(Received for publication, April 21, 1997, and in revised form, May 30, 1997)
From ICOS Corporation, Bothell, Washington 98021
Lysophosphatidic acid (1-acyl-sn-glycero-3-phosphate (LPA)) is a phospholipid with diverse biological activities. The mediator serves as an intermediate in membrane phospholipid metabolism but is also produced in acute settings by activated platelets. LPA is converted to phosphatidic acid, itself a lipid mediator, by an LPA acyltransferase (LPAAT). A human expressed sequence tag was identified by homology with a coconut LPAAT and used to isolate a full-length human cDNA from a heart muscle library. The predicted amino acid sequence bears 33% identity with a Caenorhabditis elegans LPAAT homologue and 23-28% identity with plant and prokaryotic LPAATs. Recombinant protein produced in COS 7 cells exhibited LPAAT activity with a preference for LPA as the acceptor phosphoglycerol and arachidonyl coenzyme A as the acyl donor. Northern blotting demonstrated that the mRNA is expressed in most human tissues including a panel of brain subregions; expression is highest in liver and pancreas and lowest in placenta. The human LPAAT gene is contained on six exons that map to chromosome 9, region q34.3.
Lysophosphatidic acid (1-acyl-sn-glycero-3-phosphate (LPA))1is a potent bioactive lipid with diverse biological activities that range from the physiologic to the pathophysiologic. Natural physiologic functions include mitogenesis, cell differentiation, platelet aggregation, actin cytoskeleton remodeling, monocyte chemotaxis, smooth muscle contraction, and neurite retraction (1). In vitro experiments suggest that LPA can also impact immune cell functions such as proliferation and IL-2 production (2). The phospholipid may also participate in the pathophysiology of neurodegenerative processes by causing vasoconstriction as well as impairment of glutamate and glucose uptake by astrocytes (3, 4). In addition, LPA is a potent promoter of tumor cell growth and invasion (5, 6). LPA exerts its biological effects via at least one, but perhaps multiple, specific G protein-coupled receptors (7-9). Engagement of the receptor results in activation of Ras and the Raf/mitogen-activated protein kinase pathway, stimulation of phospholipases C and D, inhibition of adenylyl cyclase, and tyrosine phosphorylation of focal adhesion proteins along with actin cytoskeleton remodeling (1).
Given the breadth of its biological impact, LPA metabolism has been a subject of intense study. During membrane phospholipid biosynthesis, LPA is formed by acylation of sn-glycerol-3-phosphate or by acylation of dihydroxyacetone phosphate followed by reduction of the acyl-dihydroxyacetone phosphate (10). In contrast, LPA that is rapidly generated in the plasma membrane of thrombin-activated platelets and growth factor-stimulated fibroblasts (11, 12) appears to arise from hydrolysis of phosphatidic acid (PA) by a phospholipase A2 (13-15). Additionally, Fourcade et al. (16) have demonstrated that a secretory phospholipase A2 acts upon membrane microvesicles shed from activated cells to convert PA to LPA. PA is a key intermediate in membrane phospholipid biosynthesis (10), but it can also serve as a second messenger in activated cells (17). PA can be converted to CDP-diacylglycerol or to diacylglycerol by the action of PA phosphatase or to LPA by the phospholipase A2.
LPA is present in serum at physiologically active concentrations. Because activated platelets copiously secrete the mediator, it has been suggested that aggregated platelets are the primary source of the serum LPA (14, 18). This source, coupled with the mitogenic and chemotactic properties of LPA, has prompted the hypothesis that the phospholipid is an important mediator of wound healing (12). Additionally, several of the known effects of LPA are consistent with a potential proinflammatory or proimmune function. The fact that the phospholipid is present in serum at functional concentrations implies the necessary presence of an "anti-LPA" mechanism to preclude inappropriate activation of LPA-sensitive cells. Consistent with this, there appear to be at least three mechanisms whereby LPA bioactivity might be attenuated. (i) LPA can be converted to PA in cells by the action of LPA acyltransferase (LPAAT) (7, 10). This enzyme is found in microsomes and the plasma membrane (29) and may be regulated by phosphorylation in response to interleukin-1 (IL-1) (29, 34). (ii) An ecto-(lyso) PA phosphatase can generate mono- or diacylglycerol from LPA or PA with a short sn-2 acyl chain (19). (iii) An LPA-specific lysophospholipase activity has been purified from rat brain (20).
