Bone, 11, 133-139 (1990) Printed in the USA. All rights reserved.
8756-3282190 $3.00 + .OO Copyright 0 1990 Pergamon Press plc
A Monoclonal Antibody Against the Surface of Osteoblasts Recognizes Alkaline Phosphatase Isoenzymes in Bone, Liver, Kidney, and Intestine S. P. BRUDER Skeletal Research
and A. I. CAPLAN Center, Department
of Biology, Case Western Reserve
University, Cleveland,
Ohio
Address for correspondence
and reprints: Dr. Arnold I. Caplan, Skeletal Research Center, Department of Biology, Case Western Reserve University, 2080 Adelbert Road, Cleveland, Ohio 44106, USA.
Abstract
clonal antibodies against stage-specific antigens on the surface of osteogenic cells has allowed us to further define the developmental transitions, or lineage states, through which embryonic osteoblasts proceed during differentiation (Bruder and Caplan 1989a, in press). Immunostaining of developing chick limbs with a monoclonal antibody which recognizes Pre-Osteoblasts, SB-1, is both spatially and temporally coupled to the expression of alkaline phosphatase activity in developing bones. SB-1 is also reactive in extraskeletal tissues known to possess alkaline phosphatase activity, and the immunostaining pattern is identical to that observed in serial sections stained by alkaline phosphatase enzyme histochemistry (Bruder and Caplan 1989b). This observation suggests that SB-1 is directed against a common epitope on alkaline phosphatase isoenzymes found in chicken bone, liver, kidney, and intestine. Biochemical characterization of alkaline phosphatases from a variety of species indicate that there are multiple forms, including a bone/liver/kidney/placental form and an intestinal form (Hass et al. 1979; Moak and Harris 1979; Goldstein et al. 1980). Differences in apparent K,,,s, heat lability, response to specific inhibitors, and electrophoretic mobility have all been used to distinguish alkaline phosphatase isoenzymes (Wilkinson 1970). Additional studies in mice indicate that a single gene locus is responsible for the bone, liver, kidney and placental isoenzymes, while a separate locus is responsible for the intestinal alkaline phosphatase (Wilcox et al. 1979; Wilcox and Taylor 1981). Although the vast majority of investigations regarding alkaline phosphatase have relied upon mammalian systems, the existence of multiple avian alkaline phosphatase isoenzymes in bone, liver, serum, yolk, and intestine have also been reported (Kuan et al. 1%6; Chang and Moog 1972; Debruyne and Stockx 1979). While monoclonal antibodies against alkaline phosphatase from rat bone (Nair et al. 1987) and human intestine (Vockley and Harris 1984) have been generated, monoclonal antibodies are not available which react with all alkaline phosphatase isoenzymes from a single species. In this study, information is presented to describe a monoclonal antibody, SB-1, generated against the surface of osteoblasts, which is shown to recognize alkaline phosphatase isoenzymes from chicken bone, liver, kidney, cartilage, and intestine.
Monoclonal antibodies against the surface of embryonic osteogenic cells have been used to characterize the osteoblastic lineage. One antibody, SB-1, reacts in frozen sections with a family of cells in bone, liver, kidney, and intestine which are identically stained by the histochemical substrate for alkaline phosphatase. In this report, biochemical and immunochemical evidence is presented to indicate that SB-1 is directed against an epitope on alkaline phosphatase which is shared by isoenzymes in a variety of chick tissues. In a solid-phase assay system, high dilutions (l:lOs) of ascites fluid were found to give significant binding of SB-1 to alkaline phosphatase extracted from chick limb or intestine. Partial puritlcation of intestinal alkaline phosphatase on a Sepharose CL-6B column results in the co-elution of alkaline phosphatase enzyme activity and antibody-binding material; this indicates that SB-1 recognizes intestinal alkaline phosphatase rather than an impurity in the crude preparation. Furthermore, Western immunoblots of chick calvarial bone extract electrophoresed on a S-20% SDS-polyacrylamide gel show that SB-1 reacts with a single 155 kD band which also is stained by the alkaline phosphatase histochemical substrate. In a similar set of experiments, SB-1 reacts with an intestinal alkaline phosphatase isoenzyme whose molecular weight is approximately 185 kD. From these studies, we conclude that SB-1 is specifically reactive with alkaline phosphatase. isoenzymes present in bone, liver, kidney, cartilage, and intestine. The reactive epitope is stable to SDS denaturation, not associated with the active site of the enzyme, and dependent on disulfide bonds which impart secondary structure to the protein. Key Words: Alkaline phosphatase-Monoclonal bodies-Isoenzymes-Osteoblast-ELISA-Immunoblot.
anti-
Introduction Studies on the differentiation of osteogenic cells have relied, in large part, on detecting the initial expression of bone-specific molecules. The generation of a set of mono-
133
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase
134
Tissue extraction
Materials and Methods Alkaline phosphatase
and monoclonal
antibodies
Partially purified chicken intestinal alkaline phosphatase was obtained from Sigma. Purified and partially purified alkaline phosphatase from chicken epiphyseal cartilage was generously provided by Dr. Roy Wuthier, University of South Carolina, Columbia, South Carolina (Cyboron and Wuthier 1981). Antibodies SB-1, SB-2, and SB-3 have been previously characterized, and details regarding their generation are reported elsewhere (Bruder and Caplan 1989b). Immunohistochemical
staining
Unfixed frozen sections of embryonic chick limbs were immunostained as previously described (Bruder and Caplan 1989b). Culture supernatant was used undiluted, while ascites fluid was diluted I:100 in PBS with 1% BSA (BPBS). Following the primary incubation, slides were rinsed three times with B-PBS and then incubated for 1 h as previously described with fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin (FITC-GAMIg) secondary antibody (Organon Teknika Corp.) diluted 1:lOO in B-PBS. Control experiments consisted of tissue sections that were incubated with non-immune mouse serum, culture supernatant from SP2/0 Ag-14 cells, or B-PBS, followed by FITC-GAMIg secondary antibody as described above. None of these control tissue sections displayed a selective immunofluorescent staining pattern. Alkaline phosphatase enzyme activity was histochemically visualized in unfixed frozen sections as outline in Sigma Kit No. 85. Enzyme linked immunosorbent
assays
Hybridoma culture supematant or ascites fluid from antibodies SB-1, SB-2, and SB-3 were separately used to coat wells of vinyl microtiter plates (GIBCO) for these modified solid-phase assays. For titration experiments, ascites samples were serially diluted in B-PBS, and 50 uL of the appropriate dilutions were left in wells for 16 hours at 4°C. The wells were washed three times with B-PBS and then blocked by completely tilling each well with B-PBS and incubating for 1 h at 25°C. Intestinal, purified, or partially purified alkaline phosphatase was dissolved in B-PBS, and 50 p,L were incubated in the appropriate wells for 1 h at 25°C. For some experiments, fractions from the molecular sieve chromatographic separations (see below) were applied rather than the aforementioned samples of commercial or purified alkaline phosphatase. After rinsing the wells three times with B-PBS, alkaline phosphatase substrate, composed of 3.5 mM p-nitrophenylphosphate, 50 mM glycine, 1 mM MgCl*, pH 10.5, was added to each well. The product formed was read at an absorbance of 405 nm with a microtiter plate reader. The above ELISA configuration was developed because primary coating of the microtiter plates with alkaline phosphatase proved inconsistent. Therefore, the antibody was first bound to the plate and then exposed to the suspected antigen. Because the suspected antigen itself is enzymatically active, substrate was then supplied to directly test for the presence of antigen bound to the antibody. Appropriate negative controls were performed with every series of experiments.
