Cloning of a Phosphatidic Acid-preferring Phospholipase A1 from Bovine Testis
1998; Elsevier BV; Volume: 273; Issue: 10 Linguagem: Inglês
10.1074/jbc.273.10.5468
ISSN1083-351X
AutoresHenry N. Higgs, May Han, Guy E. Johnson, John A. Glomset,
Tópico(s)Enzyme Catalysis and Immobilization
ResumoWe report the molecular cloning and expression of a phosphatidic acid-preferring phospholipase A1 from bovine testis. The open reading frame encoded an 875-amino acid protein with a calculated molecular mass of 97,576 daltons and a pI of 5.61. The sequence included a region similar to a lipase consensus sequence containing the putative active site serine and also included a potential, coiled-coil-forming region. Expression of the open reading frame in COS1 cells resulted in a 20–44-fold increase in phosphatidic acid phospholipase A1 activity over that of control cells. Mutation of the putative active site serine (amino acid 540) demonstrated that it was essential for this increase in enzyme activity. Northern blot analysis revealed at least five different messages with the highest overall message levels in mature testis, but detectable message in all tissues examined. Two possible alternately spliced regions in the open reading frame also were identified. Finally, a search of the data base identified six related proteins: a potential counterpart of the phospholipase A1 in Caenorhabditis elegans, two putative lipases in yeast, and three proteins separately encoded by the Drosophila retinal degeneration B gene and its mouse and human homologues. We report the molecular cloning and expression of a phosphatidic acid-preferring phospholipase A1 from bovine testis. The open reading frame encoded an 875-amino acid protein with a calculated molecular mass of 97,576 daltons and a pI of 5.61. The sequence included a region similar to a lipase consensus sequence containing the putative active site serine and also included a potential, coiled-coil-forming region. Expression of the open reading frame in COS1 cells resulted in a 20–44-fold increase in phosphatidic acid phospholipase A1 activity over that of control cells. Mutation of the putative active site serine (amino acid 540) demonstrated that it was essential for this increase in enzyme activity. Northern blot analysis revealed at least five different messages with the highest overall message levels in mature testis, but detectable message in all tissues examined. Two possible alternately spliced regions in the open reading frame also were identified. Finally, a search of the data base identified six related proteins: a potential counterpart of the phospholipase A1 in Caenorhabditis elegans, two putative lipases in yeast, and three proteins separately encoded by the Drosophila retinal degeneration B gene and its mouse and human homologues. Cellular membranes are highly dynamic structures, and this dynamic nature plays a major role in cellular physiology. In response to extracellular or intracellular signals, a number of phospholipases can be activated, including phosphoinositide-specific phospholipase C, phosphatidylcholine-specific phospholipase C, phospholipase D, phosphatidic acid (PA) 1The abbreviations used are: PA, phosphatidic acid; CELG, predicted sequence of an unknown protein from C. elegans; COS1 cells, SV40-transformed African Green monkey kidney cells; 5′-RACE, 5′-rapid amplification of cDNA ends; kb, kilobase pair(s); MALDI, matrix-assisted laser desorption ionization; MOPS, 3-(N-morpholino)propanesulfonic acid; ORF, open reading frame; PA-PLA1, phosphatidic acid-preferring phospholipase A1; PCR, polymerase chain reaction; PITP, phosphatidylinositol transfer protein; PLA1, phospholipase A1; RDGBd, RDGBh, and RDGBm: sequences encoded by the retinal degeneration B gene in Drosophila and the corresponding genes in humans and mice; SCERV, predicted sequence of an unknown protein from S. cerevisiae; SPOMB, predicted sequence of an unknown protein from S. pombe; PAGE, polyacrylamide gel electrophoresis; 3′-UTR, 3′-untranslated region. 1The abbreviations used are: PA, phosphatidic acid; CELG, predicted sequence of an unknown protein from C. elegans; COS1 cells, SV40-transformed African Green monkey kidney cells; 5′-RACE, 5′-rapid amplification of cDNA ends; kb, kilobase pair(s); MALDI, matrix-assisted laser desorption ionization; MOPS, 3-(N-morpholino)propanesulfonic acid; ORF, open reading frame; PA-PLA1, phosphatidic acid-preferring phospholipase A1; PCR, polymerase chain reaction; PITP, phosphatidylinositol transfer protein; PLA1, phospholipase A1; RDGBd, RDGBh, and RDGBm: sequences encoded by the retinal degeneration B gene in Drosophila and the corresponding genes in humans and mice; SCERV, predicted sequence of an unknown protein from S. cerevisiae; SPOMB, predicted sequence of an unknown protein from S. pombe; PAGE, polyacrylamide gel electrophoresis; 3′-UTR, 3′-untranslated region.phosphohydrolase, phospholipase A2, and sphingomyelinase (1Liscovitch M. Signal-activated Phospholipases. R. G. Landes Co., Austin, TX1994Google Scholar, 2Singer W.D. Brown H.A. Sternweiss P.C. Annu. Rev. Biochem. 1997; 66: 475-509Crossref PubMed Scopus (347) Google Scholar). These enzymes produce second messengers that function in cellular regulation.One second messenger that has received increasing attention is PA. This phospholipid can be produced either by the phospholipase D-catalyzed degradation of phosphatidylcholine or by a two-step pathway involving the phospholipase C-catalyzed degradation of phosphatidylinositol 4,5-bisphosphate to diacylglycerol and the subsequent, diacylglycerol kinase-catalyzed phosphorylation of this diacylglycerol (3Bocckino S.B. Exton J., H. Bell R.M. Exton J.H. Prescott S.M. Handbook of Lipid Research: Lipid Second Messengers. 