University of Ant- werp, Antwerp,. Belgium. ...... IgG and. Vectastain. ABC reagent, the absorbance was read on calibration curves constructed with purified. PLAP ...
CLIN.
CHEM.
38/12,
2493-2500
(1992)
Allelic Amino Acid Substitutions Germ-Cell Alkaline Phosphatase Marc
F. Hoylaerts,”2
Thomas
Affect the Conformation Phenotypes
Manes,’
and
Jos#{233} Luis
gene
(PLAP)
gene encoding placental alkaline phosphatase displays a well-documented allelic polymorphism. Likewise, different phenotypes exist for the PLAP-related germ-cell alkaline phosphatase (GCAP). We investigated the extent to which various allelic GCAP positions are critical in determining the enzymatic, structural, and immunological properties of GCAP phenotypes. Three homozygous GCAP phenotypes [JEG3, BeWo, and wildtype (wt) GCAP] were analyzed and compared with a “core” GCAP mutant that contains the seven amino acid substitutions that are consistently different between PLAP and GCAP but are common to the three known allelic GCAP genotypes. Although some substitutions could
in
influence
the electrophoretic
behavior
PLAP.
The
selective
of the phenotypes,
the kinetic properties the immunoreactivity detected with a panel antibodies (MAbs) to
immunoreactivity
of
the
PLAP/
GCAP-discnminating MAb C2 was critically dependent on the nature of the allelic residues 133 and 361 in GCAP. Residue 133 was also importantfor the general stability of the molecule because BeWo and wt GCAP, which have Asn1 and Val1, respectively, instead of Met1, showed a consistently reduced heat stability compared to core GCAP and JEG3. Because the core GCAP mutant consistently shows the characteristics of wt GCAP, its use as an antigen should allow the generation of monoclonal antibodies
to GCAP
that
will not cross-react
with
and whose immunoreactivity will only marginally enced by allelic GCAP variation. Additional .
Keyphrases:
monoclonal
allelic polymorphism
antibodies
.
PLAP
be influ-
genetic
variation
isoenzymes
Human alkaline phosphatases (ALPs) are encoded by a gene family composed offour loci [for review, see (1)]. Whereas the three tissue-specific ALP (TSAP) genes, placental (PLAP), germ-cell (GCAP), and intestinal ALPs, are composed ofli exons and occupy less than 5.0 kb of DNA, the single tissue-nonspecific ALP (TNAP) 1 La Jolla Cancer Research Foundation, Cancer Research Center, 10901 North Torrey Pines Road, La Jolla, CA 92037. 2 Department of Nephrology-Hypertension, University of Ant-
werp,
Antwerp, Belgium. Author for correspondence. 4 Nonstandard abbreviations: TSAP, tissue-specific alkaline 3
ALP, phosphatase;
alkaline TNAP,
phosphatase; tissue-nonspe-
cific alkaline phosphatase; PLAP, placental alkaline phosphatase; GCAP, germ-cell alkaline phosphatase; wt, wild type; MAb, monoclonal antibody; DEA diethanolamine; and pNPP, p-nitrophenyl phosphate. Received
May
13, 1992;
accepted
August
3, 1992.
