HA derivative. ACKNOWLEDGIbWNlS We thank Shirley Chapman for preparation of hepatocytes and Lisa Raney for help preparing the manuscript. This work was supported by National Institutes of Health grant GM 35978. REFERENCES 1.
Evered, D. and Whelan, J., eds., (1989) The Bioloe
Chichester, UK 1596
of Hyaluronan. John Wiley,
Vol. 189, No. 3, 1992
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Hascall, V.C. and Heinegard, D. (1974) J. Biol. Chem. 249, 4242-4249. Delpech, B. and Halavant, C. (1981) J. Neurochem. 3, 855-859. Underhill, C.B. and Toole, B.P. (1980) J. Biol. Chem. 255, 4544-4549. Truppe, W., Basner, R., Von Figura, K. and Kresse, H. (1977) Biochem. Biophys. Res. Commun. a, 713-719. Smedsrod, B., Pertoft, H., Eriksson, S., Fraser, J.R.E. and Laurent, T.C. (1984) Biochem. J. 223.617-626. Frost, S.J., McGary, C.T., Raja, R.H. and Weigel, P.H. (1988) Biochim. Biophys. Acta 946, 66-74. Raja, R.H., McGary, CT. and Weigel, P.H. (1988) J. Biol. Chem. a, 16661-16668. Green, S.J., Tarone, G. and Underhill, C.B. (1988) Exp. Cell Res. m , 224-232. Lesley, J., Schulte, R. and Hyman, R. (1990) Exp. Cell Res. 187, 224-233. M iyake, K., Underhill, C.B., Lesley, J. and Kincade, P.W. (1990) J. Exp. Med. 172, 69-75. Aruffo, A., Stamenkovic, I., Melnick, M ., Underhill, C.B. and Seed,B. (1990) Cell f& 1303-1313. Ripellino, J.A., Margolis, R.U., and Margolis, R.K. (1988) J. Cell Biol. 105, 845-855. Fedarko, N.S. and Conrad, H.E. (1986) J. Cell Biol. _142,587-599. Fromme, H.G., Buddecke, E., Von Figura, K. and Kresse, H. (1976) Exp. Cell Res. m , 445-449. Frost, S.J., Raja, R.H. and Weigel, P.H. (1990) Biochemistry a, 10425-10432. Raja, R., LeBoeuf, R., Stone, G. and Weigel, P.H. (1984) Anal. Biochem. 139, 168177. Yannariello-Brown, J., Frost, S.J. and Weigel, P.H. (1992) J. Biol. Chem. Z,2045 l20456. Clarke, B.L., Oka, J.A. and Weigel, P.H. (1987) J. Biol. Chem. a, 17384-17392. Dalet, C., Fehlmann, M . and Debey, P. (1982) Anal. Biochem. 122, 119-123. Laemmli, U.K. (1970) Nature 227, 680-685. Bradford, M .M. (1976) Anal. Biochem. 2, 248-254. Labarca, C. and Paigen, K. (1980) Anal. Biochem. l& 344-352.
1597