8. Plenum Press, New York1996: 75-123Google Scholar). PA appears to function in a number of cellular processes, including superoxide production in neutrophils (4Cockcroft S. Biochim. Biophys. Acta. 1992; 1113: 135-160Crossref PubMed Scopus (243) Google Scholar, 5Bauldry S.A. Bass D.A. Cousart S.L. McCall C.E. J. Biol. Chem. 1991; 266: 4173-4179Abstract Full Text PDF PubMed Google Scholar) and actin polymerization (6Ha K.S. Exton J.H. J. Cell Biol. 1993; 123: 1789-1796Crossref PubMed Scopus (154) Google Scholar). Other roles for PA are suggested by the facts that (a) mammalian tissues contain a large (10 isoforms) family of diacylglycerol kinases as well as several phospholipase D enzymes (7Kanoh H. Kai M. Wada I. J. Lipid Mediators Cell. Signalling. 1996; 14: 245-250Crossref PubMed Scopus (6) Google Scholar, 8Hodgkin M.N. Pettitt T.R. Martin A. Wakelam M.J. Biochem. Soc. Trans. 1996; 24: 991-994Crossref PubMed Scopus (10) Google Scholar, 9Exton J. Physiol. Rev. 1997; 77: 303-320Crossref PubMed Scopus (386) Google Scholar), and (b) phospholipase D activity can be activated by numerous factors, including phosphatidylinositol 4,5-bisphosphate, protein kinase C, p60src, and the low molecular weight GTP-binding proteins (ADP-ribosylation factor, cdc42, and Rho A) (9Exton J. Physiol. Rev. 1997; 77: 303-320Crossref PubMed Scopus (386) Google Scholar). Although much is known about how PA is produced in response to cell stimulation, little is known about how the PA signal functions or is attenuated.We recently showed that bovine testis contains a phospholipase A1 (PLA1) that preferentially catalyzes the hydrolysis of PA in mixed micelle assay systems. The activity of this PA-preferring PLA1 (PA-PLA1) was present in high speed supernatant fractions from mature testis and brain, but not from newborn testis, liver, kidney, spleen, heart, or blood (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar). This restricted localization implied that the PLA1 might be taking part in a signaling pathway in brain and testis. We subsequently purified the enzyme to the point where only one major band of 110 kDa was observed on silver-stained SDS-PAGE, found evidence that the enzyme was a homotetramer in solution, and studied its enzymology (11Higgs H.N. Glomset J.A. J. Biol. Chem. 1996; 271: 10874-10883Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). Like the impure PA-PLA1, the purified enzyme catalyzed the preferential hydrolysis of PA. In addition, PA appeared to activate the PA-PLA1 reaction by promoting the enzyme's binding to micelles.To learn more about the enzymology and physical role of PA-PLA1, we attempted to clone and sequence its cDNA. This paper describes the primary structure of PA-PLA1, identifies a serine residue that is critical for its activity, examines the expression pattern of its mRNA in mammalian tissues, explores the possibility that splice variants for the protein exist, and identifies six related protein sequences. With this information, a detailed study of PA-PLA1 function can be undertaken.DISCUSSIONIn this study, we cloned and sequenced the cDNA for bovine testis PA-PLA1. In support of this conclusion, the sequences of six regions in the putative ORF corresponded to those of peptides isolated from digests of the purified bovine testis enzyme, the calculated molecular mass of the ORF (97,576 daltons) agreed well with the molecular mass of purified bovine testis PA-PLA1determined by MALDI, expression of the ORF in COS1 cells was accompanied by a 20–40-fold increase in PA-PLA1 activity, and serine 540, which is located in a region of the ORF that resembles a conserved sequence in lipases, was shown to be required for PA-PLA1 activity.With knowledge of the sequence of PA-PLA1 in hand, several important questions can now be addressed. For example, one set of questions concerns the structural basis of PA-PLA1activity. Are other specific amino acids in addition to serine 540 required for catalysis? What is the basis for the enzyme's substrate specificity? What regions of the PA-PLA1 sequence influence the enzyme's association with the membrane lipid bilayer? It may be possible to address these questions by taking advantage of some of the other observations made in the course of this study.Catalysis by several known lipases has been shown to require the presence of a "catalytic triad" of amino acids: the conserved serine nucleophile, a histidine residue, and an aspartic acid residue (see, for example, Ref. 32Wang A. Loo R. Chen Z. Dennis E.A. J. Biol. Chem. 1997; 272: 22030-22036Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). The results of the data base search with Block Maker that are shown in Fig. 5 may provide clues concerning the location of potential catalytic triad histidine and aspartic acid residues in PA-PLA1 and the putative lipases from C. elegans and yeast. Several of the regions of similarity that are shown contained aligned residues of histidine and aspartic acid that could be candidates for catalytic triad function, and it should be possible to examine the functional significance of these residues by experimentation with point mutants of PA-PLA1.As mentioned earlier, experiments with mixed micelles have suggested that PA-PLA1 may have both a substrate-binding site for PA, involved in catalysis, and multiple binding sites for PA involved in enzyme activation (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar, 11Higgs H.N. Glomset J.A. J. Biol. Chem. 1996; 271: 10874-10883Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). Experiments with well defined, unilamellar liposomes, currently under way in our laboratory, have supported this possibility. 