- Email: [email protected]
Differential effects of phospholipids on skeletal alkaline phosphatase activity in extracts, in situ and in circulation
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 2.21, No. 2, March, pp. 477-488, 1983
Differential
Effects of Phospholipids on Skeletal Alkaline Phosphatase Activity in Extracts, in Situ and in Circulation JOHN
Departments
of Biodemtitry
R. FARLEY’ and Medicine, Received
URSULA
AND bma
Linda
September
University,
M. JORCH Loma
Linda,
Cal&&a
92550
21, 1982
Human skeletal alkaline phosphatase (ALP) purified from human bone was subject to competitive inhibitions by phospholipids including cephalins, lecithins, and phosphatidylinositol. Ki values ranged from 0.7 to 1.5 mM, at pH 9.5. As previously shown, the enzyme was subject to uncompetitive inhibition by imidazole. The inhibitory phospholipids potentiated this effect, and altered the nature of the imidazole inhibition, from uncompetitive to mixed type, suggesting that imidazole was bound more efficiently to the enzyme-phospholipid complex than to the enzyme-substrate complex. No interactions were observed between phospholipids and other uncompetitive inhibitors of ALP. The skeletal ALP activity of cultured chick calvarial cells was assayed both in situ and in extracts. Like the extracted human isoenzyme, the extracted chick ALP was subject to competitive inhibition by cephalin (Ki = 0.3 mM at pH 9.3) and an inhibitory interaction between cephalin and imidazole, but the same isoenzyme showed neither effect in situ. The value of K ,QN~P at pH 9.5 for chick skeletal ALP was 1.5 mM in extracts and 7.1 mM in situ. When embryonic chick bones were cultured in vitro, skeletal ALP activity was released into the serum-free medium. Unlike the same isoenzyme extracted from the bones, the ALP activity in the medium was not inhibited by cephalin and showed no inhibitory interaction between cephalin and imidazole. Similarly, human serum ALP activities were not as sensitive to phospholipid inhibition as the same isoenzymes extracted from tissues. Human skeletal ALP extracted from bone was inhibited by cephalin, but the skeletal isoenzyme in Pagetic serum was not, suggesting that the potential for phospholipid interaction was altered during or after release from osteoblast cell membranes. The observation that extracted human skeletal ALP lost its potential for inhibition by phospholipids after treatment with phospholipase C further suggests that ALP activity may be released from cells during membrane turnover.
There are two reasons why skeletal ALP2 (EC 3.1.3.1.) is thought to play a role in
bone formation (l-3). (a) In bone, skeletal ALP is concentrated in the plasma membranes of osteoblasts, and these cells are responsible for bone matrix formation. (b) In serum, the activity of skeletal ALP tends to vary with the bone formation rate (4, 5). The evidence is circumstantial. After almost 60 years of investigation the function of skeletal ALP in vivo is unknown.
1 To whom correspondence should be addressed: Research Service (151), Jerry L. Pettis Veterans Hospital, 11201 Benton Street, Loma Linda, California 92357. 2 Abbreviations used: ALP, alkaline phosphatase; PNPP, pnitrophenylphosphate; PNP, pnitrophenol; cCMP, cytidine 3’,5’-cyclic monophosphate; PE, phosphatidylethanolamine (cephalin); PC, phosphatidylcholine (lecithin); PI, phosphatidylinositol; PEA, phosphorylethanolamine; SDS, sodium dodecyl sul-
fate; CI, competitive inhibition; UCI, uncompetitive inhibition; MT, mixed-type inhibition; Knsppp, Kiaw, etc., apparent value of kinetic constant. 477
0003-9861/83/040477-12$03.00/O Copyright All rights
Q 1983 by Academic Press. Inc. of reproduction in any form reserved.
478
FARLEY
We have attempted to determine skeletal ALP function in wivo by characterizing the binding sites that the isoenzyme possesses, reasoning that the most efficient substrates and inhibitors (i.e., the most tightly bound) should most closely resemble the effecters of skeletal ALP in wivo. We had previously shown that human skeletal ALP possesses a primary binding site for Pi or phosphoryl substrates, as well as secondary sites that will accommodate Tris (or imidazole, etc.) and theophylline (or cCMP, etc.) (6). These secondary sites are only expressed when the primary site has been filled. (For a general review of binding sites on ALP isoenzymes see Ref. (7). We now report (a) the results of multiple inhibition analyses to further characterize the relations between these binding sites; (b) the effects of lipid-enzyme interactions, both in extracts and in situ, to modulate ALP activity and to affect the binding sites; and (c) the differential effects of phospholipids on ALP activity in tissue extracts and in circulation. MATERIALS
AND METHODS
Materials. PNPP, lecithin, cephalin, imidazole, levamisole, placental ALP and phosphatidylinositol, phospholipase C, and phospholipase A were obtained from Sigma Chemical Company. Lysolecithin (palmitoyl), lecithin (dimyristoyl), cephalin (dipalmitoyl), and lecithin (fl-dicaproyl) were purchased from Calbiochem/Mannheim. Collagenase and BGJb culture medium were obtained from GIBCO. Tissue culture plates (35-mm dishes and multiwells) were purchased from Falcon. Assay for ALP activity. The standard assay mixture (used during purification of the isoenzyme) consisted of 0.15 M carbonate buffer, pH 10.3.1 mM MgClz, 10 mM PNPP, and ALP in a total volume of 1.5 ml at 22°C (ambient room temperature). The reaction was initiated by the addition of the enzyme (or the substrate) and monitored by the change in absorbance at 410 nm during a 15- to 120-min incubation. Controls, without ALP, were included with each assay to correct for the nonspecific hydrolysis of PNPP. Variations in reaction conditions (ie., pH, substrate concentration, addition of inhibitors, etc.) are detailed in the text. Partial purification of ALP from human bone. This method has been previously described in detail (6). Briefly, human femoral heads were obtained at hip replacement surgery, frozen, crushed, rinsed (to re-
AND JORCH move contaminating marrow and serum), and extracted with 20% butanol (v/v) in carbonate buffer (25 mM, pH 8.3) for 24 h at 4°C. ALP activity was obtained in the aqueous phase, and after dialysis against the same carbonate buffer, the activity was further purified by ammonium sulfate precipitation, DEAE-cellulose column chromatography, and filtration through a column of agarose-linked Cibacron blue (Bio-Rad). The skeletal ALP thus prepared had a specific activity of about 0.3 Wmg in our standard assay. ALP activity in human serum. Serum samples were obtained from two patients with active Paget’s disease of bone (total serum ALP > 390 U/liter for each) and two patients with obstructive jaundice (total serum ALP > 356 U/liter for each). Isoenzyme analysis (8) confirmed that the Pagetic and jaundiced sera contained, almost exclusively, skeletal and hepatic ALP activities, respectively. Aliquots of the samples were extracted with 26% butanol (v/v) in 25 rnrd carbonate buffer (pH 8.3) for 24 h at 4°C and then dialyzed against the same carbonate buffer for kinetic analyses and comparison with their unextracted counterparts. Additional aliquota were incubated in imidazole buffer (10 mM, pH 7.4) containing 2 mM CaClz and 0.15 M NaCl with various amounts of phospholipase C for 4 h at 4’C. The reaction mixtures were then dialyzed against 25 mM carbonate buffer (pH 8.0) for kinetic analyses. Isolation and &re of ernbrgontk chick calvarial cells. These methods have been described in detail (9). Briefly, cells (predominately osteoblasta and osteoblast progenitors) were released from the calvariae of lbday embryonic chicks by sequential collagenage digestion, plated at an initial density of 250/ mmz in 16-mm-diameter multiwell plates, and cultured in serum-free BGJb medium in an atmosphere of 5% COz and at 37°C. Cells were also cultured in 35-mm-diameter culture dishes at the same initial density and under identical conditions. After an 18h incubation, the medium was changed, and the cells were either assayed immediately or cultured for an additional l-3 days before use. Specific conditions for each experiment are given in the text. Partial pur&adon of ALP from chick bone cdls. Embryonic chick calvarial cells, cultured as described above in 35-mm-diameter culture dishes, were rinsed and extracted with 20% butanol (v/v) in 25 rnrd carbonate buffer (pH 8.3) for 4-8 h at 4°C. The extract was dialyzed against the same carbonate buffer and the insoluble material was removed by centrifugation. Activity was determined by variations of the standard assay (i.e., changes in pH, substrate concentration, addition of inhibitors and effecters, etc.) as detailed in the text. No kinetic differences were discernible between ALP prepared from cells cultured overnight and those cultured for 3 additional days.
