Characterization of human PLD2 and the analysis of PLD isoform splice variants

^a Novartis Institute for Biomedical Research, Summit, New Jersey 07901, USA

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Phospholipase D (PLD) cleaves phosphatidylcholine in response to a variety of cell stimuli to release phosphatidic acid, which is associated with a number of cellular responses including regulated secretion, mitogenesis, and cytoskeletal changes. Recent advances in this field include the reports of cDNA sequences for two mammalian PLD isoforms: human PLD1 and rodent PLD2. We report the characterization of cDNA encoding human PLD2. In these experiments, we uncovered alternate splice variants of both human isoforms and evaluated the relative abundance of these messages by reverse transcriptase polymerase chain reaction, thereby indicating the physiologically relevant forms. Further, Northern hybridization experiments defined the tissue distribution of the human PLD messages. Human PLD1 does not appear to be an abundant message in any tissue tested whereas levels of human PLD2 mRNA apparently were higher and more variable. The specific activity and regulation of recombinant human PLD2 are indistinguishable from that of recombinant mouse PLD2. Analysis of the amino acid sequences of both human isoforms revealed important putative Pleckstrin homology domains and identified additional members of the PLD gene family that help to delimit the catalytic domain. The presence of Pleckstrin homology domains in the PLDs resolves several contradictory observations regarding PLD regulation and the domain structure of the proteins.—Steed, P. M., Clark, K. L., Boyar, W. C., Lasala, D. J. Characterization of human PLD2 and the analysis of PLD isoform splice variants. FASEB J. 12, 1309–1317 (1998)

Key Words: PH domains • tissue distribution • alternate splicing • enzyme regulation

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PHOSPHOLIPASE D (PLD)² is an intracellular signaling enzyme that, in response to a variety of extracellular signals, cleaves the most abundant phospholipid in mammalian membranes—phosphatidylcholine—to liberate choline and the second messenger molecule phosphatidic acid (for reviews, see refs 1–3). Phosphatidic acid can be converted to diacylglycerol via phosphatidic acid phosphohydrolase (4), and elevated levels of PLD-derived phosphatidic acid and diacylglycerol have been associated with numerous physiological responses, including regulated secretion (5), mitogenesis (6), phagocyte activation (7, 8), and cytoskeletal changes (9). Indicative of the central role for PLD in intracellular signaling, PLD activity is regulated by many proteins, including small GTP binding proteins (10), protein kinase C, trimeric GTP binding proteins, cytoskeletal regulators (11), tyrosine phosphorylation (12), and proteins involved in the regulation of synaptic vesicles (13, 14), among others.

PLD signaling has been investigated for many years, but molecular studies were impossible until the recent cloning of human PLD1 (hPLD1; ref 15), rat PLD1 (16), and rodent PLD2 (17, 18). Though the two isoforms are approximately 50% identical, PLD1 exhibits low basal activity that is activated by numerous factors whereas PLD2 exhibits high basal activity that is apparently controlled by repression. PLD1 is regulated in a manner consistent with the numerous reports on the control of PLD catalysis by the ADP-ribosylation factor (ARF) and Rho families of GTP binding proteins, as well as protein kinase C (15). The regulatory controls on PLD2 activity are unknown, though overexpression of this isoform affects cytoskeletal structure (17).

Studies of human PLD1 indicate the presence of an alternate splice form, and both isoforms have been reported to lack known homology motifs such as SH2, SH3, or Pleckstrin homology (PH) domains, unlike isoforms of phospholipase C PH domains (10, 17). PH domains are recognized as membrane adaptors for signaling proteins via the binding of inositol lipids to the PH domain, and this process is driven by the action of phosphatidylinositol-3' kinase (19). The reported lack of PH domains in the PLDs is unexpected, since both isoforms are potently activated by inositol lipids. PLD activity has been shown to be both positively and negatively affected by wortmannin, a specific inhibitor of phosphatidylinositol-3' kinase (20–22).