of
Milan”3
The
the allelic differences did not affect of GCAP. However, they did affect and conformation of the variants as of 1 8 epitope-mapped monoclonal
and Immunoreactivity
contains an additional, differentially spliced exon 5’ region and significantly larger introns that 40-50 kb of DNA. The TSAP genes are colocalized in the long arm of chromosome 2, but the TNAP gene resides at the end of the short arm of chromosome 1. The PLAP gene is subject to a high degree of polymorphism. Three common alleles (Plo, pF’, and P1’) account the occupy
for over 90% of the PLAP phenotypes (2). However, more than 20 rare allozymes have been described, often only in heterozygous combinations because of their low allelic frequency (3). The two most common PLAP phenotypes, S (slow) and F (fast), differ only in an Arg#{176} to p2o9 substitution (4), but additional amino acid replacements have been identified for the I (intermediate) variant (5). Although GCAP is encoded by a different gene from PLAP, the primary structure of GCAP shows 98% sequence identity with PLAP (6, 7). However, when various GCAP samples were analyzed with respect to their immunoreactivity with a number of monoclonal antibodies raised against PLAP, some of the antibodies reacted with high, intermediate, or low affinity (8). From these findings, the existence of a GCAP polymorphism with up to nine allelic GCAP variants was anticipated (8-10). Determination of the sequence of GCAP, derived from JEG3 choriocarcinoma cells (11) and BeWo cells (12), confirmed the existence of allelic variation in the GCAP gene. Using a series of site-directed PLAP mutants, we recently showed that individual amino acid substitutions in the PLAP isoenzyme had a considerable effect on the immunoreactivity and conformation ofthe resulting mutants, as detected by a panel of 18 monoclonal antibodies (MAbs) to PLAP (4). These results emphasized the importance of characterizing in great detail the reactivity of those MAbs used clinically for the serological evaluation of ALP isoenzymes. PLAP and GCAP are useful tumor markers in the management of patients with adenocarcinoma of the ovary and seminoma of the testis (13-1 7). The existence of allelic GCAP differences can influence the accuracy of the immunochemical detection of GCAP phenotypes, their electrophoretic identification, and the molecular stability of the GCAP allotypes. To evaluate these variables, we used site-directed mutagenesis to construct a series of PLAP and GCAP mutants. We then compared these mutants with those GCAP phenotypes for which the sequence is known. Although the presence of Gly at position 429 in GCAP is the key element determining the enzymatic properties ofGCAP (18), our results show that different amino acids at certain allelic positions can CLINICAL
CHEMISTRY,
Vol. 38, No. 12, 1992
2493
influence significantly Materials PLAP
the
and
Mutants
conformation affect antibody
of the GCAP recognition.
molecule
and
with
tively,
low
a rabbit
Genotypes
The study
PLAP (F phenotype) and GCAP used in this have been described previously (18) and are referred to as wild-type (wt) PLAP and wt GCAP. A 2.0-kb Eco RI-Kpn I fragment of the PLAP cDNA (6) was used as the source of template DNA to generate a series of PLAP mutants. Site-directed mutagenesis experiments were performed according to Kunkel (19) by using the mutagene M13 in vitro mutagenesis kit (Bio-Rad Laboratories, Richmond, CA). The generation of the single amino acid mutants, [Gln’5]PLAP, [Thr67]PLAP, [PhessIPLAP, [Ser84]PLAP, [His’]PLAP, [Leu]PLAP, and [G1y429JPLAP, was described previously (4). These cDNAs were used as a source of fragments to reassemble the more complex mutants. [HisZl, Leu, G1y429]PLAP ([HLG]PLAP) was constructed by ligating a 568-bp BamHI-SacI fragment containing the [His’1 mutation (BamHI-[His’]-SacI), the 154-bp Sac I-[Leu254]-SacII, and the 937-bp SacII-[Gly4]-KpnI