990), 62, Suppi. 62, Suppl. (1990), Br. J. Cancer (I Br. J. Cancer
6 II X, X, 6-11
'." ©
Macmillan Press Ltd., 1990 Macmillan
1990
Purification and characterisation of the placental-like alkaline phosphatase from ovarian epithelial tumours I.
Koyama', K. Hirano2, R. Makiyal, U. Stendahl3 & T. Stigbrand'
Departments of 'Physiological Chemistry and 3Gynecological Oncology, University of Umed, S-901 87 Umed, Sweden and 2Gifu Pharmaceutical University, 5-6-1 Mitahora-higashi, Gifu 502, Japan. Summary The placental alkaline phosphatase was purified by immunoaffinity chromatography from ovarian epithelial tumours to homogeneity. Up to 40% of the catalytical phosphatase activity in these tumours was derived from this placental type alkaline phosphatase (PLAP). The purified enzyme had a subunit molecular mass of 63,500. Inhibition by L-leucine and mobility tests by starch gel electrophoresis revealed the purified enzyme to the PLAP-like enzyme. Several catalytic and immunochemical properties of the enzyme were similar to those of PLAP, whereas the PLAP-like isozyme was more heat-stable and resistant to 8 M urea than PLAP. The amino terminal sequence of the PLAP-like enzyme demonstrated heterogeneity at position three in the N-terminal end compared with PLAP. Phenyl-Sepharose affinity chromatography and different lectin chromatographies demonstrated the tumour-derived enzyme to be microheterogeneous, both with regard to concanavalin A binding and hydrophobicity properties.
Human alkaline phosphatases (EC.3.1.3.1) are highly polymorphic in man and several isozymes have been characterised by both catalytic and immunochemical approaches (McComb et al., 1979). Recently, the cloning of the cDNAs coding for the tissue-unspecific alkaline phosphatase (AP) (Weiss et al., 1986), intestinal alkaline phosphatase (IAP) (Berger et al., 1987) and placental alkaline phosphatase (PLAP) (Kam et al., 1985; Millan, 1986) have been reported. The data revealed 52% homology between the AP and the PLAP sequence, and 86% homology between IAP and PLAP
at the amino acid level. Another isozyme, closely related to PLAP, belonging to the same multigene family has been derived from some tumours (Stigbrand, 1984). The identification of this PLAPlike enzyme has been described by starch gel electrophoresis and sensitivity to inhibition by L-leucine (Nakayama & Kitamura, 1975). The cDNA sequencing of this seminomaderived PLAP-like gene indicated 98% homology between the PLAP and the PLAP-like enzyme (Millan & Manes, 1988), and a very close evolutionary relationship. PLAP or PLAP-like enzymes are of interest as useful tumour markers. Elevated serum levels have been reported in the sera of patients with testicular and ovarian tumours (Fishman, 1987). No report has so far been given on the purification and characterisation of the PLAP-like enzyme from malignant ovarian tumours. In the present paper we report the purification of the PLAP-like enzyme derived from ovarian tumour tissues and some molecular and catalytic properties of the purified enzyme.
Materials and methods
Assay for enzymatic activity Catalytic activity was determined by 5 mM disodium pnitrophenyl phosphate (Sigma Chemicals, St Louis, USA) in 50 mM carbonate-bicarbonate buffer, pH 10.0, supplemented with 5 mm MgCl2 at 37°C unless otherwise indicated. One unit of activity was defined as 1 JLmol of hydrolysed substrate min-' using the molar absorption of p-nitrophenolate of 1.87 x 104 at 405 nm as previously described (Koyama et al., 1983). An alternative assay was also used (Sigma), with 50 mM sodium phenyl phosphate as substrate and the absorbance at 500 nm was measured. Different monoclonal immunocatalytic assays (MICAs) for the different isozymes
Correspondence: T. Stigbrand.
carried out as described (Hirano et al., 1987). The results obtained by the MICA assay are dependent upon both intact immunoreactivity and catalytic activity. Protein content was determined by the method of Lowry et al. (1951) using bovine serum albumin as standard. The AP and IAP isozymes were purified and used as standards in the MICA in order to compare the different catalytic activities. All enzymatic assays were performed in triplicate. were
Purification of the ovarian epithelial tumour PLAP-like enzyme A malignant ovarian epithelial tumour, a serous cystadenocarcinoma, 200 g wet weight, was obtained from a patient undergoing surgery at Umea University Hospital, Sweden. The tissue was homogenised with physiological saline supplemented with 0.3 mM phenylmethylsulphonylfluoride, and extracted with 50% n-butanol for 3 h. The resulting aqueous layer was precipitated by 60% cold acetone to remove the lipid fraction, dissolved in 50mM Tris-HCI buffered saline (TBS), and dialyzed against this buffer. All purification steps were carried out at 4°C. This crude extract
was directly subjected to monoclonal immunoaffinity chromatography. A column with immobilised anti-PLAP monoclonal antibody HPMS-1 (Hirano et al., 1986) on Sepharose (I x 6 cm) was equilibrated with 50 mM TBS, pH 7.5. The crude extract, approximately 200 ml, was applied and after washing with the equilibration buffer the enzyme was eluted with 0.2 M Na2CO3, pH 11.4, containing 0.5 M
NaCI. Fractions were collected in tubes containing 1.0 M Tris-HCI buffer, pH 7.5. Gel filtration of this enzyme fraction was then carried out (1.8 x 70 cm) on a Sephadex G-200 column (1.8 x 70 cm) purchased from Pharmacia Fine Chemicals (Uppsala, Sweden), in order to eliminate the components which were non-specifically adsorbed to the antiPLAP-Sepharose column and eluted by high pH. The eluted fractions containing the enzyme activity were pooled, dialyzed against 10 mm TBS, pH 7.5, and concentrated by ultrafiltration (Diaflo UMIO, MW 10,000). The resulting preparation was stored at -20°C until use. The S and F homozygous variants of PLAP were also purified from placentas from full term pregnancies by the same methods described above in order to compare the different isozymes.
Analytical methods The procedure for 5% polyacrylamide slab gel electrophoresis and SDS-polyacrylamide gel electrophoresis have been described earlier (Laemmli, 1970; Koyama et al., 1988).
ALKALINE PHOSPHATASE FROM OVARIAN EPITHELIAL TUMOURS
Apparent molecular weights were determined by coelectrophoresis of the marker protein purchased from Bio-Rad Laboratories (CA, USA), myosin (200,000), betagalactosidase (116,000), phosphorylase B (92,000), bovine serum albumin (66,000), and ovalbumin (43,000). Starch gel electrophoresis was performed using two different buffer systems, one alkaline (pH 8.6) and one acid (pH 5.6), to obtain full separation of all the phenotypes as described by Beckman et al. (1966). Proteins were stained with Commassie Brilliant Blue, and catalytically active bands of the enzymes were detected with 5 mg ml1 5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt (Sigma) in 1.0 M ammediol buffer, pH 10.2, containing 1 mM MgC12. To evaluate the immunoreactivities of the enzymes, isozyme-specific monoclonal antibodies against AP, IAP and PLAP, designated HLMS-1, 2HIMS-1 and HPMS-1, respectively, were employed (Hirano et al., 1987). These antibodies are not cross-reactive with the other isozymes (Hirano et al., 1986). All other reagents of analytical grade were obtained from KEBO-Lab (Stockholm, Sweden).