Here we report the cloning, analysis, and expression of the cDNA sequence encoding hPLD2, which, like human PLD1, has multiple alternate splice forms. The relative abundance of the splice forms for both human PLD isoforms was evaluated along with the tissue distributions of the messages. Our analysis of these sequences revealed the presence of putative PH domains and other important features in both isoforms.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Cloning and expression of human PLD2
Using the mouse sequence of PLD2 (mPLD2) reported by Colley et al. (17), seven human expressed sequence tags (est) from the Genbank database were found to represent parts of the human PLD2, covering approximately 900 bp of the cDNA at the 3' end of the coding region and 3' untranslated region (accession numbers: W75418, W61990, W66743, R93485, R83570, H02092, R69739, H01995, R93432, R97756). These sequences were used to generate approximately 930 bp of the human cDNA sequence by reverse transcriptase polymerase chain reaction (rtPCR) with human brain Quick-Clone cDNA (Clontech, Palo Alto, Calif.) as a template. PCR and cloning procedures were carried out using standard methods. Two independent clones of the 3' end of human PLD2 were sequenced (both strands) with an ABI 377 automated sequencer (Perkin-Elmer, Norwalk, Conn.). Using the sequence obtained from the hPLD2 3' clones, the bulk of the hPLD2 cDNA was obtained via 5' RACE, with human brain Marathon-Ready cDNA (Clontech) used as a template. A single identifiable fragment of approximately 2.7 kb was isolated, cloned, and sequenced as described above. Since there was no est information available for this region of the cDNA, five independent clones were analyzed to obtain reliable sequence data. Using this information, six full-length cDNA clones were obtained by rtPCR and sequenced for further confirmation. The sequence for human PLD2 has been deposited into GenBank, accession number AF033850. A cDNA encoding hPLD2 was engineered by PCR and introduced into a pFASTBAC-ht vector (Life Technologies, Gaithersburg, Md.) for expression with a 5' 6X His tag. For comparison, a cDNA encoding rmPLD2 was obtained by rtPCR using the published sequence (GenBank accession # U87557) with mouse lung RNA as template (kindly provided by Dr. S. I. Hu). The cDNA encoding mPLD2 was introduced into the pFASTBAC-ht vector as described above. Baculovirus stocks were obtained and used to generate rhPLD2s in Sf9 cells by using standard procedures. Recombinant proteins were obtained from the insect cells as previously described (10), but purified using the Talon Cobolt chelating resin (Clontech). For these experiments, the standard PLD in vitro assay was used (15).

Cloning and expression of human PLD1 and ARF1
Using the sequence of human ARF-dependent PLD1 reported by Hammond et al. (15), the gene was cloned in two overlapping (the unique HindIII site) fragments. PCR fragments were generated, cloned, and sequenced as described above. Several tissue sources were used for these amplifications, but cDNA from human cultured chondrocytes (kindly furnished by Dr. S. I. Hu) provided by far the highest yield of product, which was subsequently used to obtain two independent clones. The sequence obtained from both clones agreed with the reported sequence of hPLD1. Recombinant human PLD1 (rhPLD1) was expressed as described above. In a similar manner, human ARF1 was cloned and expressed in Escherichia coli using pPROEX-ht vectors (Life Technologies) with purification via the 6X His tag as described above.

Analysis of hPLD1 and hPLD2 splice variants
Two regions of variability in the DNA sequence were identified from the sequenced clones. Oligonucleotides corresponding to bases 2530–2549 and the complement of 2901–2920 of hPLD2 were used to amplify the sequence around one variable sequence and the 5' end of the coding region under the conditions described above. Additional sources of template were the Quick-Screen human cDNA library panel (Clontech) along with HeLa and human leukocyte Quick-Clone cDNA (Clontech). PCR products were separated on 10% polyacrylamide TBE gels (Bio Rad, Hercules, Calif.), stained with SYBR Green I DNA gel stain (Molecular Probes, Eugene, Oreg.), and visualized on a Molecular Dynamics Storm fluorescence scanning system (Molecular Dynamics, Sunnyvale, Calif.). For the other variable region, oligonucleotides corresponding to bases 1024–1043 and the complement of 1356–1375 were used. Representatives of each class of PCR product were isolated, cloned, and sequenced. Oligonucleotides corresponding to bases 1251–1271 and the complement of (1986–2002 of hPLD1 were used to amplify the sequence around this putative variable region with the conditions and templates described above. PCR products were separated, visualized, isolated, cloned, and sequenced as described above.

Northern analysis of hPLD1 and hPLD2
Human multiple tissue Northern blots (I, II, and III; Clontech) were probed according to manufacturer's instructions. DNA fragments corresponding to the first 2.7 kb of hPLD1a and 1.7 kb hPLD2 were labeled with the Rediprime random primer labeling kit (Amersham, Arlington Heights, Ill.) and used as probes. Hybridization was visualized with a Phosphor screen (Amersham Life Sciences) on a Molecular Dynamics Storm fluorescence scanning system.