fragments into pSVT7-PLAP digested with BamHI and KpnI. A 1414-bp BstEll-KpnI fragment from the [HLG] PLAP construct was then isolated and ligated with a 276-bp BamHI-[SerJ-BstEII fragment into either cut with BamHI and KpnI to create Leu254, Gly429]PLAP ([SHLG]PLAP) or digested with BamHI and KpnI to SerM, His241, Leu2M, G1y429]PLAP ([QSHLG]PLAP). Finally, a 276-bp Barn HI-[Ser, r67, PhessJBstEll fragment was ligated with the 1414-bp BstEII-[His1, Leu, Gly4}-KpnI fragment into pSVT7-[Gln’5IPLAP to generate core GCAP. The sequence of the mutagenesis primer used to generate [Leu361]PLAP was as follows: 5’-GGA GAA GAj GTG GGA GTG GTC-3’ (the underlined base indicates the change). The wt and mutagenized PLAP cDNAs were
antiserum
concentrations
tants, core GCAP, the insolubilized
Methods and GCAP
coated
and GCAP MAbs. Upon
fraction was measured total enzyme concentration surements were carried age concentration divided by the fraction the
binding
step
increasing that
expressed
in the
in triplicates
(60%) in the activity of the PLAP mutants and the GCAP phenotypes (Figure 4A). These results show that the conformational change that occurs in PLAP when Glu429 is substituted for Gly4 has a major impact on the general stability of the isoenzyme. They also show that further substitutions can partially correct for this decreased stability. However, because all PLAP mutants and GCAP phenotypes were not inactivated to the same degree, we also measured residual activities after mactivation at a less critical temperature (56 #{176}C). Exposure to 56 #{176}C for 30 mm (Figure siB) caused a comparable destabilization for wt PLAP (80% residual activity), but differences within the group of PLAP mutants and GCAP phenotypes were also evident. The stability ofthe PLAP mutants and GCAP phenotypes, expressed relative to that of wt PLAP, is greater at 56 #{176}C than at 65 #{176}C, even when the enzymes are exposed for longer time intervals. The addition of ZnCl2 during incubation at elevated temperatures could partially protect some GCAP phenotypes from denaturation (not shown). Therefore, to describe the kinetics of heat inactivation in more detail, we analyzed the residual activity of wt PLAP time-dependently in 1 molJL DEA buffer, pH 9.8, containing 20 moI/L ZnCl2 and 0.5 mmol/L MgCl2, both at 56 #{176}C and at 65 #{176}C (Figure 5A). At 65 #{176}C, the wt PLAP inactivation showed a biphasic process, with the first phase lasting approximately 10 assessment
and GCAP
CLINICAL CHEMISTRY,
Vol. 38, No. 12, 1992
2497
100
A
100
65C
A
80
50
60
20 wt PLAP
56 C : 0
wtPLAP
65C:I I
) 100
I
10
I
20
I
I
30
40
50
60
30
40
50
60
40
50
60
56C
B
.80
B
100 >1
u6O a 4O
50 20
0
0
Q
Q9
20 C,
[GJPLAP:o
[HLG] PLAP : [SHLG] PLAP : o
b
0
Fig. 4. Residual enzyme activity of the wt PLAP, PLAP mutants, and GCAP phenotypes after heat inactivation in 1 mol/L DEA buffer, pH 9.8, containing 0.5 mmol/L MgCI2 at (A) 65 #{176}C for 6 mm and at (B)
10
[QSHLG]PLAP:u
Cl)
10
I
20
56 #{176}C for 30 mm
Values expressed
mm; tion
relative to the activity of the unheated
samples
100
at 56 #{176}C, a slower process
was
heat
inactivation
PLAP
mutants
turation
but almost monophasic inactivaapparent. Likewise, kinetic analysis of of [GIPLAP and the more complex at 56 #{176}C (Figure 5B) showed that dena-
occurred
by
way
of a monophasic
mechanism.
The introduction of additional mutations in [GIPLAP increased the stability of the more complex mutants. The heat inactivation behavior ofthe GCAP phenotypes at 56 #{176}C (Figure 5C) could also be described as a monophasic process. Core GCAP and JEG3 GCAP had a general stability comparable with that of the multiply substituted PLAP mutants. However, wt GCAP and BeWo GCAP consistently showed lower stability. This difference
must
be
related
to
the
single
amino
acid
substitution