3Q. Lin, unpublished experiments. However, the amino acids or domains that contribute to the PA-binding sites have yet to be identified. If expression studies of the ORFs from C. elegans and yeast confirm that they encode lipases, it might be possible to identify the relevant PA-binding sites in PA-PLA1 by carefully comparing the properties and sequences of the four enzymes.The experiments with unilamellar liposomes that have done so far have also provided evidence that non-substrate lipids influence the ability of PA-PLA1 to interact with PA. Therefore, it is possible that the enzyme may bind to liposome surfaces in additional, yet-to-be-identified ways. Domains that might be responsible for this binding remain to be identified, but the regions of sequence similarity, regions 5 and 6, shown in Fig. 5, might be good candidates for study because of their similarity to two regions identified in the putative membrane association domains of the RDGB proteins (Fig. 6).Another set of questions concerns the relation between PA-PLA1 and other lipases: Does PA-PLA1 belong to a special family of lipases? What defining structural and functional characteristics do members of this family share? How do members of the family differ from one another? Evidence related to the first of these questions is already beginning to accumulate.The sequence of PA-PLA1 definitely differs from the sequences of many other lipases, even though the region that surrounds serine 540 in PA-PLA1 resembles a conserved region in many of these lipases. For example, PA-PLA1 lacks further sequence similarity to types I–IV phospholipase A2 (33Dennis E.A. J. Biol. Chem. 1994; 269: 13057-13060Abstract Full Text PDF PubMed Google Scholar,34Tang J. Kriz R.W. Wolfman N. Shaffer M. Seehra J. Jones S.S. J. Biol. Chem. 1997; 272: 8567-8575Abstract Full Text Full Text PDF PubMed Scopus (273) Google Scholar), a phosphatidylserine-specific PLA1 from rat platelets (35Sato T. Aoki J. Nagai Y. Dohmae N. Taiko K. Doi T. Arai H. Inoue K. J. Biol. Chem. 1997; 272: 2192-2198Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar), lecithin:cholesterol acyltransferase (36McLean J. Fielding C. Drayna D. Dieplinger H. Baer B. Kohr V. Henzel W. Lawn R. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 2335-2339Crossref PubMed Scopus (143) Google Scholar), lysophospholipases (27Wang A. Deems R.A. Dennis E.A. J. Biol. Chem. 1997; 272: 12723-12729Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar, 37Sugimoto H. Hayashi H. Yamashita S. J. Biol. Chem. 1996; 271: 7705-7711Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar), and triacylglycerol lipases (38Wion K.L. Kirchgessner T.G. Lusis A.J. Schotz M.C. Lawn R.M. Science. 1987; 235: 1638-1641Crossref PubMed Scopus (349) Google Scholar, 39Bodmer M.W. Angal S. Yarranton G.T. Harris T.J. Lyons A. King D.J. Pieroni G. Riviere C. Verger R. Lowe P.A. Biochim. Biophys. Acta. 1987; 909: 237-244Crossref PubMed Scopus (138) Google Scholar). However, the sequence of PA-PLA1 does resemble that of CELG, SCERV, and SPOMB, as shown in Fig. 5; and SCERV and SPOMB appear to correspond to different, but closely related family members. We are currently attempting to prepare and express the cDNA that corresponds to SCERV to examine the catalytic properties of the putative SCERV lipase by direct experimentation.If PA-PLA1 does indeed belong to a special family of lipases, splice variants of PA-PLA1 might well be included. The three bovine testis cDNAs that contained the 123-base deletion shown in Fig. 1 may correspond to such splice variants. The 40 amino acids that would be eliminated by the deletion would normally be located between regions 2 and 3 identified by Block Maker (Fig. 4), but their functional significance is unknown. We have also obtained evidence for the existence of a second type of PA-PLA1splice variant. 4H. N. Higgs, M. H. Han, G. E. Johnson, and J. A. Glomset, unpublished results. We sequenced a cDNA clone from human infant brain, which corresponded to an expressed sequence tag (accession no. R13928), and found that it contained a 64-base deletion in another region of the ORF. Further work with these variants might suggest functions for the deleted regions.If splice variants of PA-PLA1 that have different substrate specificities are present in mammalian cells, an apparent discrepancy between the results of an earlier study (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar) and the expression patterns of human PA-PLA1 mRNA that are shown in Fig. 3(A and C) might conceivably be explained. We previously found high PA-PLA1 activity in high speed supernatant fractions from homogenates of mature bovine testis and brain, but little or no PA-PLA1 activity in corresponding fractions from newborn testis, liver, spleen, heart, kidney, and blood (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar). In good agreement with the measurements of enzyme activity in bovine tissues, expression of PA-PLA1 mRNA was highest by far in human testis, strong in brain, and very low in liver. However, unexpectedly, there also appeared to be expression of human PA-PLA1 mRNA in the spleen and kidney (Fig. 3 A) as well as in the heart (data not shown). One possible explanation of this unexpected result could be that the majority of the message detected in the spleen, kidney, and heart was due to PA-PLA1 splice variants, and that the enzymes that corresponded to these splice variants had substrate specificities that differed from that of PA-PLA1 and so escaped detection in our assay system.The presence of splice variants of PA-PLA1 might also account, at least in part, for the multiple message sizes for PA-PLA1 observed in this study. The >10-kb message is of particular interest because it was absent from testis but strongly present in lung, spleen, cerebellum, and fetal brain (Fig. 3,A and B). Do these different transcripts correspond to particular splice variants, such as the ones described above? Another possibility, not mutually exclusive with the first, is that differences in message lengths depend on differences in the 3′-untranslated regions of the messages, which confer different properties to the transcripts, such as variations in RNA half-life, translational control, or cellular localization (40Jackson R.J. Cell. 1993; 74: 9-14Abstract Full Text PDF PubMed Scopus (371) Google Scholar, 41Steward O. Neuron. 