Sequence analysis of human PLDs and their splice variants
Sequence data bases were searched using standard search engines, and potential homologues were further analyzed with the SEQALIGN program package (23), using the scoring matrix of Henikoff and Henikoff (24). For each alignment, the observed sequence alignment score was compared to the distribution of alignment scores from 250 pairs of randomly shuffled sequences. The resulting number of standard deviations from the mean of the distribution of scores from shuffled sequences is a measure of statistical significance and is referred to as the Z-value. Z-values greater than 6 indicate a statistically significant sequence similarity and hence a conserved 3-dimensional fold; however, alignments with Z-values of less than 10 are less likely to correctly identify equivalence of every structural element of a 3-dimensional fold (25). Families of aligned sequences were built in a pairwise manner, starting with the most similar sequences.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The cloning and expression of human PLD2 and analysis of alternate splice forms
A small portion of the coding region of hPLD2 was cloned via rtPCR using several overlapping est sequences as described in Methods. This information was used to amplify the bulk of the coding region by 5' RACE, providing the cDNA sequence of the entire coding region. To verify the sequence, six full-length clones were generated by rtPCR and sequenced, revealing two regions of variability (designated hPLD2b and hPLD2c) in the cDNA sequence leading to alternate amino acid sequences ( Fig. 1). One variant, PLD2b, contained a small deletion at the 3' end of the coding region resulting in a deletion of 11 amino acids; this deletion was also present in one of the est sequences corresponding to hPLD2. The other variant, PLD2c, is a 56 bp insertion that results in premature termination of the polypeptide ( Fig. 1). To verify that these sequence differences represented alternate splice forms and not cloning or PCR artifacts, the relevant regions were amplified, visualized, cloned, and sequenced from cDNAs derived from multiple tissues ( Fig. 2). The identification of the PCR products from alternate splice forms was verified by DNA sequencing. All of the alternately spliced forms were present at comparable concentrations in all tissues tested with the exception of PLD2c, which is present in low abundance in brain and skeletal muscle.

View larger version (138K): [in this window] [in a new window]   Figure 1. Alignment of amino acid sequences. Regions of sequence identity are marked by a black background, and regions of sequence similarity are marked with a grey background. Alignment of mammalian PLD domains. Accession numbers: human PLD1, U38545; Norway rat, D88672; mouse PLD1, U87557; human PLD2; Chinese hamster, U94995.

View larger version (24K): [in this window] [in a new window]   Figure 2. Relative abundance of human PLD2 splice variants. PCR reactions were carried out as described in Materials and Methods. A) 3' End splice variant distinguishing PLD2b from PLD2a. B) Splice variant distinguishing PLD2c. The Clontech Quick-Screen panel of cDNA clones was used to determine relative abundance of messages: lane 1, brain; 2, placenta; 3, skeletal muscle; 4, kidney; 5, heart; 6, lung; 7, liver; 8, pancreas; 9, universal.

Recombinant human and mouse PLD2 were produced, purified, and assayed in parallel with no detectable differences in activity or response to ARF (data not shown). As reported previously for the mouse protein (17), PLD2 has high constitutive activity that is unaffected by ARF.

The analysis of hPLD1 alternate splice forms
Human PLD1 was amplified from human chondrocyte cDNA using primers designed from the published sequence in two fragments overlapping the HindIII site in the middle of the coding region (base 1492). Two independent clones of the 5' end of the coding region, bases 59 to 1574, agreed with the reported sequence of hPLD1. However, both of the independent clones of the 3' end of the gene, bases 1849–1962, were lacking 114 bp, 38 amino acid residues corresponding to amino acids 585–622. We then sequenced two additional clones of this region amplified from HeLa cDNA, the source used to clone PLD1 (15), both of which had the identical sequence to our original clones. This deletion is identical to the alternate splice form of PLD1, PLD1b, reported by Hammond et al. (10) while this manuscript was in preparation. Since we were unable to isolate any clones of PLD1 that were full-length, we performed PCR on this region of the sequence to determine the relative abundance of the two splice variants using the same techniques as used for hPLD2 ( Fig. 3). In all tissues tested, with the possible exception of HeLa, the most abundant form is the short variant, PLD1b. We also found additional potential alternate splice forms, one larger than PLD1a and PLD1b (approximately 800 bp) and one smaller (approximately 250 bp). At least two independent clones of each of these fragments were analyzed by DNA sequencing. The bands migrating to the expected sizes of PLD1a and PLD1b were cloned from HeLa and skeletal muscle, sequenced, and verified in this manner. The smaller potential splice variant was amplified from brain, cloned, and sequenced. This fragment represents an alternate splice variant with the same 3' junction as PLD1b but a 5' junction at base 1634 of PLD1, in contrast to the 5' junction of base 1850 for PLD1b ( Fig. 1). We have designated this form of PLD message as PLD1c, but it is unlikely to be physiologically relevant because the splice junction causes a change in reading frame, resulting in premature transcriptional termination ( Fig. 1), and deletions of this region are catalytically inactive (26).