Chromatographic studies of the PLAP-like enzyme from ovarian tumours
Phenyl-Sepharose purchased from Pharmacia was equilibrated with 0.1 M sodium acetate diethanolamine HCI buffer, pH 8.2, (buffer A), containing 400 mM ammonium sulphate. The affinity of the PLAP and PLAP-like enzymes derived from the tumours to phenyl-Sepharose was carried out according to the methods of Seiffert et al. (1984) with minor modifications. Elution buffers were used with stepwise elution, containing different concentrations of ammonium sulphate (200, 100, 50, 25 and 0 buffer A) respectively. The isozyme preparation which contained 400 mM ammonium sulphate was applied to the phenyl-Sepharose column (0.5 x 5 cm), which was stepwise eluted by the different buffers. All tubes were assayed for catalytic activities. Recovery of enzyme activities on the column was more than 90%. The different isozymes were also digested with trypsin (Jemmerson & Stigbrand, 1984) or bromelain (Jemmerson et al., 1985). Briefly, 1 mg of trypsin or bromelain (Sigma) was conjugated to 1 ml gel of CNBr-activated Sepharose (Pharmacia). The trypsin-conjugated Sepharose, 250 gAl, was incubated at 37°C for 3 h with the enzyme preparation in 50 M phosphate-buffered saline, pH 8.3. Bromelain-Sepharose was prepared in the same way as trypsin-Sepharose but incubated in 50 mM HEPES, pH 6.8, at 37°C for 3 h instead. Concanavalin A (Con A)-Sepharose (Pharmacia) was used to examine the binding properties of the enzyme to this lectin. The purified enzyme was applied to the Con ASepharose column (0.5 x 5 cm) equilibrated with 10 mM TBS, pH 8.0, containing 1I gM each of MgCI2, CaCl2 and 1O lM ZnC12. After washing with the equilibrating buffer, elution was performed by 0.5 M a-methyl-D-mannoside as described previously (Koyama et al., 1987). Recovery of the enzyme activities from this column was close to 100%.
specific activity for PLAP in the crude extract of 0.03 unitsmg-' of protein as shown in Table I. In the purification procedure, the crude extract was directly subjected to an immunoaffinity chromatography. Figure la shows the elution profile when 10 ml of the extract
a
applied to the anti-PLAP-Sepharose column. The bound fraction accounted for about 40% of the total activity in good agreement with the results of the MICA assays. The eluted fraction from the immunoaffinity column was concentrated and applied on a Sephadex G-200 gel filtration column. Low molecular weight contaminating proteins were removed by this procedure (Figure lb). The purity of the enzyme was significantly increased by these procedures as shown in Table I. The finally obtained enzyme preparation had a specific activity of 138 units mg-' protein, corresponding to a 4,600fold purification factor and was homogeneous as judged from SDS-polyacrylamide gel electrophoresis (Figure 2). The subunit molecular weight of the purified enzyme was determined as 63,500. was
Characterisation of the purified enzyme Evidence that the isozyme derived from the ovarian tumour was the PLAP-like enzyme was obtained from starch gel electrophoresis and sensitivity to inhibition by amino acids. The mobility of the PLAP-like enzyme was compared with that of two common phenotypes, the S and F homozygous variants (Figure 3). The mobility of the enzyme with a broad Table I Purification of the PLAP-like isozyme derived from a malignant ovarian tumour Total Enzyme activity protein Specific activity Total Purification Yield (mg) (units mg-') (units) factor (%) Crude extract 776 0.03 23.2 1.0 100 Purified enzyme 0.056 138 7.7 4.600 33.2 a
01)
E
c Cn
0
-
a) co
U)
a) ._)
n
D .0
0
Co
b
a)
0 .0
i
E
N
I
o.
1
0.2~ o0.010
Results
Purification of the PLAP-like enzyme from ovarian tumour tissue The crude extract from the ovarian tumour had a specific activity of 0.075 units mg-' of protein as measured by use of
p-nitrophenyl phosphate. This crude preparation contains not only PLAP but is also known to contain significant amounts of AP or IAP. The different MICA assays were used to determine the contents of the different isozymes. The intestinal isozyme (IAP) could not be detected in the crude extract. The AP accounted for approximately 60% and the PLAP isozyme for 40% of the total catalytic activity, giving
7
\
,
,
0.005
n)_j
,
,
,, -
10
-ea S *--~~~~~~-.E *,-
20
30
Fraction numbers
Figure 1 Column chromatographies of the PLAP-like isozyme derived from an ovarian tumour. a Chromatography of the crude extract on anti-PLAP-Sepharose. After application of 10ml of the extract, the column was washed with 10 mM TBS until the absorbance of 280 nm was below 0.001. Elution is indicated with an arrow using high pH. b Gel filtration of the enzyme fraction on Sephadex G-200. Elution was carried out at a rate of 10 ml h-' and 2.5 ml fractions were collected.
I. KOYAMA et al.
8
1
2
(X10-3)
200
U-b
116 97
3-*. U-b'
c
66 U-b_
0
I.0O
43 a_
u10
1001 0
50Figure 2 SDS-polyacrylamide gel electrophoresis of the purified 7.5% polyacrylamide gel was used and the gel was stained with Coomassie Brilliant Blue. Lane 1, marker proteins; lane 2, the purified enzyme. enzyme.
_
0
0-
0.1
1.0
10
Inhibitors (mM)
J
Figure 4 Inhibitory effects of a L-homoarginine, b Lphenylalanine and c L-leucine on the catalytic activities. The reaction was carried out using p-nitrophenyl phosphate as substrate in the presence of the different amino acids, respectively. 0, PLAP; *, the purified PLAP-like enzyme; A, IAP; A, AP.
k
1
2
3
Figure 3 Starch gel electrophoresis (pH 8.6) of PLAP and the PLAP-like enzyme. The gel was stained for catalytic activity. Lanes and 2, the F and S variants, respectively; lane 3, the purified PLAP-like enzyme.
band revealing microheterogeneity was more retarded than the S variant, and the PLAP-like enzyme could not be classified in any of the normal placental phenotypes. Figure 4 shows the inhibition of catalytic activity by Lhomoarginine, L-phenylalanine and L-leucine, respectively, and the PLAP-like isozyme was compared with the PLAP and AP isozymes. The AP isozyme is sensitive to inhibition by L-homoarginine, and PLAP comparatively resistant, whereas the purified PLAP-like enzyme was more sensitive to inhibition by L-homoarginine than PLAP (Figure 4). There were no significant differences in the inhibition by Lphenylalanine for the PLAP and PLAP-like enzymes (Figure 4b). The PLAP-like enzyme, however, was more sensitive to inhibition by L-leucine than PLAP (Figure 4c), but less sensitive to inhibition by the peptide L-Leu-Gly-Gly or L-PheGly-Gly which inhibited the PLAP activity (Jeppson et al., 1984) (results not shown). These results taken together justify
the conclusion that the purified enzyme from this malignant ovarian tumour is the PLAP-like isozyme. The catalytic and physicochemical properties of the PLAPlike enzyme are summarised in Table II. The F variant of PLAP was used for comparison with the PLAP-like enzyme. The Km values were determined by use of p-nitrophenyl phosphate as substrate. The pH optima were determined in 100 mM carbonate-bicarbonate buffer at a pH range of 9.6-10.5, containing 3 mM p-nitrophenyl phosphate and I mM MgCl2. Although the Km values and optimal pH values Table II Catalytic and physicochemical properties of PLAP and the PLAP-like isozyme PLAP PLAP-like enzyme Km value (mM) 3.2 3.0 (for pNPP) Optimal pH 9.9-10.0 9.9-10.0
(for pNPP) Substrate specificity 1.48 1.32 (pNPP/PhP ratio) Remaining activity (%) 96 108 after exposure to heat (56'C, 30 min) EDTA (l mM) 17 86 Urea (8M) 75 118 Reactivity (%) against anti-PLAP 100 100 anti-IAP 0 0 anti-AP 0 0 Binding affinity 100 82 for Con A (%) pNPP, p-nitrophenyl phosphate; PhP, phenyl phosphate.