Genetic and molecular biological approaches have facilitated cloning of genes encoding the LPAAT from nonmammalian species; however, no cloning of the mammalian counterparts of any of the key regulatory proteins has been reported. In plants, storage triacylglycerols are synthesized via a four-step pathway that involves acylation of glycerol 3-phosphate at the sn-1 position to form LPA, acylation of LPA by LPAAT, and then conversion of the PA to diacylglycerol by PA phosphatase followed by sn-3 acylation to form triacylglycerol (21). Interest in this metabolic pathway from the perspective of understanding and manipulating plant oil production has resulted in the cloning of several plant cDNAs that encode enzymes with LPAAT activity (22-25). Interestingly, these cDNAs have extensive sequence homology with each other as well as with LPAAT cDNAs from prokaryotic organisms, yeast, and nematodes. This prompted us to explore the Expressed Sequence Tag (EST) data base for uncharacterized mammalian homologues of the plant and microbial LPAATs. In this report we describe the EST-based cloning of a human LPAAT cDNA. The enzyme preferentially acylates LPA with arachidonate and is expressed in most tissues.
A TBLASTN search of the
GenBankTM dbest data base using the coconut LPAAT sequence
(25) identified two human ESTs deposited by the WashU-Merck EST project
with accession numbers H39628 and H44282. Based upon the EST
sequences, two oligonucleotide primers (forward:
5-GGGCCTCATCATGTACCTCGGGGGCG-3
; reverse: 5
-CTGCCCTCCCCCAGGTC-3
) were designed and used in a polymerase chain reaction (PCR) to obtain a fragment of the LPAAT homologue cDNA from a human
macrophage cDNA library (26). The fragment was used to generate a
radiolabeled probe by random priming (Random Primed Labeling Kit,
Boehringer Mannheim) that was applied to both a human heart muscle
cDNA library in Lambda Zap II and a genomic DNA library in Lambda
Fix II (both from Stratagene, La Jolla, CA). Approximately 5 × 105 to 1 × 106 phage were blotted onto
nitrocellulose and screened in 50% formamide, 0.75 M
sodium chloride, 75 mM sodium citrate, 50 mM
sodium phosphate (pH 6.5), 1% polyvinylpyrrolidine, 1% Ficoll, 1%
bovine serum albumin (BSA), and 100 µg/ml sonicated salmon sperm DNA.
After overnight hybridization at 42 °C, blots were washed
extensively in 3 mM sodium chloride, 0.3 mM
sodium citrate, 0.1% SDS at 50 °C. Following a secondary screen
under identical conditions, individual hybridizing plaques were
selected for DNA purification. The nucleotide sequence of both strands
of the cDNA was determined. Genomic DNA was excised from the Lambda
Fix II vector by NotI digestion and subcloned into
pBluescript (Stratagene). Genomic organization was determined by
nucleotide sequencing of the genomic clone with oligonucleotide primers
based upon the cDNA sequence. Additional primers designed according
to intron sequence were used to sequence across the exon/intron
boundaries and through the exons. Sequencing was accomplished by the
automated dideoxy chain termination method.
An expression construct
was engineered to contain the LPAAT coding region flanked at the 3 end
by a sequence encoding the FLAG tag (Eastman Kodak Co., Rochester, NY;
5
-GACTACAAGGACGACGATGACAAG) to facilitate detection of recombinant
protein. Oligonucleotide primers located 53 base pairs upstream of the
LPAAT cDNA translation initiation codon (Start R1:
5
-ATCAGAATTCCGGGAGCGGGAGCGGGAGCGAGCTGGCGGCGC) and at the termination
codon (FLAG Stop Xba:
5
-ATTCTCTAGACTACTTGTCATCGTCGTCCTTGTAGTCCTGGGCCGGCTGCACGCCAGA-CCCCGCAGTGGCCCCGTTC) were synthesized and used to generate a PCR fragment that could be digested with EcoRI and XbaI. The digested
fragment was cloned into the corresponding restriction sites of the
eukaryotic expression vector, pcDNA3 (Invitrogen, Carlsbad, CA).