One dozen 17-day chick embryos were removed from their shells and decapitated, and the heads immediately placed in 4°C Tyrode’s salt solution containing protease inhibitors (0.1 M 6-aminohexanoic acid, 0.005 M benzamide, 0.01 M ethylenediaminetetraacetic acid, and 0.1 mM phenylmethylsulfonyl fluoride). The calvarial bones were then dissected, placed in 1 mL of fresh Tyrode’s containing inhibitors and 0.1% ‘Biton X-100, homogenized with a Polytron (Brinkman) homogenizer for 1 min, and sonicated for 1 min with 15-s bursts. The resulting homogenate was placed in an Eppendorf tube and centrifuged at 8800 x g for 5 min to remove the insoluble material. The supernatant was separated into small aliquots, frozen, and stored at - 20°C until further use. All tissue extraction steps were performed at 4°C. Protein content was determined by the method of Bradford (1976), and standards were prepared with BSA (Sigma). Electrophoretic
separation of proteins
Sodium dodecyl sulfate (SDS)-polyacrylamide slab gels of 5-20% gradient were prepared and run as described elsewhere (Laemmli 1970). In all cases, a 5% stacking gel was formed on top of the analytical gel. For nonreduced samples, the sample buffer contained 10% glycerol, 2% SDS, and 0.01% bromophenol blue. For reduced samples, freshly diluted 5% B-mercaptoethanol was added to the above buffer. Selected samples were also boiled for 3 min immediately before loading. In some experiments, nondenaturing continuous 10% slab gels were prepared by substituting Triton X-100 for SDS in the polyacrylamide gel. A 5% stacking gel was also formed on top of the analytical gel. Electrophoresis buffer for these native gels consisted of 50 mM Bis and 192 mM glycine, pH 8.3. Sample buffer for these gels contained 10% glycerol and 0.01% bromophenol blue. Immediately after electrophoresis, gels were fixed in 40% methanol, 10% acetic acid, and subsequently stained with Coomassie Blue. For demonstration of alkaline phosphatase activity within the gels, immediately after electrophoresis the gels were soaked in 37°C 50 mM glycine, 1 mM MgCl,, pH 10.5, for 30 min. Gels were then soaked in 37°C alkaline phosphatase substrate as outlined in Sigma Kit No. 85 and fixed in 40% methanol, 10% acetic acid. Western immunoblot analysis
Samples separated by gel electrophoresis were electrophoretically transferred to Immobilon PVDF Transfer Membrane (Millipore Corp.) with a Hoefer Transphor electrophoresis apparatus according to Millipore technical protocol TP-006, except methanol was omitted from the transfer buffer. PVDF membranes were cut into strips, blocked with 3% BSA in Tiis-buffered saline (TBS), 0.05% ‘Iween-20 overnight at 4°C. Some strips were not blocked, but immediately incubated with alkaline phosphatase substrate as outlined in Sigma Kit No. 85. BSA-blocked strips were then washed with TBS-Tween and incubated with hybridoma culture supematant diluted 1: 10 in 3% BSA, TBSTween for 30 min on a shaker plate at 25°C. Subsequent immunoblotting steps, including preparation of fresh diaminobenzidine substrate, were performed according to the Vectastain ABC Kit instructions (Vector Labs).
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase Molecular sieve chromatography
A Sepharose CL-6B gel filtration column (50 x 1.5 cm) was prepared and eluted with 1.0 M NaCl, 50 mM Tris, pH 8.0. Effluent fractions of 0.6 mL were collected and analyzed for protein content by optical density measurements at 280 nm, alkaline phosphatase enzyme activity, and specific binding to SB-1 in the ELISA described above. Alkaline phosphatase
enzyme activity
Fractions eluted from the Sepharose CLdB column were tested for alkaline phosphatase activity by adding small aliquots to tubes containing p-nitrophenylphosphate substrate buffer as prepared above. Triplicate samples were incubated at 37°C for 1.5min, and the product formed was read at an absorbance of 405 nM on a Shimadzu dual beam spectrophotometer.
Results Immunohistochemical
staining
A detailed description of the immunohistochemical staining by antibody SB-1 during first bone formation of the embryonic chick tibia has already been presented (Bruder and Caplan 1989a, 1989b). In brief, antibody SB-1 stains the surfaces of osteoblasts which comprise the inner layer of the bilayered periosteum. In serial sections of bone, liver, kidney, and intestine, an identical family of cells is stained by antibody SB-1 and the histochemical stain for alkaline phosphatase (Bruder and Caplan 1989b). No reactivity was observed when cryosections from rat, rabbit, or man were immunostained. Figure 1 illustrates the immunostaining pattern of SB-1 in a longitudinal section of an embryonic day 10 tibia. The surfaces of cuboidal cells within the inner layer of the periosteum are immunoreactive, but cells entrapped within bone matrix are neither stained by SB-1 nor the histochemical stain for alkaline phosphatase. Occasional flattened chondrocytes along the periphery of the cartilage core are similarly reactive (Pechak et al. 1986; Bruder and Caplan 1989b). Antibodies SB-2 and SB-3 each stain a different family of cells than
135
those stained by SB-1 (Bruder and Caplan 1989a, 1989b). For this reason, SB-2 and SB-3 serve as negative controls for several of the experiments in this study. ELZSA binding titration curves
The binding of several alkaline phosphatase preparations to immobilized aliquots of serially diluted SB-1 ascites samples was evaluated in the solid-phase assay described in Materials and Methods. In addition, antibodies SB-2 and SB-3 were similarly coated on wells of microtiter plates at serial IO-fold dilutions. When commercially available alkaline phosphatase was used as the antigen (and indicator) in this assay, significant binding by SB-1 was observed with ascites fluid dilutions as great as 1: 10s. The titration curve for this antigen:antibody interaction is illustrated in Fig. 2A. Control experiments performed with antibodies SB-2 and SB-3 demonstrate that alkaline phosphatase is not bound by these antibodies. A similar series of experiments was performed with both purified and partially purified alkaline phosphatase from chicken epiphyseal cartilage (Cyboron and Wuthier 1981). The titration curve for SB-1 binding with partially purified material (Fig. 2B) indicates that, like intestinal alkaline phosphatase, binding is observed with ascites fluid dilutions as great as 1:loS. Purified alkaline phosphatase samples from epiphyseal cartilage similarly were bound by SB-l-coated microtiter plates; however, the limited quantity of purified material precluded the preparation of detailed titration curves. Control experiments performed with antibodies SB-2 and SB-3 again demonstrate that alkaline phosphatase is not nonspecifically bound to antibodies in this solid-phase assay. Molecular sieve chromatography
of intestinal
alkaline phosphatase
To further characterize the binding of SB-1 to the commercial intestinal alkaline phosphatase preparation, a sample of the commercial preparation was applied to a Sepharose CL-6B column in order to separate the enzyme from contaminating components. Figure 3A compares the protein content of the eMuent fractions with the alkaline phospha-
Fig. 1. Longitudinal frozen section of the tibia of a lo-day-old chick embryo stained with monoclonal antibody SB-1 followed by FITCGAMIgG and viewed by fluorescent (A) or phase (B) microscopy. Stacked cells (st) in the outer layer of the periosteum are not reactive, but all cells of the inner periosteal layer react with SB-1 at their surfaces. An immunoreactive flattened chondrocyte (arrowhead) can be seen at the periphery of the cartilaginous core. Note that neither bone matrix (M), nor cells trapped within bone matrix (osteocytes) are immunoreactive. V, Blood vessel. (X 310)
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase
136
1
2
3
4
5
6
7
3.0
0.6
2.0
0.4
1.0
0.2
8
Log dilution 0.0
0.0 10
20
30
50
Fraction #
? 2.0
40
60
70 T vt
vo
0.6
0.2
1.0
0.0 1
2
3
4
5
6
7
0.4
8
Log dilution Fig. 2. Titration curves for antibody reactivity with various forms of alkaline phosphatase. Log dilutions of ascites from antibodies SB-1 (O), SB-2 (0) and SB-3 (+) were coated onto microtiter plates and reacted with intestinal alkaline phosphatase (A) or partially purified epiphyseal cartilage alkaline phosphatase (B). Binding was assayed by reading the microtiter plates at 405 nm after applying p-nitrophenylphosphate substrate solution as described in Materials and Methods.
