A natural disruption of the secretory group II phospholipase A2 gene in inbred mouse strains.

The synovial fluid or group II secretory phospholipase A2 (sPLA2) has been implicated as an important agent involved in a number of inflammatory processes. In an attempt to determine the role of sPLA2 in inflammation, we set out to generate sPLA2-deficient mice. During this investigation, we observed that in a number of inbred mouse strains, the sPLA2 gene was already disrupted by a frameshift mutation in exon 3. This mutation, a T insertion at position 166 from the ATG of the cDNA, terminates out of frame in exon 4, resulting in the disruption of the calcium binding domain in exon 3 and loss of both activity domains coded by exons 4 and 5. The mouse strains C57BL/6, 129/Sv, and B10.RIII were found to be homozygous for the defective sPLA2 gene, whereas outbred CD-1:SW mice had variable genotype at this locus. BALB/c, C3H/HE, DBA/1, DBA/2, NZB/B1N, and MRL lpr/lpr mice had a normal sPLA2 genotype. The sPLA2 mRNA was expressed at very high levels in the BALB/c mouse small intestine, whereas in the small intestine of the sPLA2 mutant mouse strains, sPLA2 mRNA was undetectable. In addition, PLA2 activity in acid extracts of the small intestine were approximately 40 times higher in BALB/c than in the mutant mice. Transcription of the mutant sPLA2 gene resulted in multiple transcripts due to exon skipping. None of the resulting mutant mRNAs encoded an active product. The identification of this mutation should not only help define the physiological role of sPLA2 but also has important implications in mouse inflammatory models developed by targeted mutagenesis.

The synovial fluid or group II secretory phospholipase A 2 (sPLA 2 ) has been implicated as an important agent involved in a number of inflammatory processes. In an attempt to determine the role of sPLA 2 in inflammation, we set out to generate sPLA 2 -deficient mice. During this investigation, we observed that in a number of inbred mouse strains, the sPLA 2 gene was already disrupted by a frameshift mutation in exon 3. This mutation, a T insertion at position 166 from the ATG of the cDNA, terminates out of frame in exon 4, resulting in the disruption of the calcium binding domain in exon 3 and loss of both activity domains coded by exons 4 and 5. The mouse strains C57BL/6, 129/Sv, and B10.RIII were found to be homozygous for the defective sPLA 2 gene, whereas outbred CD-1:SW mice had variable genotype at this locus. BALB/c, C3H/HE, DBA/1, DBA/2, NZB/B1N, and MRL lpr/lpr mice had a normal sPLA 2 genotype. The sPLA 2 mRNA was expressed at very high levels in the BALB/c mouse small intestine, whereas in the small intestine of the sPLA 2 mutant mouse strains, sPLA 2 mRNA was undetectable. In addition, PLA 2 activity in acid extracts of the small intestine were approximately 40 times higher in BALB/c than in the mutant mice. Transcription of the mutant sPLA 2 gene resulted in multiple transcripts due to exon skipping. None of the resulting mutant mRNAs encoded an active product. The identification of this mutation should not only help define the physiological role of sPLA 2 but also has important implications in mouse inflammatory models developed by targeted mutagenesis.
Secretory nonpancreatic phospholipase A 2 (sPLA 2 ) 1 has been implicated as one of the important pathogenetic agents in both local and systemic inflammatory conditions (for reviews, see Refs. [1][2][3]. Levels of circulating sPLA 2 in both adult (4) and juvenile (5) rheumatoid arthritis as well as in septic shock (6) and multiorgan failure (7) correlate positively with disease severity and poor prognosis. sPLA 2 gene expression has also been shown to be induced by various inflammatory stimuli, like endotoxin, IL-1␤, TNF-␣, and IL-6 (1, 2). Nevertheless, the exact mode of action of sPLA 2 and the role it plays in these diseases is not very well understood. sPLA 2 catalyzes the release of fatty acid from the sn-2 position of phospholipids producing free fatty acid and lysophospholipid. When the fatty acid is arachidonic acid, which is usually esterified at this position in phospholipids, it can lead to the production of the various leukotriene and prostaglandin proinflammatory mediators. The other product of this reaction, lysophospholipid, can also be metabolized into a potent inflammatory mediator (platelet-activating factor). However, since sPLA 2 has very little preference for the type of fatty acid in the sn-2 position, it is not known to what extent the proinflammatory activity of sPLA 2 is mediated through the release of arachidonic acid and its subsequent metabolism into proinflammatory mediators. One approach to determine the role of sPLA 2 in these various disease states would be to develop potent specific inhibitors of this enzyme. Recently, there have been a number of sPLA 2 inhibitors reported (8 -13). However, none of these inhibitors have been able to convincingly determine the role sPLA 2 plays in inflammation. In an attempt to circumvent this problem and investigate the role of sPLA 2 in inflammation, we initiated studies to generate mice with a targeted disruption of the sPLA 2 gene. Provided, the mutation was not lethal, sPLA 2deficient mice could be used in inflammatory models to determine the contribution of sPLA 2 in the inflammatory process.
