Cloning and recombinant expression of phospholipase A2 present in rheumatoid arthritic synovial fluid.

Synovial fluid from arthritic patients contains multiple forms of phospholipase A2 (PLA2), as resolved by high performance liquid chromatography (Seilhamer, J.J., Plant, S., Pruzanski, W., Schilling, J., Stefanski, E., Vadas, P., and Johnson, L. K. (1989) J. Biochem. (Tokyo), submitted for publication). Here we describe the cloning of a human 4.5-kilobase gene and 800-base pair cDNA encoding the form representing the major peak of activity and protein mass (peak A). The clones encode a mature peptide of 124 amino acids, which follows a prepeptide of 20 residues. The deduced amino acid sequence constitutes an enzyme of the "Type II" class of PLA2s, and resembles PLA2s from other mammalian sources. This represents the first report of a full length mammalian non-pancreatic PLA2 sequence. Active transcription of this PLA2 gene was detected in two different inflammatory cell sources. Recombinant human peak A PLA2 was expressed in vaccinia as a secreted protein which accumulated in conditioned medium.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s)
tures (5). Inhibition of the specific PLAz enzymeb) present in arthritic synovial fluid could represent a possible point of therapeutic intervention in such inflammatory disorders.
Despite the extensive knowledge of structure and enzymatic mechanism of the relatively abundant venom and mammalian pancreatic PLAzs (7-lo), relatively little is known about other mammalian PLAz enzymes. The amino-terminal portion of the amino acid sequences of several mammalian non-pancreatic PLAzs have been determined, including enzymes isolated from porcine intestine (ll), rabbit ascites (12), rat platelet (13), rat spleen (14), and rat peritoneal exudate (15). Recently, amino-terminal sequences for human synovial fluid PLAz (16)2s and another putative PLAz sequence of unknown function (19) have become the first human non-pancreatic PLA2 sequences available. This information has yielded much insight into relative form, function, and identity of these mammalian PLAz isolates. On the other hand, information about the other two-thirds of these molecules, which contains both the active site regions and many important structural and functional determinants has been nonexistent. Furthermore, the extrapolation of the biochemical properties and sequences of other mammalian enzymes to a form relevant to human disease has been, at best, ambiguous. An additional limitation is that without PLAz form-specific probes, it has been difficult to attribute biological effects to specific enzyme forms.
The isolation and characterization of PLAz from human arthritic synovial fluid has been detailed previously (20-21). Upon further purification of the enzyme on high performance liquid chromatography, we were able to resolve the enzyme into multiple peaks of activity, each with distinct biochemical properties.' The fraction eluting earliest, peak A, constituted the majority of activity and protein mass present in the extract. A second most abundant fraction (peak B), could readily be distinguished from the former by its enhanced activity in 0.5 M Tris and/or 0.2% sodium deoxycholate. In the present work we describe the isolation of both genomic and cDNA clones encoding the peak A PLAz sequence. Also, we have examined transcription of this gene in cells obtained from inflamed sources. Finally, we have demonstrated recombinant expression from the peak A cDNA of a secreted active PLAp with properties similar to the native enzyme.
RNA Blot Hybridization-Total cell RNAs were isolated using the method of Gubler and Hoffman (22) and were electrophoresed on a A 1 16  1.0% agarose, 18% formaldehyde gel, transfer blotted to a nitrocellulose filter, baked at 80 "C 2 h, and processed as above with lo7 cpm of 32P-labeled oligonucleotide probe. The filter was washed in 1 x SSC, 0.1% SDS at 60 "C for 60 min and autoradiographed for 3 days. cDNA Library Production-Human peritoneal exudate cDNA was synthesized using a kit (Boehringer Mannheim) and was cloned into X g t l O vector arms (Stratagene Inc.) via EcoRI linkers. The resulting clones were screened as described above.

ACCACGCCCATCCCCAGCCGTGCCTCACCTIICCCCCMCCTCCAGAGGGAGCAG~~GGGAG C A G C A G T G C A G M C A M C M G A C G G C C I ' G G G G A T A C M C T C T G A G C C
PLA2 Actiuity Assay-PLA2 activity was measured using a phosphatidylcholine/sodium deoxycholate mixed micele assay system as described.2