ALKALINE PHOSPHATASE FROM OVARIAN EPITHELIAL TUMOURS
showed no differences for these enzymes, the substrate specificity in terms of p-nitrophenyl phosphate/phenyl phosphate ratio differed slightly. The PLAP-like enzyme from the ovarian tumour, however, exhibited a pronounced heat stability. Preincubation of the enzyme preparations was performed at 56°C for 30 min with 1 mg ml-' bovine serum albumin and 1 mM MgCl2. PLAP is a heat-stable enzyme with 96% remaining activity after heat treatment for 30 min. The activity of the PLAP-like enzyme was above 100% of the control. Effects of ethylenediamine tetraacetic acid (EDTA) or urea on the PLAP-like catalytic activity are also shown in Table II. Since preincubation of all alkaline phosphatases with EDTA leads to a time-dependent denaturation (Congers et al., 1967), both the enzymes were appropriately diluted to unify the catalytic activity. Preincubation for exactly 5 min was performed with EDTA at a final concentration of 1 mM. The remaining catalytic activities were measured using pnitrophenylphosphate without MgCl2. Preincubation with 1 mM EDTA significantly reduced the catalytic activity of PLAP, while the PLAP-like enzyme was comparatively resistant .o EDTA. The remaining activity was 42% of the control even in the presence of 5 mM EDTA (results not shown). Preincubation of alkaline phosphatases with urea is known to result in a time-dependent denaturation (Birkett et al., 1967). Both enzymes with unified catalytic activity were preincubated with 8 M urea for 30 min. PLAP was slightly inactivated by this preincubation with 8 M urea, whereas the PLAP-like enzyme demonstrated a slightly increased catalytic activity, again demonstrating the stability of the PLAP-like isozyme. In order to describe the immunoreactivity of the PLAP-like enzyme, the isozyme specific MICAs were employed. Both PLAP and the PLAP-like enzymes were reactive with the anti-PLAP monoclonal antibody, not to anti-AP or anti-IAP, indicating that the purified enzyme does not contain any AP which is the dominating isozyme in the ovary (Table II). In order to characterise the microheterogeneity of the PLAP isozyme with the PLAP-like enzyme, phenyl-Sepharose affinity chromatography was performed and the chromatographic patterns are shown in Figure 5. Heterogeneity of the isozymes in this affinity chromatography are dependent upon hydrophobicity differences in the molecules (Seiffert et al., 1984). The catalytic PLAP activities were mainly obtained in the 100mM ammonium sulphate buffer (Figure 5a). In contrast, the PLAP-like enzyme demonstrated, besides the 100mM fraction, also minor fractions eluted by 25, OmM ammonium sulphate and detergent buffer. No differences could be identified in the elution pattern between intact PLAP and PLAP treated with trypsin, whereas the pattern of the PLAP-like enzyme treated with trypsin was changed (Figure Sb) and minor fractions eluted by low concentrations of ammonium sulphate were no longer identified and the pattern was similar to that of PLAP. Treatment with bromelain changed the patterns of both PLAP and the PLAP-like enzyme from their precursor isozymes (Figure Sc). The minor fractions in the PLAP-like preparation were still detected as well as the intact major PLAP-like isozyme. Thus, both PLAP and the PLAP-like isozyme demonstrated heterogeneities, and treatment of the PLAP-like enzyme with trypsin but not with bromelain made the PLAP-like enzyme behave similarly to PLAP in the phenyl-Sepharose affinity chromatography. The N-terminal amino acid sequences up to the 16th residue of the PLAP-like enzyme from the ovarian tumour were determined. The results were as follows: Ile-Ile-Pro (Leu)-Val-X-X-X-Asn-Pro-Asp-Phe-Trp-Asn-Ar-Glu-Ala. Although the third amino acid is Pro for variant F and S, and Leu for variant I (Henthorn et al., 1986), the PLAP-like enzyme exhibited both Pro and Leu in a ratio 1:1. Except for the third residue, there was no difference in sequence between the PLAP and PLAP-like isozyme (Henthorn et al., 1986). Finally, the PLAP-like isozyme was subjected to Con A chromatography to evaluate the binding properties to this lectin (Table II). PLAP was bound 100% to the Con A,
9
a
400
200
Ammonium sulphate (mM) 50 100 25 0
Detergent
0.4
0.2
O,
* iv
;
E -1
v L.Q
lq
b
CD
to
6
6 oe
O nnt n15Be n9ooo
0.4
0
8
CU
o 0.2
E
N
w
* 00 n -0 P.
t
n
1,
,"a0b
N \. (F &
(lb
O.-.a c
0.4 0.2 \1
uj
r] v Ion
n
nno
ooooQQ crsnannooo,oss
lb
\
1; o1°!'°a~-v0voooonn.oono *..\* -e .
20
30
Og
40
Fraction number
Figure 5 Affinity chromatographies of different preparations of PLAP and PLAP-like isozymes on phenyl-Sepharose. The enzyme preparation was applied on the column and eluted stepwise with different concentrations of ammonium sulphate as described in the text; a intact enzymes; b enzymes digested with trypsin; c enzymes digested with bromelain. 0, PLAP; *, the PLAP-like enzyme.
whereas 18% of the PLAP-like isozyme did not bind to the lectin. Furthermore, polyacrylamide gel electrophoresis of both enzymes digested with neuroaminidase exhibited differences in mobility of these isozymes, indicating that the sensitivity to neuraminidase of PLAP and PLAP-like enzyme is different (results not shown). Discussion Human PLAP was one of the first oncofetal antigens found
to be ectopically expressed by tumour cells and this antigen is now established as a useful tumour marker for gonadal
malignancies (Fishman, 1987). The isozyme derived from tumours demonstrates identifiable but discrete differences compared with the isozyme derived from placenta. The expression of heat-stable alkaline phosphatases or determination by radioimmunoassays of PLAP in sera from patients with ovarian tumours has confirmed the synthesis of isozymes derived also from ovarian tumours (Mano et al., 1986; Haije et al., 1987). To our knowledge, no report on the purification of the PLAP-like enzyme from ovarian tumour tissues has been presented. The quantitatively dominant form of alkaline phosphatase in the ovary is the 'tissue-unspecific' alkaline phosphatase and minor but significant amounts of PLAP are present in the malignant ovary. In the present study, monoclonal immunoaffinity chromatography was used as the first step to purify the PLAP-like enzyme from the ovarian tumour tissue. The yield of the enzyme was approximately 33% and the enzyme was purified to homogeneity as revealed by SDS-polyacrylamide gel electrophoresis. The purified enzyme was confirmed to be the PLAP-like enzyme from the results of both starch gel electrophoresis and sensitivity to inhibition by L-leucine. The catalytic pro-
10
I. KOYAMA et al.
perties were similar to those of PLAP, but deviations in heat stability were observed for the PLAP-like enzyme. The Nagao isozyme is sensitive to EDTA (Nakayama & Kitamura, 1975), but Cox et al. (1971) indicated that the steroid-induced alkaline phosphatase in HeLa 65 cells retained catalytic activity in the presence of EDTA compared with the base levels of enzyme, an observation which may be related to the present findings. PLAP is known to be resistant to EDTA and urea when compared with AP and IAP (Birkett et al., 1967; Congers et al., 1967). The PLAP-like enzyme from the ovarian tumour, however, was more resistant to these inhibitors than PLAP. These findings reflect biochemical differences between these closely related isozymes. The hydrophobic characteristics of the protein were studied by affinity chromatography on phenylSepharose for evaluation of the heterogeneities of the purified enzyme. The elution profile of the PLAP-like enzyme on the phenyl-Sepharose column was different from that of PLAP, demonstrating differences in hydrophobic properties. Treatment with trypsin of the PLAP-like enzyme gave identical elution patterns with PLAP. Trypsin is known to cleave native PLAP at amino acid residue 62 from the amino terminal end without affecting the catalytic activity of the enzyme (Jemmerson & Stigbrand, 1984). Bromelain cleaves a 2,000 Da hydrophobic fragment from the C-terminal of the protein, a fragment which includes the membrane-anchoring domain (Jemmerson et al., 1985). The digestion with bromelain did not affect the observed difference of the PLAP-like and PLAP isozymes. These results suggest that the differences in hydrophobic heterogeneity between PLAP and the PLAP-like enzyme are related to the N-terminal and not to the C-terminal part of the molecule. The primary structure of the first 16 residues of the PLAPlike enzyme was determined. No exchanges of amino acids which could explain the differences in hydrophobicity were identified. However, the third amino acid of the PLAP-like enzyme from a single tumour revealed Pro and Leu in a molar ratio of 1:1. Residue no. 3 in PLAP is Pro for the variants F and S, and Leu for variant I (Henthorn et al., 1986). Furthermore, the PLAP-like enzyme derived from seminomas have Gln at residue 15 instead of Glu as is the case for PLAP (Millan & Manes, 1988). Residue no. 15 in the PLAP-like enzyme from the ovarian tumour was Glu which is similar to PLAP. The presence of equimolar amounts of both Pro and Leu at residue no. 3 is intriguing and may reflect either the simultaneous existence of heterodimers of the F or S and I variants or the existence of an even more pronounced heterogeneity at the PLAP locus,
not detectable by electrophoretic techniques. Differences in the glycosylation of the oligosaccharide chains in glycoproteins may contribute to the structural differences identified. It is well known that structural changes of the asparagine-linked oligosaccharide chains in many glycoproteins are partially modified following malignant transformation (Ogata et al., 1976; Tamashita et al., 1983). Digestion of the PLAP-like enzyme with neuraminidase yielded a somewhat different mobility pattern at polyacrylamide gel electrophoresis compared with PLAP. Furthermore, the tumour-derived PLAP-like enzyme contained a fraction which was not binding to Con A, which is clearly different from PLAP. It is concluded that the PLAP-like enzyme is at least partially differently glycosylated than is PLAP. The mechanisms behind this observation can be speculated upon. The unbound fraction may contain bisecting N-acetylglucosamine at the asparagine-linked complextype sugar chains as described for a-fetoprotein obtained from transformed cells which also contains a fraction not binding to Con A, whereas a-fetoprotein from untransformed cells is completely bound to Con A (Yoshima et al., 1980). PLAP purified from term placenta has mainly biantennary complex-type sugar chains (Endo et al., 1988). From the present results it can be inferred that the PLAP-like enzyme from ovarian tumours contains a bisected complex-type oligosaccharide chain. Recent reports indicate the membrane-anchoring domain of PLAP to be a glycosylphospholipid constituting the membrane anchor of PLAP (Ogata et al., 1988). Previous reports have demonstrated that the attachment of the enzyme to tumour membrane fragments is somewhat different from that observed in placental membranes (Taylor et al., 1980), although the chemical nature of the membrane anchor of the PLAP-like enzyme is unknown. These results are of significant clinical impact since the appearance in the circulation of tumour-derived enzymes detected by different immunoassays serves as useful tumour markers. Furthermore, it may be possible to generate monoclonal antibodies capable of discriminating between PLAP and the PLAP-like isozyme which could increase the specificity in the monitoring of some malignant conditions. We are grateful to Dr T. Komoda and Professor Y. Sakagishi, Saitama Medical School, Saitama, Japan who encouraged us throughout the course of this work. This work was supported by the Swedish Cancer Research Council (Project No. 1387), Lion's Foundation, Umea and the Medical Faculty, University of Umea.
