Cloning and Recombinant Expression of a Structurally Novel Human Secreted Phospholipase A2

The nucleotide sequence reported in this paper has been submitted to the GenBank"/EBI Data Bank with accession number AF306567. Abbreviations : : Secreted phospholipase A 2 ; sPLA 2 ; POPC/G/S, 1-palmitoyl-2-oleoyl- sn -glycero-3-phospho-choline/glycerol/serine. A comprehensive abbreviation system for the various mammalian sPLA2s is used: Each sPLA2 is abbreviated with a lowercase letter indicating the sPLA2 species (m and h for mouse and human, respectively), followed by uppercase letters identifying the sPLA2 group (GIB, GIIA, GIIC, GIID, GIIE, GIIF, GIII, GV, GIX, GX, GXI, and GXII). Mammals contain a diverse set of secreted phospholipases A2 (sPLA 2 s) which liberate arachidonic acid from phospholipids for the production of eicosanoids and exert a variety of physiological and pathological effects. We report the cloning, recombinant expression and kinetic properties of a novel human sPLA 2 which defines a new structural class of sPLA2s called group XII. The human group XII (hGXII) cDNA contains a putative signal peptide of 22 residues followed by a mature protein of 167 amino acids that displays homology to all known sPLA2s only over a short stretch of amino acids in the active site region. Northern blot and RT-PCR analyses show that the tissue distribution of hGXII is distinct from the other human sPLA2s with strong expression in heart, skeletal muscle, kidney, and pancreas and weaker expression in brain, liver, small intestine, lung, placenta, ovaries, testis, and prostate. Catalytically active hGXII was produced in E. coli and shown to be Ca2+-dependent despite the fact that it is predicted to have an unusual Ca2+ binding loop. Like for the previously characterized mouse group IIE sPLA2s, the specific activity of hGXII is low in comparison to those of other mammalian sPLA2, suggesting that hGXII could have novel functions that are independent of its phospholipase A2 activity.

Secreted phospholipases A 2 (sPLA 2 ) are Ca 2+ -dependent, disulfide-rich, 14-18 kDa enzymes that catalyze the hydrolysis of phospholipids at the sn 2 position to release fatty acids and lysophospholipids (1)(2)(3). In mammalian cells stimulated with proinflammatory agonists, a subset of sPLA 2 s are involved in the release of arachidonic acid for eicosanoid production (4,5). The first mammalian sPLA 2 to be identified was the pancreatic sPLA 2 . This sPLA 2 is found at high levels in pancreatic juice, where it has a well-known function in the digestion of dietary phospholipids (6), but also at lower levels in lung, liver, spleen, kidney, and ovary where it has been proposed to play a role in cell proliferation, acute lung injury, cell migration, and endotoxic shock (7)(8)(9). The first non-pancreatic mammalian sPLA 2 to be identified was the group IIA enzyme which is expressed at high levels during inflammation (10), and is the principal bactericidal agent against Gram-positive bacteria in human tears (11).
In addition to the above evidence, it is becoming clear that sPLA 2 s are involved in a diverse set of physiological functions (7,(12)(13)(14). In the last few years, 6 mouse and 5 human sPLA 2 s structurally related to GIB and GIIA sPLA 2 s (mGIIC, hGIID, mGIID, hGIIE, mGIIE, mGIIF, hGIIF, hGV, mGV, hGX, and mGX) have been identified (15)(16)(17)(18)(19)(20) (Valentin et al., submitted for publication). All of these group I/II/V/X sPLA 2 s have similar primary structures, including identical catalytic site residues and partially overlapping sets of disulfides (21). However, they are not closely related isoforms since the level of amino acid identity is typically 20-50% among these sPLA 2 s. More recently, a novel human group III sPLA 2 was identified (22), which is structurally distinct from the group I/II/V/X sPLA 2 s but related to the group III sPLA 2 s found in bee and lizard venoms. This diversity of sPLA 2 structures and the fact that the tissue distribution of the different sPLA 2 s are distinct argue for a diversity of physiological functions for these lipolytic enzymes.