During this investigation, we observed that the sPLA 2 gene, cloned from a 129/Sv-derived genomic library, already contained a natural mutation. This mutation causes a disruption in the reading frame of exon 3 resulting in an inactive gene product. This paper describes the nature of this mutation, its consequence, and its prevalence in a number of inbred mouse strains.

MATERIALS AND METHODS
Animals-Male mice, 20 -25 g, were obtained from the following companies; CD-1:Swiss Webster, which is an outbred strain and BALB/c were from Charles River Laboratories (St. Constant, Quebec); 129/Sv were from Taconic (Germantown, NY); C57BL/6 were obtained from three suppliers, Jackson Laboratories (Bar Harbor, ME), Charles River Laboratories, and Taconic; and B10.RIII and MRL lpr/lpr were from Jackson Laboratories.
Cloning of the Mouse sPLA 2 Gene-The mouse sPLA 2 gene was cloned from a CC1.2 Sau3A genomic library in DASHII (Stratagene) using standard molecular biology methods (14). The probe used for * This work was supported in part by a Medical Research Council/ Pharmaceutical Manufacturer Association of Canada health program grant. 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.
screening was a random primed (Pharmacia Biotech) 32 P-labeled rat sPLA 2 cDNA, a generous gift of Dr. H. Van den Bosch (15). When the library was initially screened with the full-length sPLA 2 cDNA, approximately 1 ϫ 10 6 phage were plated, and 30 positive clones were picked. These clones were all identical and were missing the 3Ј-end of the gene. The library was replated and screened using a 160-bp PvuII/HindIII fragment containing the 3Ј coding region of the rat sPLA 2 cDNA. This screening produced a single positive clone. The clones were mapped, and the appropriate DNA fragments containing the sPLA 2 gene were cloned and sequenced on both strands using an ABI 373A automated DNA sequencer.
Cloning of the Mouse sPLA 2 cDNA and COS7 Transfection-Primers derived from the 5Ј (5Ј-AGCTGACAGCATGAAGGTCCTCC-3Ј) and 3Ј (5Ј-TTCTGGGTGAAGACAGAAGGGCC-3Ј) ends of the coding region of the mouse sPLA 2 gene were used in RT-PCR reactions to clone the mouse sPLA 2 cDNA. These primers will also amplify the rat sPLA 2 cDNA. The RT-PCR reactions were carried out on 1 g of total RNA using conditions described by the manufacturer (Perkin Elmer). The amplification conditions were 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s for 35 cycles. The PCR products were electrophoresed on a 1% low melt agarose gel, recovered, and cloned into the pCRII cloning vector (Invitrogen). The cDNA clones were sequenced on both strands using an ABI 373A automated DNA sequencer. The sPLA 2 cDNA sequences were then cloned into the eukaryotic expression vector pSG5 (Stratagene). The various sPLA 2 cDNAs were then transfected into COS7 cells as described previously (16). The sPLA 2 activity released into the media 48 h after transfection was then determined.
Genomic Analysis-The DNA used in the genomic analyses were either prepared (14) or obtained from commercial sources (see the legends to Fig. 3 and Table I). DNA was digested with BamHI and analyzed by Southern blotting using the genomic 2.5-kb HindIII/NotI (the NotI site is in the Dash II multiple cloning sequence) probe from 3 (see Fig. 1). PCR analysis of exon 3 was carried out using the following primers: 5Ј-primer (5Ј-CTGGCTTTCCTTCCTGTCAGCCTG-GCC-3Ј); 3Ј-primer (5Ј-GGAAACCACTGGGACACTGAGGTAGTG-3Ј). The PCR reaction was performed using 1 g of genomic DNA, 50 pmol of each primer, and the following cycling conditions: 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s for 35 cycles. The amplified exon 3 DNA fragment was cloned and sequenced as described above. For genotype analysis, the PCR product was digested with BamHI and analyzed on 1% low melt agarose gels.