References
BECKMAN, L., BJORLING, G. & CHRISTODOULOU, G. (1966). Pregnancy enzyme and placental polymorphism. 1. Alkaline phosphatase. Acta Genet., 16, 59. BERGER, J., GARANTTINI, E., HUA, J.-C. & UDENFRIEND, S. (1987).
Cloning and sequence of human intestinal alkaline phosphatase cDNA. Proc. Natl Acad. Sci. USA, 84, 695.
BIRKETT, D.J., CONGERS, R.A.J., NEALS, F.C., POSEN, S. &
BRUDENELL-WOODS, J. (1967). Action of urea on human alkaline phosphatases: With a description of some automated techniques for the study of enzyme kinetics. Arch. Biochem. Biophys., 121, 470. CONGERS, R.A.J., BIRKETT, D.J., NEALS, F.C., POSEN, S. & BRUDENELL-WOODS, J. (1967). The action of EDTA on human alkaline phosphatase. Biochim. Biophys. Acta, 139, 363. COX, R.P., ELSON, N.A., TU, S.H. & GRIFFIN, M.J. (1971). Hormonal induction of alkaline phosphatase by an increase in catalytic efficiency of the enzyme. J. Mol. Biol., 58, 197. ENDO, T., OHBAYASHI, H., HAYASHI, Y., IKEHARA, Y., KOCHIBE,
N. & KOBATA, A. (1988). Structural studies on the carbohydrate moiety of human placental alkaline phosphatase. J. Biochem., 103, 182. FISHMAN, W.H. (1987). Oncotrophoblast gene expression. Placental alkaline phosphatase. Adv. Cancer Res., 48, 1.
HAIJE, W.G., VAN DRIEL, J. & VAN DER BURG, M.E. (1987).
Catalytic and immunologic activities of placental-like alkaline phosphatase in clinical studies. The value of PLAP in follow-up of ovarian cancer. Clin. Chim. Acta, 165, 165. HENTHORN, P.S., KNOLL, B.J., RADUCHA, M. & 6 others (1986). Products of two common alleles at the locus for human placental alkaline phosphatase differ by seven amino acids. Proc. Natl Acad. Sci. USA, 83, 5597. HIRANO, K., IIZUMI, Y., HAYASHI, Y. & 5 others (1986). A highly sensitive assay method for human placental alkaline phosphatase involving a monoclonal antibody bound to a paper disc. Anal. Biochem., 154, 624. HIRANO, K., MATSUMOTO, H., TANAKA, T., IINO, S., DOMAR, U. &
STIGBRAND, T. (1987). Specific assays for human alkaline phosphatase isozyme. Clin. Chim. Acta, 166, 265. JEMMERSON, R. & STIGBRAND, T. (1984). Monoclonal antibodies block the trypsin cleavage site on human placental alkaline phosphatase. FEBS Lett., 173, 357. JEMMERSON, R., MILLAN, J.L., KLIER, F.G. & FISHMAN, W.H. (1985). Monoclonal antibodies block the bromelain-mediated release of human placental alkaline phosphatase from cultured cancer cells. FEBS Lett., 179, 316.
ALKALINE PHOSPHATASE FROM OVARIAN EPITHELIAL TUMOURS JEPPSON, A., WAHREN, B., BREHMER-ANDERSSON, E., SILF-
VERSWARD, C., STIGBRAND, T. & MILLAN, J.L. (1984). Eutopic expression of placental-like alkaline phosphatase in testicular tumors. Int. J. Cancer, 34, 757. KAM, W., CLAUSER, E., KIM, Y.S., KAN, Y.W. & RUTTER, W.J.
(1985). Cloning, sequencing and chromosomal localization of human term placental alkaline phosphatase cDNA. Proc. Natl Acad. Sci. USA, 82, 8715. KOYAMA, I., ARAI, Y., MUIRA, M., MATSUZAKI, H., SAKAGISHI, Y.
& KOMODA, T. (1988). Similarity of the sugar moiety of human alkaline phosphatases between kidney cortex and duodenum, or medulla and ileum. Clin. Chim. Acta., 179, 139. KOYAMA, I., KOMODA, T., SAKAGISHI, Y. & KURATA, M. (1983). A possible mechanism for the changes in hepatic and intestinal alkaline phosphatase activity in bile-duct-ligated rats or guinea pigs. Biochim. Biophys. Acta, 760, 169. KOYAMA, I., MUIRA, M., MATSUZAKI, H., SAKAGISHI, Y. & KOMODA, T. (1987). Sugar-chain heterogeneities of human alkaline phosphatases. Difference between normal and tumorassociated isozymes. J. Chromatogr., 413, 65. LAEMMLI, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680. LOWRY, O.H., ROSENBROUGH, N.J., FARR, A.L. & RANDALL, R.J.
(1951). Protein measurement with the Folin phenol reagent. J. Biol. Chem., 192, 265. MANO, H., FURUHASHI, Y., MORIYAMA, Y., HATTORI, E., GOTO, S.
& TOMODA, T. (1986). Radioimmunoassay of placental alkaline phosphatase in ovarian cancer and tissues. Obstet. Gynecol., 68, 759. MCCOMB, R.B., BOWERS, Jr, G.N. & POSEN, S. (1979). Alkaline Phosphatases, p. 373. Plenum Press: New York. MILLAN, J.L. (1986). Molecular cloning and sequence analysis of human placental alkaline phosphatase. J. Biol. Chem., 261, 3112. MILLAN, J.L. & MANES, T. (1988). Seminoma-derived Nagao isozyme is encoded by a germ-cell alkaline phosphatase gene. Proc. Natl Acad. Sci. USA, 85, 3024.
11
NAKAYAMA, T. & KITAMURA, M. (1975). L-Leucine sensitive alkaline phosphatase isozyme from cancer tissue. Ann. N. Y. Acad. Sci., 259, 325. OGATA, S., HAYASHI, Y., TAKAMI, N. & IKEHARA, Y. (1988). Chemical characterization of the membrane-anchoring domain of
human placental alkaline phosphatase. J. Biol. Chem., 263, 10489. OGATA, S., MURAMATSU, T. & KOBATA, A. (1976). New structural characteristics of the large glycopeptides from transformed cells. Nature, 259, 580. SEIFFERT, U.B., SIEDE, W.H., WELSH, G.J. & OREMEK, G. (1984). Multiple forms of alkaline phosphatases in human liver tissue. Clin. Chim. Acta, 144, 17. STIGBRAND, T. (1984). Present status and future trends of human alkaline phosphatases. In Human Alkaline Phosphatases, Stigbrand, T. & Fishman, W. (eds) p. 3. Alan R. Liss Inc.: New York.
TAMASHITA, K., HITOI, A., TANIGUCHI, N., YAKASAWA, N.,
TSUKADA, T. & KOBATA, A. (1983). Comparative study of the sugar chains of y-glutamyl transpeptidase purified from rat liver and rat AH-66 cells. Cancer Res., 43, 5059. TAYLOR, D.D., HOMESLEY, H.D. & DOELLGAST, G.J. (1980). Binding of specific peroxidase-labeled antibody to placental-type phosphatase on tumour-derived membrane fragment. Cancer Res., 40, 4064. WEISS, M.J., HENTHORN, P.S., LAFFERTY, M.A., SLAUGHTER, C.,
RADUCHA, M. & HARRIS, H. (1986). Isolation and characterization of a cDNA encoding a human liver/bone/kidney-type alkaline phosphatase. Proc. Natl Acad. Sci. USA, 83, 7182. YOSHIMA, H., MIZUOCHI, TISHII, M. & KOBATA, A. (1980). Structure of the asparagine-linked sugar chains of a-fetoprotein purified from ascites. Cancer Res., 40, 4276.