The nucleotide sequence of the expression construct insert was
determined to ensure that no sequence changes were introduced by the
PCR. COS 7 cells were seeded at a density of 600,000 per 60-mm tissue
culture dish in Dulbecco's modified Eagle's medium supplemented with
10% fetal calf serum, 1 mM L-glutamine, 100 units/ml penicillin, and 100 µg/ml streptomycin. The following day
the cells were transfected in Dulbecco's modified Eagle's medium
containing 0.5 mg/ml DEAE-dextran, 0.1 mM chloroquine, and
10 µg of plasmid DNA for 1.5 h. The cells then were treated with
10% Me2SO in phosphate-buffered saline for 45 s,
washed with serum-free medium, and incubated in Dulbecco's modified
Eagle's medium supplemented with 1 mM
L-glutamine, 100 units/ml penicillin, 100 µg/ml
streptomycin, and 10% fetal calf serum. After 4 days, the medium was
removed, and cells were rinsed twice with phosphate-buffered saline and
harvested by scraping in 0.2 ml of 20 mM Tris-HCl, pH 7.5, and 5 mM NaCl. Cells were lysed by sonication.
Mock-transfected cells served as a negative control.
COS 7 cell lysates
containing recombinant LPAAT were assayed for acyltransferase activity
with various combinations of acyl donors and acceptors. Donors included
myristoyl coenzyme A, palmitoyl coenzyme A, stearoyl coenzyme A, and
arachidonyl coenzyme A (all from Sigma). Acyl acceptors tested included
1-palmitoyl-2-hydroxy-sn-glycero-3-phosphate, 1-myristoyl-2-hydroxy-sn-glycero-3-phosphocholine,
1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine, 1-alkyl-2-hydroxy-sn-glycero-3-phosphocholine (all from
Avanti Polar Lipids, Alabaster, AL). Each assay contained 10 µl of
cell lysate diluted 10-fold added to 240 µl of assay buffer and
substrate. Final concentrations of assay components are as follows: 100 mM HEPES-NaOH (pH 7.5), 200 mM NaCl, 5% (w/v)
glycerol, 10 mM EDTA, and 5 mM
-mercaptoethanol. Acyl acceptor species were added to a final
concentration of 20 µM. Also included was 1.3 µM 14C-radiolabeled acyl-CoA donor substrate
(American Radiolabeled Chemicals, St. Louis, MO) and 40 µM of the corresponding nonlabeled acyl-CoA. After a
5-min incubation at 37 °C, reactions were terminated by the addition
of 250 µl of 1 M KCl, 0.2 M
H3PO4, and 40 µl of 1 mg/ml BSA. Lipids were
extracted by the addition of 0.75 ml of chloroform:methanol (2:1). The
resulting organic phase was recovered, and 400 µl was dried to 30 µl and applied to a silica gel thin layer chromatography plate (TLC;
0.25-mm layer). Ascending TLC was performed in
chloroform:pyridine:formic acid (50:30:7). Radioactive spots resolved
by TLC were quantitated using a PhosphoImager with ImageQuant 3.0 software (Molecular Dynamics, Sunnyvale, CA). The authenticity of the
labeled product as PA was confirmed by comigration with PA in the TLC
system; in addition, PA was not produced in the absence of LPA. The
amount of PA generated was expressed as phosphoimaging units/µg of
protein. All assays were conducted in triplicate.
Proteins from COS 7 cell lysates were separated by SDS-polyacrylamide gel electrophoresis (12% gel; Novex, San Diego, CA) and electrotransferred to a polyvinylidene fluoride membrane (Pierce) for 1 h at 125 V in 192 mM glycine, 25 mM Tris base, 20% methanol, and 0.01% SDS. The membrane was incubated in 20 mM Tris, 100 mM NaCl (TBS) containing 5% BSA overnight at 4 °C. The blot was subsequently incubated 1 h at room temperature with 1 µg/ml of an anti-FLAG antibody (Kodak) in TBS containing 5% BSA, then washed with TBS, and incubated with a horseradish peroxidase-conjugated goat anti-mouse IgG in TBS containing 5% BSA for 1 h at room temperature. The blot was again washed with TBS and the signal was developed by electrochemiluminescence (NEN Life Science Products) according to the supplier's protocol.
Tissue mRNA Expression AnalysisNorthern blots containing poly(A)-selected human RNA from various tissues were obtained from CLONTECH (Palo Alto, CA). The blots were prehybridized for 6 h at 42 °C in 50% formamide, 0.75 M sodium chloride, 75 mM sodium citrate, 50 mM sodium phosphate (pH 6.5), 1% polyvinylpyrrolidine, 1% Ficoll, 1% BSA, 2% SDS, and 100 µg/ml sonicated salmon sperm DNA. After overnight hybridization in the same solution containing 1 × 106 dpm/ml 32P-labeled LPAAT coding region DNA, blots were washed extensively in 30 mM sodium chloride, 3 mM sodium citrate, 0.1% SDS at 50 °C and exposed to film overnight.