PLA 2 Activity-PLA 2 activity was measured as described previously (17) using radiolabeled E. coli membrane phospholipid substrate. One unit of PLA 2 activity is defined as the hydrolysis of 56 pmol of phospholipid substrate (representing 1% of total E. coli phospholipid) in 30 min at 37°C. sPLA 2 was prepared from the mouse small intestine using acid extraction as described by Kramer et al. (18). Briefly, mouse small intestine (0.23-0.38 g wet weight) was homogenized in 2 ml of homogenization buffer (30 mM Tris, pH 7.5, 120 mM NaCl, and 2 mM EGTA) for 3 ϫ 20 s on ice. An equal volume of cold 0.36 N H 2 SO 4 was added, mixed by vortexing, and placed on ice for 1 h. The acid extracts were centrifuged for 30 min at 10,000 ϫ g, the supernatant was recovered, and the protein concentration was determined by Bradford assay (Pierce). A small aliquot of the supernatant (200 l) was neutralized with 1 M Tris base (35 l), and the PLA 2 activity was determined. The PLA 2 activity in the media of transfected COS7 cells was also determined.
Lipopolysaccharide (LPS) Treatment of Mice and RNA Isolation-LPS (0111:B4, Sigma) in phosphate-buffered saline (PBS) was used for all injections. Mice were injected intraperitoneally with either 5 or 10 g of LPS, and 5 h later tissues were removed. LPS up to 500 g was also injected intravenously into the tail vein, and tissues were collected 18 h after injection. In an attempt to enhance the effect of LPS on sPLA 2 gene expression, mice were first primed with 5 mg of carrageenan (FMC Corp., Rockport ME) (in 0.5 ml of PBS) injected intraperitoneally, and 24 h later injected intravenously with varying concentrations of LPS up to 10 g as described by Ogata et al. (19). Tissues were removed from these animals 5 h after the LPS injection. Total RNA was prepared from tissue samples using Trizol (Life Technologies, Inc.) reagent as described by the manufacturer. Northern blot analysis of total RNA was performed using 1% agarose formaldehyde gels to separate the RNA, which was then transferred to nitrocellulose. Blots were probed with the mouse sPLA 2 cDNA, stripped, and reprobed with a human glyceraldehyde 3-phosphate dehydrogenase cDNA (Clontech).

RESULTS
Mouse sPLA 2 Gene-As a first step in generating mice with a targeted disruption in the sPLA 2 gene, a CC1.2 embryonic stem cell genomic library was screened to obtain an isogenic genomic clone of the mouse sPLA 2 gene (20). The CC1.2 line was originally derived from the mouse strain 129/Sv (21). The preliminary screening of the library using a rat sPLA 2 cDNA probe resulted in 30 positive clones, all of them identical to 3 (Fig. 1). The sPLA 2 gene was located at one end of this genomic clone, subcloned, sequenced, and was found to be missing exon 5. In order to obtain the complete gene, the library was rescreened with a 3Ј rat sPLA 2 cDNA probe (see "Materials and Methods"). One positive clone, 6 was obtained from this screening, and it contained the complete gene (Fig. 1). The sequence of the mouse sPLA 2 gene is shown in Fig. 2. It is not a large gene, the complete coding sequence including introns and exons spans about 3.5 kb. The organization of the mouse gene is identical to that of the human and rat genes (18,(22)(23)(24). It consists of five exons, each representing a specific domain of the enzyme; exon 1 is a 5Ј-untranslated sequence, exon 2 codes for the signal peptide sequence, exon 3 codes for the calcium binding domain, and exons 4 and 5 code for catalytic activity domains. The mouse and rat sPLA 2 gene sequences including introns are nearly 80% identical.