It is also clear that mammals contain a collection of proteins that tightly bind sPLA 2 s. Two types of sPLA 2 receptors (M-and N-type) and some other soluble sPLA 2 binding proteins have been identified (7,13,21,(23)(24)(25) and are likely to play a role in the physiological functions of mammalian sPLA 2 s and in the toxicity of a wide variety of myotoxic and neurotoxic sPLA 2 s found in reptile and invertebrate venoms. Very recently, the cell surface proteoglycan glypican was also identified as a sPLA 2 binding protein able to facilitate arachidonic acid release by GIIA and GV sPLA 2 s in fibroblastic cells (26).
Because of the presence of a large collection of sPLA 2 s in both mammals and many reptile and invertebrate venoms, we have been searching nucleic acid databases for the presence of novel mammalian sPLA 2 s with homology to all known types of these enzymes including structurally distinct ones like the group IX sPLA 2 (Conodipine-M) from the venom of the cone snail Conus magus (27). In this paper, we report the cloning, recombinant expression, tissue distribution, and enzymatic properties of a novel human sPLA 2 . Because this sPLA 2 clearly belongs to a new structural class, we propose to name it human group XII sPLA 2 (hGXII) to follow the recently identified group XI plant sPLA 2 s (21,28,29). flanking the above open reading frame and containing appropriate restriction sites were used to amplify by RT-PCR the cDNA fragment coding for hGXII sPLA 2 . The expected 717 nucleotide hGXII cDNA fragment could be amplified from human fetal lung, pancreas and testis cDNAs (Clontech) using a Taq Pwo polymerase mixture (Hybaid, UK). The PCR fragments were digested with Not I and Xba I, ligated into the mammalian expression vector pRc/CMV neo (Invitrogen), and entirely sequenced. Several clones were found to be identical to the consensus sequence described above.

Recombinant Expression of hGXII sPLA 2
˙ The pRc/CMVneo-hGXII S transferase (~10 kDa) encoded by the modified pGEX-2T vector (pAB3), which has been previously used to express several sPLA 2 s in E. coli (17). Protein production in E. coli BL21, purification of inclusion bodies, and refolding and cleavage of the fusion protein with factor Xa were carried out as described (17). suggesting that this protein belongs to the sPLA 2 family. The homology however appears to be weak (< 35% identity with blast scores lower than 35) and restricted to a short stretch of less than 60 amino acid residues containing the active site domain and the HD catalytic diad, indicating that the hGXII sPLA 2 is unique among all known sPLA 2 s (Fig. 1B). The histidine of HD is thought to function as a general base to deprotonate a water molecule as it attacks the substrate ester carbonyl carbon, and the ß-carboxyl group of the adjacent aspartate coordinates directly to the catalytic Ca 2+ cofactor (6,33). Except for 3 cysteines in the active site consensus sequence CCXXHDXC which match those of other groups of sPLA 2 s, the location of the other 11 cysteines residues in hGXII is distinct from that of other sPLA 2 s (Fig. 1B). Since the structural arrangement of disulfides has been the main basis for designating the different sPLA 2 group numbers, the naming of the new sPLA 2 as hGXII seems appropriate.