0.1
0.2
0.0 10
0.0 20
30 ?
tase enzyme activity of each fraction. These data reveal that enzyme activity is concentrated within 10% of the effluent, fractions 35 through 38, while protein is distributed from the V, to the V, of the column. When the effluent fractions were tested for their ability to bind to SB-l-
coated microtiter plates in the modified ELISA, the fractions which bound were those which contained the enriched alkaline phosphatase enzyme activity (Fig. 3B). The binding of proteins which eluted either before or after the peak of enzyme activity was not detected. Molecular weight determination Both freshly extracted alkaline phosphatase from embryonic calvaria and commercially prepared alkaline phosphatase from chicken intestine were subjected to electrophoresis on a 5-20% gradient SDS-polyacrylamide gel (Fig. 4). Coomassie Blue staining indicates that the calvarial extract contains a heterogeneous mixture of protein which ranges from over 200 kD to 14 kD in molecular weight (Lane A). By contrast, the commercial preparation of intestinal alkaline phosphatase is highly enriched in one protein which migrates at a position corresponding to approximately 185 kD (Lane B), as determined by calculated Rf values. Reduction of the intestinal preparation does not alter its electrophoretic mobility (Lane C), although boiling of a reduced sample results in the elimination of the 185 kD species and the emergence of a doublet protein band with lower molecular weights (Lane D). The major protein band is present at a position which corresponds to approximately 110 kD, and the minor band migrates to a position which corresponds to approximately 97 kD. Identically prepared gels were also stained for alkaline
40
50
Fraction #
vo
60
70 3L vt
Fig. 3. (A) Elution profile of a Sepharose CLdB chromatographic separation of commercially prepared intestinal alkaline phosphatase. Alkaline phosphatase activity (m) was measured at 485 nm and protein content was determined (0) for each fraction. (B) Binding of each effluent fraction to antibody-coated microtiter plates was determined (0), and plotted against the alkaline phosphatase activity (m) from A. In panel B, the left axis refers to the solid-phase immunoassay and the right axis refers to the alkaline phosphatase activity assay. V, = void volume, V, = total volume. phosphatase enzyme activity (Fig. 5). Analysis of this gel indicates that the bone and intestinal isoenzymes of alkaline phosphatase are of significantly different molecular weights. Enzyme activity in both reduced (Lane A) and unreduced (Lane B) samples of calvarial bone extract is present at a position corresponding to approximately 155 kD, although reduction results in compromised enzyme activity evidenced by the lower intensity of staining in Lane A compared to Lane B. By contrast, both reduced (Lane D) and unreduced (Lane C) intestinal alkaline phosphatase migrates to a position corresponding to approximately 185 kD. Again, the intensity of the alkaline phosphatase enzymatic reaction in the reduced sample is less than that of the unreduced sample. Boiling of the alkaline phosphatase samples results in the elimination of enzyme activity (data not shown).
Western immunoblot
analvsis
Crude extract from day-17 embryonic chick calvaria was electrophoresed on a 5-20% SDS-polyacrylamide gel as described above. The material within the gel was then
137
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase
ABCD
ABCD
A BC
ABC +
200* 11w 97,
* >‘ f,!; f$
200* L. 3
66*
116~ 97*
z
66, 45*
31-
31*
22*
22*
14*
14*
4
'. .r ',,'-T
5
6
7
Fig. 4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of calvarial extract and intestinal alkaline phosphatase stained with Coomassie Blue. Samples and amounts loaded were: Lane A, reduced cabarial extract, 40 pg of protein; Lane B, unreduced intestinal alkaline phosphatase, 20 pg of protein; Lane C, reduced intestinal alkaline phosphatase, 20 pg of protein; Lane D, boiled and reduced intestinal alkaline phosphatase, 20 pg of protein. Molecular weight markers are Bio-Rad high- and low-molecular weight standards. Fig. 5. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of alkaline phosphatase samples stained to demonstrate enzyme activity. Samples and amounts loaded were: Lane A, reduced calvarial extract, 40 ug of protein; Lane B, unreduced calvarial extract, 40 ug of protein; Lane C, unreduced intestinal alkaline phosphatase, 20 ug of protein; Lane D, reduced intestinal alkaline phosphatase, 20 kg of protein. Molecular weight markers are Bio-Rad high- and low-molecular weight standards. The single-head arrow denotes the 185 kD alkaline phosphatase isoenzyme from intestine. The double-head arrow denotes the 155 kD alkaline phosphatase isoenzyme from bone. Fig. 6. Immunoblot display of SB-1 reactivity in a freshly prepared calvarial extract. Forty ug of reduced protein was applied to Lane A, while Lanes B and C were loaded with forty ug of unreduced extract. Samples were electrophoresed on a 5-20% SDS-polyacrylamide gel, transferred to Immobilon PVDF, and visualized by the following techniques: Lane A, alkaline phosphatase histochemical stain; Lane B, antibody SB-1 followed by secondary antibody and substrate; Lane C, secondary antibody and substrate without the SB-I antibody. The arrowhead denotes the single protein band which is both immunoreactive with antibody SB-1 (Lane B) and stained by alkaline phosphatase histochemicai substrate (Lane A). Fig. 7. Immunoblot display of SB-1 reactivity in an intestinal alkaline phosphatase preparation. tienty pg of unreduced protein were applied to each lane. Samples were electrophoresed on a continuous 10% ‘Riton X-100~polyacrylamide gel, transferred to ImmobiIon PVDF, and visualized by the following techniques: Lane A, alkaline phosphatase histochemical stain; Lane B, antibody SB-1 followed by secondary antibody and substrate; Lane C, secondary antibody and substrate without the SB-1 antibody. The arrowhead refers to the sinale nrotein band which is both immunoreactive with antibody SB-1 (Lane B) and stained by alkaline phosphatase histochemicai
substrate (Lane A). transferred to an Immobilon PVDF membrane and reacted with antibody SB-1 (Fig. 6). Lane A, stained for alkaline phosphatase activity, indicates where the enzymatically active protein is localized on the membrane under reducing conditions. When this membrane was reacted with antibody SB-1 followed by the secondary antibody and substrate, predominant immunostaining was observed (Lane B) at a position corresponding to the band possessing alkaline phosphatase enzyme activity in Lane A. Although several other bands are observed at intermediate and low molecular weights in Lane B, these bands are identically reactive in Lane C, which was exposed only to the secondary antibody and substrate. The immunoreactivity of SB-1 with intestinal alkaline phosphatase was additionally assessed by Western immunoblotting. Samples were electrophoresed on a continuous 10% Triton X-lOO-polyacrylamide gel, electrotransferred onto Immobilon PVDF, and immunoreacted as described in Materials and Methods. Alkaline phosphatase activity, under nonreducing (Lane A) conditions, is observed as a single band in Fig. 7. When the nonreduced sample was reacted with antibody SB-1, and subsequently exposed to secondary antibody and substrate, a single band was visualized (Lane B) corresponding to the region of alkaline phosphatase activity present in Lane A. The immunoblot control with only secondary antibody and substrate produced no specific immunoreactive product (Lane C).