However, between the two mouse sPLA 2 genomic clones there appeared to be a BamHI polymorphism. When 3 was digested with BamHI, two DNA fragments hybridized with the sPLA 2 cDNA probe, a 5.7-kb 5Ј-fragment and a 600-bp 3Јfragment representing the end of this clone. Based on the sequence from the sPLA 2 gene in 3, this BamHI site occurs in exon 3. This appears to be a conserved sequence since this BamHI site is also present in exon 3 of the rat and human genes. A similar BamHI digest of 6 or of embryonic stem cell genomic DNA resulted in a single 8.2-kb fragment. An explanation for this discrepancy is that 6 originated from the stem cell, whereas 3 was from contaminating STO feeder cell genomic DNA. The STO feeder cells are mouse fibroblasts derived from the inbred Swiss strain SIM (25). Subsequent sequencing of the sPLA 2 gene in 6 revealed that the BamHI site in exon 3 was lost because of an insertion of an extra T residue (GGATCC 3 GGATTCC). This insertion not only The sequence shown is from 6, with the mutation in exon 3 removed, in order to show the correct translation of the gene. Highlighted within the sequence are the putative TATA box, which is overlined; the five exons and the deduced amino acid sequence, which are underlined; the BamHI site in exon 3, which contains the T insertion mutation is in boldface and has an asterisk below it; and a putative polyadenylation site, which is in boldface. The TATA box, exon 1, and the polyadenylation site are based on the structure of the rat sPLA 2 gene (23, 24). would disrupt the BamHI recognition sequence, but it would also cause a frameshift in the expressed mRNA. The resulting sPLA 2 enzyme would have a disrupted calcium binding domain and terminate out of frame in exon 4, losing both the exon 4 and 5 sequences that make up the active site (see Fig. 5). This sPLA 2 enzyme would be totally inactive since both the calcium and catalytic activity domains are absolutely required for activity. This result would indicate that the full-length sPLA 2 gene in 6 isolated from the ES stem cell genomic library codes for an inactive gene product.
There are at least four possible explanations for this result. First, the mutation was a sequencing error (this is unlikely since this region was sequenced with a number of different primers on both strands, and in addition this sequence is not cleaved by BamHI); second, the mutation was generated during cloning; third, it could be an artifact of the ES cell line; and fourth, the mutation originated from the 129/Sv mouse strain from which the ES cells were originally derived. If this last possibility was correct, then there already existed a mouse strain that had a natural disruption of the sPLA 2 gene.
Disruption of the sPLA 2 Gene in Inbred Mouse Strains-In order to determine if this sPLA 2 gene mutation originated from the 129/Sv mouse strain or was due to a cloning or cell line artifact, genomic DNA was prepared from these mice, digested with BamHI, and analyzed by Southern blotting (Fig. 3A). The presence of the 8.2-kb BamHI fragment would be indicative of this mutation, whereas hybridization to 5.7-and 2.5-kb fragments would suggest a wild-type sPLA 2 gene structure. In addition to the 129/Sv mouse strain, genomic DNA from, C57BL/6, BALB/c, and CD-1:Swiss Webster mice was also analyzed (Fig. 3). The Southern blotting indicates that both the 129/Sv and the C57BL/6 mouse strains are homozygous for this sPLA 2 mutation. The BALB/c mouse strain has a wild-type genotype and the CD-1 mouse chosen for this experiment was heterozygous for the mutation. The heterozygosity of the sPLA 2 gene in CD-1 mice is due to the fact that this is an outbred mouse strain.
To confirm the Southern blotting result and prove that the loss of the BamHI site in exon 3 was due to the T insertion, exon 3 was amplified from each of these mice strains, cloned, and sequenced. An example of this analysis is shown in Fig. 3B, in which exon 3 clones from these mouse strains were digested with BamHI. The C57BL/6 exon 3 clones were not cleaved by BamHI, whereas all of the BALB/c exon 3 clones contained the BamHI site and gave the appropriate cleavage products. The CD-1 exon 3 clones, as expected, consisted of both types. Analysis of the exon 3 DNA sequences revealed that 129/Sv, C57BL/6 and the mutant allele in the CD-1 all had the T insertion mutation.