6 II X, X, 6-11
'." ©
Macmillan Press Ltd., 1990 Macmillan
1990
Purification and characterisation of the placental-like alkaline phosphatase from ovarian epithelial tumours I.
Koyama', K. Hirano2, R. Makiyal, U. Stendahl3 & T. Stigbrand'
Departments of 'Physiological Chemistry and 3Gynecological Oncology, University of Umed, S-901 87 Umed, Sweden and 2Gifu Pharmaceutical University, 5-6-1 Mitahora-higashi, Gifu 502, Japan. Summary The placental alkaline phosphatase was purified by immunoaffinity chromatography from ovarian epithelial tumours to homogeneity. Up to 40% of the catalytical phosphatase activity in these tumours was derived from this placental type alkaline phosphatase (PLAP). The purified enzyme had a subunit molecular mass of 63,500. Inhibition by L-leucine and mobility tests by starch gel electrophoresis revealed the purified enzyme to the PLAP-like enzyme. Several catalytic and immunochemical properties of the enzyme were similar to those of PLAP, whereas the PLAP-like isozyme was more heat-stable and resistant to 8 M urea than PLAP. The amino terminal sequence of the PLAP-like enzyme demonstrated heterogeneity at position three in the N-terminal end compared with PLAP. Phenyl-Sepharose affinity chromatography and different lectin chromatographies demonstrated the tumour-derived enzyme to be microheterogeneous, both with regard to concanavalin A binding and hydrophobicity properties.
Human alkaline phosphatases (EC.3.1.3.1) are highly polymorphic in man and several isozymes have been characterised by both catalytic and immunochemical approaches (McComb et al., 1979). Recently, the cloning of the cDNAs coding for the tissue-unspecific alkaline phosphatase (AP) (Weiss et al., 1986), intestinal alkaline phosphatase (IAP) (Berger et al., 1987) and placental alkaline phosphatase (PLAP) (Kam et al., 1985; Millan, 1986) have been reported. The data revealed 52% homology between the AP and the PLAP sequence, and 86% homology between IAP and PLAP
at the amino acid level. Another isozyme, closely related to PLAP, belonging to the same multigene family has been derived from some tumours (Stigbrand, 1984). The identification of this PLAPlike enzyme has been described by starch gel electrophoresis and sensitivity to inhibition by L-leucine (Nakayama & Kitamura, 1975). The cDNA sequencing of this seminomaderived PLAP-like gene indicated 98% homology between the PLAP and the PLAP-like enzyme (Millan & Manes, 1988), and a very close evolutionary relationship. PLAP or PLAP-like enzymes are of interest as useful tumour markers. Elevated serum levels have been reported in the sera of patients with testicular and ovarian tumours (Fishman, 1987). No report has so far been given on the purification and characterisation of the PLAP-like enzyme from malignant ovarian tumours. In the present paper we report the purification of the PLAP-like enzyme derived from ovarian tumour tissues and some molecular and catalytic properties of the purified enzyme.
Materials and methods
Assay for enzymatic activity Catalytic activity was determined by 5 mM disodium pnitrophenyl phosphate (Sigma Chemicals, St Louis, USA) in 50 mM carbonate-bicarbonate buffer, pH 10.0, supplemented with 5 mm MgCl2 at 37°C unless otherwise indicated. One unit of activity was defined as 1 JLmol of hydrolysed substrate min-' using the molar absorption of p-nitrophenolate of 1.87 x 104 at 405 nm as previously described (Koyama et al., 1983). An alternative assay was also used (Sigma), with 50 mM sodium phenyl phosphate as substrate and the absorbance at 500 nm was measured. Different monoclonal immunocatalytic assays (MICAs) for the different isozymes
Correspondence: T. Stigbrand.
carried out as described (Hirano et al., 1987). The results obtained by the MICA assay are dependent upon both intact immunoreactivity and catalytic activity. Protein content was determined by the method of Lowry et al. (1951) using bovine serum albumin as standard. The AP and IAP isozymes were purified and used as standards in the MICA in order to compare the different catalytic activities. All enzymatic assays were performed in triplicate. were
Purification of the ovarian epithelial tumour PLAP-like enzyme A malignant ovarian epithelial tumour, a serous cystadenocarcinoma, 200 g wet weight, was obtained from a patient undergoing surgery at Umea University Hospital, Sweden. The tissue was homogenised with physiological saline supplemented with 0.3 mM phenylmethylsulphonylfluoride, and extracted with 50% n-butanol for 3 h. The resulting aqueous layer was precipitated by 60% cold acetone to remove the lipid fraction, dissolved in 50mM Tris-HCI buffered saline (TBS), and dialyzed against this buffer. All purification steps were carried out at 4°C. This crude extract
was directly subjected to monoclonal immunoaffinity chromatography. A column with immobilised anti-PLAP monoclonal antibody HPMS-1 (Hirano et al., 1986) on Sepharose (I x 6 cm) was equilibrated with 50 mM TBS, pH 7.5. The crude extract, approximately 200 ml, was applied and after washing with the equilibration buffer the enzyme was eluted with 0.2 M Na2CO3, pH 11.4, containing 0.5 M
NaCI. Fractions were collected in tubes containing 1.0 M Tris-HCI buffer, pH 7.5. Gel filtration of this enzyme fraction was then carried out (1.8 x 70 cm) on a Sephadex G-200 column (1.8 x 70 cm) purchased from Pharmacia Fine Chemicals (Uppsala, Sweden), in order to eliminate the components which were non-specifically adsorbed to the antiPLAP-Sepharose column and eluted by high pH. The eluted fractions containing the enzyme activity were pooled, dialyzed against 10 mm TBS, pH 7.5, and concentrated by ultrafiltration (Diaflo UMIO, MW 10,000). The resulting preparation was stored at -20°C until use. The S and F homozygous variants of PLAP were also purified from placentas from full term pregnancies by the same methods described above in order to compare the different isozymes.
Analytical methods The procedure for 5% polyacrylamide slab gel electrophoresis and SDS-polyacrylamide gel electrophoresis have been described earlier (Laemmli, 1970; Koyama et al., 1988).
ALKALINE PHOSPHATASE FROM OVARIAN EPITHELIAL TUMOURS
Apparent molecular weights were determined by coelectrophoresis of the marker protein purchased from Bio-Rad Laboratories (CA, USA), myosin (200,000), betagalactosidase (116,000), phosphorylase B (92,000), bovine serum albumin (66,000), and ovalbumin (43,000). Starch gel electrophoresis was performed using two different buffer systems, one alkaline (pH 8.6) and one acid (pH 5.6), to obtain full separation of all the phenotypes as described by Beckman et al. (1966). Proteins were stained with Commassie Brilliant Blue, and catalytically active bands of the enzymes were detected with 5 mg ml1 5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt (Sigma) in 1.0 M ammediol buffer, pH 10.2, containing 1 mM MgC12. To evaluate the immunoreactivities of the enzymes, isozyme-specific monoclonal antibodies against AP, IAP and PLAP, designated HLMS-1, 2HIMS-1 and HPMS-1, respectively, were employed (Hirano et al., 1987). These antibodies are not cross-reactive with the other isozymes (Hirano et al., 1986). All other reagents of analytical grade were obtained from KEBO-Lab (Stockholm, Sweden).
Chromatographic studies of the PLAP-like enzyme from ovarian tumours
Phenyl-Sepharose purchased from Pharmacia was equilibrated with 0.1 M sodium acetate diethanolamine HCI buffer, pH 8.2, (buffer A), containing 400 mM ammonium sulphate. The affinity of the PLAP and PLAP-like enzymes derived from the tumours to phenyl-Sepharose was carried out according to the methods of Seiffert et al. (1984) with minor modifications. Elution buffers were used with stepwise elution, containing different concentrations of ammonium sulphate (200, 100, 50, 25 and 0 buffer A) respectively. The isozyme preparation which contained 400 mM ammonium sulphate was applied to the phenyl-Sepharose column (0.5 x 5 cm), which was stepwise eluted by the different buffers. All tubes were assayed for catalytic activities. Recovery of enzyme activities on the column was more than 90%. The different isozymes were also digested with trypsin (Jemmerson & Stigbrand, 1984) or bromelain (Jemmerson et al., 1985). Briefly, 1 mg of trypsin or bromelain (Sigma) was conjugated to 1 ml gel of CNBr-activated Sepharose (Pharmacia). The trypsin-conjugated Sepharose, 250 gAl, was incubated at 37°C for 3 h with the enzyme preparation in 50 M phosphate-buffered saline, pH 8.3. Bromelain-Sepharose was prepared in the same way as trypsin-Sepharose but incubated in 50 mM HEPES, pH 6.8, at 37°C for 3 h instead. Concanavalin A (Con A)-Sepharose (Pharmacia) was used to examine the binding properties of the enzyme to this lectin. The purified enzyme was applied to the Con ASepharose column (0.5 x 5 cm) equilibrated with 10 mM TBS, pH 8.0, containing 1I gM each of MgCI2, CaCl2 and 1O lM ZnC12. After washing with the equilibrating buffer, elution was performed by 0.5 M a-methyl-D-mannoside as described previously (Koyama et al., 1987). Recovery of the enzyme activities from this column was close to 100%.
