HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 51, 49808-49814, December 20, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/51/49808    most recent M207394200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Le Du, M.-H. Articles by Millán, J. L. Search for Related Content PubMed PubMed Citation Articles by Le Du, M.-H. Articles by Millán, J. L. Social Bookmarking                      What's this?

Structural Evidence of Functional Divergence in Human Alkaline Phosphatases^*

Marie-Hélène Le Du^§ and José Luis Millán^¶

From the  Département d'Ingénierie et d'Etudes des Protéines (DIEP), CEA, Bat 152 C. E. Saclay, 91191 Gif-sur-Yvette, Cedex, France and the ^¶ The Burnham Institute, La Jolla, California 92037

Received for publication, July 23, 2002, and in revised form, September 11, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

The evolution of the alkaline phosphatase (AP) gene family has lead to the existence in humans of one tissue-nonspecific (TNAP) and three tissue-specific isozymes, i.e. intestinal (IAP), germ cell (GCAP), and placental AP (PLAP). To define the structural differences between these isozymes, we have built models of the TNAP, IAP, and GCAP molecules based on the 1.8-Å structure of PLAP (1) and have performed a comparative structural analysis. We have examined the monomer-monomer interface as this area is crucial for protein stability and enzymatic activity. We found that the interface allows the formation of heterodimers among IAP, GCAP, and PLAP but not between TNAP with any of the three tissue-specific isozymes. Secondly, the active site cleft was mapped into three regions, i.e. the active site itself, the roof of the cleft, and the floor of the cleft. This analysis led to a structural fingerprint of the active site of each AP isozyme that suggests a diversification in substrate specificity for this isozyme family.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

Alkaline phosphatases (EC 3.1.3.1) (APs)¹ are dimeric enzymes present in most, if not all, organisms (2). They catalyze the hydrolysis of phosphomonoesters with release of inorganic phosphate (3). Mammalian APs have low sequence identity with the Escherichia coli enzyme (25-30%), but the residues involved in the active site of the enzyme and those coordinating the two zinc atoms and the magnesium ion are largely conserved, and the catalytic mechanism deduced from the structure of the E. coli AP was proposed to be similar in eukaryotic APs (4). In humans, APs are encoded by four distinct loci. Three isozymes are tissue-specific, i.e. intestinal AP (IAP), placental AP (PLAP), and germ cell AP (GCAP). They are 90-98% homologous, and their genes are clustered on chromosome 2, bands q34-q37. The fourth AP isozyme is tissue-nonspecific (TNAP) and is expressed in a variety of tissues throughout development. TNAP is about 50% identical to the other three isozymes, and its gene is located on chromosome 1, bands p36.1-p34 (5).

Our current understanding of the functional properties of mammalian APs comes largely from studies using PLAP and TNAP as paradigms. Isozyme-specific properties, such as the characteristic uncompetitive inhibition properties of mammalian APs (6-8), their variable heat stability (9), and even their allosteric properties (10), have been attributed to a top, flexible loop (or crown domain) unique to mammalian APs. This domain is also responsible for collagen binding in the case of TNAP (9, 11) but does not appear to mediate the reported binding of PLAP to IgG (12, 13).

The recent elucidation of the 1.8-Å resolution structure of human PLAP (1) has facilitated further studies on the structure and function of mammalian APs. An analysis of the structural-functional relationship of residues conserved between the E. coli AP and the PLAP structure revealed a conserved function for those residues that stabilize the active site zinc and magnesium metal ions, whereas the non-homologous disulphide bonds differ in their structural significance and non-conserved residues take part in determining the heat stability and uncompetitive inhibition properties of mammalian alkaline phosphatases (14). Deactivating mutations in the TNAP gene cause the inborn error of metabolism known as hypophosphatasia (15), characterized by poorly mineralized cartilage and bones. The severity and expressivity of hypophosphatasia depends on the nature of the TNAP mutation (16). The mapping of hypophosphatasia mutations to specific three-dimensional locations on the TNAP molecule has provided clues as to the structural significance of these areas for enzyme structure and function (17). It appears clear that the function of TNAP in bone tissue consists of hydrolizing inorganic pyrophosphate to maintain a proper concentration of this mineralization inhibitor to ensure proper bone mineralization (18). However, the physiological role of the three tissue-specific human APs remains to be clarified.