of Met’ (core GCAP) for Asn’ (BeWo Met’33 for Val’33 (wt GCAP) substitution effect on the stability of wt GCAP. This already described as being conformationally critical, causing a 10-fold loss in immunoreactivity with MAb C2. Therefore, from the antibody affinity studies and the heat inactivation analysis, we can attribute a structural role to the allelic amino acid residue at position 133. The double Val36’ (for Leu361) and Pro (for Arg479) substitution in core GCAP is silent in terms of heat inactivation behavior. Yet, from the pronounced effect of the Val36’ to Leu36’ substitution in wt PLAP on the immunoreactivity of C2, we can attribute a structural role to this amino acid position that is evident only when it is combined with additional amino acid substitutions that confer the general GCAP structure. GCAP). The had a similar residue was
Active
Site Properties
of GCAP
Allelic Variants
We showed recently that the Glu429 for Gly429 substitution in wt PLAP is accompanied by a small decrease 2498
CLINICAL
CHEMISTRY,
50
Vol. 38, No. 12, 1992
20
CoreGCAP: 0 JEG3 GCAP : wtGCAP: 0
10
IBeW0GCAP:
a
io
20 Time
Fig. 5. Kinetics
containing
ofthe
heat inactivation
30
(mm) in 1 mol/L
DEA buffer,
pH 9.8,
20 moVL
ZnCI2 and 0.5 mmol/L MgCI2 of (A) wt PLAP at 56 #{176}C and 65 #{176}C, (B) the different PLAP mutants at 56 #{176}C, and (C) the different GCAP phenotypes at 56 #{176}C Michaelis constant (Km) from 0.35 mmol/L to 0.1 when measured in 1 mol/L DEA containing 0.5 mmol/L MgCl2 (18, 24). This substitution only slightly affected the turnover number (k) from 460 s’ (wt PLAP) to’344 s’ ([Gly429]PLAP). Additional substitutions in wt PLAP did not further affect these kinetic parameters. We have now confirmed that this result also holds for th core GCAP mutant and the three GCAP phenotypes studied; they have very similar Km (0.1 mmolJL) and (280-300 1) values. The substitution in wt PLAP of Glu429 for Gly429 accounts for the differential inhibition ofGCAP by L-Leu (18), a phenomenon explained by steric hindrance exerted by the Glu4 side chain in PLAP, but absent in GCAP, during positioning of the inhibitor in the active site of the enzyme (24). Our present comparison ofthe inhibition of the different GCAP phenotypes by increasing concentrations of L-Leu (Figure 6A) confirms that all GCAPs in
mniol/L
#{149}
A
100
These
mutant we phenotype displaying the consistent characteristics of GCAP, including immunoreactivity, molecular stability, and inhibition properties. The use ofcore GCAP as an antigen for the production of monoclonal antibodies is likely to allow the generation ofreagents that will show very low cross-reactivity with PLAP and will enable us to produce antibodies for which immunoreactivity is only constructed
80
40 >1
20 0
0.05
0
0.1
1
10
studies
show
behaves
that
as
the
core
a prominent
GCAP GCAP
marginally determined by allelic amino in GCAP. These monoclonal antibodies valuable in the specific determination serum of patients.
acid variations would prove of GCAP in the
100 801-
Supported by grant CA42595 from the National Health and by the Veremging voor Kankerbestrijding, We thank Elisabeth Bossi for help with the transfections
60
DNAS.
0)
40
References
(GI PLAP: [HLG] PIAP:
201-
1. Harris H. The human and what we don’t know. 2. Beckman G, Beckman
PLAP: PLAP:
I.
0L 0.05
0.1
1
10
[L-Leu] (mM) Fig. 6. Inhibition of the enzymatic activity of (A) the different GCAP phenotypes and (B) the different PLAP mutants by increasing L-Leu concentrations (0.05-i 0 mmol/L) in comparison with wt PLAP
investigated efficiencies
are than
Significance
of Core
inhibited by L-Leu with 10-fold higher wt PLAP. The [Gly41PLAP and the triple [His’, Leu4, Gly429]PLAP mutant are inhibited with even slightly higher affinities (Figure 6B). We have shown previously that Ser in GCAP plays a modulating role on the L-Leu inhibition. As soon as this mutation is superimposed on top of the other PLAP mutations, L-Leu inhibition profiles are identical to those obtained for GCAP.