specific activity for PLAP in the crude extract of 0.03 unitsmg-' of protein as shown in Table I. In the purification procedure, the crude extract was directly subjected to an immunoaffinity chromatography. Figure la shows the elution profile when 10 ml of the extract
a
applied to the anti-PLAP-Sepharose column. The bound fraction accounted for about 40% of the total activity in good agreement with the results of the MICA assays. The eluted fraction from the immunoaffinity column was concentrated and applied on a Sephadex G-200 gel filtration column. Low molecular weight contaminating proteins were removed by this procedure (Figure lb). The purity of the enzyme was significantly increased by these procedures as shown in Table I. The finally obtained enzyme preparation had a specific activity of 138 units mg-' protein, corresponding to a 4,600fold purification factor and was homogeneous as judged from SDS-polyacrylamide gel electrophoresis (Figure 2). The subunit molecular weight of the purified enzyme was determined as 63,500. was
Characterisation of the purified enzyme Evidence that the isozyme derived from the ovarian tumour was the PLAP-like enzyme was obtained from starch gel electrophoresis and sensitivity to inhibition by amino acids. The mobility of the PLAP-like enzyme was compared with that of two common phenotypes, the S and F homozygous variants (Figure 3). The mobility of the enzyme with a broad Table I Purification of the PLAP-like isozyme derived from a malignant ovarian tumour Total Enzyme activity protein Specific activity Total Purification Yield (mg) (units mg-') (units) factor (%) Crude extract 776 0.03 23.2 1.0 100 Purified enzyme 0.056 138 7.7 4.600 33.2 a
01)
E
c Cn
0
-
a) co
U)
a) ._)
n
D .0
0
Co
b
a)
0 .0
i
E
N
I
o.
1
0.2~ o0.010
Results
Purification of the PLAP-like enzyme from ovarian tumour tissue The crude extract from the ovarian tumour had a specific activity of 0.075 units mg-' of protein as measured by use of
p-nitrophenyl phosphate. This crude preparation contains not only PLAP but is also known to contain significant amounts of AP or IAP. The different MICA assays were used to determine the contents of the different isozymes. The intestinal isozyme (IAP) could not be detected in the crude extract. The AP accounted for approximately 60% and the PLAP isozyme for 40% of the total catalytic activity, giving
7
\
,
,
0.005
n)_j
,
,
,, -
10
-ea S *--~~~~~~-.E *,-
20
30
Fraction numbers
Figure 1 Column chromatographies of the PLAP-like isozyme derived from an ovarian tumour. a Chromatography of the crude extract on anti-PLAP-Sepharose. After application of 10ml of the extract, the column was washed with 10 mM TBS until the absorbance of 280 nm was below 0.001. Elution is indicated with an arrow using high pH. b Gel filtration of the enzyme fraction on Sephadex G-200. Elution was carried out at a rate of 10 ml h-' and 2.5 ml fractions were collected.
I. KOYAMA et al.
8
1
2
(X10-3)
200
U-b
116 97
3-*. U-b'
c
66 U-b_
0
I.0O
43 a_
u10
1001 0
50Figure 2 SDS-polyacrylamide gel electrophoresis of the purified 7.5% polyacrylamide gel was used and the gel was stained with Coomassie Brilliant Blue. Lane 1, marker proteins; lane 2, the purified enzyme. enzyme.
_
0
0-
0.1
1.0
10
Inhibitors (mM)
J
Figure 4 Inhibitory effects of a L-homoarginine, b Lphenylalanine and c L-leucine on the catalytic activities. The reaction was carried out using p-nitrophenyl phosphate as substrate in the presence of the different amino acids, respectively. 0, PLAP; *, the purified PLAP-like enzyme; A, IAP; A, AP.
k
1
2
3
Figure 3 Starch gel electrophoresis (pH 8.6) of PLAP and the PLAP-like enzyme. The gel was stained for catalytic activity. Lanes and 2, the F and S variants, respectively; lane 3, the purified PLAP-like enzyme.
band revealing microheterogeneity was more retarded than the S variant, and the PLAP-like enzyme could not be classified in any of the normal placental phenotypes. Figure 4 shows the inhibition of catalytic activity by Lhomoarginine, L-phenylalanine and L-leucine, respectively, and the PLAP-like isozyme was compared with the PLAP and AP isozymes. The AP isozyme is sensitive to inhibition by L-homoarginine, and PLAP comparatively resistant, whereas the purified PLAP-like enzyme was more sensitive to inhibition by L-homoarginine than PLAP (Figure 4). There were no significant differences in the inhibition by Lphenylalanine for the PLAP and PLAP-like enzymes (Figure 4b). The PLAP-like enzyme, however, was more sensitive to inhibition by L-leucine than PLAP (Figure 4c), but less sensitive to inhibition by the peptide L-Leu-Gly-Gly or L-PheGly-Gly which inhibited the PLAP activity (Jeppson et al., 1984) (results not shown). These results taken together justify
the conclusion that the purified enzyme from this malignant ovarian tumour is the PLAP-like isozyme. The catalytic and physicochemical properties of the PLAPlike enzyme are summarised in Table II. The F variant of PLAP was used for comparison with the PLAP-like enzyme. The Km values were determined by use of p-nitrophenyl phosphate as substrate. The pH optima were determined in 100 mM carbonate-bicarbonate buffer at a pH range of 9.6-10.5, containing 3 mM p-nitrophenyl phosphate and I mM MgCl2. Although the Km values and optimal pH values Table II Catalytic and physicochemical properties of PLAP and the PLAP-like isozyme PLAP PLAP-like enzyme Km value (mM) 3.2 3.0 (for pNPP) Optimal pH 9.9-10.0 9.9-10.0
(for pNPP) Substrate specificity 1.48 1.32 (pNPP/PhP ratio) Remaining activity (%) 96 108 after exposure to heat (56'C, 30 min) EDTA (l mM) 17 86 Urea (8M) 75 118 Reactivity (%) against anti-PLAP 100 100 anti-IAP 0 0 anti-AP 0 0 Binding affinity 100 82 for Con A (%) pNPP, p-nitrophenyl phosphate; PhP, phenyl phosphate.