PHOSPHOLIPID
INHIBITION
OF SKELETAL
Assay for chick bone Ceu ALP in situ. ALP activity was measured in situ using embryonic chick calvarial cells cultured in 16-mm-diameter multiwells as described above. The culture medium was removed and the adherent cells were rinsed with 0.5 ml of 25 mM carbonate buffer, pH 9.5, containing 0.14 Y NaCl. The ALP assay mixture (usually at pH 9.5) was then added to the wells in a total volume of 1.5 ml and the supernatant solutions were transferred (by Pasteur pipet) sequentially to cuvettes in the spectrophotometer for determinations of absorbance at 404 nm, at the beginning and at the end of the incubation period (usually 30 to 90 min). Most of the cells remain attached to the tissue culture wells during this procedure. Repeat assays using the same cells (rinsed between assays) showed that more than 80% of the original activity remained after an initial 60-min incubation. Organ culture of embryonic chick bae. Frontal and parietal bones were asceptically dissected from 13day embryonic chicks, rinsed in serum-free BGJb medium, and cultured individually on lens paper rafts in 24-place multiwell tissue culture plates in a volume of 0.5 ml of serum-free BGJ,, medium. (A total of 48 bones were used for these experiments.) The medium was changed at 36-h intervals, and conditioned medium was pooled and frozen for subsequent determination of ALP activity. After ‘72 h, ALP activity was extracted from the bones with 20% butanol (v/ v) in 25 rnrd CO8 buffer, pH 8.3. The butanol extracts
TABLE
ALKALINE
479
PHOSPHATASE
were pooled, and both the butanol extracts and the pooled, conditioned media were lyophilized and resuspended in one-tenth their original volumes, before dialysis against 25 mM COIl buffer, pH 8.3, and subsequent kinetic analyses. Additional methods. Protein concentrations were determined by the dye-binding method of Bradford (10). Spectral determinations were made on a Beckman DU-Gilford spectrophotometer. All kinetic experiments were performed with duplicates of individual data points; average values are reported. The kinetic nomenclature of Segel (11) is used throughout. RESULTS
Characterization of the cCMP-binding site. Previous studies had shown that human skeletal ALP was subject to uncompetitive inhibition by cCMP, Ki = 3.2 m&q from pH 8.1 to pH 9.6 (6). Data summarized in Table I compare inhibitions by cCMP and other cytidine nucleotides. (It is interesting to note that inhibition by cCMP is uncompetitive with respect to PNPP, but inhibitions by CMP and CTP are competitive.) Since ALP is a membrane-bound enzyme and since cytidine I
HUMAN SKELETAL ALP: CHARACTERIZATION OF THE cCMP-BINDING SITE Inhibitor cCMP CTP CMP Cytidine Sialic acid CMP and sialic acid CTP and sialic acid Glycerol CMP and glycerol CTP and glycerol PEA CMP and PEA CTP and PEA Cephalin CMP and cephalin
Effect” UC1 CI CI None None Noneb Noneb None Noneb None” CI None b None” CI None”
Gpp at PHI hM) at pH 9.6 16 at pH 8.8 2.0 at pH 8.8 >30 at pH 8.8” >40 at pH 8.8” 3.2
at pH 8.1 0.5 at pH 8.2 3.3
~40 at
-
pH 8.8”
30 at
pH 9.5
0.5 at
pH 8.0
0.8 at
pH 9.5 -
0.7 at pH 8.8 -
“All inhibitors and combinations were tested at pH 8.8 and at pH 10.0 (i.e., minimum pH range), with both saturating and nonsaturating concentrations of PNPP. b No additional inhibition resulting from the combination. ’ Since no inhibition was observed, we have calculated a minimum possible value for KGpp.
480
FARLEY
AND JORCH
120
/
T
,__
T
Inhibitor Ki (VHM)
none
SDS
-
-
PC 0.7mM
PE 0.65mM
PEA 30mM
-
Pi 2.5mM
FIG. 1. Relative potency of ALP inhibitors. Percentage of control activity (shown as mean with range of four determinations) for human skeletal ALP assayed in the presence of various potential inhibitors. Each was tested at a concentration of 1.0 mM with 0.3 mM PNPP and 1 mM MgClr in CO* buffer at pH 9.3. The value of Kicpp at pH 9.3 (determined by replots of initial velocity data) is also given for each inhibitory compound.
nucleotides are involved in lipid and carbohydrate metabolism in membranes, we considered the possibility that ALP might be involved in one of these reaction pathways. We therefore examined the effects of other compounds involved in these reactions (sialic acid, glycerol, phosphoethanolamine, and cephalin) on the activity of human skeletal ALP with and without added CMP and CTP. The results are shown in Table I. Inhibition of human skektal ALP by phospholipids. The inhibitory effect of PE (phosphatidylethanolamine, cephalin) on human skeletal ALP was compared with the effects of PC, PEA, Pi, and SDS. The results are shown in Fig. 1. The inhibition by PE was linear competitive with respect to PNPP; this is shown in Fig. 2. A variety of lipids were found to be similarly inhibitory; these results are summarized in Table II. These lipids form micelles at the concentrations we have used. The background optical density of the ALP assay mixtures (i.e., absorbance and light scattering at 410 nm before the phosphatase
reaction has begun) was therefore increased by addition of the lipids. This background was, however, stable over the
8 l/V
Volume
PE
$!I)
FIG. 2. Inhibition of ALP activity by cephalin. Dixon plot: reciprocal of initial velocity of human skeletal ALP (relative units) vs volume of cephalin (PE) added. (Stock PE = 11.9 rnr& reaction volume = 1.5 ml; 100 ~1 added cephalin corresponds to a concentration of 0.79 maa). Activity assayed in COll buffer, pH 9.3, with 2 rnM added MgCla and 2.5 mM (O), 0.4 rnM (m), or 0.2 mM (V) added PNPP.
PHOSPHOLIPID
INHIBITION
OF
SKELETAL
course of the reaction, whether the lipids were incubated alone or in combination with either PNPP or skeletal ALP. To ensure that we were observing lipidenzyme interactions we investigated the possibilities that phospholipids could be (a) sequestering the substrate and thereby reducing its effective concentration or (b) interfering with the absorbance of the reaction product, PNP. Regarding the second possibility, Fig. 3a demonstrates that cephalin, at inhibitory concentrations, did not interfere with the absorbance of PNP (i.e., the lines are parallel). Both lecithin and cephalin were tested over a wide range of PNP concentrations, corresponding to the range of enzyme activities assayed, and we could not detect interference. Regarding the other possible source of methodological artifact ((a) above), Fig. 3b shows the time course of PNP distribution between dialysis half-cells (Thomas Scientific, volume = 1.5 ml per half-cell) separated by dialysis membrane with a nominal i& cutoff of 12,000, where one chamber initially contained 1.2 mM cephalin (or buffer) and the other contained 0.06 mM PNP. (Separate controls established that cephalin, at the concentrations tested, did not cross the dialysis membrane in significant amounts during the time required for equilibration of PNP.) The fact that the predicted equilibrium distribution was not exceeded in the presence of the lipid indicates that cephalin does not sequester PNP, and, since PNPP has a greater charge/mass ratio at pH 9.5 than PNP, it is reasonable to assume that PNPP is not sequestered either. We have also tested the possibility that phospholipids might inhibit ALP activity by competing with the enzyme for MgZ+. We saw no difference in the kinetic characteristics of PE inhibition when the [Mgzf] was varied between 1 and 10 rnM. Eflects of phap?wlipi& on the secmdarg binding sites. Since the competitive inhibitions seen with phospholipids were not methodological artifacts and since they involved the primary, phosphoryl, binding site, we examined the interactions between phospholipids and uncompetitive inhibitors (i.e., secondary site effecters) of
ALKALINE
TABLE HUMAN
SKELETAL
Lipid SDS Cephalin (crude) Cephalin (dipalmitoyl) Lecithin (crude) Lecithin (dimyristoyl) Lecithin (dicaproyl) Lysolecithin (palmitoyl) Phosphatidyl inositol Myristic acid with 1 mM SDS
481
PHOSPHATASE
ALP:
II INHIBITION
BY LIPIDS
5w of inhibition”
KicPP @f)
None Competitive Competitive Competitive Competitive None Competitive Competitive
>20b 0.8 1.1 0.8 1.2 >5b 1.5 0.9
Competitive
0.7
D At pH 9.5, evaluated by Lineweaver-Burke Dixon plots of initial velocity data. bSince no inhibition was observed, we have culated a minimum possible value for Kinpp.
or cal-
human skeletal ALP, such as imidazole, homoarginine, and theophylline (6). Figure 4 demonstrates uncompetitive inhibition of human skeletal ALP activity by imidazole, and Fig. 5 demonstrates how PE (cephalin) can potentiate this inhibition. The replot in Fig. 5b shows that the value of K+pp for imidazole decreases from 7 mM with no added PE to 0.7 mM with 0.5 mM added PE. We saw a similar inhibitory synergism between lecithin and imidazole and between phosphatidylinositol and imidazole. No such interaction was observed between cephalin and homoarginine (K..*,nm.dam,I?= 8.0 and 7.5 mM with 0 and 1 mM added PE, respectively) or between cephalin and theophylline (Ki,thkphylline = 0.34 and 0.36 m&r with 0 and 1 mM added PE, respectively). We have already shown (in Table I) the lack of interaction between cephalin and CMP. It is interesting to note that the inhibitory effect of histamine on skeletal ALP was reduced by added cephinhibition inalin (K,iapp for histamine creased from 11.3 to 16.4 mM as the cephalin concentration was increased from 0 to 0.5 mM). Chick calvarial cell ALP activity tracts and in situ. As an approach
termining teractions
whether phospholipid-ALP might be physiologically
in exto deinsig-
FARLEY
0
AND JORCH
60 Volume Cephalin
Time
120 (~1)
(mid
FIG. 3. Potential interactions of cephalin with PNPP or PNP. (a) Test for possible interference with PNP light absorption. Absorbance at 404 nm vs volume of cephalin added (stock cephalin at 11.9 mM; 100 ~1 added cephalin gives a final concentration of 0.79 mM) in CO8 buffer, pH 9.3, with 2 mM MgClc, no added ALP, and 0 (0) or 0.02 mM (V) added PNP. (b) Test for possible sequestering of PNP by cephalin micelles. Equilibrium dialysis half-cells (volume = 1.5 ml) separated by dialysis membrane (12,000 AZ,cutoff) were filled with CO8 buffer (pH 9.3) containing 2 mM MgCI,, or cephalin in the same buffer on one side and 0.06 rnM PNP in buffer on the other. The ratio of absorbance (and light scattering) at 404 nm on the cephalin (or buffer) side to absorbance on the PNP side is shown as a function of time for cells with 1.2 rnrd added cephalin (V) and for controls (0). The dotted line indicates the theoretical limit for equilibration without interaction.
nificant we compared the ALP activity of cultured embryonic chick bone cells in situ with the activity of the same isoenzyme in butanol extracts of cells. Figure 6 shows that the value of Km,pNppwas about 4.7fold lower in the extracts than in the cells, at pH 9.5. Figure 7 shows the effect of pH on Vm and KQNPP for ALP in situ and in extracts. The data in Table III demonstrate that the activity of skeletal ALP is
identical in cultured chick calvarial cells and in extracts of those cells with respect to all tested inhibitors except for cephalin and the cephalin/imidazole combination. When cultured chick calvarial cell ALP was assayed in situ at pH 9.5 and the cells were rinsed and then reassayed, we found that 8590% of the initial activity remained. However, this value was reduced to 53-62% when cephalin was added to the
PHOSPHOLIPID
INHIBITION
0
OF
SKELETAL
2
lI[pFILP]
mM“
FIG. 4. Uncompetitive inhibition of ALP by azole. Reciprocal of human skeletal ALP activity tial velocity, relative units) vs reciprocal of concentration. Activity was measured at pH CO8 buffer with 2 mM added MgClz and 0 (O), (H), or 10 mrd (0) added imidazole.
imid(iniPNPP 9.9 in 5 m&i
initial assay mixture (0.5 and 1 InM). There was no detectable change in cell number during this latter incubation.