It has been suggested that PLAP may be involved in the transfer of maternal IgG to the fetus (12, 13, 19), and evidence has accumulated indicating a role of PLAP in cell division in normal and transformed cells (20-23). Of considerable interest is the fact that human APs are abundantly expressed in tumor cells and that their serum levels are often used as tumor markers (24). Plasma TNAP levels can indicate the presence of osteosarcomas (25), Paget's disease (26), and osteoblastic bone metastates (27). PLAP is a marker of cancer of the ovary, testis, lung, and the gastrointestinal tract (28-30). GCAP is a particularly good marker of carcinoma (in situ) of the testis (31-33), and IAP is a marker of hepatocellular carcinoma (34).

Although APs are homodimeric molecules, the re-expression in cancer cells of more than one AP isozyme often results in the formation and release into body fluids of heterodimeric enzymes. The Kasahara AP isoform was identified in a variety of human cancer cell lines (35, 36) and cancer sera and was later found to consist of heterodimers of the IAP and PLAP (37). The human postnatal intestine also contains heterodimers of IAP and PLAP (38). Ovarian cancer cells often express both PLAP and GCAP (39), and cell lines derived from these tumors have been shown to express PLAP/GCAP heterodimers (40, 41). However, no heterodimers have ever been reported between any of the tissue-specific APs and TNAP. The fact that APs can form heterodimers is of structural significance since APs are non-cooperative allosteric enzymes in which the stability and the catalytic properties of each monomer are controlled by the conformation of the second subunit (10). This means that the properties of the heterodimeric enzymes do not correspond to the weighted average of each homodimeric counterpart. Understanding the behavior of AP heterodimers is also of biological significance since, in tissues such as the bovine intestine where up to seven IAP isozymes with differing kinetic properties are co-expressed (42), the formation of heterodimers can give rise to significant functional complexity and novel substrate specificities.

In this study, we have built three-dimensional models of GCAP, IAP, and TNAP based on the 1.8-Å PLAP structure. We have analyzed the homodimer interface and the active site cleft of each modeled isozyme structure. This analysis has allowed us to understand the restrictions observed in AP heterodimer formation, whereas also defining a fingerprint of the active site characteristic of each AP isozyme

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

The PLAP dimer was generated from the coordinates of the PLAP structure (Protein Data Bank accession code 1EW2 (1)) by using the symmetry operation corresponding to the C222₁ space group. The GCAP, IAP, and TNAP sequences were aligned to the PLAP sequence using the program BLAST (43; www.ncbi.nlm.nih.gov/gorf/bl2.html). The PLAP and GCAP molecules display 98% identity with no insertion or deletion relative to PLAP. The IAP and PLAP molecules show 87% identity and 91% homology with no insertion or deletion relative to PLAP. The TNAP and PLAP molecules, however, display 57% identity and 74% homology, and TNAP has four insertions of one residue, one insertion of three residues, and one deletion of two residues relative to PLAP.

The GCAP, IAP, and TNAP homodimeric models and the PLAP/GCAP, PLAP/IAP, and PLAP/TNAP heterodimeric models were constructed using the sequence alignment as found with BLAST, the coordinates of the PLAP dimer, and the program MODELLER (44). The quality of the model geometries was checked with PROCHECK (45). Each model was superimposed to the structure of PLAP using the program ALIGN (46).