Discussion The results presented above show that SB-I, a monoclonal antibody generated against the surface of avian osteogenic cells, recognizes an epitope on alkaline phosphatase which is present on the bone, liver, kidney, cartilage, and intestinal isoenzymes. Previous immunohistochemical data illustrate the binding of SB-1 to the above tissues in a pattern which is identical to that obtained from alkaline phosphatase enzyme histochemistry of serial sections (Bruder and Caplan 1989b). While these data alone do not confirm that SB-1 is directed against alkaline phosphatase, they suggest that this is a tenable possibility. The biochemical and immunochemical evidence presented above strongly support the identity of the molecules recognized by antibody SB-1 as alkaline phosphatase isoenzymes. The modified ELISA designed for this study demonstrates several important points about the reactivity of SB-1 with alkaline phosphatase. First, the shape of the titration curve indicates that antibody binding is saturable (Fig. 2). The binding signal strength is maximal at an ascites dilution of l:lO“, with no further increase in antibody binding at even lower ascites dilutions. Furthermore, the high dilution (1:105) of ascites fluid used to obtain a significant binding signal in this solid-phase assay indicates that such binding is of a high affinity. Together, these data suggest that the binding of SB-1 to alkaline phosphatase is
138
S. P. Bruder and A. 1. Caplan: Monoclonal antibody against alkaline phosphatase
of a specific nature. The absence of binding in control studies with either no antibody or monoclonal antibodies SB-2 and SB-3 eliminates the possibility that alkaline phosphatase nonspecifically binds to either the plate itself or to mouse immunoglobulins. Partial purification of intestinal alkaline phosphatase on a Sepharose CL-6B column results in the co-elution of protein having alkaline phosphatase enzyme activity and antibody-binding reactivity (Fig. 3). This observation provides evidence that SB-1 recognizes intestinal alkaline phosphatase rather than an impurity in the crude preparation which is separated during molecular sieve chromatography. Thus, the data from this series of experiments indicate that alkaline phosphatase isoenzymes from both developing limb and intestine are specifically recognized by antibody SB- 1. Standard immunoblot techniques were used with chick calvarial extracts which had been electrophoresed on SDS-polyacrylamide gels. When the transfer membrane was exposed to antibody SB- 1, one predominant band was visualized at a position which matches the single band of alkaline phosphatase activity at an apparent molecular weight of 155 kD (Fig. 6). Although additional bands were present at lower molecular weights, the same bands were observed in the secondary antibody control experiment. The reasons for the secondary antibody’s binding with several species in the extract is not clear, but experience with this system has shown that several preparations of secondary antibody from a variety of sources nonspecifically react with different extract components. While the molecular weight of chick bone alkaline phosphatase has not been rigorously studied and reported, the apparent subunit molecular mass of rat bone alkaline phosphatase has been shown to be 82 kD (Nair et al. 1987). Our estimate of 155 kD for chick bone alkaline phosphatase is consistent with the dimeric structure of yolk alkaline phosphatase, the molecular weight of which is 150 kD (Debruyne and Stockx 1979). It is possible that chick bone alkaline phosphatase also has a dimeric structure of subunits whose molecular weight is similar to that reported for rat bone. Additional characterization of the chick bone alkaline phosphatase is required to confirm this postulate. Additional immunoblot experiments were performed with intestinal alkaline phosphatase from chick. These experiments reveal that a single band of SB-1 reactivity is present at a position corresponding to the band of alkaline phosphatase enzyme activity within the gel. This immunoblot experiment was performed on a nondenaturing gel in order to demonstrate SB-I reactivity while minimizing nonspecific background staining. In the case of intestinal alkaline phosphatase, the single band of enzyme activity has an apparent molecular weight of 185 kD. This estimate is similar to that reported for duodenal alkaline phosphatase from 4-day-old chicks (Chang and Moog 1972). SDSpolyacrylamide gel electrophoresis of boiled and reduced samples suggest that the native protein is composed of two subunits with molecular weights of 97 kD and 100 kD. Taken together, the various lines of experimentation indicate that SB-1 is specifically reactive with alkaline phosphatase isoenzymes present in bone, liver, kidney, cartilage, and intestine. The predominant cell surface immunostaining by SB-1 in cryosections of developing bone is consistent with the plasma membrane-bound localization of alkaline phosphatase on osteoblasts (Rodan and Rodan 1983). The data further indicate that the epitope recognized by SB-1 is not a part of the active site based on the observation that in the modified ELISA, alkaline phosphatase
which is bound to SB-1 on the microtitration plate retains its enzymatic activity. If the reactive epitope was associated with the active site of the enzyme, one would expect interference in the catalytic process. In another series of experiments, the addition of SB-1 to solutions of alkaline phosphatase was found to not hinder the cleavage of the p-nitrophenylphosphate substrate (unpublished data). While the reactive epitope is stable to SDS denaturation (Fig. 6), reduction of the disulfide linkages results in abolition of antibody binding in both the immunoblot assay and a dot-blot assay (data not shown). For this reason, the reactive epitope seems to be dependent on disulfide bondmediated secondary structure of the protein. Although further work is required to characterize the exact details of the epitope, the data presented above provide strong evidence that SB- 1 specifically recognizes alkaline phosphatase isoenzymes from a variety of chick tissues.
The authors would like to thank Mr. James Holecek, Drs. David Carrino, Louis Gerstenfeld and Stephen Haynesworth for their valuable assistance during the course of this study. SPB was supported by a Medical Scientist Training Grant GMO-7250-13. This work was supported by grants from the National Institutes of Health.
Acknowledgments:
References Bradford, M. A rapid and sensitive method for the quantitation of p-gram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem.
72~248-254;
1976.
Bruder. S. P.; Caplan, A. I. Discrete stages within the osteogenic lineage are revealed by alterations in the cell surface architecture of embryonic bone cells. Glimcher, M. J.; Lian, J. B., eds. The chemistry and biology of mineralized tissues. New York: Gordon and Breach; 1989a:pp. 73-79. Bruder. S. P.; Caplan, A. 1. First bone formation and the dissection of an osteogenic lineage in the embryonic chick tibia is revealed by monoclonal antibodies against osteoblasts. Bone 10:359-375; 1989b. Bruder, S. P.; Caplan, A. I. Terminal differentiation of osteogenic cells in the embryonic chick tibia is revealed by a monoclonal antibody against osteocytes. Bone (in press). Chang, C.; Moog, F. Alkaline phosphatases of the chicken duodenum. I. Isolation and partial characterization of the multiple forms of duodenal phosphatase in pre- and post-hatching stages. Biochim. Biophys. Acra. 258:154-165; 1972. Cyboron, G. W.; Wuthier, R. E. Purification and initial characterization of intrinsic membrane-bound alkaline phosphatase from chicken epiphyseal cartilage. JBC 256(14):7262-7268; 1981. Debruyne, I.; Stockx, J. Alkaline phosphatases from hen’s egg yolk, liver, blood plasma and intestinal mucosa: aftinity chromatographic puritication and comparison. ht. J. Biochem. 10:981-988: 1979. Goldstein, D. S.; Rogers, C. E.; Harris, H. Expression of alkaline phosphatase loci in mammalian tissues. Proc. Natl. Acad. Sci. USA 77:2860-2957;
1980.
Hass. P. E.; Wada, H. G.; Horman, M. M.; Sussman, H. H. AIkaIine phosphatase of mouse teratoma stem cells: immunochemical and structural evidence for its identity as a somatic gene product. Proc. Nat/. Aced. Sci. WA 76:1164-1168; 1979. Kuan, S. S.; Martin, W. G.; Patrick, H. Alkaline phosphatases of the chick. Partial characterization of the tissue isozymes. Proc. Sot. Exp. Biol. Med. 122:172-177; 1966. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227680-685; 1970. Moak, G.; Harris, H. Lack of homology between dog and human placental alkaline phosphatases. Proc. Nat/. Acnd. Sci. USA 761948-1951; 1979. Nair, B. C.; Majeska, R. J.; Rodan, G. Rat alkaline phosphatase. I. Purifi-
S. P. Bruder
and A. I. Caplan:
Monoclonal
antibody
against
alkaline
cation and characterization of the enzyme from osteosarcoma: generation of monoclonal and polyclonal antibodies. Arch. Biochem. Biophys. 254~18-27; 1987. Pechak, D. G.; Kujawa, M. K.; Caplan, A. I. Morphological and histochemical events during first bone formation in embryonic chick limbs. Bone 7:441-458; 1986. Rodan, G.; Rodan, S. Expression of the osteoblast phenotype. Peck, W. A., ed. Bone and mineral research. New York: Elsevier; 1983:~~. 244-285. (Vol. 2). Vockley, J.; Harris, H. Purification of human adult and foetal intestinal alkaline phosphatases by monoclonal antibody immunoaffinity chromatography. B&hem. J. 217:535-541; 1984. Wilcox, F. H.; ‘lbylor, B. A. Genetics of the Akp-2 locus for alkaline phosphatase of liver, kidney, bone, and placenta in the mouse. J. Hered. 72:387-390;
1981.
phosphatase
139
Wilcox, F. H.; Hirschhorn, L.; ‘lhylor, B. A.: Womack, J. E.; Roderick, T. H. Genetic variation in alkaline phosphatase of the house mouse (Mus musculus) with emphasis on a manganese-requiring isozyme. Biothem. Gem?. 17: 1093- 1107; 1979. Wilkinson, J. Phosphatase isoenzymes. Wilkinson, London: Chapman & Hall; 1970:~~. 69-102.
J., ed. Isoenomes.