ALKALINE
PHOSPHATASE
Human ALP ksoewyme.s-inhibition by phospholipids. Since skeletal ALP obtained from human bone was subject to competitive inhibition by phospholipids, we wanted to determine whether a similar inhibition could be observed (a) with skeletal ALP in serum or (b) with other ALP isoenzymes, either purified from organs or in serum. The results of these studies are summarized in Table IV. A commercial preparation of placental ALP, which had been purified by butanol extraction, was inhibited by cephalin (linear, competitive: Ki = 1.1 mM at pH 9.5). In contrast, the skeletal ALP activity in Pagetic serum was not inhibited by cephalin at pH 9.5, even using 0.25 mM PNPP (Km.rNPP for this isoenzyme was 0.5 mM at pH 9.5). Butanol extraction and subsequent dialysis had no detectable effect on the value of Km.rNPP
25
t
0
-4
4
8
-%pp [Imidazole]
h
mM
(b)
6
ha
FIG. 5. Inhibition of ALP by imidazole and cephalin. (a) Dixon plot: reciprocal of human skeletal ALP activity (initial velocity, relative units) vs imidazole concentration. Activity was measured in CO3 buffer at pH 9.5 with 1mM PNPP, 2 mM MgClz, and 0 (m), 0.08 mM (O), or 0.25 mM (A) cephalin. Values of Kiapp for different cephalin concentrations are indicated by the arrows. (b) Replot of Ktipp for imidazole inhibition (with supplemental data) vs concentration of cephalin.
.
483
484
FARLEY
AND JORCH
for this circulating skeletal ALP; however, this treatment did result in a significant, but nonlinear, competitive inhibition by cephalin. (The value of &,rp for cephalin inhibition decreased from 1.6 to 1.1 to 0.5 mM as the PNPP concentration decreased from 5 to 0.5 to 0.25 mM, respectively.) A similar, nonlinear competitive inhibition by cephalin was also observed for hepatic ALP in serum, even without butanol extraction. (In this case, Ki,app decreased from 1.9 to 1.8 to 0.9 mM as the PNPP concentration decreased from 5 to 0.5 to 0.25 IIIM, respectively.) Again, in contrast to the skeletal ALP activity extracted from bone, the skeletal ALP in Pagetic serum did not allow an inhibitory interaction between cephalin and imidazole, even after butanol extraction, although the hepatic ALP activity in jaundiced serum did. (The value of Ki for imidazole inhibition of hepatic ALP activity in serum decreased from 6 to 1.6 mM when cephalin was added at 1.2 mM, at pH 9.5 with 1.7 IIIM PNPP.) These results suggest differences, with respect to phospholipid interactions, between skeletal ALP in extracts of bone and the same isoenzyme in serum, and also differences between circulating skeletal and hepatic isoenzymes. Chick bone ALP in extracts and in wnditicmed culture medium. As an approach to determining the basis of the phospholipid interaction differences between human ALP isoenzymes in tissue extracts and in circulation (Table IV), we cultured em-
FIG. 6. Effect of extraction on the value of Kdmp for chick skeletal ALP. Reciprocal of chick calvarial cell ALP activity (initial velocity, relative units) vs reciprocal of PNPP concentration in CO8 buffer, pH 9.5, with 2 rnM MgCl,, in situ (0) or in extracts (0). The arrows indicate the values of -l/Kwp.
(0)
PH
g
io
FIG. ‘7. Effect of pH on chick skeletal ALP activity in extracts and in situ. (a) V,, for chick calvarial cell ALP (determined from reciprocal plots of initial velocity data) shown as percentage of V,, at pH 9.5 vs pH. The reactions were performed in CO8 buffer with 2 my Mg and various concentrations of PNPP (from l/2 K,,, to 2K,,,), using chick calvarial cell ALP in situ (V) or in extracts (0). (b) Log Kvs pH, conditions and symbols as in (a).
bryonic chick bone in serum-free BGJb medium and recovered skeletal ALP activities from the bones (by butanol extraction) and from the conditioned medium (i.e., activity released from bone during culture). Our kinetic comparison of these two forms of the same isoenzyme is summarized in Table V. The enzyme activities were identical with respect to Km,pNpp and Kcimible, but the extracted enzyme was much more sensitive to inhibition by cephalin than the enzyme in conditioned medium. The extracted enzyme also showed an inhibitory interaction between imidazole and cephalin (in fact, the pattern of inhibition by imidazole became mixed type,3 but the en* In the following schematic reaction E = enzyme (ALP), S = substrate (PNPP), I = inhibitori (imidaxole), X = inhibitors (cephalin), P = product (Pi and PNP).
PHOSPHOLIPID
INHIBITION
OF
SKELETAL
zyme in conditioned medium did not, even after butanol extraction). Human shdetul ALP-efect of phosphe &nose C. The studies detailed above indicated that the potential for ALP-phospholipid interactions was altered during or after release of the isoenzyme from osteoblast cell membranes. Since previous studies had shown that the ability of extracted ALP to reassociate with phospholipid micelles was inhibited by treatment with phospholipase C (17), we examined the effect of treating extracted human skeletal ALP with phospholipase C in our kinetic system. The results are summarized in Table VI. Incubation with phospholipase C affected subsequent interactions between human skeletal ALP and both imidazole and cephalin in a dose-dependent manner. The treated isoenzyme was less sensitive to inhibitions by imidazole, cephalin, and the combination. In these respects it resembled the circulating isoenzyme. The value of Km,app for PNPP was not affected, and phospholipase C had no direct effect on ALP activity. DISCUSSION
Our previous studies had shown that partially purified human skeletal ALP expresses secondary binding sites (i.e., for uncompetitive inhibitors-homoarginine, levamisole, etc.) when the primary, phosphoryl, binding site is filled (6). Histolog-
E+S + X Kx
Jt EX + I
Ki
K.
kp
=ES-E+P + I Jt K, ES1
Jt EXI
X is a competitive inhibitor and I is uncompetitive (each with respect to S). If Ki < KI we expect increased inhibition by I (imidasole) in the presence of X (cephalin), and KbW will approach Ki as a limit, as the concentration of X (cephalin) increases. The pattern of inhibition by I will become mixed type with respect to PNPP (in the limiting use it will be competitive) as the concentration of X increases.
ALKALINE
485
PHOSPHATASE TABLE
III
CHICK BONE CELL ALP: INHIBITION PROFILE IN EXTRACTS AND in Situ K;
Inhibitor Vanadate Homoarginine Theophylline Levamisole Cephalin Imidazole Imidazole with 0.5 rn~ cephalin
Type of inhibition”
at pH 9.5 (mu)
In extracts
In situ
CI UC1 UC1 UC1 CI UC1
0.04 5-6 0.15 0.20 0.3 7-9
0.07 5-7 0.16 0.17 >lOb 6-7
UC1
1.3
6-7
a Inhibition with respect to PNPP. b Since no inhibition was observed, we have calculated a minimum possible value for K,.
ical studies of ALP activity, using these same uncompetitive inhibitors, have demonstrated that the secondary binding sites are also expressed in situ (12,13). Because one of these secondary binding sites will accommodate cCMP and other cytidine nucleotides, because derivatives of cytidine nucleotides are involved in lipid and carbohydrate metabolism in membranes, and because ALP is membrane bound, we investigated the possibility that ALP was involved in one of these reaction pathways. To this end we applied the method of multiple inhibition analysis. For example, if skeletal ALP were involved in reaction sequence CDP-choline + 1,2-diglyceride % PC + CMP, we would expect to see an interaction between PC and CMP as inhibitors. We also looked at the possibility of an interaction between sialic acid and CMP or CTP because sialic acid has been reported to inhibit ALP activity (14). We did not observe an inhibition by sialic acid, nor did we detect an interaction between any of the inhibitors tested. What we did find was surprising: skeletal ALP activity was inhibited by cephalin. This inhibition was competitive with respect to PNPP and was mimicked by a variety of phospholipids. The inhibition could not be attributed to (a) lipid-substrate/product interactions, (b) detergent effects (the inhibition was reversible and SDS was not
486
FARLEY
AND JORCH
TABLE
IV
EFFECTS OF CEPHALIN ON HUMAN ALP ISOENZYMES ALP Isoenzyme (source)a Skeletal Skeletal Skeletal Hepatic Placental
(hone extracted) (Pagetic serum) (Pagetic serum extracted) (serum) (placenta extracted)
Effect of cephalin on imidazole inhibition
Effect of cephalin* CI (linear) None CI (nonlinear)” CI (nonlinear)” CI (linear)
4?.PP None None &PP Noned
“ALP activity was measured in sera and in butanol extracts of tissues and sera. * Cephalin was tested at concentrations up to 1 mM with minimum PNPP concentrations pH 9.5. ‘Inhibition significant only at 0.25 mM PNPP. d Imidazole does not inhibit placental ALP.