Chromosomal LocalizationA 20-kilobase genomic fragment containing the LPAAT gene was labeled with digoxigenin by nick translation and used as a probe for fluorescence in situ hybridization (FISH) of human chromosomes (Genome Systems, Inc., St. Louis, MO). The labeled probe was hybridized to normal metaphase chromosomes derived from phytohemagglutinin-stimulated peripheral blood lymphocytes. Reactions were carried out in the presence of sheared human DNA in 50% formamide, 10% dextran sulfate, 30 mM sodium chloride, 3 mM sodium citrate, 0.1% SDS. Hybridization signals were detected by treating slides with fluoresceinated antidigoxigenin antibodies followed by counterstaining with 4,6-diamidino-2-phenylindole. Initial labeling implicated a group C chromosome. A p16 INK4A probe that specifically hybridizes to chromosome 9, band p21 was used to demonstrate cohybridization of chromosome 9 with the LPAAT probe. A total of 80 metaphase cells were analyzed with 45 exhibiting specific labeling.
Sequence and Genetic Data AnalysisNucleotide sequences were analyzed with Geneworks (IntelliGenetics, Mt. View, CA). Amino acid sequence alignments were conducted using the ClustalW1 algorithm as found in the BCM Search Launcher-Multiple Sequence Alignments (http://dot.imgen.bcm.tmc.edu:9331/multi-align/multi-align.html). Individual pairwise alignments were conducted using Align Query (http://vega.crbm.cnrs-mop.fr/bin/align-guess.cgi). Transmembrane domain predictions were carried out by TMpred (http://ulrec3.unil.ch/cgi-bin/print_hit_bold.pl/software/TMPRED_form.html?tmpred#first_hit) and PSort (http://psort.nibb.ac.jp/form.html). O-Linked glycosylation sites were predicted by NetOglyc according to Hansen et al. ((27) CBS; http://genome.cbs.dtu.dk/netOglyc/cbsnetOglyc.html). N-Linked glycosylation sites and phosphorylation sites were predicted by Geneworks. A list of genes present on chromosome 9q34.3 was procured from Online Mendelian Inheritance in Man Gene Map (http://www3.ncbi.nlm.nih.gov/htbin-post/Omim/getmap?d1900).
The coconut LPAAT cDNA
sequence (25) was used to conduct a TBLASTN search of the
GenBankTM EST data base for mammalian homologues. Five ESTs
derived from human tissue cDNA libraries displayed significant
sequence homology with the plant sequence. The sequences of two of
these (accession numbers H39628 and H44282) were used to generate PCR
primers for screening a human macrophage cDNA plasmid library (26). A partial macrophage cDNA was used to screen a human heart muscle cDNA library for additional clones. The nucleotide sequence of a
hybridizing heart clone contains an open reading frame predicted to
encode a polypeptide of 278 amino acids with a molecular mass of 30.9 kDa (Fig. 1A). The predicted
protein sequence exhibits 23.2% identity with the coconut LPAAT, but
its identity with other members of the LPAAT family ranges as high as
33% (Table I). Like the other LPAATs,
the human enzyme is predicted to have multiple hydrophobic regions
which may serve as transmembrane domains (Fig. 1B). While
complete sequence identity at any given amino acid position between all
members of the family is relatively infrequent, a core region of highly
conserved amino acids is found from positions 167-205 of the human
sequence (Fig. 2).
|
Enzymatic Activity of Recombinant LPAAT
We subcloned the
coding region of the human LPAAT cDNA into a eukaryotic expression
vector and expressed the enzyme in COS 7 cells. Western analysis of
cell lysates showed high level expression of the recombinant
FLAG-tagged protein in cells transfected with the expression construct
but not in mock-transfected cells (not shown). As shown in Fig.