1997; 18: 9-12Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar).A different set of questions concerns the possible biological functions of PA-PLA1 and its close relatives. Answers to many of these questions will have to await the accumulation of further information about the PA-PLA1 lipase family. However, it may be possible to address questions related to the biological function of PA-PLA1 itself in the near future. As mentioned earlier, we have been focusing attention on PA-PLA1 because experiments with mixed micelle systems and unilamellar liposomes have shown that it hydrolyzes and is activated by PA, and because it is present in high levels in the mature testis and brains. This has suggested that it may function in phospholipase D- or diacylglycerol kinase-dependent signaling systems required for spermatogenesis and neuronal interactions. Now that the bovine testis PA-PLA1 has been sequenced and an antibody to the enzyme has been prepared, it may be possible to obtain further clues concerning the biological role of the enzyme using the tools of molecular biology and immunocytochemistry.In summary, this paper announces the cloning of the cDNA for a bovine testis PA-PLA1. Knowledge of the primary structure of PA-PLA1 will be of great utility in the characterization of its enzymatic properties and physiologic function. In addition, several different splice variants of the protein appear to exist, suggesting modulations of its function in different cells or sub-cellular regions. Finally, the PA-PLA1 ORF contains regions that are similar to regions in six other ORFs, suggesting interesting possibilities for further experimentation. Cellular membranes are highly dynamic structures, and this dynamic nature plays a major role in cellular physiology. In response to extracellular or intracellular signals, a number of phospholipases can be activated, including phosphoinositide-specific phospholipase C, phosphatidylcholine-specific phospholipase C, phospholipase D, phosphatidic acid (PA) 1The abbreviations used are: PA, phosphatidic acid; CELG, predicted sequence of an unknown protein from C. elegans; COS1 cells, SV40-transformed African Green monkey kidney cells; 5′-RACE, 5′-rapid amplification of cDNA ends; kb, kilobase pair(s); MALDI, matrix-assisted laser desorption ionization; MOPS, 3-(N-morpholino)propanesulfonic acid; ORF, open reading frame; PA-PLA1, phosphatidic acid-preferring phospholipase A1; PCR, polymerase chain reaction; PITP, phosphatidylinositol transfer protein; PLA1, phospholipase A1; RDGBd, RDGBh, and RDGBm: sequences encoded by the retinal degeneration B gene in Drosophila and the corresponding genes in humans and mice; SCERV, predicted sequence of an unknown protein from S. cerevisiae; SPOMB, predicted sequence of an unknown protein from S. pombe; PAGE, polyacrylamide gel electrophoresis; 3′-UTR, 3′-untranslated region. 1The abbreviations used are: PA, phosphatidic acid; CELG, predicted sequence of an unknown protein from C. elegans; COS1 cells, SV40-transformed African Green monkey kidney cells; 5′-RACE, 5′-rapid amplification of cDNA ends; kb, kilobase pair(s); MALDI, matrix-assisted laser desorption ionization; MOPS, 3-(N-morpholino)propanesulfonic acid; ORF, open reading frame; PA-PLA1, phosphatidic acid-preferring phospholipase A1; PCR, polymerase chain reaction; PITP, phosphatidylinositol transfer protein; PLA1, phospholipase A1; RDGBd, RDGBh, and RDGBm: sequences encoded by the retinal degeneration B gene in Drosophila and the corresponding genes in humans and mice; SCERV, predicted sequence of an unknown protein from S. cerevisiae; SPOMB, predicted sequence of an unknown protein from S. pombe; PAGE, polyacrylamide gel electrophoresis; 3′-UTR, 3′-untranslated region.phosphohydrolase, phospholipase A2, and sphingomyelinase (1Liscovitch M. Signal-activated Phospholipases. R. G. Landes Co., Austin, TX1994Google Scholar, 2Singer W.D. Brown H.A. Sternweiss P.C. Annu. Rev. Biochem. 1997; 66: 475-509Crossref PubMed Scopus (347) Google Scholar). These enzymes produce second messengers that function in cellular regulation. One second messenger that has received increasing attention is PA. This phospholipid can be produced either by the phospholipase D-catalyzed degradation of phosphatidylcholine or by a two-step pathway involving the phospholipase C-catalyzed degradation of phosphatidylinositol 4,5-bisphosphate to diacylglycerol and the subsequent, diacylglycerol kinase-catalyzed phosphorylation of this diacylglycerol (3Bocckino S.B. Exton J., H. Bell R.M. Exton J.H. Prescott S.M. Handbook of Lipid Research: Lipid Second Messengers. 8. Plenum Press, New York1996: 75-123Google Scholar). PA appears to function in a number of cellular processes, including superoxide production in neutrophils (4Cockcroft S. Biochim. Biophys. Acta. 1992; 1113: 135-160Crossref PubMed Scopus (243) Google Scholar, 5Bauldry S.A. Bass D.A. Cousart S.L. McCall C.E. J. Biol. Chem. 1991; 266: 4173-4179Abstract Full Text PDF PubMed Google Scholar) and actin polymerization (6Ha K.S. Exton J.H. J. Cell Biol. 1993; 123: 1789-1796Crossref PubMed Scopus (154) Google Scholar). Other roles for PA are suggested by the facts that (a) mammalian tissues contain a large (10 isoforms) family of diacylglycerol kinases as well as several phospholipase D enzymes (7Kanoh H. Kai M. Wada I. J. Lipid Mediators Cell. Signalling. 1996; 14: 245-250Crossref PubMed Scopus (6) Google Scholar, 8Hodgkin M.N. Pettitt T.R. Martin A. Wakelam M.J. Biochem. Soc. Trans. 