All
the
raised
immunological
MAbs
GCAP
and 130) used in this study Therefore, it is not surprising that GCAP was generally recognized with lower affinities than PLAP. However, it is clear from the present study that amino acids in allelic GCAP positions can strongly affect the conformation and immunoreactivity ofthe enzyme. Only two MAbs (17E3 and C2) consistently showed a low reactivity with the different GCAP phenotypes, a property that has been exploited in the clinical evaluation and quantitation of PLAP and GCAP in serum (9, 10). A third MAb (H317), with properties comparable with those of 17E3 and C2, has also been described (27); however, to date there are no were
(except
against
reagents
Institutes of Belgium. of mutant
151
PLAP.
that
allow
the
specific
mea-
surement of GCAP in the presence of PLAP. Yet, a correct assessment ofthe GCAP concentration would be clinically valuable because in seminoma (9, 10, 13-15, 25), other testicular tumors (14, 15), and ovarian cancer (16, 1 7) GCAP concentrations increase in serum and fluids.
alkaline phosphatases: what we know Clin Chim Acta 1990;186:133-50. L. The placental alkaline phosphatase polymorphism. Hum Hered 1969;19:524-9. 3. Donald U, Robson EB. Rare variants of placental alkaline phosphatase. Ann Hum Genet Lond 1974;37:303-13. 4. Hoylaerts MF, Mill#{225}nJL. Site-directed mutagenesis and epitope-mapped monoclonal antibodies define a catalytically important conformational difference between human placental and germ cell alkaline phosphatase. Eur J Biochem 1991;202:605-16. 5. Henthorn PS, Knoll BJ, Raducha M, et a!. Products of two common alleles at the locus for human placental alkaline phosphatase differ by 7 amino acids. Proc Natl Acad Sd USA 1986;83: 5597-601. 6. Millan
JL Molecular cloning and sequence analysis of human placental alkaline phosphatase. J Biol Chem 1986;261:3112-5; J Biol Chem [Letter] 1991;266:4023. 7. Millan JL, Manes T. Seminoma-derived Nagao isozyme is encoded by a germ-cell alkaline phosphatase gene. Proc Nat! Acad Sci USA 1988;85:3024-8. 8. Millan JL, Stigbrand T. Antigenic determinants of human placenta! and testicular placental-like alkaline phosphatases as mapped by monoclonal antibodies. Eur J Biochem 1983;136:1-7. 9. Wahren B, HinkulaJ, Stigbrand T, et a!. Phenotypes of placental-type alkaline phosphatase in seminoma sara as defIned by monoclonal antibodies. hit J Cancer 1986;37:595-600. 10. Hendrix PG, Hoylaerts MF, Nouwen EJ, De Bros ME. Enzyme immunoassay ofhuman placental and germ-cell alkaline phosphatase in serum. Cim Chem 1990;36:1793-9. 11. Watanabe 5, Watanabe T, Li WB, Soong BW, Choy JY. Expression of the germ cell alkaline phosphatase gene in human choriocarcinoma cells. J Biol Chem 1989;264:12611-9. 12. Lowe ME, Strauss AW. Expression ofa Nagao-type, phosphatidylinositol-glycan anchored alkaline phosphatase in human choriocarcinoma. Cancer Res 1990;50:3956-62. 13. Lange PH, Mill#{225}nJL, Stigbrand T, Vesse!la RI, Ruoslahti E, Fishman WH. Placental alkaline phosphatase as a tumor marker for seminoma. Cancer Rae 1982;42:3244-7. 14. PaivaJ, Damjanov I, Lange PH, Harris ical localization of placental-like alkaline
H. Immunohistochem-
phosphatase in testis and germ cell tumors using monoclonal antibodies. Am J Pathol 1983;111:156-65. 15. Jeppsson A, Wahren B, Brehmer-Andersson E, Si!fverswArd C, Stigbrand T, Millan JL. Eutopic expression of placental-like alkaline phosphatase in testicular tumors. mt J Cancer 1984;34: 757-61. 16. Vergote
I, Onarud M, Nustad K. Placental alkaline phosphatase as a tumor marker in ovarian cancer. Obstet Gynecol 1987; 69:228-32. 17. De Bros ME, Pollet DE. Multicenter evaluation of human placental alkaline phosphatase as a possible tumor-associated antigen in serum. Clin Chem 1988;34:1995-9. CLINICAL
CHEMISTRY,
Vol. 38, No. 12, 1992
2499
18. Hummer uncompetitive tase. Biochem
C, Mill#{225}n JL. Gly4 is the major determinant of inhibition of human germ cell alkaline phosphaJ 1991;274:91-5. 19. KunkelTA. Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc Nat! Acad Sci USA 1985;82:488-92. 20. Bird P, Gething M-J, Sambrook J. Translocation in yeast and mammalian cells: not all signal sequences are functionally equivalent. J Cell Biol 1987;105:2905-14. 21. German CM, Moffat LF, Howard BH. Recombinant genomes which express chloramphethcol acetyltransferase in mammalian cells. Mo! Cell Biol 1982;2:1044-51. 22. Ito F, Chou JY. Induction of placental alkaline phosphatase biosynthesis by sodium butyrate. J Biol Chem 1984;259:2526-30. 23. Poulik MD. Starch gel electrophoresis in a discontinuous system and buffers. Nature (London) 1957;180:1477-9. 24. Hoylaerts MF, Manes T, Mill#{225}n JL. Molecular mechanism of uncompetitive inhibition of human placental and germ cell alkaline
25.