Received: June 30, 1989 Revised: November 15, 1989 Accepted: November 16, 1989
8756-3282190 $3.00 + .OO Copyright 0 1990 Pergamon Press plc
A Monoclonal Antibody Against the Surface of Osteoblasts Recognizes Alkaline Phosphatase Isoenzymes in Bone, Liver, Kidney, and Intestine S. P. BRUDER Skeletal Research
and A. I. CAPLAN Center, Department
of Biology, Case Western Reserve
University, Cleveland,
Ohio
Address for correspondence
and reprints: Dr. Arnold I. Caplan, Skeletal Research Center, Department of Biology, Case Western Reserve University, 2080 Adelbert Road, Cleveland, Ohio 44106, USA.
Abstract
clonal antibodies against stage-specific antigens on the surface of osteogenic cells has allowed us to further define the developmental transitions, or lineage states, through which embryonic osteoblasts proceed during differentiation (Bruder and Caplan 1989a, in press). Immunostaining of developing chick limbs with a monoclonal antibody which recognizes Pre-Osteoblasts, SB-1, is both spatially and temporally coupled to the expression of alkaline phosphatase activity in developing bones. SB-1 is also reactive in extraskeletal tissues known to possess alkaline phosphatase activity, and the immunostaining pattern is identical to that observed in serial sections stained by alkaline phosphatase enzyme histochemistry (Bruder and Caplan 1989b). This observation suggests that SB-1 is directed against a common epitope on alkaline phosphatase isoenzymes found in chicken bone, liver, kidney, and intestine. Biochemical characterization of alkaline phosphatases from a variety of species indicate that there are multiple forms, including a bone/liver/kidney/placental form and an intestinal form (Hass et al. 1979; Moak and Harris 1979; Goldstein et al. 1980). Differences in apparent K,,,s, heat lability, response to specific inhibitors, and electrophoretic mobility have all been used to distinguish alkaline phosphatase isoenzymes (Wilkinson 1970). Additional studies in mice indicate that a single gene locus is responsible for the bone, liver, kidney and placental isoenzymes, while a separate locus is responsible for the intestinal alkaline phosphatase (Wilcox et al. 1979; Wilcox and Taylor 1981). Although the vast majority of investigations regarding alkaline phosphatase have relied upon mammalian systems, the existence of multiple avian alkaline phosphatase isoenzymes in bone, liver, serum, yolk, and intestine have also been reported (Kuan et al. 1%6; Chang and Moog 1972; Debruyne and Stockx 1979). While monoclonal antibodies against alkaline phosphatase from rat bone (Nair et al. 1987) and human intestine (Vockley and Harris 1984) have been generated, monoclonal antibodies are not available which react with all alkaline phosphatase isoenzymes from a single species. In this study, information is presented to describe a monoclonal antibody, SB-1, generated against the surface of osteoblasts, which is shown to recognize alkaline phosphatase isoenzymes from chicken bone, liver, kidney, cartilage, and intestine.
Monoclonal antibodies against the surface of embryonic osteogenic cells have been used to characterize the osteoblastic lineage. One antibody, SB-1, reacts in frozen sections with a family of cells in bone, liver, kidney, and intestine which are identically stained by the histochemical substrate for alkaline phosphatase. In this report, biochemical and immunochemical evidence is presented to indicate that SB-1 is directed against an epitope on alkaline phosphatase which is shared by isoenzymes in a variety of chick tissues. In a solid-phase assay system, high dilutions (l:lOs) of ascites fluid were found to give significant binding of SB-1 to alkaline phosphatase extracted from chick limb or intestine. Partial puritlcation of intestinal alkaline phosphatase on a Sepharose CL-6B column results in the co-elution of alkaline phosphatase enzyme activity and antibody-binding material; this indicates that SB-1 recognizes intestinal alkaline phosphatase rather than an impurity in the crude preparation. Furthermore, Western immunoblots of chick calvarial bone extract electrophoresed on a S-20% SDS-polyacrylamide gel show that SB-1 reacts with a single 155 kD band which also is stained by the alkaline phosphatase histochemical substrate. In a similar set of experiments, SB-1 reacts with an intestinal alkaline phosphatase isoenzyme whose molecular weight is approximately 185 kD. From these studies, we conclude that SB-1 is specifically reactive with alkaline phosphatase. isoenzymes present in bone, liver, kidney, cartilage, and intestine. The reactive epitope is stable to SDS denaturation, not associated with the active site of the enzyme, and dependent on disulfide bonds which impart secondary structure to the protein. Key Words: Alkaline phosphatase-Monoclonal bodies-Isoenzymes-Osteoblast-ELISA-Immunoblot.
anti-
Introduction Studies on the differentiation of osteogenic cells have relied, in large part, on detecting the initial expression of bone-specific molecules. The generation of a set of mono-
133
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase
134
Tissue extraction
Materials and Methods Alkaline phosphatase
and monoclonal
antibodies
Partially purified chicken intestinal alkaline phosphatase was obtained from Sigma. Purified and partially purified alkaline phosphatase from chicken epiphyseal cartilage was generously provided by Dr. Roy Wuthier, University of South Carolina, Columbia, South Carolina (Cyboron and Wuthier 1981). Antibodies SB-1, SB-2, and SB-3 have been previously characterized, and details regarding their generation are reported elsewhere (Bruder and Caplan 1989b). Immunohistochemical
staining
Unfixed frozen sections of embryonic chick limbs were immunostained as previously described (Bruder and Caplan 1989b). Culture supernatant was used undiluted, while ascites fluid was diluted I:100 in PBS with 1% BSA (BPBS). Following the primary incubation, slides were rinsed three times with B-PBS and then incubated for 1 h as previously described with fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin (FITC-GAMIg) secondary antibody (Organon Teknika Corp.) diluted 1:lOO in B-PBS. Control experiments consisted of tissue sections that were incubated with non-immune mouse serum, culture supernatant from SP2/0 Ag-14 cells, or B-PBS, followed by FITC-GAMIg secondary antibody as described above. None of these control tissue sections displayed a selective immunofluorescent staining pattern. Alkaline phosphatase enzyme activity was histochemically visualized in unfixed frozen sections as outline in Sigma Kit No. 85. Enzyme linked immunosorbent
assays
Hybridoma culture supematant or ascites fluid from antibodies SB-1, SB-2, and SB-3 were separately used to coat wells of vinyl microtiter plates (GIBCO) for these modified solid-phase assays. For titration experiments, ascites samples were serially diluted in B-PBS, and 50 uL of the appropriate dilutions were left in wells for 16 hours at 4°C. The wells were washed three times with B-PBS and then blocked by completely tilling each well with B-PBS and incubating for 1 h at 25°C. Intestinal, purified, or partially purified alkaline phosphatase was dissolved in B-PBS, and 50 p,L were incubated in the appropriate wells for 1 h at 25°C. For some experiments, fractions from the molecular sieve chromatographic separations (see below) were applied rather than the aforementioned samples of commercial or purified alkaline phosphatase. After rinsing the wells three times with B-PBS, alkaline phosphatase substrate, composed of 3.5 mM p-nitrophenylphosphate, 50 mM glycine, 1 mM MgCl*, pH 10.5, was added to each well. The product formed was read at an absorbance of 405 nm with a microtiter plate reader. The above ELISA configuration was developed because primary coating of the microtiter plates with alkaline phosphatase proved inconsistent. Therefore, the antibody was first bound to the plate and then exposed to the suspected antigen. Because the suspected antigen itself is enzymatically active, substrate was then supplied to directly test for the presence of antigen bound to the antibody. Appropriate negative controls were performed with every series of experiments.