3, transfected cells exhibited 17-fold greater LPAAT activity than the mock-transfected cells when arachidonyl coenzyme A was used as substrate. Under the conditions of the assay,
acyl chain transfer tended to be reduced with the shorter, saturated
acyl substrates when comparing transfected cells with the corresponding
mock-transfected controls (6- to 10-fold increase in PA formation). The
enzyme was unable to utilize lysophosphatidylcholine as the acyl chain
acceptor.
mRNA Tissue Expression Patterns
Northern blot analyses
demonstrated that the human LPAAT mRNA is differentially expressed
(Fig. 4). Highest expression was seen in
the liver and pancreas; particularly low levels were found in a section
of brain and in placental tissue. Low expression in the brain was
unexpected since the LPA receptor is expressed at high levels in neural
tissue (8). To determine if regional localization might account for
this result, we hybridized a Northern blot containing multiple brain
subregions with the human LPAAT cDNA. While the message was found
in every region represented on the blot, expression was greater in the
spinal cord and subthalamic nucleus and least in the cerebellum,
caudate nucleus, corpus callosum, and hippocampus (Fig. 4B).
Northern blot analysis of LPAAT expression in the human myeloid-like
cell lines THP-1, HL-60, and U937 suggested that the gene is
constitutively expressed in these cells. Treatment of any of these cell
lines with phorbol 12-myristate 13-acetate or treatment of HL-60 cells
with Me2SO had no effect on message levels (not shown).
Genomic Structure and Organization
We examined a human
genomic DNA library and found that the entire LPAAT gene consists of
six exons (Table II), all of which are
located on a single 20-kb genomic clone. Exon 1 contains 66 bp of
5-noncoding sequence and encodes the first 61 amino acids of the
predicted polypeptide. The average size of the exons, exclusive of exon
6, is 145 nucleotides. Exon 6 is much larger (784 nucleotides) and
contains all of the 3
-untranslated sequence as well as sequence encoding the 58 most C-terminal amino acids. With the exception of the
splice donor sequence of the intron 3
of exon 1, all splice junctions
agree favorably with the known consensus splice sites for mammalian
genes (28).
|
To establish the chromosomal
location of the human LPAAT gene, FISH analysis was performed using the
20-kb human genomic clone as a probe. An initial hybridization
experiment localized the gene to the q terminus of a group C
chromosome. Subsequent cohybridization experiments with a p16 INK4A
probe implicated the q terminus of chromosome 9. Further evaluation of
10 specifically hybridized chromosomes pinpointed the location of the
LPAAT gene to band 9q34.3 (Fig. 5).
The LPAAT gene shares the 9q34.3 address with numerous genes of varied
functions. These include a number of enzymes such as the
fucosyltransferase gene, FUT7, the dopamine--hydroxylase gene, as well as the lipid-metabolizing enzymes carboxyl ester lipase
and prostaglandin D2 synthase. Genes encoding receptors include the
N-methyl-D-aspartate receptor 1 and the retinoid
X receptor. Diseases associated with this region include type II Ehlers-Danlos syndrome, which arises from a collagen V translocation, and a T-cell acute lymphoblastoid leukemia associated with a Notch homologue 1 translocation. Guanine nucleotide-releasing factor 2, orosomucoid 1 and 2 glycoproteins, and the homeodomain protein LHX3
also map to this region. There are no obvious functional relationships
between the LPAAT and other gene products or associated diseases that
map to the 9q34.3 location.
This manuscript describes the first cloning of a mammalian LPAAT cDNA. The significant homology with LPAATs from organisms as diverse as plants, nematodes, and prokaryotes supports our observations that the enzyme catalyzes the conversion of LPA to PA. The enzyme appears to be expressed in most tissues, suggesting it is an essential component of phospholipid metabolic pathways. The enzyme may also participate in intracellular signaling by generating PA, a known second messenger, or by regulating levels of LPA.
The predicted amino acid sequence of the human LPAAT is homologous to other LPAATs and thus displays a very similar hydropathy profile. Analysis of the human sequence for potential transmembrane domains strongly predicted four transmembrane helices (Fig. 1B). The first putative transmembrane domain may in fact represent a signal peptide (residues 1-23). However, the very short hydrophilic sequence between this putative signal peptide and the following hydrophobic helix is unusual, perhaps supporting a role for the first hydrophobic region being an integral transmembrane domain. The TMpred algorithm predicts the first (residues 3-21) and third (residues 123-141) putative transmembrane helices to be oriented within the membrane outside to inside and the second (residues 30-50) and fourth (residues 189-207) helices to be oriented inside to outside. This orientation places the predicted N-glycosylation site (Asn-59) and O-glycosylation sites (Thr-233, Thr-262) outside of the membrane. Interestingly, one of the predicted protein kinase phosphorylation sites (Thr-174) orients internal to the membrane in this configuration. This threonine lies within the most highly conserved region of the protein and is itself completely conserved in the eight most closely related members of the LPAAT family (Fig. 2). These observations, while speculative, are consistent with an integral plasma membrane localization of the enzyme such as has been proposed (29). Similarly, the PA metabolizing enzyme, PA phosphatase 2, is predicted to have five or six transmembrane domains and is thought to localize to the plasma membrane (30). It will be of interest to examine the spatial relationship between the two enzymes; perhaps they comprise part of a multienzyme locus that mediates LPA/PA metabolism.