RESULTS AND DISCUSSION
Using the 25-residue amino-terminal amino acid sequence we obtained from material purified from human synovial fluid,' recently confirmed elsewhere (16), and by making allowances for conserved Cys and Gly residues, two partially overlapping 45-mer oligonucleotide probes were designed (Fig.  L4). Since the cellular source of synovial fluid PLAz was unknown, a genomic library was screened first to obtain the nucleotide sequence, and then non-degenerate probes made specific for the sequence would subsequently be used to find an mRNA source. A human genomic library was screened for coincident hybridization signals to both probes. Six such signals were detected, and the clones were purified through additional rounds of screening. Restriction digest analysis of the cloned DNAs revealed that they were identical. An AluI fragment containing the hybridizing DNA (Fig. 123, bases 202-287) was subcloned into M13 and the DNA sequence was determined. The DNA sequence of the AluI fragment was found to contain a region encoding the correct amino acid sequence. It also contained the remainder of the coding exon (bases 176-320), including a sequence resembling the Ca'++binding loop (bases 265-282), conserved in known PLA's (7,8). The presence of the remainder of the exon confirmed that the clone contained a PLAz gene.
In order to screen various sources for the presence of transcripts from this gene, a 60-base oligonucleotide matching the DNA sequence of the genomic clone was synthesized, RNAs from various sources, including phorbol ester-induced human cell lines U937 and HL60, inflamed human synovial tissue, and a cell pellet from a human peritoneal exudate (containing mononuclear and polymorphonuclear cells), along with RNA from several other mammalian tissues and cell lines were prepared and blotted onto nitrocellulose filters. As seen in Fig. 2, the presence of an -800-base transcript was detected only in the RNA from synovial tissue and the peritoneal exudate cells. In the exudate cell RNA the signal was observed in much greater relative abundance. From this, it would appear that the gene encoding this PLAz form is actively transcribed in cells associated with two different inflammatory disorders. Interestingly, no signal was detected in RNA from the two induced human promyelocytic cell lines, U937 and HL60, suggesting these cells could contain a different PLA'.
The relative abundance of signal in the RNA from peritoneal exudate cells suggested it would make the best cDNA source. A cDNA library was prepared and then screened using the 60-base oligonucleotide probe described in Fig 2. Sixteen signals were obtained from 2 x lo5 plaques, and 4 of the clones were purified further. The DNA sequence of the longest of these clones (number 4) is represented in Fig. lB, along I 2 3 4 5 6 7 8 9 1 0 I I 12 in the hydrolytic mechanism of other PLAz enzymes (8,9), including Phe', Ile', His4', Asp4', Aspw, Ala'", and Ala'03 are present in the clone 4 sequence. The Ca2++-binding loop (residues T~r~' -G l y~~) is entirely conserved except for His3', the only variable position within the loop. Overall, its spacing and Cys residues exactly match the sequence from Crotalus venom, except for the short a-helical region immediately following the active site (residues 52-56). In this region, the spacing of residues more closely resembles the pancreatic enzyme. This sequence contains, like other PLAzs with strong anticoagulant activity (

CCFVEDCCYG---KAT&-----tjPKTVS~~SEENGEIIC-
" " " " " " " " " " " " " " " " " " " " " " " " " " " " " " " 1.25% dimethyl sulfoxide ( l a n e 2), inflamed human synovial tissue ( l a n e 6), a cell pellet from a peritoneal exudate from a nonsterile peritonitis patient ( l a n e 7), porcine jejunum ( l a n e 9), pancreas ( l a n e 121, and spleen (lane I I ) , and rat liver ( l a n e IO). The  with the encoded sequence of the enzyme. The coding regions of the genomic clone were determined by matching the sequence with the cDNA (see Fig. 1C). The gene spans 4.6 kilobases and contains 5 exons. Introns 2, 3, and 4 fall in precisely the same positions within the sequence as they occur in the gene encoding pancreatic PLAz (23). The first intron occurred 107 bases upstream from the initiating Met, and interrupts the 5"noncoding region of the cDNA. In the gene sequence (Fig. lD), a TATA-like sequence (TATTTAA) was found 41 bases upstream from the start of the cDNA sequence. Another putative transcriptional control signal, CCAAT, was located 121 bases upstream. The precise localization of the transcriptional start site will require primer extension analysis; however, the presence of these putative transcriptional signals suggests that it is unlikely to be significantly upstream from the start of the cDNA clone.

E F V~---~T~-----@~D R~S~W
The cDNA clone 4 encoded a mature protein with a calculated molecular mass of 13,939 daltons, taking into account the amino terminus observed from the protein sequence. The amino acid sequence was aligned in Fig. 3 with other relevant published PLA2 sequences. Its sequence bears a striking resemblance to those of other known PLAZs. First, like the other non-pancreatic mammalian enzymes, the placement of its Cys residues follows a Type I1 pattern, consistent with its proinflammatory nature. Notably, the key residues implicated
Recombinant expression of the cloned gene fragments in the vaccinia virus system (17) represented an appropriate method to test whether this gene encoded an active secreted PLAZ. The cDNA clone 4 was trimmed of its 5'-and 3'noncoding DNA through digestion at Sac1 and Hind111 sites (bases 125-130 and 588-593, respectively), both of which were engineered into the sequence via oligonucleotide-directed mutagenesis. The resulting coding segment was blunted by fillin synthesis and cloned into the SmaI site of plasmid pSC11 (18). Purified plasmid DNA was co-transfected into CV-1 cells with wild type vaccinia viral DNA, and the resulting recombinants were screened by selection in medium containing 5'-bromodeoxyuridine. Four recombinant vaccinia clones were selected for analysis by quantitating secreted and cellassociated PLA2 activity (Fig. 4, inset). Clone 5B, showing the highest media PLA2 levels was chosen for further study. In a subsequent larger culture, media samples taken at various times after viral infection were assayed for PLAz activity. As seen in Fig. 4, PLAz activity from the PLAz/vaccinia recombinant virus was primarily extracellular and accumulated linearly in the medium over 72 h, after which viral cell lysis became significant. The expressed PLAz activity retained the property of sensitivity to Tris inhibition exhibited by the native enzyme.' In summary, a cloned cDNA encoding the major PLAz species present within rheumatoid synovial fluid has been obtained, and when expressed in the vaccinia expression system yields a secreted enzyme retaining properties of the native enzyme. While this form may not be the only PLAz present in all forms of arthritis, it clearly represents the most prevalent and active form found in the types examined. Its abundance in peritoneal exudate cells suggests this enzyme occurs systemically and could play a role in many types of acute inflammatory disorders. Further work will be necessary to determine the cell type(s) from which this enzyme originates, and to elucidate its precise role in the perpetuation of inflammatory disorders.