One dozen 17-day chick embryos were removed from their shells and decapitated, and the heads immediately placed in 4°C Tyrode’s salt solution containing protease inhibitors (0.1 M 6-aminohexanoic acid, 0.005 M benzamide, 0.01 M ethylenediaminetetraacetic acid, and 0.1 mM phenylmethylsulfonyl fluoride). The calvarial bones were then dissected, placed in 1 mL of fresh Tyrode’s containing inhibitors and 0.1% ‘Biton X-100, homogenized with a Polytron (Brinkman) homogenizer for 1 min, and sonicated for 1 min with 15-s bursts. The resulting homogenate was placed in an Eppendorf tube and centrifuged at 8800 x g for 5 min to remove the insoluble material. The supernatant was separated into small aliquots, frozen, and stored at - 20°C until further use. All tissue extraction steps were performed at 4°C. Protein content was determined by the method of Bradford (1976), and standards were prepared with BSA (Sigma). Electrophoretic
separation of proteins
Sodium dodecyl sulfate (SDS)-polyacrylamide slab gels of 5-20% gradient were prepared and run as described elsewhere (Laemmli 1970). In all cases, a 5% stacking gel was formed on top of the analytical gel. For nonreduced samples, the sample buffer contained 10% glycerol, 2% SDS, and 0.01% bromophenol blue. For reduced samples, freshly diluted 5% B-mercaptoethanol was added to the above buffer. Selected samples were also boiled for 3 min immediately before loading. In some experiments, nondenaturing continuous 10% slab gels were prepared by substituting Triton X-100 for SDS in the polyacrylamide gel. A 5% stacking gel was also formed on top of the analytical gel. Electrophoresis buffer for these native gels consisted of 50 mM Bis and 192 mM glycine, pH 8.3. Sample buffer for these gels contained 10% glycerol and 0.01% bromophenol blue. Immediately after electrophoresis, gels were fixed in 40% methanol, 10% acetic acid, and subsequently stained with Coomassie Blue. For demonstration of alkaline phosphatase activity within the gels, immediately after electrophoresis the gels were soaked in 37°C 50 mM glycine, 1 mM MgCl,, pH 10.5, for 30 min. Gels were then soaked in 37°C alkaline phosphatase substrate as outlined in Sigma Kit No. 85 and fixed in 40% methanol, 10% acetic acid. Western immunoblot analysis
Samples separated by gel electrophoresis were electrophoretically transferred to Immobilon PVDF Transfer Membrane (Millipore Corp.) with a Hoefer Transphor electrophoresis apparatus according to Millipore technical protocol TP-006, except methanol was omitted from the transfer buffer. PVDF membranes were cut into strips, blocked with 3% BSA in Tiis-buffered saline (TBS), 0.05% ‘Iween-20 overnight at 4°C. Some strips were not blocked, but immediately incubated with alkaline phosphatase substrate as outlined in Sigma Kit No. 85. BSA-blocked strips were then washed with TBS-Tween and incubated with hybridoma culture supematant diluted 1: 10 in 3% BSA, TBSTween for 30 min on a shaker plate at 25°C. Subsequent immunoblotting steps, including preparation of fresh diaminobenzidine substrate, were performed according to the Vectastain ABC Kit instructions (Vector Labs).
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase Molecular sieve chromatography
A Sepharose CL-6B gel filtration column (50 x 1.5 cm) was prepared and eluted with 1.0 M NaCl, 50 mM Tris, pH 8.0. Effluent fractions of 0.6 mL were collected and analyzed for protein content by optical density measurements at 280 nm, alkaline phosphatase enzyme activity, and specific binding to SB-1 in the ELISA described above. Alkaline phosphatase
enzyme activity
Fractions eluted from the Sepharose CLdB column were tested for alkaline phosphatase activity by adding small aliquots to tubes containing p-nitrophenylphosphate substrate buffer as prepared above. Triplicate samples were incubated at 37°C for 1.5min, and the product formed was read at an absorbance of 405 nM on a Shimadzu dual beam spectrophotometer.
Results Immunohistochemical
staining
A detailed description of the immunohistochemical staining by antibody SB-1 during first bone formation of the embryonic chick tibia has already been presented (Bruder and Caplan 1989a, 1989b). In brief, antibody SB-1 stains the surfaces of osteoblasts which comprise the inner layer of the bilayered periosteum. In serial sections of bone, liver, kidney, and intestine, an identical family of cells is stained by antibody SB-1 and the histochemical stain for alkaline phosphatase (Bruder and Caplan 1989b). No reactivity was observed when cryosections from rat, rabbit, or man were immunostained. Figure 1 illustrates the immunostaining pattern of SB-1 in a longitudinal section of an embryonic day 10 tibia. The surfaces of cuboidal cells within the inner layer of the periosteum are immunoreactive, but cells entrapped within bone matrix are neither stained by SB-1 nor the histochemical stain for alkaline phosphatase. Occasional flattened chondrocytes along the periphery of the cartilage core are similarly reactive (Pechak et al. 1986; Bruder and Caplan 1989b). Antibodies SB-2 and SB-3 each stain a different family of cells than
135
those stained by SB-1 (Bruder and Caplan 1989a, 1989b). For this reason, SB-2 and SB-3 serve as negative controls for several of the experiments in this study. ELZSA binding titration curves
The binding of several alkaline phosphatase preparations to immobilized aliquots of serially diluted SB-1 ascites samples was evaluated in the solid-phase assay described in Materials and Methods. In addition, antibodies SB-2 and SB-3 were similarly coated on wells of microtiter plates at serial IO-fold dilutions. When commercially available alkaline phosphatase was used as the antigen (and indicator) in this assay, significant binding by SB-1 was observed with ascites fluid dilutions as great as 1: 10s. The titration curve for this antigen:antibody interaction is illustrated in Fig. 2A. Control experiments performed with antibodies SB-2 and SB-3 demonstrate that alkaline phosphatase is not bound by these antibodies. A similar series of experiments was performed with both purified and partially purified alkaline phosphatase from chicken epiphyseal cartilage (Cyboron and Wuthier 1981). The titration curve for SB-1 binding with partially purified material (Fig. 2B) indicates that, like intestinal alkaline phosphatase, binding is observed with ascites fluid dilutions as great as 1:loS. Purified alkaline phosphatase samples from epiphyseal cartilage similarly were bound by SB-l-coated microtiter plates; however, the limited quantity of purified material precluded the preparation of detailed titration curves. Control experiments performed with antibodies SB-2 and SB-3 again demonstrate that alkaline phosphatase is not nonspecifically bound to antibodies in this solid-phase assay. Molecular sieve chromatography
of intestinal
alkaline phosphatase
To further characterize the binding of SB-1 to the commercial intestinal alkaline phosphatase preparation, a sample of the commercial preparation was applied to a Sepharose CL-6B column in order to separate the enzyme from contaminating components. Figure 3A compares the protein content of the eMuent fractions with the alkaline phospha-
Fig. 1. Longitudinal frozen section of the tibia of a lo-day-old chick embryo stained with monoclonal antibody SB-1 followed by FITCGAMIgG and viewed by fluorescent (A) or phase (B) microscopy. Stacked cells (st) in the outer layer of the periosteum are not reactive, but all cells of the inner periosteal layer react with SB-1 at their surfaces. An immunoreactive flattened chondrocyte (arrowhead) can be seen at the periphery of the cartilaginous core. Note that neither bone matrix (M), nor cells trapped within bone matrix (osteocytes) are immunoreactive. V, Blood vessel. (X 310)
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase
136
1
2
3
4
5
6
7
3.0
0.6
2.0
0.4
1.0
0.2
8
Log dilution 0.0
0.0 10
20
30
50
Fraction #
? 2.0
40
60
70 T vt
vo
0.6
0.2
1.0
0.0 1
2
3
4
5
6
7
0.4
8
Log dilution Fig. 2. Titration curves for antibody reactivity with various forms of alkaline phosphatase. Log dilutions of ascites from antibodies SB-1 (O), SB-2 (0) and SB-3 (+) were coated onto microtiter plates and reacted with intestinal alkaline phosphatase (A) or partially purified epiphyseal cartilage alkaline phosphatase (B). Binding was assayed by reading the microtiter plates at 405 nm after applying p-nitrophenylphosphate substrate solution as described in Materials and Methods.