inhibitory), (c) competition for Mgz+, or (d) contaminating Pi. The fact that the inhibition was competitive implied that the phospholipids could prevent PNPP hydrolysis by direct interaction with skeletal ALP. The phospholipids were binding to or affecting a primary site. Apparently, the interaction was not dependent on the phosphate ester region of the lipids, since myristic acid (assayed in the presence of SDS for optical clarity), which is not a phospholipid, was also a competitive inhibitor. The observation that both dimyristoyl lecithin and palmitoyl lysolecithin, which are long-chain fatty acid phosphatides (Cl4 and C16, respectively) were efTABLE CHICK BONE
of 0.25 mM at
fective as inhibitors, but dicaproyl lecithin (which has a shorter fatty acid chain, C,) was not, suggests that the lipid moeity is primarily responsible for the effect. The observation that cephalin, lecithin, and phosphatidylinositol can potentiate the inhibitory effect of imidazole on skeletal ALP activity indicates that imidazole can bind to the enzyme-phospholipid complex more easily than it binds to the enzyme-PNPP complex. (The usual inhibition by imidazole may actually represent binding to the phosphoryl-enzyme reaction intermediate; the actual point of this addition is not known.) The apparent effect of cephalin to decrease the inhibition V
ALP: EXTRACTED
AND RELEASED
Effect/kinetic Effector*
Extracted ALP”
PNPP Imidazole Cephalin Imidazole (in the presence of 1 mM cephalin)
Substrate/K, = 0.81 mM UCI/Ki = 4 mM CI/K, = 0.6 mM MT/Ki = 2.6 mMC
constant Conditioned
medium ALP”
Substrate/K,,, = 1.1 mM UCI/K, = 4.6 mM CI/K, = 4.5 mM UCI/K, = 5 mM
’ ALP activity was extracted from cultured embryonic chick bones (i.e., with butanol) and recovered from the conditioned serum-free culture medium. *Effecters tested in carbonate buffer, pH 9.5, with 1 rnrd MgCIP. c Since the pattern of inhibition here is different from that seen without added cephalin, comparisons of K
PHOSPHOLIPID
INHIBITION
OF
SKELETAL
caused by histamine could mean that histamine also binds to the enzyme-phospholipid complex, but with less affinity than it has in the normal reaction sequence. Neither homoarginine nor theophylline inhibition of skeletal ALP activity was affected by phospholipids. It is interesting to speculate that if ALP activity is affected by phospholipids in viva, the interaction with imidazole may mean that a histidyl residue, or some related compound, may also be involved. But is cellular ALP activity affected by phospholipids? Our kinetic evidence suggests that it may be. The value of K7n,app for PNPP is greater for chick calvarial cell ALP in situ than for the same isoenzyme in extracts. A similar difference has been reported for Escherichia coli ALP in cell suspensions and in sonicated extracts (15). Our hypothesis assumes that the lipid-enzyme interaction is responsible for those differences. Although this would be consistent with our kinetic observations of extracted isoenzymes, it has not been established. In terms of our assumption the simplest explanation for these differences in substrate afhnity is that although ALP is an ectoenzyme (16), the substrate (PNPP) may have a sterically limited access to membrane-bound skeletal ALP, relative to its access to the same enzyme in solution. The inhibition of extracted skeletal ALP activity by phospholipids might then be due to the incorporation of ALP into phospholipid micelles, with a similarly limited access to the substrate. Whatever the explanation, our data indicate that phospholipids inhibit extracted ALP activity in solution and that a similar phospholipidALP interaction could modulate ALP activity in situ. Butanol-extracted alkaline phosphatases prepared from human bone and human placental tissue were both subject to linear competitive inhibitions by cephalin, whereas the skeletal and hepatic isoenzymes in serum were either not inhibited or less efficiently inhibited (and in a nonlinear fashion), even after butanol extractions. We interpret this to mean that the extracted ALPS retained more of their ability to reassociate with phospholipids
ALKALINE
TABLE HUMAN
487
PHOSPHATASE
SKELETAL
VI ALP:
EFFECT
OF
C
PHOSPHOLIPASE
Inhibitors of ALP (percentage remaining activity)b Preincubation phospholipase (Units) None 1 2.5 10 25
with C”
10 mM Imidazole 42 41 43 58 63
1 mM PE 61 69 77 34 103
10 mrd Imidazole and 1 mM PE 13 17.6 22 27.6 44
0 Human skeletal ALP (extracted from bones) was incubated for 4 h at 4°C with various amounts of phospholipase C (Sigma Type V, 460 unita/mg) as described under Materials and Methods. Incubation without phospholipase C did not affect activity and phospholipase C did not directly affect the assay for ALP activity. b Remaining ALP activity (shown as percentage of control) in the presence of inhibitors in carbonate buffer, pH 9.5, containing 1 mM MgQ and 2 mM PNPP. Values shown are the means of triplicate determinations.
than the circulating ALP isoenzymes. In this regard, Low and Zilversmit (17) have examined the differences between the ALP activity released from pig kidney microsomes by butanol extraction and the ALP activity released by phospholipase C. They reported that the butanol-extracted ALP could reassociate with phospholipid vesicles but the enzymatically released ALP could not. When butanol-extracted ALP was incubated with phospholipase C, it lost its ability to reassociate with phospholipid vesicles. Low and Zilversmit, therefore, suggested that ALP was covalently bound to membrane phospholipids and could be released either with attached phospholipid, by membrane dissolution (i.e., butanol), or without attached phospholipid, by phospholipase C activity. Our own observation that treating extracted human skeletal ALP with phospholipase C affected the potential for phospholipid inhibitions is consistent with the possibility that the appearance of skeletal ALP activity in serum is partially regulated by phospholipase activity in osteoblasts. With respect to phospholipid inhibition, treating human skeletal ALP with phospholipase C converted the extracted isoenzyme
488
FARLEY
AND JORCH
to the circulating form. Alternatively, it is possible that ALP activities may be released from cells by hydrophobic extraction (as we observed with chick calvarial cells in culture)4 or by matrix vesicle formation (20) or normal membrane turnover. Studies are in progress to evaluate these possibilities. Our kinetic studies comparing the skeletal ALP activity extracted from cultured chick bones with the ALP activity released from the growing bones into the serumfree medium during culture (Table V) showed the same pattern of difference, with respect to lipid interactions, as we had seen with human skeletal ALP in tissue extracts and in circulation. These data, therefore, support the premise that the potential for interaction between skeletal ALP and phospholipids (i.e., phospholipid inhibition of ALP activity or reassociation of ALP with phospholipid vesicles) is irreversibly altered when the enzyme is released to circulation. These data also suggest that the chick bone organ culture system may provide a model for determining the mechanism of ALP release. In summary, we have shown (a) that phospholipids can inhibit the activity of extracted human skeletal ALP (or chick bone ALP) much more efficiently than they can affect the activity of the same isoenzyme in serum (or in serum-free culture medium); (b) that the interaction between ALP and phospholipids could account for the observed difference in substrate affinities between skeletal ALP in extracts and skeletal ALP in situ; and (c) that the lack of interaction between phospholipids and skeletal ALP in serum and in conditioned organ culture medium is consistent with ’ The possibility that hydrophobic extraction is partially responsible for the release of ALP activity from cell membranes into circulation is supported by observations that an increase in dietary lipids can increase the circulating level of intestinal ALP in humans (19) and in rata (20).
the possibility that ALP is released from bone cells by phospholipase activity. ACKNOWLEDGMENTS Richard Widstrom, Kathleen King, Elmer Feist, and Robert Haller deserve thanks for their technical assistance, as do Penny Nasabal and Peggy Duane for their help in preparing this manuscript. This research was supported by NIH Grant AM 31061-01. REFERENCES 1. ROBINSON,R. (1923) &o&em J. 17,268-293. 2. KAY, H. D. (1932) Physiol Rev. 12, 384-422. 3. FUN, M., WHYTE, M., AND TEITLEBAUM, S. (1980) Lab. Invest.
43.489-494.
4. LAUFFENBURGER,T., OLAH, A. J., DAMBACHER, J., et al. (1977) Metabolism 26, 589-597. 5. CLARK, L. C., ANLI BECK, E. (1950) J. Pediutr. 36, 335-341. 6. FARLEY, J. R., IVEY, J. L., AND BAYLINK, D. J. (1980) J. Biol Chxm. 255,4680-4686. 7. MCCOMB, R. B., BOWERS, G. N., JR., AND POSEN, S. (1979) Alkaline Phosphatase, Plenum, New York. 8. FARLEY, J. R., CHESNUT, C. H. III, AND BAYLINK, D. J. (1981) CZin. Chem. 27,2002-2007. 9. HOWARD, G. A., BO~EMILLER, B. L., AND BAYLINK, D. J. (1980) Metab. Bone LX-s. Rel Res. 2, 131-136. 10. BRADFORD, M. M. (1976) And Biochem 72,248255. 11. SEGEL, I. H. (1975) Enzyme Kinetics, Wiley-Interscience/Wiley, New York. 12. SUGIMURA, K., AND MIZUTANI, A. (1979) Hi&+ chemist?y 61,X3-129. 13. MOOG, F., AND GLAZIER, H. S. (1972) Cmp. Biochem Ph@ol 42. 321-336. 14. KOMODA, T., AND SAKAGISHI, Y. (1977) B&&m. Biophys Ada 482.79-84. 15. BROCKMAN, R. W., AND HEPPEL, L. A. (1968) Bi+ ch.emi&q 7.2554-2559. 16. DEPIERRE, J. W., AND KARNOVSKY, M. L. (1974) J. Biol Chem. 249, 7111-7120. 17. Low, M. G., AND ZILVERSMIT, D. B. (1980) Bti chemistry 19,3913-3918. 18. KLEEREKOPER, M., HORNE, M., CORNISH,C. J., AND POSEN, S. (1970) CZin Sci 38, 339-346. 19. SAINI, P. K., AND POSEN, S. (1969) Biochem Bi+ phys. Actu 177,50-57. 20. MATLISZAWA, T., AND ANDERSON, H. C. (1971) J. Hi&&em Cytochem 19, 801-809.