ALKALINE PHOSPHATASE FROM OVARIAN EPITHELIAL TUMOURS
showed no differences for these enzymes, the substrate specificity in terms of p-nitrophenyl phosphate/phenyl phosphate ratio differed slightly. The PLAP-like enzyme from the ovarian tumour, however, exhibited a pronounced heat stability. Preincubation of the enzyme preparations was performed at 56°C for 30 min with 1 mg ml-' bovine serum albumin and 1 mM MgCl2. PLAP is a heat-stable enzyme with 96% remaining activity after heat treatment for 30 min. The activity of the PLAP-like enzyme was above 100% of the control. Effects of ethylenediamine tetraacetic acid (EDTA) or urea on the PLAP-like catalytic activity are also shown in Table II. Since preincubation of all alkaline phosphatases with EDTA leads to a time-dependent denaturation (Congers et al., 1967), both the enzymes were appropriately diluted to unify the catalytic activity. Preincubation for exactly 5 min was performed with EDTA at a final concentration of 1 mM. The remaining catalytic activities were measured using pnitrophenylphosphate without MgCl2. Preincubation with 1 mM EDTA significantly reduced the catalytic activity of PLAP, while the PLAP-like enzyme was comparatively resistant .o EDTA. The remaining activity was 42% of the control even in the presence of 5 mM EDTA (results not shown). Preincubation of alkaline phosphatases with urea is known to result in a time-dependent denaturation (Birkett et al., 1967). Both enzymes with unified catalytic activity were preincubated with 8 M urea for 30 min. PLAP was slightly inactivated by this preincubation with 8 M urea, whereas the PLAP-like enzyme demonstrated a slightly increased catalytic activity, again demonstrating the stability of the PLAP-like isozyme. In order to describe the immunoreactivity of the PLAP-like enzyme, the isozyme specific MICAs were employed. Both PLAP and the PLAP-like enzymes were reactive with the anti-PLAP monoclonal antibody, not to anti-AP or anti-IAP, indicating that the purified enzyme does not contain any AP which is the dominating isozyme in the ovary (Table II). In order to characterise the microheterogeneity of the PLAP isozyme with the PLAP-like enzyme, phenyl-Sepharose affinity chromatography was performed and the chromatographic patterns are shown in Figure 5. Heterogeneity of the isozymes in this affinity chromatography are dependent upon hydrophobicity differences in the molecules (Seiffert et al., 1984). The catalytic PLAP activities were mainly obtained in the 100mM ammonium sulphate buffer (Figure 5a). In contrast, the PLAP-like enzyme demonstrated, besides the 100mM fraction, also minor fractions eluted by 25, OmM ammonium sulphate and detergent buffer. No differences could be identified in the elution pattern between intact PLAP and PLAP treated with trypsin, whereas the pattern of the PLAP-like enzyme treated with trypsin was changed (Figure Sb) and minor fractions eluted by low concentrations of ammonium sulphate were no longer identified and the pattern was similar to that of PLAP. Treatment with bromelain changed the patterns of both PLAP and the PLAP-like enzyme from their precursor isozymes (Figure Sc). The minor fractions in the PLAP-like preparation were still detected as well as the intact major PLAP-like isozyme. Thus, both PLAP and the PLAP-like isozyme demonstrated heterogeneities, and treatment of the PLAP-like enzyme with trypsin but not with bromelain made the PLAP-like enzyme behave similarly to PLAP in the phenyl-Sepharose affinity chromatography. The N-terminal amino acid sequences up to the 16th residue of the PLAP-like enzyme from the ovarian tumour were determined. The results were as follows: Ile-Ile-Pro (Leu)-Val-X-X-X-Asn-Pro-Asp-Phe-Trp-Asn-Ar-Glu-Ala. Although the third amino acid is Pro for variant F and S, and Leu for variant I (Henthorn et al., 1986), the PLAP-like enzyme exhibited both Pro and Leu in a ratio 1:1. Except for the third residue, there was no difference in sequence between the PLAP and PLAP-like isozyme (Henthorn et al., 1986). Finally, the PLAP-like isozyme was subjected to Con A chromatography to evaluate the binding properties to this lectin (Table II). PLAP was bound 100% to the Con A,
9
a
400
200
Ammonium sulphate (mM) 50 100 25 0
Detergent
0.4
0.2
O,
* iv
;
E -1
v L.Q
lq
b
CD
to
6
6 oe
O nnt n15Be n9ooo
0.4
0
8
CU
o 0.2
E
N
w
* 00 n -0 P.
t
n
1,
,"a0b
N \. (F &
(lb
O.-.a c
0.4 0.2 \1
uj
r] v Ion
n
nno
ooooQQ crsnannooo,oss
lb
\
1; o1°!'°a~-v0voooonn.oono *..\* -e .
20
30
Og
40
Fraction number
Figure 5 Affinity chromatographies of different preparations of PLAP and PLAP-like isozymes on phenyl-Sepharose. The enzyme preparation was applied on the column and eluted stepwise with different concentrations of ammonium sulphate as described in the text; a intact enzymes; b enzymes digested with trypsin; c enzymes digested with bromelain. 0, PLAP; *, the PLAP-like enzyme.
whereas 18% of the PLAP-like isozyme did not bind to the lectin. Furthermore, polyacrylamide gel electrophoresis of both enzymes digested with neuroaminidase exhibited differences in mobility of these isozymes, indicating that the sensitivity to neuraminidase of PLAP and PLAP-like enzyme is different (results not shown). Discussion Human PLAP was one of the first oncofetal antigens found
to be ectopically expressed by tumour cells and this antigen is now established as a useful tumour marker for gonadal
malignancies (Fishman, 1987). The isozyme derived from tumours demonstrates identifiable but discrete differences compared with the isozyme derived from placenta. The expression of heat-stable alkaline phosphatases or determination by radioimmunoassays of PLAP in sera from patients with ovarian tumours has confirmed the synthesis of isozymes derived also from ovarian tumours (Mano et al., 1986; Haije et al., 1987). To our knowledge, no report on the purification of the PLAP-like enzyme from ovarian tumour tissues has been presented. The quantitatively dominant form of alkaline phosphatase in the ovary is the 'tissue-unspecific' alkaline phosphatase and minor but significant amounts of PLAP are present in the malignant ovary. In the present study, monoclonal immunoaffinity chromatography was used as the first step to purify the PLAP-like enzyme from the ovarian tumour tissue. The yield of the enzyme was approximately 33% and the enzyme was purified to homogeneity as revealed by SDS-polyacrylamide gel electrophoresis. The purified enzyme was confirmed to be the PLAP-like enzyme from the results of both starch gel electrophoresis and sensitivity to inhibition by L-leucine. The catalytic pro-
10
I. KOYAMA et al.
perties were similar to those of PLAP, but deviations in heat stability were observed for the PLAP-like enzyme. The Nagao isozyme is sensitive to EDTA (Nakayama & Kitamura, 1975), but Cox et al. (1971) indicated that the steroid-induced alkaline phosphatase in HeLa 65 cells retained catalytic activity in the presence of EDTA compared with the base levels of enzyme, an observation which may be related to the present findings. PLAP is known to be resistant to EDTA and urea when compared with AP and IAP (Birkett et al., 1967; Congers et al., 1967). The PLAP-like enzyme from the ovarian tumour, however, was more resistant to these inhibitors than PLAP. These findings reflect biochemical differences between these closely related isozymes. The hydrophobic characteristics of the protein were studied by affinity chromatography on phenylSepharose for evaluation of the heterogeneities of the purified enzyme. The elution profile of the PLAP-like enzyme on the phenyl-Sepharose column was different from that of PLAP, demonstrating differences in hydrophobic properties. Treatment with trypsin of the PLAP-like enzyme gave identical elution patterns with PLAP. Trypsin is known to cleave native PLAP at amino acid residue 62 from the amino terminal end without affecting the catalytic activity of the enzyme (Jemmerson & Stigbrand, 1984). Bromelain cleaves a 2,000 Da hydrophobic fragment from the C-terminal of the protein, a fragment which includes the membrane-anchoring domain (Jemmerson et al., 1985). The digestion with bromelain did not affect the observed difference of the PLAP-like and PLAP isozymes. These results suggest that the differences in hydrophobic heterogeneity between PLAP and the PLAP-like enzyme are related to the N-terminal and not to the C-terminal part of the molecule. The primary structure of the first 16 residues of the PLAPlike enzyme was determined. No exchanges of amino acids which could explain the differences in hydrophobicity were identified. However, the third amino acid of the PLAP-like enzyme from a single tumour revealed Pro and Leu in a molar ratio of 1:1. Residue no. 3 in PLAP is Pro for the variants F and S, and Leu for variant I (Henthorn et al., 1986). Furthermore, the PLAP-like enzyme derived from seminomas have Gln at residue 15 instead of Glu as is the case for PLAP (Millan & Manes, 1988). Residue no. 15 in the PLAP-like enzyme from the ovarian tumour was Glu which is similar to PLAP. The presence of equimolar amounts of both Pro and Leu at residue no. 3 is intriguing and may reflect either the simultaneous existence of heterodimers of the F or S and I variants or the existence of an even more pronounced heterogeneity at the PLAP locus,
not detectable by electrophoretic techniques. Differences in the glycosylation of the oligosaccharide chains in glycoproteins may contribute to the structural differences identified. It is well known that structural changes of the asparagine-linked oligosaccharide chains in many glycoproteins are partially modified following malignant transformation (Ogata et al., 1976; Tamashita et al., 1983). Digestion of the PLAP-like enzyme with neuraminidase yielded a somewhat different mobility pattern at polyacrylamide gel electrophoresis compared with PLAP. Furthermore, the tumour-derived PLAP-like enzyme contained a fraction which was not binding to Con A, which is clearly different from PLAP. It is concluded that the PLAP-like enzyme is at least partially differently glycosylated than is PLAP. The mechanisms behind this observation can be speculated upon. The unbound fraction may contain bisecting N-acetylglucosamine at the asparagine-linked complextype sugar chains as described for a-fetoprotein obtained from transformed cells which also contains a fraction not binding to Con A, whereas a-fetoprotein from untransformed cells is completely bound to Con A (Yoshima et al., 1980). PLAP purified from term placenta has mainly biantennary complex-type sugar chains (Endo et al., 1988). From the present results it can be inferred that the PLAP-like enzyme from ovarian tumours contains a bisected complex-type oligosaccharide chain. Recent reports indicate the membrane-anchoring domain of PLAP to be a glycosylphospholipid constituting the membrane anchor of PLAP (Ogata et al., 1988). Previous reports have demonstrated that the attachment of the enzyme to tumour membrane fragments is somewhat different from that observed in placental membranes (Taylor et al., 1980), although the chemical nature of the membrane anchor of the PLAP-like enzyme is unknown. These results are of significant clinical impact since the appearance in the circulation of tumour-derived enzymes detected by different immunoassays serves as useful tumour markers. Furthermore, it may be possible to generate monoclonal antibodies capable of discriminating between PLAP and the PLAP-like isozyme which could increase the specificity in the monitoring of some malignant conditions. We are grateful to Dr T. Komoda and Professor Y. Sakagishi, Saitama Medical School, Saitama, Japan who encouraged us throughout the course of this work. This work was supported by the Swedish Cancer Research Council (Project No. 1387), Lion's Foundation, Umea and the Medical Faculty, University of Umea.