0.1
0.2
0.0 10
0.0 20
30 ?
tase enzyme activity of each fraction. These data reveal that enzyme activity is concentrated within 10% of the effluent, fractions 35 through 38, while protein is distributed from the V, to the V, of the column. When the effluent fractions were tested for their ability to bind to SB-l-
coated microtiter plates in the modified ELISA, the fractions which bound were those which contained the enriched alkaline phosphatase enzyme activity (Fig. 3B). The binding of proteins which eluted either before or after the peak of enzyme activity was not detected. Molecular weight determination Both freshly extracted alkaline phosphatase from embryonic calvaria and commercially prepared alkaline phosphatase from chicken intestine were subjected to electrophoresis on a 5-20% gradient SDS-polyacrylamide gel (Fig. 4). Coomassie Blue staining indicates that the calvarial extract contains a heterogeneous mixture of protein which ranges from over 200 kD to 14 kD in molecular weight (Lane A). By contrast, the commercial preparation of intestinal alkaline phosphatase is highly enriched in one protein which migrates at a position corresponding to approximately 185 kD (Lane B), as determined by calculated Rf values. Reduction of the intestinal preparation does not alter its electrophoretic mobility (Lane C), although boiling of a reduced sample results in the elimination of the 185 kD species and the emergence of a doublet protein band with lower molecular weights (Lane D). The major protein band is present at a position which corresponds to approximately 110 kD, and the minor band migrates to a position which corresponds to approximately 97 kD. Identically prepared gels were also stained for alkaline
40
50
Fraction #
vo
60
70 3L vt
Fig. 3. (A) Elution profile of a Sepharose CLdB chromatographic separation of commercially prepared intestinal alkaline phosphatase. Alkaline phosphatase activity (m) was measured at 485 nm and protein content was determined (0) for each fraction. (B) Binding of each effluent fraction to antibody-coated microtiter plates was determined (0), and plotted against the alkaline phosphatase activity (m) from A. In panel B, the left axis refers to the solid-phase immunoassay and the right axis refers to the alkaline phosphatase activity assay. V, = void volume, V, = total volume. phosphatase enzyme activity (Fig. 5). Analysis of this gel indicates that the bone and intestinal isoenzymes of alkaline phosphatase are of significantly different molecular weights. Enzyme activity in both reduced (Lane A) and unreduced (Lane B) samples of calvarial bone extract is present at a position corresponding to approximately 155 kD, although reduction results in compromised enzyme activity evidenced by the lower intensity of staining in Lane A compared to Lane B. By contrast, both reduced (Lane D) and unreduced (Lane C) intestinal alkaline phosphatase migrates to a position corresponding to approximately 185 kD. Again, the intensity of the alkaline phosphatase enzymatic reaction in the reduced sample is less than that of the unreduced sample. Boiling of the alkaline phosphatase samples results in the elimination of enzyme activity (data not shown).
Western immunoblot
analvsis
Crude extract from day-17 embryonic chick calvaria was electrophoresed on a 5-20% SDS-polyacrylamide gel as described above. The material within the gel was then
137
S. P. Bruder and A. I. Caplan: Monoclonal antibody against alkaline phosphatase
ABCD
ABCD
A BC
ABC +
200* 11w 97,
* >‘ f,!; f$
200* L. 3
66*
116~ 97*
z
66, 45*
31-
31*
22*
22*
14*
14*
4
'. .r ',,'-T
5
6
7
Fig. 4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of calvarial extract and intestinal alkaline phosphatase stained with Coomassie Blue. Samples and amounts loaded were: Lane A, reduced cabarial extract, 40 pg of protein; Lane B, unreduced intestinal alkaline phosphatase, 20 pg of protein; Lane C, reduced intestinal alkaline phosphatase, 20 pg of protein; Lane D, boiled and reduced intestinal alkaline phosphatase, 20 pg of protein. Molecular weight markers are Bio-Rad high- and low-molecular weight standards. Fig. 5. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of alkaline phosphatase samples stained to demonstrate enzyme activity. Samples and amounts loaded were: Lane A, reduced calvarial extract, 40 ug of protein; Lane B, unreduced calvarial extract, 40 ug of protein; Lane C, unreduced intestinal alkaline phosphatase, 20 ug of protein; Lane D, reduced intestinal alkaline phosphatase, 20 kg of protein. Molecular weight markers are Bio-Rad high- and low-molecular weight standards. The single-head arrow denotes the 185 kD alkaline phosphatase isoenzyme from intestine. The double-head arrow denotes the 155 kD alkaline phosphatase isoenzyme from bone. Fig. 6. Immunoblot display of SB-1 reactivity in a freshly prepared calvarial extract. Forty ug of reduced protein was applied to Lane A, while Lanes B and C were loaded with forty ug of unreduced extract. Samples were electrophoresed on a 5-20% SDS-polyacrylamide gel, transferred to Immobilon PVDF, and visualized by the following techniques: Lane A, alkaline phosphatase histochemical stain; Lane B, antibody SB-1 followed by secondary antibody and substrate; Lane C, secondary antibody and substrate without the SB-I antibody. The arrowhead denotes the single protein band which is both immunoreactive with antibody SB-1 (Lane B) and stained by alkaline phosphatase histochemicai substrate (Lane A). Fig. 7. Immunoblot display of SB-1 reactivity in an intestinal alkaline phosphatase preparation. tienty pg of unreduced protein were applied to each lane. Samples were electrophoresed on a continuous 10% ‘Riton X-100~polyacrylamide gel, transferred to ImmobiIon PVDF, and visualized by the following techniques: Lane A, alkaline phosphatase histochemical stain; Lane B, antibody SB-1 followed by secondary antibody and substrate; Lane C, secondary antibody and substrate without the SB-1 antibody. The arrowhead refers to the sinale nrotein band which is both immunoreactive with antibody SB-1 (Lane B) and stained by alkaline phosphatase histochemicai
substrate (Lane A). transferred to an Immobilon PVDF membrane and reacted with antibody SB-1 (Fig. 6). Lane A, stained for alkaline phosphatase activity, indicates where the enzymatically active protein is localized on the membrane under reducing conditions. When this membrane was reacted with antibody SB-1 followed by the secondary antibody and substrate, predominant immunostaining was observed (Lane B) at a position corresponding to the band possessing alkaline phosphatase enzyme activity in Lane A. Although several other bands are observed at intermediate and low molecular weights in Lane B, these bands are identically reactive in Lane C, which was exposed only to the secondary antibody and substrate. The immunoreactivity of SB-1 with intestinal alkaline phosphatase was additionally assessed by Western immunoblotting. Samples were electrophoresed on a continuous 10% Triton X-lOO-polyacrylamide gel, electrotransferred onto Immobilon PVDF, and immunoreacted as described in Materials and Methods. Alkaline phosphatase activity, under nonreducing (Lane A) conditions, is observed as a single band in Fig. 7. When the nonreduced sample was reacted with antibody SB-1, and subsequently exposed to secondary antibody and substrate, a single band was visualized (Lane B) corresponding to the region of alkaline phosphatase activity present in Lane A. The immunoblot control with only secondary antibody and substrate produced no specific immunoreactive product (Lane C).