The homology between hGXII and all known sPLA 2 s is so low that it is difficult to find the Ca 2+ binding loop, which is usually highly conserved and provides 3 of the 4 amino acid ligands for the catalytic Ca 2+ (34). All mammalian group I, II, V, and X sPLA 2 s contain 19 amino acid residues between the most N-terminal residue that serves as a ligand to the active site Ca 2+ (i.e. His-27 of hGIIA) and the catalytic histidine (i.e. His-47 of hGIIA). In contrast, the corresponding distances for hGIII and plant GXI sPLA 2 s are 25 and 23 residues, respectively. hGXII contains a potential Ca 2+ binding segment GCGSP with 23 residues between the N-terminal glycine and the putative catalytic histidine as shown in Fig. 1. This segment is perfectly conserved among all of the GXII proteins found in gene databases. The xray structures of groups I, II, and III sPLA 2 s reveal that the Ca 2+ loop contains the consensus segment X 1 CG 1 X 2 G 2 . The backbone carbonyl oxygens of residues X 1 , G 1 , and G 2 coordinate to Ca 2+ , and the backbone NH of G 1 is proposed to donate a hydrogen bond to the carbonyl oxygen of the enzyme-susceptible substrate ester (33,35). The fact that this residue is glycine in catalytically active sPLA 2 s and that mutating this residue to serine lowers catalytic activity by about 10-to 20-fold (35) argues that steric bulk is poorly tolerated at this position. The putative Ca 2+coordinating segment of hGXII shown in Fig. 1B fits the consensus sequence of other sPLA 2 s with the exception that G 2 is a proline in hGXII. The prediction based on examination of the x-ray structures of sPLA2 s is that the hGXII Ca 2+ binding segment should be functional. It contains G 1 , and the backbone carbonyl of the Cterminal proline can coordinate to Ca 2+ since its three extra methylenes, compared to glycine, are sterically allowed because of the location of this residue on the enzyme´s surface away from the substrate binding cavity. Interestingly, sPLA 2 isozymes with relatively low sPLA 2 activity from the venom of the banded krait also contain proline in place of G 2 (36).
Tissue Distribution of hGXII sPLA 2 ˙ The tissue distribution of hGXII was first analyzed by hybridization at high stringency to a human northern blot (Fig. 2).
hGXII is expressed as several transcripts including a major one of ~1.4 kilobase, which is abundant in heart, skeletal muscle and kidney. hGXII transcripts are also present at lower levels in brain, liver, small intestine, lung and placenta, and expressed poorly, if at all, in colon, thymus, spleen and peripheral blood leukocytes.
Furthermore, analysis by RT-PCR with commercial human tissue cDNA panels indicates a pattern of hGXII expression that is consistent with the northern blot data and additionally shows that this sPLA 2 is strongly expressed in pancreas, and weakly in ovaries, testis, and prostate (not shown).  Fig. 1A. This result indicates that all 14 cysteines are engaged in disulfide bonds, and thus it is assumed that recombinant hGXII is properly folded.
protonation of the active site histidine. As for all mammalian sPLA 2 s examined so far, the enzymatic activity of hGXII on phosphatidylglycerol vesicles is highest (Fig.   3C), which probably reflects the tighter binding of hGXII to anionic vesicles (37).
Although hGXII hydrolyzes POPC at only ~7% of the rate of POPG, this difference is small compared to the greater than 10 5 -fold preference of hGIIA for POPG versus POPC (18). POPS is also a good substrate for hGXII (Fig. 3C). of hGXII shows that it is a catalytically active, Ca 2+ -dependent sPLA 2 . However, the specific enzymatic activity of hGXII appears very low compared to those of other mammalian sPLA 2 s (for example hGIB, hGIIA, hGV, hGX) and is comparable to the low specific activity reported for mGIIE sPLA 2 (18). This may be the reason why hGXII sPLA 2 was not detected in earlier biochemical studies, despite the fact that this protein may be expressed at significant levels, as the transcripts coding for this later are present in fairly high amounts in several human tissues (Fig. 2). It is also interesting to note that the putative GXII sPLA 2 from zebrafish (Danio rerio) is represented in gene databases by several ESTs that all contain a leucine in place of histidine in the catalytic HD segment. This type of mutation suggests that the zebrafish GXII sPLA 2 has very low or no catalytic activity, probably lower than that of hGXII sPLA 2 . This in turn suggests that the catalytic activity of group XII sPLA 2 s may not be critical for their physiological functions, and that they may act by serving as ligands for sPLA 2 binding proteins rather than by acting as lipolytic enzymes (13).