Differential
Effects of Phospholipids on Skeletal Alkaline Phosphatase Activity in Extracts, in Situ and in Circulation JOHN
Departments
of Biodemtitry
R. FARLEY’ and Medicine, Received
URSULA
AND bma
Linda
September
University,
M. JORCH Loma
Linda,
Cal&&a
92550
21, 1982
Human skeletal alkaline phosphatase (ALP) purified from human bone was subject to competitive inhibitions by phospholipids including cephalins, lecithins, and phosphatidylinositol. Ki values ranged from 0.7 to 1.5 mM, at pH 9.5. As previously shown, the enzyme was subject to uncompetitive inhibition by imidazole. The inhibitory phospholipids potentiated this effect, and altered the nature of the imidazole inhibition, from uncompetitive to mixed type, suggesting that imidazole was bound more efficiently to the enzyme-phospholipid complex than to the enzyme-substrate complex. No interactions were observed between phospholipids and other uncompetitive inhibitors of ALP. The skeletal ALP activity of cultured chick calvarial cells was assayed both in situ and in extracts. Like the extracted human isoenzyme, the extracted chick ALP was subject to competitive inhibition by cephalin (Ki = 0.3 mM at pH 9.3) and an inhibitory interaction between cephalin and imidazole, but the same isoenzyme showed neither effect in situ. The value of K ,QN~P at pH 9.5 for chick skeletal ALP was 1.5 mM in extracts and 7.1 mM in situ. When embryonic chick bones were cultured in vitro, skeletal ALP activity was released into the serum-free medium. Unlike the same isoenzyme extracted from the bones, the ALP activity in the medium was not inhibited by cephalin and showed no inhibitory interaction between cephalin and imidazole. Similarly, human serum ALP activities were not as sensitive to phospholipid inhibition as the same isoenzymes extracted from tissues. Human skeletal ALP extracted from bone was inhibited by cephalin, but the skeletal isoenzyme in Pagetic serum was not, suggesting that the potential for phospholipid interaction was altered during or after release from osteoblast cell membranes. The observation that extracted human skeletal ALP lost its potential for inhibition by phospholipids after treatment with phospholipase C further suggests that ALP activity may be released from cells during membrane turnover.
There are two reasons why skeletal ALP2 (EC 3.1.3.1.) is thought to play a role in
bone formation (l-3). (a) In bone, skeletal ALP is concentrated in the plasma membranes of osteoblasts, and these cells are responsible for bone matrix formation. (b) In serum, the activity of skeletal ALP tends to vary with the bone formation rate (4, 5). The evidence is circumstantial. After almost 60 years of investigation the function of skeletal ALP in vivo is unknown.
1 To whom correspondence should be addressed: Research Service (151), Jerry L. Pettis Veterans Hospital, 11201 Benton Street, Loma Linda, California 92357. 2 Abbreviations used: ALP, alkaline phosphatase; PNPP, pnitrophenylphosphate; PNP, pnitrophenol; cCMP, cytidine 3’,5’-cyclic monophosphate; PE, phosphatidylethanolamine (cephalin); PC, phosphatidylcholine (lecithin); PI, phosphatidylinositol; PEA, phosphorylethanolamine; SDS, sodium dodecyl sul-
fate; CI, competitive inhibition; UCI, uncompetitive inhibition; MT, mixed-type inhibition; Knsppp, Kiaw, etc., apparent value of kinetic constant. 477
0003-9861/83/040477-12$03.00/O Copyright All rights
Q 1983 by Academic Press. Inc. of reproduction in any form reserved.
478
FARLEY
We have attempted to determine skeletal ALP function in wivo by characterizing the binding sites that the isoenzyme possesses, reasoning that the most efficient substrates and inhibitors (i.e., the most tightly bound) should most closely resemble the effecters of skeletal ALP in wivo. We had previously shown that human skeletal ALP possesses a primary binding site for Pi or phosphoryl substrates, as well as secondary sites that will accommodate Tris (or imidazole, etc.) and theophylline (or cCMP, etc.) (6). These secondary sites are only expressed when the primary site has been filled. (For a general review of binding sites on ALP isoenzymes see Ref. (7). We now report (a) the results of multiple inhibition analyses to further characterize the relations between these binding sites; (b) the effects of lipid-enzyme interactions, both in extracts and in situ, to modulate ALP activity and to affect the binding sites; and (c) the differential effects of phospholipids on ALP activity in tissue extracts and in circulation. MATERIALS
AND METHODS
Materials. PNPP, lecithin, cephalin, imidazole, levamisole, placental ALP and phosphatidylinositol, phospholipase C, and phospholipase A were obtained from Sigma Chemical Company. Lysolecithin (palmitoyl), lecithin (dimyristoyl), cephalin (dipalmitoyl), and lecithin (fl-dicaproyl) were purchased from Calbiochem/Mannheim. Collagenase and BGJb culture medium were obtained from GIBCO. Tissue culture plates (35-mm dishes and multiwells) were purchased from Falcon. Assay for ALP activity. The standard assay mixture (used during purification of the isoenzyme) consisted of 0.15 M carbonate buffer, pH 10.3.1 mM MgClz, 10 mM PNPP, and ALP in a total volume of 1.5 ml at 22°C (ambient room temperature). The reaction was initiated by the addition of the enzyme (or the substrate) and monitored by the change in absorbance at 410 nm during a 15- to 120-min incubation. Controls, without ALP, were included with each assay to correct for the nonspecific hydrolysis of PNPP. Variations in reaction conditions (ie., pH, substrate concentration, addition of inhibitors, etc.) are detailed in the text. Partial purification of ALP from human bone. This method has been previously described in detail (6). Briefly, human femoral heads were obtained at hip replacement surgery, frozen, crushed, rinsed (to re-
AND JORCH move contaminating marrow and serum), and extracted with 20% butanol (v/v) in carbonate buffer (25 mM, pH 8.3) for 24 h at 4°C. ALP activity was obtained in the aqueous phase, and after dialysis against the same carbonate buffer, the activity was further purified by ammonium sulfate precipitation, DEAE-cellulose column chromatography, and filtration through a column of agarose-linked Cibacron blue (Bio-Rad). The skeletal ALP thus prepared had a specific activity of about 0.3 Wmg in our standard assay. ALP activity in human serum. Serum samples were obtained from two patients with active Paget’s disease of bone (total serum ALP > 390 U/liter for each) and two patients with obstructive jaundice (total serum ALP > 356 U/liter for each). Isoenzyme analysis (8) confirmed that the Pagetic and jaundiced sera contained, almost exclusively, skeletal and hepatic ALP activities, respectively. Aliquots of the samples were extracted with 26% butanol (v/v) in 25 rnrd carbonate buffer (pH 8.3) for 24 h at 4°C and then dialyzed against the same carbonate buffer for kinetic analyses and comparison with their unextracted counterparts. Additional aliquota were incubated in imidazole buffer (10 mM, pH 7.4) containing 2 mM CaClz and 0.15 M NaCl with various amounts of phospholipase C for 4 h at 4’C. The reaction mixtures were then dialyzed against 25 mM carbonate buffer (pH 8.0) for kinetic analyses. Isolation and &re of ernbrgontk chick calvarial cells. These methods have been described in detail (9). Briefly, cells (predominately osteoblasta and osteoblast progenitors) were released from the calvariae of lbday embryonic chicks by sequential collagenage digestion, plated at an initial density of 250/ mmz in 16-mm-diameter multiwell plates, and cultured in serum-free BGJb medium in an atmosphere of 5% COz and at 37°C. Cells were also cultured in 35-mm-diameter culture dishes at the same initial density and under identical conditions. After an 18h incubation, the medium was changed, and the cells were either assayed immediately or cultured for an additional l-3 days before use. Specific conditions for each experiment are given in the text. Partial pur&adon of ALP from chick bone cdls. Embryonic chick calvarial cells, cultured as described above in 35-mm-diameter culture dishes, were rinsed and extracted with 20% butanol (v/v) in 25 rnrd carbonate buffer (pH 8.3) for 4-8 h at 4°C. The extract was dialyzed against the same carbonate buffer and the insoluble material was removed by centrifugation. Activity was determined by variations of the standard assay (i.e., changes in pH, substrate concentration, addition of inhibitors and effecters, etc.) as detailed in the text. No kinetic differences were discernible between ALP prepared from cells cultured overnight and those cultured for 3 additional days.