The protein surface, interface surface, and residue accessibility were calculated with the program AREAIMOL as implemented in the CCP4 package (47). The interactions between the two monomers were calculated with the program CONTACT as implemented in CCP4. The definition of a secondary structure of proteins given a set of three-dimensional coordinates (DSSP) algorithm (48) as implemented in TURBO was used to calculate secondary structures.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

Geometry of the GCAP, IAP, and TNAP Models-- The models of GCAP, IAP, and TNAP, listed in the order of structural similarity, were built from the PLAP structure using the sequence alignment shown in Fig. 1. This figure also shows the secondary structures, the residues involved in the homodimer interface, the surface residues, and those residues in the active site. As expected, the overall structure of each model is very close to that of PLAP (Table I). The program used to build the models optimizes the geometry, and the quality of the Ramachandran plot is equivalent for each model to that of the PLAP structure. The secondary structures are conserved with those of PLAP, except for one -helix in TNAP, which is lost due to a 2-residue deletion (Fig. 1). Finally, the overall surface of each model, as well as the surface buried at the interface of each homodimer, is similar to that of PLAP, showing that residue accessibility has not been affected by the substitutions performed in MODELLER (Table I).

View larger version (104K): [in this window] [in a new window]   Fig. 1.   Sequence alignment of human placental (PLAP), germ cell (GCAP), intestinal (IAP), and tissue-nonspecific (TNAP) alkaline phosphatases. The alignment shows the secondary structures of PLAP (<->, -helix; <**>, -sheet). Residues highlighted in yellow are buried by more than 10 Å upon dimerization; residues in blue have an accessibility between 10 and 100 Å2; residues in red have an accessibility higher than 100 Å2; and residues underlined are located in the 12-Å sphere around the phosphate group in the active site.

Monomer-Monomer Interface in Human APs-- The overall surface buried at the interface varies between 4134 and 4244 Å² per monomer (Table I), which corresponds to about 25% of the overall protein surface and comprises about 90 residues per monomer that bury more than 10 Å² of their surface upon dimerization. The residues involved in hydrogen bonds, salt bonds, hydrophobic stacking, or cation- interactions are summarized in Table II. This table shows that the stabilizing interactions found in the four proteins are quite similar, although the IAP interface is slightly better stabilized than the others. The distribution of the residues at the interface shows two large clusters and two small clusters, fully conserved between the four isozymes (Fig. 2). The comparison of the interfaces reveals high conservation in the tissue-specific APs, compatible with the fact that PLAP/GCAP or PLAP/IAP heterodimers form readily in nature. In TNAP, however, the substitutions E7K, R117E, N13R, and P68R (PLAP numbering) would lead to repulsive forces at the interface, incompatible with the formation of heterodimers between TNAP and any of the tissue-specific APs. The formation of a heterodimer between TNAP and PLAP or any of the three tissue-specific APs would bring Lys-7 in front of Arg-117, Glu-118 in front of Glu-7, Arg-13 in front of Arg-135, and Arg-69 in front of Lys-81 (Fig. 3). Three of these repulsive interactions are clustered in the same area and can probably not be compensated by neighboring residues. The existence of these repulsive interactions is the most likely explanation for the absence of heterodimer formation between TNAP and any tissue-specific isozyme. The models of PLAP/GCAP and PLAP/IAP hetrodimers confirm that the interactions at the interface do not lead to sterical hindrance or repulsive forces (Table III). The type and number of interactions observed in both heterodimers are very similar to those present in the corresponding homodimers with only slight differences due to sequence differences or side chain flexibility. The PLAP/GCAP heterodimer is difficult to distinguish from a PLAP or GCAP homodimer since there are only nine substitutions between these two isozymes, and none of those differences affect the monomer-monomer interface (Fig. 3b). In contrast, PLAP/IAP heterodimers lead to a protein in which one monomer is enriched in ionic residues, and the other monomer is enriched in neutral or aliphatic residues (Fig. 3c). Our model predicts the testable hypothesis that introducing the E7K, N13R, P68R mutations in PLAP would prevent PLAP homodimer formation but would allow PLAP/TNAP heterodimer formation. The corresponding substitutions in TNAP, i.e. K7E, R13N, R69P, and E118R, should allow heterodimer formation between TNAP and any of the three tissue-specific APs.

View this table: [in this window] [in a new window]   Table II Hydrogen bonds, salt bonds, hydrophobic stacking, and cation- interactions in AP homodimer interfaces Underlined residues correspond to an interaction that involves the side chain of the residue; residues in italic correspond to an interaction that involves the main chain of the residue, and residues in bold are involved in a hydrophobic stacking or cation- interaction. For the interactions corresponding to the homodimeric interfaces, the symmetrical interactions are not given.