We have also analyzed the sPLA 2 gene structure for a number of additional mouse strains, and the results are presented in Table I. Most of the mouse strains tested had a homozygous wild-type sPLA 2 BamHI digestion pattern. The heterozygosity of the B6/D2F1 mice would indicate that the DBA/2 mouse strain has the wild-type sPLA 2 gene, since C57BL/6 is homozygous for the sPLA 2 mutation. The congenic strain B10.RIII has the mutated sPLA 2 BamHI digestion pattern, which would mean that the C57BL/10 mouse strain also has the mutated sPLA 2 gene. This is very likely since C57BL/6 and C57BL/10 are substrains of C57BL. Based on the known genealogy (26) of inbred mouse strains, it is possible that 129/Sv and C57BL could have originated from a common ancestor that passed on the sPLA 2 mutation. The origin of the other mouse strains listed in Table I are much further removed from these two strains.
Mutant and Wild-type sPLA 2 Gene Expression-The expression of sPLA 2 mRNA and enzyme activity was next analyzed to confirm that the sPLA 2 gene mutation results in an inactive enzyme. Mulherkar et al. (27,28) have cloned the BALB/c mouse sPLA 2 cDNA from the small intestine where it is present at very high levels and localized in the Paneth cell (27)(28)(29)(30)(31). sPLA 2 mRNA expression and enzyme activity were determined in the small intestine of control and LPS-injected, BALB/c mice (sPLA 2 ϩ/ϩ or wild-type) and C57BL/6 mice (sPLA 2 Ϫ/Ϫ or mutant). Northern blot analysis of total RNA isolated from these animals and hybridized with a mouse sPLA 2 cDNA probe is shown in Fig. 4A. High levels of sPLA 2 mRNA can be detected in both control and LPS-treated BALB/c mice. In control and LPS-treated C57BL/6 mice, sPLA 2 mRNA was undetectable.  (Fig. 1). The two C57BL/6 lanes represent DNA isolated from two separate animals. The BALB/c genomic DNA was from Clontech. All other genomic DNAs were prepared. B, analysis of PCR clones of exon 3 from the mice indicated. Exon 3 of the sPLA 2 gene was amplified from mouse genomic DNA as described under "Materials and Methods" and cloned into the pCR II vector. Four clones from each amplification were digested with EcoRI, which releases the inserted DNA and with BamHI. Lane 6 from the amplification of CD-1 genomic DNA contains two exon 3 fragments, one that is digested by BamHI and the other that is not. Total RNA isolated from the small intestine of the 129/Sv mice also had undetectable levels of sPLA 2 mRNA (Fig. 4A). The sPLA 2 mRNA was also not detected by RT-PCR of total RNA from the small intestine of either C57BL/6 or 129/Sv mouse strains (not shown). The high level of sPLA 2 mRNA expression in the mouse intestine would suggest that this tissue should also be a good source for the sPLA 2 enzyme. In fact the enzyme responsible for the majority of the acid-stable PLA 2 activity in the BALB/c small intestine was recently purified and shown to be sPLA 2 (32). Therefore, the acid-extractable PLA 2 activity in the small intestine of BALB/c and C57BL/6 mice was determined. The PLA 2 activity in the BALB/c control intestine (126, 594 units/ ml) was found to be at least 40 times higher than in the C57BL/6 control intestine (3,031 units/ml). The small amount of PLA 2 activity remaining in the C57BL/6 intestine is most likely due to one or a combination of the other low molecular weight secretory phospholipases A 2 , which are also acid-stable. Treatment of the mice with LPS did not increase the level of PLA 2 activity in either the BALB/c or C57BL/6 intestine, which was consistent with the lack of induction by LPS on sPLA 2 mRNA levels in this tissue. The reason for the lack of induction of sPLA 2 expression by LPS in the intestine is not clear, since LPS or endotoxin has been shown to greatly induce sPLA 2 expression in the rat (33). One likely explanation is that the level of sPLA 2 expression is already so high in the BALB/c small intestine that it would be difficult to see any amount of induction in this tissue. In other tissues like the aorta and in the small intestine of mice with the sPLA 2 mutation, sPLA 2 mRNA can be induced by LPS (see below).