Northern blot analysis (Fig. 4) demonstrated that the LPAAT message is expressed at very high levels in the liver and pancreas, at low levels in the placenta and certain regions of the brain, and at intermediate levels in numerous other tissues. This pattern of expression is in marked contrast to the pattern seen with the LPA receptor, which is expressed most abundantly in the brain, not in the liver, and at intermediate levels in other tissues (7, 8). This suggests the possibility that, at least in the brain and liver, LPA signaling and metabolism are mediated by additional receptors and enzymes. The cloning from Xenopus oocytes of an LPA receptor cDNA that bears little sequence identity to the mouse brain LPA receptor supports this possibility (9).
It is not clear if the LPAAT described in this report has a significant
role in modulating the intercellular signaling function of the LPA
produced by activated cells or if it primarily serves to generate the
PA intermediate in the membrane phospholipid synthetic pathway. The
apparent constitutive and widespread LPAAT mRNA expression patterns
may support the latter possibility. Further support is derived from
experiments in which quiescent fibroblasts degraded exogenous LPA
primarily to monoacylglycerol (31), suggesting that the LPA was
catabolized by a phosphatase rather than acylated by an LPAAT. On the
other hand, stimuli such as IL-1, tumor necrosis factor-
,
platelet-activating factor, and lipid A stimulate rapid accumulation of
intracellular PA (29, 32, 33). This appears to occur as a result of
phosphorylation-mediated activation of an intracellular pool of LPAAT
(34). It is also possible that expression of the LPAAT message is
up-regulated in response to specific stimuli or in cell types not
examined in the current study.
The biological consequences that arise from metabolic processing of
acutely generated LPA remain ambiguous. LPA at the site of inflammation
or tissue injury may mediate wound healing but, in local or temporal
excess, may participate in propagating an inflammatory response. In
this situation, there must be a mechanism to resolve the effects of the
mediator. One possible mechanism is to catabolize LPA via a phosphatase
or lysophospholipase to produce a simple glycerolipid that is subject
to rapid recycling. In contrast, the product of LPA acylation is PA. A
number of recent reports have suggested that PA may be a key
intracellular messenger in a common signaling pathway activated by
proinflammatory mediators such as IL-1, tumor necrosis factor-
,
platelet-activating factor, and lipid A (29, 32, 33). That this PA is
generated by the action of LPAAT was demonstrated by the observation
that small molecule inhibitors of the enzyme block PA formation in P388
monocytic leukemia cells stimulated with bacterial lipopolysaccharide
(35) and in hypoxia-treated human neutrophils (36). The small molecule inhibitors also protected mice from lipopolysaccharide-mediated endotoxic shock and from lung injury in a model of hemorrhage and
resuscitation (36). Others have demonstrated that formyl peptide-stimulated neutrophils can be induced to produce PA by a
phospholipase D-dependent mechanism (17). In this setting, the PA activated NADPH oxidase. In these examples, PA generated intracellularly in response to a particular extracellular stimulus mediates signaling. Many reports implicate PA as an extracellular agonist as well. The phospholipid purportedly stimulates monocyte migration (37), is mitogenic to Balb-c/3T3 cells (11), causes superoxide generation in neutrophils, activates protein
phosphorylation, and stimulates phosphatidylinositol-4-phosphate
kinase, and inactivates Ras GTPase-activating protein (reviewed in Ref.
38). It has been suggested, however, that contaminating LPA may be the
mediator of at least some of these effects (39). All of these
observations suggest that PA is an important signaling molecule, at
least intracellularly. Whether generation of PA via acylation of LPA in
the extracellular milieu at an inflammatory site leads to further
cellular activation remains to be determined experimentally.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) AF000237.
We thank Dina Leviten and Marsalina Quiggle for oligonucleotide synthesis and DNA sequencing, Joy Jarvis for assistance in preparation of the manuscript, and Thomas McIntyre, Stephen Prescott, Diana Stafforini, and Guy Zimmerman for helpful discussions.