PHOSPHOLIPID
INHIBITION
OF SKELETAL
Assay for chick bone Ceu ALP in situ. ALP activity was measured in situ using embryonic chick calvarial cells cultured in 16-mm-diameter multiwells as described above. The culture medium was removed and the adherent cells were rinsed with 0.5 ml of 25 mM carbonate buffer, pH 9.5, containing 0.14 Y NaCl. The ALP assay mixture (usually at pH 9.5) was then added to the wells in a total volume of 1.5 ml and the supernatant solutions were transferred (by Pasteur pipet) sequentially to cuvettes in the spectrophotometer for determinations of absorbance at 404 nm, at the beginning and at the end of the incubation period (usually 30 to 90 min). Most of the cells remain attached to the tissue culture wells during this procedure. Repeat assays using the same cells (rinsed between assays) showed that more than 80% of the original activity remained after an initial 60-min incubation. Organ culture of embryonic chick bae. Frontal and parietal bones were asceptically dissected from 13day embryonic chicks, rinsed in serum-free BGJb medium, and cultured individually on lens paper rafts in 24-place multiwell tissue culture plates in a volume of 0.5 ml of serum-free BGJ,, medium. (A total of 48 bones were used for these experiments.) The medium was changed at 36-h intervals, and conditioned medium was pooled and frozen for subsequent determination of ALP activity. After ‘72 h, ALP activity was extracted from the bones with 20% butanol (v/ v) in 25 rnrd CO8 buffer, pH 8.3. The butanol extracts
TABLE
ALKALINE
479
PHOSPHATASE
were pooled, and both the butanol extracts and the pooled, conditioned media were lyophilized and resuspended in one-tenth their original volumes, before dialysis against 25 mM COIl buffer, pH 8.3, and subsequent kinetic analyses. Additional methods. Protein concentrations were determined by the dye-binding method of Bradford (10). Spectral determinations were made on a Beckman DU-Gilford spectrophotometer. All kinetic experiments were performed with duplicates of individual data points; average values are reported. The kinetic nomenclature of Segel (11) is used throughout. RESULTS
Characterization of the cCMP-binding site. Previous studies had shown that human skeletal ALP was subject to uncompetitive inhibition by cCMP, Ki = 3.2 m&q from pH 8.1 to pH 9.6 (6). Data summarized in Table I compare inhibitions by cCMP and other cytidine nucleotides. (It is interesting to note that inhibition by cCMP is uncompetitive with respect to PNPP, but inhibitions by CMP and CTP are competitive.) Since ALP is a membrane-bound enzyme and since cytidine I
HUMAN SKELETAL ALP: CHARACTERIZATION OF THE cCMP-BINDING SITE Inhibitor cCMP CTP CMP Cytidine Sialic acid CMP and sialic acid CTP and sialic acid Glycerol CMP and glycerol CTP and glycerol PEA CMP and PEA CTP and PEA Cephalin CMP and cephalin
Effect” UC1 CI CI None None Noneb Noneb None Noneb None” CI None b None” CI None”
Gpp at PHI hM) at pH 9.6 16 at pH 8.8 2.0 at pH 8.8 >30 at pH 8.8” >40 at pH 8.8” 3.2
at pH 8.1 0.5 at pH 8.2 3.3
~40 at
-
pH 8.8”
30 at
pH 9.5
0.5 at
pH 8.0
0.8 at
pH 9.5 -
0.7 at pH 8.8 -
“All inhibitors and combinations were tested at pH 8.8 and at pH 10.0 (i.e., minimum pH range), with both saturating and nonsaturating concentrations of PNPP. b No additional inhibition resulting from the combination. ’ Since no inhibition was observed, we have calculated a minimum possible value for KGpp.
480
FARLEY
AND JORCH
120
/
T
,__
T
Inhibitor Ki (VHM)
none
SDS
-
-
PC 0.7mM
PE 0.65mM
PEA 30mM
-
Pi 2.5mM
FIG. 1. Relative potency of ALP inhibitors. Percentage of control activity (shown as mean with range of four determinations) for human skeletal ALP assayed in the presence of various potential inhibitors. Each was tested at a concentration of 1.0 mM with 0.3 mM PNPP and 1 mM MgClr in CO* buffer at pH 9.3. The value of Kicpp at pH 9.3 (determined by replots of initial velocity data) is also given for each inhibitory compound.
nucleotides are involved in lipid and carbohydrate metabolism in membranes, we considered the possibility that ALP might be involved in one of these reaction pathways. We therefore examined the effects of other compounds involved in these reactions (sialic acid, glycerol, phosphoethanolamine, and cephalin) on the activity of human skeletal ALP with and without added CMP and CTP. The results are shown in Table I. Inhibition of human skektal ALP by phospholipids. The inhibitory effect of PE (phosphatidylethanolamine, cephalin) on human skeletal ALP was compared with the effects of PC, PEA, Pi, and SDS. The results are shown in Fig. 1. The inhibition by PE was linear competitive with respect to PNPP; this is shown in Fig. 2. A variety of lipids were found to be similarly inhibitory; these results are summarized in Table II. These lipids form micelles at the concentrations we have used. The background optical density of the ALP assay mixtures (i.e., absorbance and light scattering at 410 nm before the phosphatase
reaction has begun) was therefore increased by addition of the lipids. This background was, however, stable over the
8 l/V
Volume
PE
$!I)
FIG. 2. Inhibition of ALP activity by cephalin. Dixon plot: reciprocal of initial velocity of human skeletal ALP (relative units) vs volume of cephalin (PE) added. (Stock PE = 11.9 rnr& reaction volume = 1.5 ml; 100 ~1 added cephalin corresponds to a concentration of 0.79 maa). Activity assayed in COll buffer, pH 9.3, with 2 rnM added MgCla and 2.5 mM (O), 0.4 rnM (m), or 0.2 mM (V) added PNPP.
PHOSPHOLIPID
INHIBITION
OF
SKELETAL
course of the reaction, whether the lipids were incubated alone or in combination with either PNPP or skeletal ALP. To ensure that we were observing lipidenzyme interactions we investigated the possibilities that phospholipids could be (a) sequestering the substrate and thereby reducing its effective concentration or (b) interfering with the absorbance of the reaction product, PNP. Regarding the second possibility, Fig. 3a demonstrates that cephalin, at inhibitory concentrations, did not interfere with the absorbance of PNP (i.e., the lines are parallel). Both lecithin and cephalin were tested over a wide range of PNP concentrations, corresponding to the range of enzyme activities assayed, and we could not detect interference. Regarding the other possible source of methodological artifact ((a) above), Fig. 3b shows the time course of PNP distribution between dialysis half-cells (Thomas Scientific, volume = 1.5 ml per half-cell) separated by dialysis membrane with a nominal i& cutoff of 12,000, where one chamber initially contained 1.2 mM cephalin (or buffer) and the other contained 0.06 mM PNP. (Separate controls established that cephalin, at the concentrations tested, did not cross the dialysis membrane in significant amounts during the time required for equilibration of PNP.) The fact that the predicted equilibrium distribution was not exceeded in the presence of the lipid indicates that cephalin does not sequester PNP, and, since PNPP has a greater charge/mass ratio at pH 9.5 than PNP, it is reasonable to assume that PNPP is not sequestered either. We have also tested the possibility that phospholipids might inhibit ALP activity by competing with the enzyme for MgZ+. We saw no difference in the kinetic characteristics of PE inhibition when the [Mgzf] was varied between 1 and 10 rnM. Eflects of phap?wlipi& on the secmdarg binding sites. Since the competitive inhibitions seen with phospholipids were not methodological artifacts and since they involved the primary, phosphoryl, binding site, we examined the interactions between phospholipids and uncompetitive inhibitors (i.e., secondary site effecters) of
ALKALINE
TABLE HUMAN
SKELETAL
Lipid SDS Cephalin (crude) Cephalin (dipalmitoyl) Lecithin (crude) Lecithin (dimyristoyl) Lecithin (dicaproyl) Lysolecithin (palmitoyl) Phosphatidyl inositol Myristic acid with 1 mM SDS
481
PHOSPHATASE
ALP:
II INHIBITION
BY LIPIDS
5w of inhibition”
KicPP @f)
None Competitive Competitive Competitive Competitive None Competitive Competitive
>20b 0.8 1.1 0.8 1.2 >5b 1.5 0.9
Competitive
0.7
D At pH 9.5, evaluated by Lineweaver-Burke Dixon plots of initial velocity data. bSince no inhibition was observed, we have culated a minimum possible value for Kinpp.
or cal-
human skeletal ALP, such as imidazole, homoarginine, and theophylline (6). Figure 4 demonstrates uncompetitive inhibition of human skeletal ALP activity by imidazole, and Fig. 5 demonstrates how PE (cephalin) can potentiate this inhibition. The replot in Fig. 5b shows that the value of K+pp for imidazole decreases from 7 mM with no added PE to 0.7 mM with 0.5 mM added PE. We saw a similar inhibitory synergism between lecithin and imidazole and between phosphatidylinositol and imidazole. No such interaction was observed between cephalin and homoarginine (K..*,nm.dam,I?= 8.0 and 7.5 mM with 0 and 1 mM added PE, respectively) or between cephalin and theophylline (Ki,thkphylline = 0.34 and 0.36 m&r with 0 and 1 mM added PE, respectively). We have already shown (in Table I) the lack of interaction between cephalin and CMP. It is interesting to note that the inhibitory effect of histamine on skeletal ALP was reduced by added cephinhibition inalin (K,iapp for histamine creased from 11.3 to 16.4 mM as the cephalin concentration was increased from 0 to 0.5 mM). Chick calvarial cell ALP activity tracts and in situ. As an approach
termining teractions
whether phospholipid-ALP might be physiologically
in exto deinsig-
FARLEY
0
AND JORCH
60 Volume Cephalin
Time
120 (~1)
(mid
FIG. 3. Potential interactions of cephalin with PNPP or PNP. (a) Test for possible interference with PNP light absorption. Absorbance at 404 nm vs volume of cephalin added (stock cephalin at 11.9 mM; 100 ~1 added cephalin gives a final concentration of 0.79 mM) in CO8 buffer, pH 9.3, with 2 mM MgClc, no added ALP, and 0 (0) or 0.02 mM (V) added PNP. (b) Test for possible sequestering of PNP by cephalin micelles. Equilibrium dialysis half-cells (volume = 1.5 ml) separated by dialysis membrane (12,000 AZ,cutoff) were filled with CO8 buffer (pH 9.3) containing 2 mM MgCI,, or cephalin in the same buffer on one side and 0.06 rnM PNP in buffer on the other. The ratio of absorbance (and light scattering) at 404 nm on the cephalin (or buffer) side to absorbance on the PNP side is shown as a function of time for cells with 1.2 rnrd added cephalin (V) and for controls (0). The dotted line indicates the theoretical limit for equilibration without interaction.
nificant we compared the ALP activity of cultured embryonic chick bone cells in situ with the activity of the same isoenzyme in butanol extracts of cells. Figure 6 shows that the value of Km,pNppwas about 4.7fold lower in the extracts than in the cells, at pH 9.5. Figure 7 shows the effect of pH on Vm and KQNPP for ALP in situ and in extracts. The data in Table III demonstrate that the activity of skeletal ALP is
identical in cultured chick calvarial cells and in extracts of those cells with respect to all tested inhibitors except for cephalin and the cephalin/imidazole combination. When cultured chick calvarial cell ALP was assayed in situ at pH 9.5 and the cells were rinsed and then reassayed, we found that 8590% of the initial activity remained. However, this value was reduced to 53-62% when cephalin was added to the
PHOSPHOLIPID
INHIBITION
0
OF
SKELETAL
2
lI[pFILP]
mM“
FIG. 4. Uncompetitive inhibition of ALP by azole. Reciprocal of human skeletal ALP activity tial velocity, relative units) vs reciprocal of concentration. Activity was measured at pH CO8 buffer with 2 mM added MgClz and 0 (O), (H), or 10 mrd (0) added imidazole.
imid(iniPNPP 9.9 in 5 m&i
initial assay mixture (0.5 and 1 InM). There was no detectable change in cell number during this latter incubation.