In order to prove that the BALB/c sPLA 2 mRNA codes for an active sPLA 2 enzyme and the mutant sPLA 2 gene codes for an inactive form, both mutant and BALB/c sPLA 2 cDNAs were cloned by RT-PCR. We had found that it was possible to induce expression of the sPLA 2 mRNA in mice carrying the mutated sPLA 2 gene if extreme conditions are used for the induction. When mice are first pretreated with carrageenan and then injected 24 h later with LPS there is a tremendous enhancement of LPS-induced TNF production and mouse mortality by endotoxin shock (19). Under these conditions, sPLA 2 mRNA can be detected in the small intestine by Northern blot analysis and in the aorta by RT-PCR (not shown) of mice with the mutated sPLA 2 gene. It should be noted that BALB/c mice do not survive this treatment; they die shortly after the carrageenan injection. We have used small intestine total RNA prepared from a CD-1 mouse that was homozygous for the sPLA 2 mutation (Ϫ/ϪCD-1) (verified by BamHI genomic Southern blotting) and treated in the above fashion and from BALB/c for RT-PCR to amplify and clone both the mutant and wild-type mouse sPLA 2 cDNAs. The PCR products from these amplifications are shown in Fig. 4B. Both the BALB/c and rat osteoblast (control) RNA produced the correct size 595-bp PLA 2 cDNA product (Fig. 4B), which could also be digested by BamHI (not shown). The sPLA 2 cDNA products from amplification of total RNA isolated from the Ϫ/ϪCD-1 mouse (or from any of the mutant sPLA 2 mice, 129/Sv, C57BL/6) gave multiple sPLA 2 cDNA products (Fig. 4B, lane 2) that could not be digested by BamHI (not shown). Both the BALB/c sPLA 2 cDNA and the mutant sPLA 2 cDNA RT-PCR products were cloned and sequenced and are shown in Fig. 5. The two more strongly amplified products from the Ϫ/Ϫ mouse were the full-length sPLA 2 cDNA with the T insertion (ImsPLA 2 ) and a sPLA 2 cDNA with exon 3 removed (X3sPLA 2 ). A third sPLA 2 cDNA with exon 4 spliced out was also found (X4sPLA 2 ). These sPLA 2 cDNAs would most likely code for inactive enzymes since they all have disrupted calcium binding and catalytic activity domains. Both the mutant and wild-type sPLA 2 cDNAs were cloned into the eukaryotic expression vector pSG5 and transiently transfected into COS7 cells. 48 h after transfection, PLA 2 activity in the media of the transfected cells was determined. Cells transfected with any of the mutant sPLA 2 cDNAs did not release any detectable PLA 2 activity into the media, confirming that mice strains carrying the mutant sPLA 2 gene do not express an active sPLA 2 enzyme. In contrast, the media of cells transfected with the BALB/c sPLA 2 cDNA contained greater than 3,000 units/ml of PLA 2 activity, confirming that the BALB/c sPLA 2 gene codes for an active enzyme. DISCUSSION In this study we have shown that the murine sPLA 2 gene is naturally disrupted in a number of inbred strains. This discovery came about following initial experiments to disrupt sPLA 2 by targeted mutagenesis in order to derive a sPLA 2 mutant model. It has been recognized that to increase the efficiency of gene targeting in ES cells, the targeting vector must be derived from DNA that is isogenic to the cell line (20). While screening a genomic DNA bacteriophage library prepared from the ES cell CC1.2, a BamHI restriction polymorphism was discovered between an isogenic sPLA 2 clone and a clone that originated from the contaminating STO feeder cells. This polymorphism was due to the insertion of a thymidine residue in exon 3 of the ES cell gene. The resulting transcribed mRNA would contain a The BALB/c and C57BL/6 mice were injected intravenously as follows, and 18 h later, the tissue was removed. Lanes 4 and 7, controls injected with PBS; lanes 5 and 8, injected with 300 g of LPS; and lanes 6 and 9, injected with 500 g of LPS. Total RNA from rat calvarial osteoblasts that had been treated with IL-1␤ and TNF-␣ (induces sPLA 2 mRNA) (48) was run as a control for hybridization. The blot was first hybridized with a mouse sPLA 2 cDNA probe and then reprobed with a human glyceraldehyde 3-phosphate dehydrogenase probe. B, RT-PCR amplification of wild-type and mutant sPLA 2 cDNAs. Total RNAs isolated from the small intestine were used in RT-PCR reactions to amplify the cPLA 2 cDNA. Lane 1, BALB/c; lane 2, CD-1 sPLA 2 -deficient mouse that was treated with carrageenan/LPS as described under "Materials and Methods"; and lane 3, rat osteoblast. The RT-PCR products were analyzed on a 2% low melt agarose gel. frameshift mutation and code for an inactive enzyme, since both the calcium binding and catalytic activity domains encoded by exons 3, 4, and 5 were absent. The CC1.2 ES cell line, having been derived from the mouse strain 129/Sv, suggested that a sPLA 2 mutant already existed in mouse populations, while STO cells, having been derived from an inbred Swiss line (25), indicated that the gene was not mutant in all strains. A survey of inbred mouse stains demonstrated that the sPLA 2 gene is disrupted in 129/Sv and C57BL/6 but not in BALB/c. It is likely that this mutation would have been overlooked had the polymorphism between CC1.2 and STO not been observed. Therefore, it may be advisable that before attempting to generate targeted mutations in mice, one should first check the expression of the particular gene in a panel of inbred strains. Based on the above observations, there exists the possibility that "knockout" mice with no overt phenotype, may have had the gene of interest already disrupted.