1996; 24: 991-994Crossref PubMed Scopus (10) Google Scholar, 9Exton J. Physiol. Rev. 1997; 77: 303-320Crossref PubMed Scopus (386) Google Scholar), and (b) phospholipase D activity can be activated by numerous factors, including phosphatidylinositol 4,5-bisphosphate, protein kinase C, p60src, and the low molecular weight GTP-binding proteins (ADP-ribosylation factor, cdc42, and Rho A) (9Exton J. Physiol. Rev. 1997; 77: 303-320Crossref PubMed Scopus (386) Google Scholar). Although much is known about how PA is produced in response to cell stimulation, little is known about how the PA signal functions or is attenuated. We recently showed that bovine testis contains a phospholipase A1 (PLA1) that preferentially catalyzes the hydrolysis of PA in mixed micelle assay systems. The activity of this PA-preferring PLA1 (PA-PLA1) was present in high speed supernatant fractions from mature testis and brain, but not from newborn testis, liver, kidney, spleen, heart, or blood (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar). This restricted localization implied that the PLA1 might be taking part in a signaling pathway in brain and testis. We subsequently purified the enzyme to the point where only one major band of 110 kDa was observed on silver-stained SDS-PAGE, found evidence that the enzyme was a homotetramer in solution, and studied its enzymology (11Higgs H.N. Glomset J.A. J. Biol. Chem. 1996; 271: 10874-10883Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). Like the impure PA-PLA1, the purified enzyme catalyzed the preferential hydrolysis of PA. In addition, PA appeared to activate the PA-PLA1 reaction by promoting the enzyme's binding to micelles. To learn more about the enzymology and physical role of PA-PLA1, we attempted to clone and sequence its cDNA. This paper describes the primary structure of PA-PLA1, identifies a serine residue that is critical for its activity, examines the expression pattern of its mRNA in mammalian tissues, explores the possibility that splice variants for the protein exist, and identifies six related protein sequences. With this information, a detailed study of PA-PLA1 function can be undertaken. DISCUSSIONIn this study, we cloned and sequenced the cDNA for bovine testis PA-PLA1. In support of this conclusion, the sequences of six regions in the putative ORF corresponded to those of peptides isolated from digests of the purified bovine testis enzyme, the calculated molecular mass of the ORF (97,576 daltons) agreed well with the molecular mass of purified bovine testis PA-PLA1determined by MALDI, expression of the ORF in COS1 cells was accompanied by a 20–40-fold increase in PA-PLA1 activity, and serine 540, which is located in a region of the ORF that resembles a conserved sequence in lipases, was shown to be required for PA-PLA1 activity.With knowledge of the sequence of PA-PLA1 in hand, several important questions can now be addressed. For example, one set of questions concerns the structural basis of PA-PLA1activity. Are other specific amino acids in addition to serine 540 required for catalysis? What is the basis for the enzyme's substrate specificity? What regions of the PA-PLA1 sequence influence the enzyme's association with the membrane lipid bilayer? It may be possible to address these questions by taking advantage of some of the other observations made in the course of this study.Catalysis by several known lipases has been shown to require the presence of a "catalytic triad" of amino acids: the conserved serine nucleophile, a histidine residue, and an aspartic acid residue (see, for example, Ref. 32Wang A. Loo R. Chen Z. Dennis E.A. J. Biol. Chem. 1997; 272: 22030-22036Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). The results of the data base search with Block Maker that are shown in Fig. 5 may provide clues concerning the location of potential catalytic triad histidine and aspartic acid residues in PA-PLA1 and the putative lipases from C. elegans and yeast. Several of the regions of similarity that are shown contained aligned residues of histidine and aspartic acid that could be candidates for catalytic triad function, and it should be possible to examine the functional significance of these residues by experimentation with point mutants of PA-PLA1.As mentioned earlier, experiments with mixed micelles have suggested that PA-PLA1 may have both a substrate-binding site for PA, involved in catalysis, and multiple binding sites for PA involved in enzyme activation (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar, 11Higgs H.N. Glomset J.A. J. Biol. Chem. 1996; 271: 10874-10883Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). Experiments with well defined, unilamellar liposomes, currently under way in our laboratory, have supported this possibility. 3Q. Lin, unpublished experiments. However, the amino acids or domains that contribute to the PA-binding sites have yet to be identified. If expression studies of the ORFs from C. elegans and yeast confirm that they encode lipases, it might be possible to identify the relevant PA-binding sites in PA-PLA1 by carefully comparing the properties and sequences of the four enzymes.The experiments with unilamellar liposomes that have done so far have also provided evidence that non-substrate lipids influence the ability of PA-PLA1 to interact with PA. Therefore, it is possible that the enzyme may bind to liposome surfaces in additional, yet-to-be-identified ways. Domains that might be responsible for this binding remain to be identified, but the regions of sequence similarity, regions 5 and 6, shown in Fig. 5, might be good candidates for study because of their similarity to two regions identified in the putative membrane association domains of the RDGB proteins (Fig. 6).Another set of questions concerns the relation between PA-PLA1 and other lipases: Does PA-PLA1 belong to a special family of lipases? What defining structural and functional characteristics do members of this family share? How do members of the family differ from one another? Evidence related to the first of these questions is already