ALKALINE
PHOSPHATASE
Human ALP ksoewyme.s-inhibition by phospholipids. Since skeletal ALP obtained from human bone was subject to competitive inhibition by phospholipids, we wanted to determine whether a similar inhibition could be observed (a) with skeletal ALP in serum or (b) with other ALP isoenzymes, either purified from organs or in serum. The results of these studies are summarized in Table IV. A commercial preparation of placental ALP, which had been purified by butanol extraction, was inhibited by cephalin (linear, competitive: Ki = 1.1 mM at pH 9.5). In contrast, the skeletal ALP activity in Pagetic serum was not inhibited by cephalin at pH 9.5, even using 0.25 mM PNPP (Km.rNPP for this isoenzyme was 0.5 mM at pH 9.5). Butanol extraction and subsequent dialysis had no detectable effect on the value of Km.rNPP
25
t
0
-4
4
8
-%pp [Imidazole]
h
mM
(b)
6
ha
FIG. 5. Inhibition of ALP by imidazole and cephalin. (a) Dixon plot: reciprocal of human skeletal ALP activity (initial velocity, relative units) vs imidazole concentration. Activity was measured in CO3 buffer at pH 9.5 with 1mM PNPP, 2 mM MgClz, and 0 (m), 0.08 mM (O), or 0.25 mM (A) cephalin. Values of Kiapp for different cephalin concentrations are indicated by the arrows. (b) Replot of Ktipp for imidazole inhibition (with supplemental data) vs concentration of cephalin.
.
483
484
FARLEY
AND JORCH
for this circulating skeletal ALP; however, this treatment did result in a significant, but nonlinear, competitive inhibition by cephalin. (The value of &,rp for cephalin inhibition decreased from 1.6 to 1.1 to 0.5 mM as the PNPP concentration decreased from 5 to 0.5 to 0.25 mM, respectively.) A similar, nonlinear competitive inhibition by cephalin was also observed for hepatic ALP in serum, even without butanol extraction. (In this case, Ki,app decreased from 1.9 to 1.8 to 0.9 mM as the PNPP concentration decreased from 5 to 0.5 to 0.25 IIIM, respectively.) Again, in contrast to the skeletal ALP activity extracted from bone, the skeletal ALP in Pagetic serum did not allow an inhibitory interaction between cephalin and imidazole, even after butanol extraction, although the hepatic ALP activity in jaundiced serum did. (The value of Ki for imidazole inhibition of hepatic ALP activity in serum decreased from 6 to 1.6 mM when cephalin was added at 1.2 mM, at pH 9.5 with 1.7 IIIM PNPP.) These results suggest differences, with respect to phospholipid interactions, between skeletal ALP in extracts of bone and the same isoenzyme in serum, and also differences between circulating skeletal and hepatic isoenzymes. Chick bone ALP in extracts and in wnditicmed culture medium. As an approach to determining the basis of the phospholipid interaction differences between human ALP isoenzymes in tissue extracts and in circulation (Table IV), we cultured em-
FIG. 6. Effect of extraction on the value of Kdmp for chick skeletal ALP. Reciprocal of chick calvarial cell ALP activity (initial velocity, relative units) vs reciprocal of PNPP concentration in CO8 buffer, pH 9.5, with 2 rnM MgCl,, in situ (0) or in extracts (0). The arrows indicate the values of -l/Kwp.
(0)
PH
g
io
FIG. ‘7. Effect of pH on chick skeletal ALP activity in extracts and in situ. (a) V,, for chick calvarial cell ALP (determined from reciprocal plots of initial velocity data) shown as percentage of V,, at pH 9.5 vs pH. The reactions were performed in CO8 buffer with 2 my Mg and various concentrations of PNPP (from l/2 K,,, to 2K,,,), using chick calvarial cell ALP in situ (V) or in extracts (0). (b) Log Kvs pH, conditions and symbols as in (a).
bryonic chick bone in serum-free BGJb medium and recovered skeletal ALP activities from the bones (by butanol extraction) and from the conditioned medium (i.e., activity released from bone during culture). Our kinetic comparison of these two forms of the same isoenzyme is summarized in Table V. The enzyme activities were identical with respect to Km,pNpp and Kcimible, but the extracted enzyme was much more sensitive to inhibition by cephalin than the enzyme in conditioned medium. The extracted enzyme also showed an inhibitory interaction between imidazole and cephalin (in fact, the pattern of inhibition by imidazole became mixed type,3 but the en* In the following schematic reaction E = enzyme (ALP), S = substrate (PNPP), I = inhibitori (imidaxole), X = inhibitors (cephalin), P = product (Pi and PNP).
PHOSPHOLIPID
INHIBITION
OF
SKELETAL
zyme in conditioned medium did not, even after butanol extraction). Human shdetul ALP-efect of phosphe &nose C. The studies detailed above indicated that the potential for ALP-phospholipid interactions was altered during or after release of the isoenzyme from osteoblast cell membranes. Since previous studies had shown that the ability of extracted ALP to reassociate with phospholipid micelles was inhibited by treatment with phospholipase C (17), we examined the effect of treating extracted human skeletal ALP with phospholipase C in our kinetic system. The results are summarized in Table VI. Incubation with phospholipase C affected subsequent interactions between human skeletal ALP and both imidazole and cephalin in a dose-dependent manner. The treated isoenzyme was less sensitive to inhibitions by imidazole, cephalin, and the combination. In these respects it resembled the circulating isoenzyme. The value of Km,app for PNPP was not affected, and phospholipase C had no direct effect on ALP activity. DISCUSSION
Our previous studies had shown that partially purified human skeletal ALP expresses secondary binding sites (i.e., for uncompetitive inhibitors-homoarginine, levamisole, etc.) when the primary, phosphoryl, binding site is filled (6). Histolog-
E+S + X Kx
Jt EX + I
Ki
K.
kp
=ES-E+P + I Jt K, ES1
Jt EXI
X is a competitive inhibitor and I is uncompetitive (each with respect to S). If Ki < KI we expect increased inhibition by I (imidasole) in the presence of X (cephalin), and KbW will approach Ki as a limit, as the concentration of X (cephalin) increases. The pattern of inhibition by I will become mixed type with respect to PNPP (in the limiting use it will be competitive) as the concentration of X increases.
ALKALINE
485
PHOSPHATASE TABLE
III
CHICK BONE CELL ALP: INHIBITION PROFILE IN EXTRACTS AND in Situ K;
Inhibitor Vanadate Homoarginine Theophylline Levamisole Cephalin Imidazole Imidazole with 0.5 rn~ cephalin
Type of inhibition”
at pH 9.5 (mu)
In extracts
In situ
CI UC1 UC1 UC1 CI UC1
0.04 5-6 0.15 0.20 0.3 7-9
0.07 5-7 0.16 0.17 >lOb 6-7
UC1
1.3
6-7
a Inhibition with respect to PNPP. b Since no inhibition was observed, we have calculated a minimum possible value for K,.
ical studies of ALP activity, using these same uncompetitive inhibitors, have demonstrated that the secondary binding sites are also expressed in situ (12,13). Because one of these secondary binding sites will accommodate cCMP and other cytidine nucleotides, because derivatives of cytidine nucleotides are involved in lipid and carbohydrate metabolism in membranes, and because ALP is membrane bound, we investigated the possibility that ALP was involved in one of these reaction pathways. To this end we applied the method of multiple inhibition analysis. For example, if skeletal ALP were involved in reaction sequence CDP-choline + 1,2-diglyceride % PC + CMP, we would expect to see an interaction between PC and CMP as inhibitors. We also looked at the possibility of an interaction between sialic acid and CMP or CTP because sialic acid has been reported to inhibit ALP activity (14). We did not observe an inhibition by sialic acid, nor did we detect an interaction between any of the inhibitors tested. What we did find was surprising: skeletal ALP activity was inhibited by cephalin. This inhibition was competitive with respect to PNPP and was mimicked by a variety of phospholipids. The inhibition could not be attributed to (a) lipid-substrate/product interactions, (b) detergent effects (the inhibition was reversible and SDS was not
486
FARLEY
AND JORCH
TABLE
IV
EFFECTS OF CEPHALIN ON HUMAN ALP ISOENZYMES ALP Isoenzyme (source)a Skeletal Skeletal Skeletal Hepatic Placental
(hone extracted) (Pagetic serum) (Pagetic serum extracted) (serum) (placenta extracted)
Effect of cephalin on imidazole inhibition
Effect of cephalin* CI (linear) None CI (nonlinear)” CI (nonlinear)” CI (linear)
4?.PP None None &PP Noned
“ALP activity was measured in sera and in butanol extracts of tissues and sera. * Cephalin was tested at concentrations up to 1 mM with minimum PNPP concentrations pH 9.5. ‘Inhibition significant only at 0.25 mM PNPP. d Imidazole does not inhibit placental ALP.