View larger version (88K): [in this window] [in a new window]   Fig. 2.   Surface of each AP monomer interface involved in dimerization. Residues involved in a hydrogen bond or ionic bond through their side chain are in red. Residues involved in a hydrogen bond or ionic bond through their main chain are in blue. Hydrophobic residues involved in a stacking interaction or cation- interaction are in violet. a, PLAP; b, GCAP; c, IAP; d, TNAP. The surfaces were drawn with MOLMOL (55).

View larger version (70K): [in this window] [in a new window]   Fig. 3.   Open book vision of a putative PLAP/TNAP (a) heterodimer showing the residues leading to repulsive forces that prevent heterodimer formation. The models of the PLAP/GCAP (b) and PLAP/IAP (c) heterodimers do not show these repulsive forces, in agreement with the fact that they form readily in biological fluids. The surfaces were drawn with MOLMOL (55).

View this table: [in this window] [in a new window]   Table III Hydrogen bonds, salt bonds, hydrophobic stacking, and cation- interactions in heterodimer interfaces Underlined residues correspond to an interaction that involves the side chain of the residue; residues in italic correspond to an interaction that involves the main chain of the residue, and residues in bold are involved in a hydrophobic stacking or cation- interaction. All the interactions from PLAP toward the second monomer are given.

Active Site Cleft-- As has been described in detail before, the single most important difference between PLAP and GCAP is the E429G substitution (PLAP numbering), which effectively converts PLAP into an enzyme with the kinetic, inhibition, and heat stability properties of GCAP (6-10, 14, 49, 50). Glu-429 is located in the immediate neighborhood of the active site Zn1 (Fig. 4), and the nature of this residue confers completely different ionic as well as steric properties to the immediate surrounding of the active site: i.e. Glu-429 in PLAP has a theoretical pK_a around 4.3; Gly-429 in GCAP has no side chain and provides important flexibility to the neighboring loop; Ser-429 in IAP is neutral and polar; and His-434 in TNAP has a theoretical pK_a of 6.0. The close proximity of this residue to the active site suggests that it is directly involved in substrate binding. The change in side chain and pK_a can therefore selectively affect the nature of the substrate favored to bind to the active site of each AP isozyme.

View larger version (149K): [in this window] [in a new window]   Fig. 4.   Corey-Pauling-Koltun (CPK) representation of the active site cleft of a homodimeric PLAP (a), GCAP (b), IAP (c), and TNAP (d) isozyme. The roof of the cleft containing the RY cluster in PLAP and GCAP is circled in yellow, and the hydrophobic/ionic pocket at the base of the cleft is circled in white. The figure was made with TURBO-FRODO (56).

On the upper part of the active site cleft, we observe a large cluster composed of Arg-314, Tyr-276, Arg-326, Arg-323, and Arg-420 surrounded by Glu-321 and Glu-418 (Fig. 4). This cluster, located at 11-17 Å from the phosphate group, exists in PLAP and GCAP but not in IAP or TNAP and is particularly interesting since it includes residues that are often found at protein-protein interfaces (51, 52). Interestingly, Glu-418 and Arg-420 were included in a region of homology that was proposed as a putative protein-protein interaction domain in PLAP (53). The clear identification of this cluster at the roof of the active site is compatible with the notion that PLAP and/or GCAP may act as phosphoprotein phosphatases or as phosphotransferases. In this case, this protein-protein interaction domain may serve to stabilize the polypeptide chain, serving either as substrate or as a phosphate acceptor.