phosphatase. Biochem K, Stigbrand
J 1992;
286:23-30.
T, Hisazumi H, Wahren B. Electrophoretic heterogeneity ofalkaline phosphatase isozymes in seminoma and normal tissue. Tumour Biol 1989;10:181-9. 26. Harris H. The principles of human biochemical genetics, 3rd ed. Amsterdam: ElseviertNorth Holland, 1980. 27. McLaughlin PJ, Johnson PM. A search for human placentaltype
Koshida
alkaline
phosphatases
using
monoclonal
antibodies.
Prog
Clin Biol Res 1984;166:67-75.
Appendix The technical issue discussed here was raised when this manuscript was being reviewed and merits consideration. Reviewer’s question: The reactivities of antibodies C2 and H7 with the different isoenzymes and mutants are not uniform, as shown in Figure 2 of this paper and in previous reports [Mill#{225}n JL, Stigbrand T. Eur J Biochem 1983;136:1-7; Hoylaerts MF, Mill#{225}nJL. Eur J Biochem 1991;202:605-16]. The concentration of antigen [El#{176} was determined by use of these monoclonal antibodies. Why can you get a true value of [E]#{176}? Authors’ response: Initially we used an ELISA procedure [Hoylaerts MF, Manes T, Mill#{225}nJL. Biochem J 1992;286:23-30] in which plates were coated with a polyclonal antiserum to PLAP. After deposition of the
2500
CLINICAL
CHEMISTRY,
Vol. 38, No. 12, 1992
samples, we used 500 gfL ofH 7/C2 for the detection of bound PLAP/GCAP mutants, using H7 preferentially for the GCAP-related enzymes (or mutants). After revealing the bound monoclonal antibody with biotinylated rabbit antiserum to mouse IgG and Vectastain ABC reagent, the absorbance was read on calibration curves constructed with purified PLAP. Concentrations thus determined yield only estimates of [El#{176}. Because, in principle, the enzyme concentrations could be estimated from their catalytic activity, we have determined Km and for each enzyme mutant. Km determinations are straightforward, but to circumvent the problems encountered in the ELISA during the determination of [El#{176}, we adapted our immunoassay as follows: A limited amount of H7 (10 ngfL) was bound onto rabbit antiserum to mouse Ig-coated plates and, during the
incubation
amount of antibody was increasing concentrations ofPLAP, GCAP, or mutant enzymes. The activity of the bound enzymes was measured at 405 nm. Plots of 1/A (405 nm) vs the dilution factor, in the range of saturating concentrations, are then linear. The A (405 nm) at infinite concentration (intersection with y-axis) represents the activity of fully saturated H7 monoclonal antibody. At saturation, independent of the mutant studied or of its affinity, the absolute amount of the enzyme is constant. Therefore, the differences in the A (405 nm) values at the intersection reflect differences in When relating these A (405 nm) values to that found for the reference PLAP (for which can be calculated easily), it is possible to calculate for the
progressively
different
step,
this
saturated
enzyme
mutants.
low
with
Thus,
we
found
that
values are only marginally influenced by the different substitutions investigated. This, we believe,justifies the choice of enzyme concentrations based on activity measurements, as was done during the measurements of relative affinities.