beginning to accumulate.The sequence of PA-PLA1 definitely differs from the sequences of many other lipases, even though the region that surrounds serine 540 in PA-PLA1 resembles a conserved region in many of these lipases. For example, PA-PLA1 lacks further sequence similarity to types I–IV phospholipase A2 (33Dennis E.A. J. Biol. Chem. 1994; 269: 13057-13060Abstract Full Text PDF PubMed Google Scholar,34Tang J. Kriz R.W. Wolfman N. Shaffer M. Seehra J. Jones S.S. J. Biol. Chem. 1997; 272: 8567-8575Abstract Full Text Full Text PDF PubMed Scopus (273) Google Scholar), a phosphatidylserine-specific PLA1 from rat platelets (35Sato T. Aoki J. Nagai Y. Dohmae N. Taiko K. Doi T. Arai H. Inoue K. J. Biol. Chem. 1997; 272: 2192-2198Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar), lecithin:cholesterol acyltransferase (36McLean J. Fielding C. Drayna D. Dieplinger H. Baer B. Kohr V. Henzel W. Lawn R. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 2335-2339Crossref PubMed Scopus (143) Google Scholar), lysophospholipases (27Wang A. Deems R.A. Dennis E.A. J. Biol. Chem. 1997; 272: 12723-12729Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar, 37Sugimoto H. Hayashi H. Yamashita S. J. Biol. Chem. 1996; 271: 7705-7711Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar), and triacylglycerol lipases (38Wion K.L. Kirchgessner T.G. Lusis A.J. Schotz M.C. Lawn R.M. Science. 1987; 235: 1638-1641Crossref PubMed Scopus (349) Google Scholar, 39Bodmer M.W. Angal S. Yarranton G.T. Harris T.J. Lyons A. King D.J. Pieroni G. Riviere C. Verger R. Lowe P.A. Biochim. Biophys. Acta. 1987; 909: 237-244Crossref PubMed Scopus (138) Google Scholar). However, the sequence of PA-PLA1 does resemble that of CELG, SCERV, and SPOMB, as shown in Fig. 5; and SCERV and SPOMB appear to correspond to different, but closely related family members. We are currently attempting to prepare and express the cDNA that corresponds to SCERV to examine the catalytic properties of the putative SCERV lipase by direct experimentation.If PA-PLA1 does indeed belong to a special family of lipases, splice variants of PA-PLA1 might well be included. The three bovine testis cDNAs that contained the 123-base deletion shown in Fig. 1 may correspond to such splice variants. The 40 amino acids that would be eliminated by the deletion would normally be located between regions 2 and 3 identified by Block Maker (Fig. 4), but their functional significance is unknown. We have also obtained evidence for the existence of a second type of PA-PLA1splice variant. 4H. N. Higgs, M. H. Han, G. E. Johnson, and J. A. Glomset, unpublished results. We sequenced a cDNA clone from human infant brain, which corresponded to an expressed sequence tag (accession no. R13928), and found that it contained a 64-base deletion in another region of the ORF. Further work with these variants might suggest functions for the deleted regions.If splice variants of PA-PLA1 that have different substrate specificities are present in mammalian cells, an apparent discrepancy between the results of an earlier study (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar) and the expression patterns of human PA-PLA1 mRNA that are shown in Fig. 3(A and C) might conceivably be explained. We previously found high PA-PLA1 activity in high speed supernatant fractions from homogenates of mature bovine testis and brain, but little or no PA-PLA1 activity in corresponding fractions from newborn testis, liver, spleen, heart, kidney, and blood (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar). In good agreement with the measurements of enzyme activity in bovine tissues, expression of PA-PLA1 mRNA was highest by far in human testis, strong in brain, and very low in liver. However, unexpectedly, there also appeared to be expression of human PA-PLA1 mRNA in the spleen and kidney (Fig. 3 A) as well as in the heart (data not shown). One possible explanation of this unexpected result could be that the majority of the message detected in the spleen, kidney, and heart was due to PA-PLA1 splice variants, and that the enzymes that corresponded to these splice variants had substrate specificities that differed from that of PA-PLA1 and so escaped detection in our assay system.The presence of splice variants of PA-PLA1 might also account, at least in part, for the multiple message sizes for PA-PLA1 observed in this study. The >10-kb message is of particular interest because it was absent from testis but strongly present in lung, spleen, cerebellum, and fetal brain (Fig. 3,A and B). Do these different transcripts correspond to particular splice variants, such as the ones described above? Another possibility, not mutually exclusive with the first, is that differences in message lengths depend on differences in the 3′-untranslated regions of the messages, which confer different properties to the transcripts, such as variations in RNA half-life, translational control, or cellular localization (40Jackson R.J. Cell. 1993; 74: 9-14Abstract Full Text PDF PubMed Scopus (371) Google Scholar, 41Steward O. Neuron. 