inhibitory), (c) competition for Mgz+, or (d) contaminating Pi. The fact that the inhibition was competitive implied that the phospholipids could prevent PNPP hydrolysis by direct interaction with skeletal ALP. The phospholipids were binding to or affecting a primary site. Apparently, the interaction was not dependent on the phosphate ester region of the lipids, since myristic acid (assayed in the presence of SDS for optical clarity), which is not a phospholipid, was also a competitive inhibitor. The observation that both dimyristoyl lecithin and palmitoyl lysolecithin, which are long-chain fatty acid phosphatides (Cl4 and C16, respectively) were efTABLE CHICK BONE
of 0.25 mM at
fective as inhibitors, but dicaproyl lecithin (which has a shorter fatty acid chain, C,) was not, suggests that the lipid moeity is primarily responsible for the effect. The observation that cephalin, lecithin, and phosphatidylinositol can potentiate the inhibitory effect of imidazole on skeletal ALP activity indicates that imidazole can bind to the enzyme-phospholipid complex more easily than it binds to the enzyme-PNPP complex. (The usual inhibition by imidazole may actually represent binding to the phosphoryl-enzyme reaction intermediate; the actual point of this addition is not known.) The apparent effect of cephalin to decrease the inhibition V
ALP: EXTRACTED
AND RELEASED
Effect/kinetic Effector*
Extracted ALP”
PNPP Imidazole Cephalin Imidazole (in the presence of 1 mM cephalin)
Substrate/K, = 0.81 mM UCI/Ki = 4 mM CI/K, = 0.6 mM MT/Ki = 2.6 mMC
constant Conditioned
medium ALP”
Substrate/K,,, = 1.1 mM UCI/K, = 4.6 mM CI/K, = 4.5 mM UCI/K, = 5 mM
’ ALP activity was extracted from cultured embryonic chick bones (i.e., with butanol) and recovered from the conditioned serum-free culture medium. *Effecters tested in carbonate buffer, pH 9.5, with 1 rnrd MgCIP. c Since the pattern of inhibition here is different from that seen without added cephalin, comparisons of K
PHOSPHOLIPID
INHIBITION
OF
SKELETAL
caused by histamine could mean that histamine also binds to the enzyme-phospholipid complex, but with less affinity than it has in the normal reaction sequence. Neither homoarginine nor theophylline inhibition of skeletal ALP activity was affected by phospholipids. It is interesting to speculate that if ALP activity is affected by phospholipids in viva, the interaction with imidazole may mean that a histidyl residue, or some related compound, may also be involved. But is cellular ALP activity affected by phospholipids? Our kinetic evidence suggests that it may be. The value of K7n,app for PNPP is greater for chick calvarial cell ALP in situ than for the same isoenzyme in extracts. A similar difference has been reported for Escherichia coli ALP in cell suspensions and in sonicated extracts (15). Our hypothesis assumes that the lipid-enzyme interaction is responsible for those differences. Although this would be consistent with our kinetic observations of extracted isoenzymes, it has not been established. In terms of our assumption the simplest explanation for these differences in substrate afhnity is that although ALP is an ectoenzyme (16), the substrate (PNPP) may have a sterically limited access to membrane-bound skeletal ALP, relative to its access to the same enzyme in solution. The inhibition of extracted skeletal ALP activity by phospholipids might then be due to the incorporation of ALP into phospholipid micelles, with a similarly limited access to the substrate. Whatever the explanation, our data indicate that phospholipids inhibit extracted ALP activity in solution and that a similar phospholipidALP interaction could modulate ALP activity in situ. Butanol-extracted alkaline phosphatases prepared from human bone and human placental tissue were both subject to linear competitive inhibitions by cephalin, whereas the skeletal and hepatic isoenzymes in serum were either not inhibited or less efficiently inhibited (and in a nonlinear fashion), even after butanol extractions. We interpret this to mean that the extracted ALPS retained more of their ability to reassociate with phospholipids
ALKALINE
TABLE HUMAN
487
PHOSPHATASE
SKELETAL
VI ALP:
EFFECT
OF
C
PHOSPHOLIPASE
Inhibitors of ALP (percentage remaining activity)b Preincubation phospholipase (Units) None 1 2.5 10 25
with C”
10 mM Imidazole 42 41 43 58 63
1 mM PE 61 69 77 34 103
10 mrd Imidazole and 1 mM PE 13 17.6 22 27.6 44
0 Human skeletal ALP (extracted from bones) was incubated for 4 h at 4°C with various amounts of phospholipase C (Sigma Type V, 460 unita/mg) as described under Materials and Methods. Incubation without phospholipase C did not affect activity and phospholipase C did not directly affect the assay for ALP activity. b Remaining ALP activity (shown as percentage of control) in the presence of inhibitors in carbonate buffer, pH 9.5, containing 1 mM MgQ and 2 mM PNPP. Values shown are the means of triplicate determinations.
than the circulating ALP isoenzymes. In this regard, Low and Zilversmit (17) have examined the differences between the ALP activity released from pig kidney microsomes by butanol extraction and the ALP activity released by phospholipase C. They reported that the butanol-extracted ALP could reassociate with phospholipid vesicles but the enzymatically released ALP could not. When butanol-extracted ALP was incubated with phospholipase C, it lost its ability to reassociate with phospholipid vesicles. Low and Zilversmit, therefore, suggested that ALP was covalently bound to membrane phospholipids and could be released either with attached phospholipid, by membrane dissolution (i.e., butanol), or without attached phospholipid, by phospholipase C activity. Our own observation that treating extracted human skeletal ALP with phospholipase C affected the potential for phospholipid inhibitions is consistent with the possibility that the appearance of skeletal ALP activity in serum is partially regulated by phospholipase activity in osteoblasts. With respect to phospholipid inhibition, treating human skeletal ALP with phospholipase C converted the extracted isoenzyme
488
FARLEY
AND JORCH
to the circulating form. Alternatively, it is possible that ALP activities may be released from cells by hydrophobic extraction (as we observed with chick calvarial cells in culture)4 or by matrix vesicle formation (20) or normal membrane turnover. Studies are in progress to evaluate these possibilities. Our kinetic studies comparing the skeletal ALP activity extracted from cultured chick bones with the ALP activity released from the growing bones into the serumfree medium during culture (Table V) showed the same pattern of difference, with respect to lipid interactions, as we had seen with human skeletal ALP in tissue extracts and in circulation. These data, therefore, support the premise that the potential for interaction between skeletal ALP and phospholipids (i.e., phospholipid inhibition of ALP activity or reassociation of ALP with phospholipid vesicles) is irreversibly altered when the enzyme is released to circulation. These data also suggest that the chick bone organ culture system may provide a model for determining the mechanism of ALP release. In summary, we have shown (a) that phospholipids can inhibit the activity of extracted human skeletal ALP (or chick bone ALP) much more efficiently than they can affect the activity of the same isoenzyme in serum (or in serum-free culture medium); (b) that the interaction between ALP and phospholipids could account for the observed difference in substrate affinities between skeletal ALP in extracts and skeletal ALP in situ; and (c) that the lack of interaction between phospholipids and skeletal ALP in serum and in conditioned organ culture medium is consistent with ’ The possibility that hydrophobic extraction is partially responsible for the release of ALP activity from cell membranes into circulation is supported by observations that an increase in dietary lipids can increase the circulating level of intestinal ALP in humans (19) and in rata (20).
the possibility that ALP is released from bone cells by phospholipase activity. ACKNOWLEDGMENTS Richard Widstrom, Kathleen King, Elmer Feist, and Robert Haller deserve thanks for their technical assistance, as do Penny Nasabal and Peggy Duane for their help in preparing this manuscript. This research was supported by NIH Grant AM 31061-01. REFERENCES 1. ROBINSON,R. (1923) &o&em J. 17,268-293. 2. KAY, H. D. (1932) Physiol Rev. 12, 384-422. 3. FUN, M., WHYTE, M., AND TEITLEBAUM, S. (1980) Lab. Invest.
43.489-494.
4. LAUFFENBURGER,T., OLAH, A. J., DAMBACHER, J., et al. (1977) Metabolism 26, 589-597. 5. CLARK, L. C., ANLI BECK, E. (1950) J. Pediutr. 36, 335-341. 6. FARLEY, J. R., IVEY, J. L., AND BAYLINK, D. J. (1980) J. Biol Chxm. 255,4680-4686. 7. MCCOMB, R. B., BOWERS, G. N., JR., AND POSEN, S. (1979) Alkaline Phosphatase, Plenum, New York. 8. FARLEY, J. R., CHESNUT, C. H. III, AND BAYLINK, D. J. (1981) CZin. Chem. 27,2002-2007. 9. HOWARD, G. A., BO~EMILLER, B. L., AND BAYLINK, D. J. (1980) Metab. Bone LX-s. Rel Res. 2, 131-136. 10. BRADFORD, M. M. (1976) And Biochem 72,248255. 11. SEGEL, I. H. (1975) Enzyme Kinetics, Wiley-Interscience/Wiley, New York. 12. SUGIMURA, K., AND MIZUTANI, A. (1979) Hi&+ chemist?y 61,X3-129. 13. MOOG, F., AND GLAZIER, H. S. (1972) Cmp. Biochem Ph@ol 42. 321-336. 14. KOMODA, T., AND SAKAGISHI, Y. (1977) B&&m. Biophys Ada 482.79-84. 15. BROCKMAN, R. W., AND HEPPEL, L. A. (1968) Bi+ ch.emi&q 7.2554-2559. 16. DEPIERRE, J. W., AND KARNOVSKY, M. L. (1974) J. Biol Chem. 249, 7111-7120. 17. Low, M. G., AND ZILVERSMIT, D. B. (1980) Bti chemistry 19,3913-3918. 18. KLEEREKOPER, M., HORNE, M., CORNISH,C. J., AND POSEN, S. (1970) CZin Sci 38, 339-346. 19. SAINI, P. K., AND POSEN, S. (1969) Biochem Bi+ phys. Actu 177,50-57. 20. MATLISZAWA, T., AND ANDERSON, H. C. (1971) J. Hi&&em Cytochem 19, 801-809.