On the lower part of the active site cleft, Lys-87, Phe-107, Glu-108, Arg-166, Asn-167, Tyr-169, and Glu-429 from one monomer, and Tyr-367 from the second monomer, form a hydrophobic pocket in PLAP (Fig. 4). This hydrophobic pocket, which involves residues from both monomers, is conserved in GCAP and IAP except at position 87 in IAP, which displays the conservative substitution K87R. In TNAP, among the 8 residues of this pocket, we observe five substitutions: K87A, F107E, Q108G, N167D, and E429H (PLAP numbering). These substitutions remove the hydrophobic character, converting it into strongly ionic in TNAP. Therefore, the properties of this pocket in the case of TNAP are completely different from those of PLAP, GCAP, or IAP. These findings correlate well with the differential behavior of the AP isozymes toward uncompetitive inhibitors, i.e. PLAP, GCAP, and IAP are inhibited by L-Phe but not by L-homoarginine, whereas TNAP is inhibited by L-homoarginine but not by L-Phe (2, 5). The location and orientation of this pocket with regard to the phosphoseryl intermediate during catalysis (4) suggests that it may participate in stabilizing the phosphate donor at the first step of the reaction. Therefore, the substrate of a tissue-specific AP or of TNAP must display ionic properties compatible with the highly divergent ionic properties of the corresponding pocket.

Thus, we have found three regions at the active site cleft, which characterize each human AP. PLAP contains Glu-429, the RY cluster at the roof, and the hydrophobic pocket at the floor of the cleft; GCAP contains Gly-429, the RY cluster, and the hydrophobic pocket; IAP contains Ser-429, has no RY cluster but has the hydrophobic pocket; whereas TNAP contains His-429, has no RY cluster and has a highly ionic pocket at the floor of the cleft. This suggests that the tissue-specific APs and TNAP are likely to have very different substrate specificities.

Concluding Remarks-- This structural analysis of the four human APs reveals important differences between the human AP isozymes that may provide clues as to their individual tissue-specific functions. Our analysis of the monomer-monomer interface provides the structural basis behind the formation of PLAP/GCAP and IAP/PLAP heterodimeric enzymes found expressed during development and in cancer cells. This analysis also provides a rational explanation for the lack of heterodimer formation between TNAP and any of the tissue-specific isozymes. At the active site cleft, we have defined a fingerprint characteristic of each AP isozyme. This fingerprint is compatible with the hypothesis of isozyme-specific specialization for a phosphate donor or phosphate acceptor in the case of transphosphorylation reaction.

	ACKNOWLEDGEMENTS

We are grateful to Prof. Marc F. Hoylaerts, Dr. Menetrey, and Marc Graille for careful reading of the manuscript.

	FOOTNOTES

^* This work was supported in part by Grants CA 42595, DE 12889, and AR 47908 from the National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom correspondence should be addressed. Tel.: 33-1-69-08-71-35; Fax: 33-1-69-08-90-71; E-mail: mhledu@cea.fr.

Published, JBC Papers in Press, October 7, 2002, DOI 10.1074/jbc.M207394200

	ABBREVIATIONS

The abbreviations used are: AP, alkaline phosphatase; GCAP, germ cell AP; IAP, intestinal AP; PLAP, placental AP; TNAP, tissue-nonspecific AP.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				T. S. Park, Z. Galic, A. E. Conway, A. Lindgren, B. J. van Handel, M. Magnusson, L. Richter, M. A. Teitell, H. K. A. Mikkola, W. E. Lowry, et al. Derivation of Primordial Germ Cells from Human Embryonic and Induced Pluripotent Stem Cells Is Significantly Improved by Coculture with Human Fetal Gonadal Cells Stem Cells, April 1, 2009; 27(4): 783 - 795. [Abstract] [Full Text] [PDF]

				T. Suzuki, K. Ishihara, H. Migaki, W. Matsuura, A. Kohda, K. Okumura, M. Nagao, Y. Yamaguchi-Iwai, and T. Kambe Zinc Transporters, ZnT5 and ZnT7, Are Required for the Activation of Alkaline Phosphatases, Zinc-requiring Enzymes That Are Glycosylphosphatidylinositol-anchored to the Cytoplasmic Membrane J. Biol. Chem., January 7, 2005; 280(1): 637 - 643. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/51/49808    most recent M207394200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Le Du, M.-H. Articles by Millán, J. L. Search for Related Content PubMed PubMed Citation Articles by Le Du, M.-H. Articles by Millán, J. L. Social Bookmarking                      What's this?