1997; 18: 9-12Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar).A different set of questions concerns the possible biological functions of PA-PLA1 and its close relatives. Answers to many of these questions will have to await the accumulation of further information about the PA-PLA1 lipase family. However, it may be possible to address questions related to the biological function of PA-PLA1 itself in the near future. As mentioned earlier, we have been focusing attention on PA-PLA1 because experiments with mixed micelle systems and unilamellar liposomes have shown that it hydrolyzes and is activated by PA, and because it is present in high levels in the mature testis and brains. This has suggested that it may function in phospholipase D- or diacylglycerol kinase-dependent signaling systems required for spermatogenesis and neuronal interactions. Now that the bovine testis PA-PLA1 has been sequenced and an antibody to the enzyme has been prepared, it may be possible to obtain further clues concerning the biological role of the enzyme using the tools of molecular biology and immunocytochemistry.In summary, this paper announces the cloning of the cDNA for a bovine testis PA-PLA1. Knowledge of the primary structure of PA-PLA1 will be of great utility in the characterization of its enzymatic properties and physiologic function. In addition, several different splice variants of the protein appear to exist, suggesting modulations of its function in different cells or sub-cellular regions. Finally, the PA-PLA1 ORF contains regions that are similar to regions in six other ORFs, suggesting interesting possibilities for further experimentation. In this study, we cloned and sequenced the cDNA for bovine testis PA-PLA1. In support of this conclusion, the sequences of six regions in the putative ORF corresponded to those of peptides isolated from digests of the purified bovine testis enzyme, the calculated molecular mass of the ORF (97,576 daltons) agreed well with the molecular mass of purified bovine testis PA-PLA1determined by MALDI, expression of the ORF in COS1 cells was accompanied by a 20–40-fold increase in PA-PLA1 activity, and serine 540, which is located in a region of the ORF that resembles a conserved sequence in lipases, was shown to be required for PA-PLA1 activity. With knowledge of the sequence of PA-PLA1 in hand, several important questions can now be addressed. For example, one set of questions concerns the structural basis of PA-PLA1activity. Are other specific amino acids in addition to serine 540 required for catalysis? What is the basis for the enzyme's substrate specificity? What regions of the PA-PLA1 sequence influence the enzyme's association with the membrane lipid bilayer? It may be possible to address these questions by taking advantage of some of the other observations made in the course of this study. Catalysis by several known lipases has been shown to require the presence of a "catalytic triad" of amino acids: the conserved serine nucleophile, a histidine residue, and an aspartic acid residue (see, for example, Ref. 32Wang A. Loo R. Chen Z. Dennis E.A. J. Biol. Chem. 1997; 272: 22030-22036Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). The results of the data base search with Block Maker that are shown in Fig. 5 may provide clues concerning the location of potential catalytic triad histidine and aspartic acid residues in PA-PLA1 and the putative lipases from C. elegans and yeast. Several of the regions of similarity that are shown contained aligned residues of histidine and aspartic acid that could be candidates for catalytic triad function, and it should be possible to examine the functional significance of these residues by experimentation with point mutants of PA-PLA1. As mentioned earlier, experiments with mixed micelles have suggested that PA-PLA1 may have both a substrate-binding site for PA, involved in catalysis, and multiple binding sites for PA involved in enzyme activation (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar, 11Higgs H.N. Glomset J.A. J. Biol. Chem. 1996; 271: 10874-10883Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). Experiments with well defined, unilamellar liposomes, currently under way in our laboratory, have supported this possibility. 3Q. Lin, unpublished experiments. However, the amino acids or domains that contribute to the PA-binding sites have yet to be identified. If expression studies of the ORFs from C. elegans and yeast confirm that they encode lipases, it might be possible to identify the relevant PA-binding sites in PA-PLA1 by carefully comparing the properties and sequences of the four enzymes. The experiments with unilamellar liposomes that have done so far have also provided evidence that non-substrate lipids influence the ability of PA-PLA1 to interact with PA. Therefore, it is possible that the enzyme may bind to liposome surfaces in additional, yet-to-be-identified ways. Domains that might be responsible for this binding remain to be identified, but the regions of sequence similarity, regions 5 and 6, shown in Fig. 5, might be good candidates for study because of their similarity to two regions identified in the putative membrane association domains of the RDGB proteins (Fig. 6). Another set of questions concerns the relation between PA-PLA1 and other lipases: Does PA-PLA1 belong to a special family of lipases? What defining structural and functional characteristics do members of this family share? How do members of the family differ from one another? Evidence related to the first of these questions is already beginning to accumulate. The sequence of PA-PLA1 definitely differs from the sequences of many other lipases, even though the region that surrounds serine 540 in PA-PLA1 resembles a conserved region in many of these lipases. For example, PA-PLA1 lacks further sequence similarity to types I–IV phospholipase A2 (33Dennis E.A. J. Biol. Chem. 1994; 269: 13057-13060Abstract Full Text PDF PubMed Google Scholar,34Tang J. Kriz R.W. Wolfman N. Shaffer M. Seehra J. Jones S.S. J. Biol. Chem. 1997; 272: 8567-8575Abstract Full Text Full Text PDF PubMed Scopus (273) Google Scholar), a phosphatidylserine-specific PLA1 from rat platelets (35Sato T. Aoki J. Nagai Y. Dohmae N. Taiko K. Doi T. Arai H. Inoue K. J. Biol. Chem. 1997; 272: 2192-2198Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar), lecithin:cholesterol acyltransferase (36McLean J. Fielding C. Drayna D. Dieplinger H. Baer B. Kohr V. Henzel W. Lawn R. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 2335-2339Crossref PubMed Scopus (143) Google Scholar), lysophospholipases (27Wang A. Deems R.A. Dennis E.A. J. Biol. Chem. 1997; 272: 12723-12729Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar, 37Sugimoto H. Hayashi H. Yamashita S. J. Biol. Chem. 1996; 271: 7705-7711Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar), and triacylglycerol lipases (38Wion K.L. Kirchgessner T.G. Lusis A.J. Schotz M.C. Lawn R.M. Science. 1987; 235: 1638-1641Crossref PubMed Scopus (349) Google Scholar, 39Bodmer M.W. Angal S. Yarranton G.T. Harris T.J. Lyons A. King D.J. Pieroni G. Riviere C. Verger R. Lowe P.A. Biochim. Biophys. Acta. 1987; 909: 237-244Crossref PubMed Scopus (138) Google Scholar). However, the sequence of PA-PLA1 does resemble that of CELG, SCERV, and SPOMB, as shown in Fig. 5; and SCERV and SPOMB appear to correspond to different, but closely related family members. We are currently attempting to prepare and express the cDNA that corresponds to SCERV to examine the catalytic properties of the putative SCERV lipase by direct experimentation. If PA-PLA1 does indeed belong to a special family of lipases, splice variants of PA-PLA1 might well be included. The three bovine testis cDNAs that contained the 123-base deletion shown in Fig. 1 may correspond to such splice variants. The 40 amino acids that would be eliminated by the deletion would normally be located between regions 2 and 3 identified by Block Maker (Fig. 4), but their functional significance is unknown. We have also obtained evidence for the existence of a second type of PA-PLA1splice variant. 4H. N. Higgs, M. H. Han, G. E. Johnson, and J. A. Glomset, unpublished results. We sequenced a cDNA clone from human infant brain, which corresponded to an expressed sequence tag (accession no. R13928), and found that it contained a 64-base deletion in another region of the ORF. Further work with these variants might suggest functions for the deleted regions. If splice variants of PA-PLA1 that have different substrate specificities are present in mammalian cells, an apparent discrepancy between the results of an earlier study (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar) and the expression patterns of human PA-PLA1 mRNA that are shown in Fig. 3(A and C) might conceivably be explained. We previously found high PA-PLA1 activity in high speed supernatant fractions from homogenates of mature bovine testis and brain, but little or no PA-PLA1 activity in corresponding fractions from newborn testis, liver, spleen, heart, kidney, and blood (10Higgs H.N. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9574-9578Crossref PubMed Scopus (73) Google Scholar). In good agreement with the measurements of enzyme activity in bovine tissues, expression of PA-PLA1 mRNA was highest by far in human testis, strong in brain, and very low in liver. However, unexpectedly, there also appeared to be expression of human PA-PLA1 mRNA in the spleen and kidney (Fig. 3 A) as well as in the heart (data not shown). One possible explanation of this unexpected result could be that the majority of the message detected in the spleen, kidney, and heart was due to PA-PLA1 splice variants, and that the enzymes that corresponded to these splice variants had substrate specificities that differed from that of PA-PLA1 and so escaped detection in our assay system. The presence of splice variants of PA-PLA1 might also account, at least in part, for the multiple message sizes for PA-PLA1 observed in this study. The >10-kb message is of particular interest because it was absent from testis but strongly present in lung, spleen, cerebellum, and fetal brain (Fig. 3,A and B). Do these different transcripts correspond to particular splice variants, such as the ones described above? Another possibility, not mutually exclusive with the first, is that differences in message lengths depend on differences in the 3′-untranslated regions of the messages, which confer different properties to the transcripts, such as variations in RNA half-life, translational control, or cellular localization (40Jackson R.J. Cell. 1993; 74: 9-14Abstract Full Text PDF PubMed Scopus (371) Google Scholar, 41Steward O. Neuron. 1997; 18: 9-12Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). A different set of questions concerns the possible biological functions of PA-PLA1 and its close relatives. Answers to many of these questions will have to await the accumulation of further information about the PA-PLA1 lipase family. However, it may be possible to address questions related to the biological function of PA-PLA1 itself in the near future. As mentioned earlier, we have been focusing attention on PA-PLA1 because experiments with mixed micelle systems and unilamellar liposomes have shown that it hydrolyzes and is activated by PA, and because it is present in high levels in the mature testis and brains. This has suggested that it may function in phospholipase D- or diacylglycerol kinase-dependent signaling systems required for spermatogenesis and neuronal interactions. Now that the bovine testis PA-PLA1 has been sequenced and an antibody to the enzyme has been prepared, it may be possible to obtain further clues concerning the biological role of the enzyme using the tools of molecular biology and immunocytochemistry. In summary, this paper announces the cloning of the cDNA for a bovine testis PA-PLA1. Knowledge of the primary structure of PA-PLA1 will be of great utility in the characterization of its enzymatic properties and physiologic function. In addition, several different splice variants of the protein appear to exist, suggesting modulations of its function in different cells or sub-cellular regions. Finally, the PA-PLA1 ORF contains regions that are similar to regions in six other ORFs, suggesting interesting possibilities for further experimentation. We thank Rudi Aebersold, Santosh Kumar, and Ken Walsh for MALDI analysis of the purified protein and peptide sequencing and Rosa Suen and Laurie Aicher for advice on molecular biology.
Referência(s)