Discussion The results presented above show that SB-I, a monoclonal antibody generated against the surface of avian osteogenic cells, recognizes an epitope on alkaline phosphatase which is present on the bone, liver, kidney, cartilage, and intestinal isoenzymes. Previous immunohistochemical data illustrate the binding of SB-1 to the above tissues in a pattern which is identical to that obtained from alkaline phosphatase enzyme histochemistry of serial sections (Bruder and Caplan 1989b). While these data alone do not confirm that SB-1 is directed against alkaline phosphatase, they suggest that this is a tenable possibility. The biochemical and immunochemical evidence presented above strongly support the identity of the molecules recognized by antibody SB-1 as alkaline phosphatase isoenzymes. The modified ELISA designed for this study demonstrates several important points about the reactivity of SB-1 with alkaline phosphatase. First, the shape of the titration curve indicates that antibody binding is saturable (Fig. 2). The binding signal strength is maximal at an ascites dilution of l:lO“, with no further increase in antibody binding at even lower ascites dilutions. Furthermore, the high dilution (1:105) of ascites fluid used to obtain a significant binding signal in this solid-phase assay indicates that such binding is of a high affinity. Together, these data suggest that the binding of SB-1 to alkaline phosphatase is
138
S. P. Bruder and A. 1. Caplan: Monoclonal antibody against alkaline phosphatase
of a specific nature. The absence of binding in control studies with either no antibody or monoclonal antibodies SB-2 and SB-3 eliminates the possibility that alkaline phosphatase nonspecifically binds to either the plate itself or to mouse immunoglobulins. Partial purification of intestinal alkaline phosphatase on a Sepharose CL-6B column results in the co-elution of protein having alkaline phosphatase enzyme activity and antibody-binding reactivity (Fig. 3). This observation provides evidence that SB-1 recognizes intestinal alkaline phosphatase rather than an impurity in the crude preparation which is separated during molecular sieve chromatography. Thus, the data from this series of experiments indicate that alkaline phosphatase isoenzymes from both developing limb and intestine are specifically recognized by antibody SB- 1. Standard immunoblot techniques were used with chick calvarial extracts which had been electrophoresed on SDS-polyacrylamide gels. When the transfer membrane was exposed to antibody SB- 1, one predominant band was visualized at a position which matches the single band of alkaline phosphatase activity at an apparent molecular weight of 155 kD (Fig. 6). Although additional bands were present at lower molecular weights, the same bands were observed in the secondary antibody control experiment. The reasons for the secondary antibody’s binding with several species in the extract is not clear, but experience with this system has shown that several preparations of secondary antibody from a variety of sources nonspecifically react with different extract components. While the molecular weight of chick bone alkaline phosphatase has not been rigorously studied and reported, the apparent subunit molecular mass of rat bone alkaline phosphatase has been shown to be 82 kD (Nair et al. 1987). Our estimate of 155 kD for chick bone alkaline phosphatase is consistent with the dimeric structure of yolk alkaline phosphatase, the molecular weight of which is 150 kD (Debruyne and Stockx 1979). It is possible that chick bone alkaline phosphatase also has a dimeric structure of subunits whose molecular weight is similar to that reported for rat bone. Additional characterization of the chick bone alkaline phosphatase is required to confirm this postulate. Additional immunoblot experiments were performed with intestinal alkaline phosphatase from chick. These experiments reveal that a single band of SB-1 reactivity is present at a position corresponding to the band of alkaline phosphatase enzyme activity within the gel. This immunoblot experiment was performed on a nondenaturing gel in order to demonstrate SB-I reactivity while minimizing nonspecific background staining. In the case of intestinal alkaline phosphatase, the single band of enzyme activity has an apparent molecular weight of 185 kD. This estimate is similar to that reported for duodenal alkaline phosphatase from 4-day-old chicks (Chang and Moog 1972). SDSpolyacrylamide gel electrophoresis of boiled and reduced samples suggest that the native protein is composed of two subunits with molecular weights of 97 kD and 100 kD. Taken together, the various lines of experimentation indicate that SB-1 is specifically reactive with alkaline phosphatase isoenzymes present in bone, liver, kidney, cartilage, and intestine. The predominant cell surface immunostaining by SB-1 in cryosections of developing bone is consistent with the plasma membrane-bound localization of alkaline phosphatase on osteoblasts (Rodan and Rodan 1983). The data further indicate that the epitope recognized by SB-1 is not a part of the active site based on the observation that in the modified ELISA, alkaline phosphatase
which is bound to SB-1 on the microtitration plate retains its enzymatic activity. If the reactive epitope was associated with the active site of the enzyme, one would expect interference in the catalytic process. In another series of experiments, the addition of SB-1 to solutions of alkaline phosphatase was found to not hinder the cleavage of the p-nitrophenylphosphate substrate (unpublished data). While the reactive epitope is stable to SDS denaturation (Fig. 6), reduction of the disulfide linkages results in abolition of antibody binding in both the immunoblot assay and a dot-blot assay (data not shown). For this reason, the reactive epitope seems to be dependent on disulfide bondmediated secondary structure of the protein. Although further work is required to characterize the exact details of the epitope, the data presented above provide strong evidence that SB- 1 specifically recognizes alkaline phosphatase isoenzymes from a variety of chick tissues.
The authors would like to thank Mr. James Holecek, Drs. David Carrino, Louis Gerstenfeld and Stephen Haynesworth for their valuable assistance during the course of this study. SPB was supported by a Medical Scientist Training Grant GMO-7250-13. This work was supported by grants from the National Institutes of Health.
Acknowledgments:
References Bradford, M. A rapid and sensitive method for the quantitation of p-gram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem.
72~248-254;
1976.
Bruder. S. P.; Caplan, A. I. Discrete stages within the osteogenic lineage are revealed by alterations in the cell surface architecture of embryonic bone cells. Glimcher, M. J.; Lian, J. B., eds. The chemistry and biology of mineralized tissues. New York: Gordon and Breach; 1989a:pp. 73-79. Bruder. S. P.; Caplan, A. 1. First bone formation and the dissection of an osteogenic lineage in the embryonic chick tibia is revealed by monoclonal antibodies against osteoblasts. Bone 10:359-375; 1989b. Bruder, S. P.; Caplan, A. I. Terminal differentiation of osteogenic cells in the embryonic chick tibia is revealed by a monoclonal antibody against osteocytes. Bone (in press). Chang, C.; Moog, F. Alkaline phosphatases of the chicken duodenum. I. Isolation and partial characterization of the multiple forms of duodenal phosphatase in pre- and post-hatching stages. Biochim. Biophys. Acra. 258:154-165; 1972. Cyboron, G. W.; Wuthier, R. E. Purification and initial characterization of intrinsic membrane-bound alkaline phosphatase from chicken epiphyseal cartilage. JBC 256(14):7262-7268; 1981. Debruyne, I.; Stockx, J. Alkaline phosphatases from hen’s egg yolk, liver, blood plasma and intestinal mucosa: aftinity chromatographic puritication and comparison. ht. J. Biochem. 10:981-988: 1979. Goldstein, D. S.; Rogers, C. E.; Harris, H. Expression of alkaline phosphatase loci in mammalian tissues. Proc. Natl. Acad. Sci. USA 77:2860-2957;
1980.
Hass. P. E.; Wada, H. G.; Horman, M. M.; Sussman, H. H. AIkaIine phosphatase of mouse teratoma stem cells: immunochemical and structural evidence for its identity as a somatic gene product. Proc. Nat/. Aced. Sci. WA 76:1164-1168; 1979. Kuan, S. S.; Martin, W. G.; Patrick, H. Alkaline phosphatases of the chick. Partial characterization of the tissue isozymes. Proc. Sot. Exp. Biol. Med. 122:172-177; 1966. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227680-685; 1970. Moak, G.; Harris, H. Lack of homology between dog and human placental alkaline phosphatases. Proc. Nat/. Acnd. Sci. USA 761948-1951; 1979. Nair, B. C.; Majeska, R. J.; Rodan, G. Rat alkaline phosphatase. I. Purifi-
S. P. Bruder
and A. I. Caplan:
Monoclonal
antibody
against
alkaline
cation and characterization of the enzyme from osteosarcoma: generation of monoclonal and polyclonal antibodies. Arch. Biochem. Biophys. 254~18-27; 1987. Pechak, D. G.; Kujawa, M. K.; Caplan, A. I. Morphological and histochemical events during first bone formation in embryonic chick limbs. Bone 7:441-458; 1986. Rodan, G.; Rodan, S. Expression of the osteoblast phenotype. Peck, W. A., ed. Bone and mineral research. New York: Elsevier; 1983:~~. 244-285. (Vol. 2). Vockley, J.; Harris, H. Purification of human adult and foetal intestinal alkaline phosphatases by monoclonal antibody immunoaffinity chromatography. B&hem. J. 217:535-541; 1984. Wilcox, F. H.; ‘lbylor, B. A. Genetics of the Akp-2 locus for alkaline phosphatase of liver, kidney, bone, and placenta in the mouse. J. Hered. 72:387-390;
1981.
phosphatase
139
Wilcox, F. H.; Hirschhorn, L.; ‘lhylor, B. A.: Womack, J. E.; Roderick, T. H. Genetic variation in alkaline phosphatase of the house mouse (Mus musculus) with emphasis on a manganese-requiring isozyme. Biothem. Gem?. 17: 1093- 1107; 1979. Wilkinson, J. Phosphatase isoenzymes. Wilkinson, London: Chapman & Hall; 1970:~~. 69-102.
J., ed. Isoenomes.
Received: June 30, 1989 Revised: November 15, 1989 Accepted: November 16, 1989