The loss of the sPLA 2 gene appears not to have a detrimental effect on the viability of these mice strains. C57BL/6 is one of the mostly widely used mouse strains and accounts for more than 14% of studies in which an inbred strain is used (34). Thus, the sPLA 2 gene does not appear to play a critical role in the development of the animal, nor does it appear to have an effect on fertilization as has been suggested (35,36). In fact, there is no obvious phenotype for the sPLA 2 mutation. However, with the identification of sPLA 2 -deficient mouse strains, the physiological role of sPLA 2 and its involvement in inflammatory conditions can now be addressed.
The high levels of sPLA 2 present in inflammatory synovial fluids of rheumatoid arthritis patients was suggestive of a proinflammatory role for sPLA 2 in this disease (1)(2)(3). The levels of circulating sPLA 2 also correlated very well with disease severity in both adult and juvenile rheumatoid arthritis (4,5). However, the function of sPLA 2 in rheumatoid arthritis is unknown. It has been suggested that its involvement in rheumatoid arthritis may be to potentiate an already established inflammatory process (1). The mouse has been used extensively as a model for arthritis. Knowing the sPLA 2 genotype of several of the inbred mouse strains, we attempted to see if there was any correlation between mouse strain, the functionality of the sPLA 2 gene, and the arthritic model. In antigen-induced arthritic models, both sPLA 2 -deficient and sPLA 2 wild-type mouse strains are susceptible to arthritis. The sPLA 2 -deficient strains C57BL/6 and B10.RIII have been used in methylated FIG. 5. DNA sequence of wild-type and mutant sPLA 2 cDNAs. The three sPLA 2 cDNAs isolated from the CD-1 sPLA 2 -deficient mouse are identified as follows: ImsPLA 2 , contains the complete sPLA 2 coding region, but it has the T insertion (asterisk) at position 166 from the ATG; X3sPLA 2 has exon 3 deleted; X4sPLA 2 has exon 4 deleted. The BALB/c sPLA 2 cDNA, which codes for an active enzyme, is identified as MsPLA 2 . The ATG initiation codon and the termination codons for each sequence are shown in boldface. The position of the exon sequences within the cDNA sequence are underlined. Two polymorphisms at positions 101 and 231 between the BALB/c and CD-1 sPLA 2 cDNA sequences were found and are underlined.
bovine serum albumin (37) and collagen (38) induced models of arthritis, respectively, whereas, BALB/c mice appear to be susceptible to proteoglycan-induced arthritis (39). Conversely, mouse strains such as DBA/1 and MRL lpr/lpr, which have a wild-type sPLA 2 gene structure, are susceptible to a spontaneous occurring arthritis (40,41). From such results, it would seem that sPLA 2 is not required for the initiation of arthritis, but it is possible that it may have a more subtle role perhaps in disease severity or prolongation, which is more difficult to decipher from these studies. Attempting to draw conclusions about the role of sPLA 2 in arthritis from previous studies is complicated by a number of factors, such as the genetic variability of each inbred strain and the complex nature of these arthritic models. A similar problem arises when trying to apply this type of analysis to other sPLA 2 proinflammatory activities such as its involvement in eicosanoid production and in septic shock. The best approach to address these issues and obtain a better understanding of the role of sPLA 2 in these diseases would be to generate sPLA 2 wild-type and sPLA 2 -deficient mice with the same genetic background. Such mouse strains could then be used to reevaluate the role of sPLA 2 in these disease models.