References
BECKMAN, L., BJORLING, G. & CHRISTODOULOU, G. (1966). Pregnancy enzyme and placental polymorphism. 1. Alkaline phosphatase. Acta Genet., 16, 59. BERGER, J., GARANTTINI, E., HUA, J.-C. & UDENFRIEND, S. (1987).
Cloning and sequence of human intestinal alkaline phosphatase cDNA. Proc. Natl Acad. Sci. USA, 84, 695.
BIRKETT, D.J., CONGERS, R.A.J., NEALS, F.C., POSEN, S. &
BRUDENELL-WOODS, J. (1967). Action of urea on human alkaline phosphatases: With a description of some automated techniques for the study of enzyme kinetics. Arch. Biochem. Biophys., 121, 470. CONGERS, R.A.J., BIRKETT, D.J., NEALS, F.C., POSEN, S. & BRUDENELL-WOODS, J. (1967). The action of EDTA on human alkaline phosphatase. Biochim. Biophys. Acta, 139, 363. COX, R.P., ELSON, N.A., TU, S.H. & GRIFFIN, M.J. (1971). Hormonal induction of alkaline phosphatase by an increase in catalytic efficiency of the enzyme. J. Mol. Biol., 58, 197. ENDO, T., OHBAYASHI, H., HAYASHI, Y., IKEHARA, Y., KOCHIBE,
N. & KOBATA, A. (1988). Structural studies on the carbohydrate moiety of human placental alkaline phosphatase. J. Biochem., 103, 182. FISHMAN, W.H. (1987). Oncotrophoblast gene expression. Placental alkaline phosphatase. Adv. Cancer Res., 48, 1.
HAIJE, W.G., VAN DRIEL, J. & VAN DER BURG, M.E. (1987).
Catalytic and immunologic activities of placental-like alkaline phosphatase in clinical studies. The value of PLAP in follow-up of ovarian cancer. Clin. Chim. Acta, 165, 165. HENTHORN, P.S., KNOLL, B.J., RADUCHA, M. & 6 others (1986). Products of two common alleles at the locus for human placental alkaline phosphatase differ by seven amino acids. Proc. Natl Acad. Sci. USA, 83, 5597. HIRANO, K., IIZUMI, Y., HAYASHI, Y. & 5 others (1986). A highly sensitive assay method for human placental alkaline phosphatase involving a monoclonal antibody bound to a paper disc. Anal. Biochem., 154, 624. HIRANO, K., MATSUMOTO, H., TANAKA, T., IINO, S., DOMAR, U. &
STIGBRAND, T. (1987). Specific assays for human alkaline phosphatase isozyme. Clin. Chim. Acta, 166, 265. JEMMERSON, R. & STIGBRAND, T. (1984). Monoclonal antibodies block the trypsin cleavage site on human placental alkaline phosphatase. FEBS Lett., 173, 357. JEMMERSON, R., MILLAN, J.L., KLIER, F.G. & FISHMAN, W.H. (1985). Monoclonal antibodies block the bromelain-mediated release of human placental alkaline phosphatase from cultured cancer cells. FEBS Lett., 179, 316.
ALKALINE PHOSPHATASE FROM OVARIAN EPITHELIAL TUMOURS JEPPSON, A., WAHREN, B., BREHMER-ANDERSSON, E., SILF-
VERSWARD, C., STIGBRAND, T. & MILLAN, J.L. (1984). Eutopic expression of placental-like alkaline phosphatase in testicular tumors. Int. J. Cancer, 34, 757. KAM, W., CLAUSER, E., KIM, Y.S., KAN, Y.W. & RUTTER, W.J.
(1985). Cloning, sequencing and chromosomal localization of human term placental alkaline phosphatase cDNA. Proc. Natl Acad. Sci. USA, 82, 8715. KOYAMA, I., ARAI, Y., MUIRA, M., MATSUZAKI, H., SAKAGISHI, Y.
& KOMODA, T. (1988). Similarity of the sugar moiety of human alkaline phosphatases between kidney cortex and duodenum, or medulla and ileum. Clin. Chim. Acta., 179, 139. KOYAMA, I., KOMODA, T., SAKAGISHI, Y. & KURATA, M. (1983). A possible mechanism for the changes in hepatic and intestinal alkaline phosphatase activity in bile-duct-ligated rats or guinea pigs. Biochim. Biophys. Acta, 760, 169. KOYAMA, I., MUIRA, M., MATSUZAKI, H., SAKAGISHI, Y. & KOMODA, T. (1987). Sugar-chain heterogeneities of human alkaline phosphatases. Difference between normal and tumorassociated isozymes. J. Chromatogr., 413, 65. LAEMMLI, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680. LOWRY, O.H., ROSENBROUGH, N.J., FARR, A.L. & RANDALL, R.J.
(1951). Protein measurement with the Folin phenol reagent. J. Biol. Chem., 192, 265. MANO, H., FURUHASHI, Y., MORIYAMA, Y., HATTORI, E., GOTO, S.
& TOMODA, T. (1986). Radioimmunoassay of placental alkaline phosphatase in ovarian cancer and tissues. Obstet. Gynecol., 68, 759. MCCOMB, R.B., BOWERS, Jr, G.N. & POSEN, S. (1979). Alkaline Phosphatases, p. 373. Plenum Press: New York. MILLAN, J.L. (1986). Molecular cloning and sequence analysis of human placental alkaline phosphatase. J. Biol. Chem., 261, 3112. MILLAN, J.L. & MANES, T. (1988). Seminoma-derived Nagao isozyme is encoded by a germ-cell alkaline phosphatase gene. Proc. Natl Acad. Sci. USA, 85, 3024.
11
NAKAYAMA, T. & KITAMURA, M. (1975). L-Leucine sensitive alkaline phosphatase isozyme from cancer tissue. Ann. N. Y. Acad. Sci., 259, 325. OGATA, S., HAYASHI, Y., TAKAMI, N. & IKEHARA, Y. (1988). Chemical characterization of the membrane-anchoring domain of
human placental alkaline phosphatase. J. Biol. Chem., 263, 10489. OGATA, S., MURAMATSU, T. & KOBATA, A. (1976). New structural characteristics of the large glycopeptides from transformed cells. Nature, 259, 580. SEIFFERT, U.B., SIEDE, W.H., WELSH, G.J. & OREMEK, G. (1984). Multiple forms of alkaline phosphatases in human liver tissue. Clin. Chim. Acta, 144, 17. STIGBRAND, T. (1984). Present status and future trends of human alkaline phosphatases. In Human Alkaline Phosphatases, Stigbrand, T. & Fishman, W. (eds) p. 3. Alan R. Liss Inc.: New York.
TAMASHITA, K., HITOI, A., TANIGUCHI, N., YAKASAWA, N.,
TSUKADA, T. & KOBATA, A. (1983). Comparative study of the sugar chains of y-glutamyl transpeptidase purified from rat liver and rat AH-66 cells. Cancer Res., 43, 5059. TAYLOR, D.D., HOMESLEY, H.D. & DOELLGAST, G.J. (1980). Binding of specific peroxidase-labeled antibody to placental-type phosphatase on tumour-derived membrane fragment. Cancer Res., 40, 4064. WEISS, M.J., HENTHORN, P.S., LAFFERTY, M.A., SLAUGHTER, C.,
RADUCHA, M. & HARRIS, H. (1986). Isolation and characterization of a cDNA encoding a human liver/bone/kidney-type alkaline phosphatase. Proc. Natl Acad. Sci. USA, 83, 7182. YOSHIMA, H., MIZUOCHI, TISHII, M. & KOBATA, A. (1980). Structure of the asparagine-linked sugar chains of a-fetoprotein purified from ascites. Cancer Res., 40, 4276.