Given the potential role of sPLA 2 in inflammation, our results suggest that caution must be exercised in generating genetic models for inflammation based on the application of targeted mutagenesis in the mouse. The targeted disruption of a gene is always carried out in ES cells derived from the 129/Sv mouse strain, while the mutation is bred into either the 129/Sv or C57BL/6 background, both of which bear the natural sPLA 2 gene disruption. Such mice will have mutations in two proinflammatory genes, which would greatly complicate the interpretation of results. Therefore, it is perhaps advisable that gene targeting experiments of proinflammatory genes be conducted in a genetic background where the sPLA 2 gene is intact.
A number of other functions have also been attributed to sPLA 2 . Mulherkar et al. (27)(28)(29)(30)(31) have localized the murine sPLA 2 to the Paneth cell of the BALB/c small intestine and have proposed that sPLA 2 acts as an enhancing factor for epidermal growth factor binding to cells, thus playing a role in cellular proliferation within the small intestine (27)(28)(29)(30)(31). Another group has recently purified sPLA 2 from the BALB/c small intestine and has shown it to have significant bactericidal properties (32). They suggest that the anti-microbial activity of sPLA 2 protects the small intestine from microbial infection. Consequently, sPLA 2 -deficient mice might be expected to be more susceptible to small intestine infections and have problems in regulating the high cell turnover that occurs in this tissue. However, sPLA 2 -deficient mouse strains appear not to have any gross observable differences in the physiology of the small intestine (34). This would suggest that either the mice survive without these sPLA 2 activities or that perhaps the minimal PLA 2 activity still present in the intestine of the mutant mice is sufficient to maintain homeostasis. There are at least three other low molecular weight sPLA 2 s identified in humans and rodents. These are the pancreatic PLA 2 and two recently cloned PLA 2 s, one of which appears to be localized in the heart (42) and the other in the testes (43). Whether or not these other secretory PLA 2 s compensate for the loss of sPLA 2 activity in these mice remains to be determined.
Nonsense or frameshift mutations have been shown to cause reduced steady-state levels of mRNA (for review, see Refs. 44 and 45). The insertion of a single T residue into exon 3 of the mouse sPLA 2 gene demonstrates this quite effectively. In mice carrying this mutation, no sPLA 2 mRNA was detectable by Northern blot analysis in the small intestine, whereas in wildtype animals, it is present at very high levels. However, the mutant sPLA 2 mRNA can be induced if more extreme conditions are used for the induction. In this case, multiple transcripts of the mutated sPLA 2 gene, due to exon skipping, are observed. These transcripts code for inactive sPLA 2 enzymes. Normally exon skipping has been observed in genes with mutations involving the splice donor or acceptor sequences. The mutation in the sPLA 2 gene occurs Ϫ24 bp from the 5Ј splice site. Recently, there have been a few reports of single base exonic mutations that result in aberrant mRNA splicing (46,47). One such mutation was identified in the mitochondrial acetoacetyl-CoA thiolase gene (T2) from a girl with T2 deficiency (46). This mutation occurs in exon 8, Ϫ13 bp from the 5Ј splice site of the intron and causes partial skipping of this exon. It was suggested that this mutation could result in an altered secondary structure of the T2 pre-mRNA around exon 8 and in this way impede normal splicing. Thus it is possible that the secondary structure of the mouse sPLA 2 pre-mRNA is altered by the single base insertion mutation in exon 3.
While this manuscript was in review, MacPhee et al. (49) published a paper describing this same sPLA 2 mutation in inbred mouse strains. They present evidence to show that the sPLA 2 gene maps to a locus, Mom1, that modifies Min-induced tumor number in the murine small and large intestine. They suggest that sPLA 2 modifies polyp number by altering the cellular microenvironment within the intestinal crypt. How sPLA 2 carries out this function remains to be determined.
In conclusion, we have shown that the sPLA 2 gene is disrupted in a number of common inbred mouse strains. This could impact on gene targeting experiments since the mouse strains 129/Sv and C57BL/6 bear this mutation. Its role in arthritis is still not clear, but it appears not to be required for the initiation of the disease. Experiments are in progress to generate sPLA 2 wild-type and sPLA 2 -deficient mice with the same genetic background and reevaluate the role of sPLA 2 in mouse inflammatory models.