View larger version (52K): [in this window] [in a new window]   Figure 3. Relative abundance of human PLD1 splice variants. PCR reactions were carried out as described in Materials and Methods. Lane 1, human chondrocyte cDNA (our PLD1 clone); lane 2, human HeLa Quick-Clone cDNA (Clontech; source for original hPLD1 clone ref 13); lane 3, blank. In addition, the Clontech Quick-Screen panel of cDNA clones was screened: lane 4, universal human cDNA library; 5, pancreas; 6, liver; 7, lung; 8, heart; 9, kidney; 10, skeletal muscle; 11, placenta; 12, brain. The larger potential splice variant from skeletal muscle was found to be a fragment of E. coli DNA corresponding to the uxu operon (involved in glucuronate metabolism, Genbank accession # D13329), an apparent contamination artifact of the cDNA source used.

Database and homology searches
Using the cDNA sequence of human PLD2, we sought to analyze the amino acid sequence for domain structures. The sequence of hPLD1 was also used in these analyses. To identify domains within the hPLD sequences, numerous database searches were conducted. The programs BLASTP and BIC_SW identified the homologues that have previously been reported (15, 26). A search of the PROSITE database with ISREC ProfileScan server identified two profiles: the TPS domain profile (transphosphatidylase signature, PS50035) and the PH domain profile (Pleckstrin homology, PS50003). Further analysis was performed to determine the extent of sequence similarity between representative members used in defining these profiles and the corresponding putative potential domains within the PLD sequences, described below.

The sequence similarity between the standard PH domain and hPLD1 is relatively weak. To investigate whether this similarity is valid and to eliminate the possibility of bias in the standard PH domain, the same sequence alignment strategy used to describe the PH domain was applied to the putative PH domain region in the PLD sequences (25). First, an aligned family of mammalian PLD sequences was generated ( Fig. 1); hPLD1b and hPLD2a were selected to represent the human PLD1 and PLD2 splice variants, respectively. A preliminary alignment of the PLD sequences family with an aligned family of 7 PH domain sequences (see ref 25 and Fig. 4) was used to identify the location and extent of the PH domain in the PLD sequence. The corresponding region of SPO14 was added to the PLD family (YA2G was excluded due to marginal statistical significance; Z-value of 2.4). Comparison of the PLD/SPO14 family to the 7 PH domain family resulted in a statistically significant Z-value of 9.5. To further support this finding, the PLD/SPO14 family was aligned with a different aligned family of 4 PH domains (aligned on the basis of their 3-dimensional structure; Fig. 4and refs 27, 28). This alignment resulted in a similar, statistically significant Z-value of 8.8, which further supports the weak but detectable sequence similarity of the putative PLD PH domain.

View larger version (58K): [in this window] [in a new window]   Figure 4. Alignment of putative PH domains in human PLD proteins. Putative PLD PH domains aligned with PLC delta 1, dynamin, spectrin, and Pleckstrin N. Dynamin, spectrin, and Pleckstrin N are aligned on the basis of their known 3-dimensional structures (Fig. 5 and ref 24).

The transphosphatidyl signature profile has already been reported to be highly conserved in all PLD1-related genes (15). Inspection of the sequences used to define the transphosphatidyl signature profile revealed the cardiolipin synthase family, which appeared to have significant sequence similarity extending well beyond the local transphosphatidyl signature motif ( Fig. 5) sequence. An aligned family of cardiolipin synthase homologues was generated by first aligning YLP2 and YWIE. Examination of the aligned sequences clearly indicated that the sequence similarity began at residue 120 of YLP2. The E. coli CLS sequence was then added to this truncated, aligned family, forming a family of three sequences restricted to a common domain that contained the transphosphatidyl signature sequences. Sequence comparison of the PLD family to the cardiolipin synthase family indicated strong sequence similarity, with a Z-value of 26.3. Although this global sequence similarity appears to define a large catalytic domain, the highly conserved transphosphatidyl signature region and its contribution to the alignment score could skew the statistical analysis. As a control experiment, the cardiolipin synthase family was `mutated' so that it no longer contained the signature sequence. The sequence similarity between this mutated version of the cardiolipin synthase family and the PLD family was still unambiguous at Z = 20.2, clearly indicating that the sequence similarity extends well beyond the signature sequence.

View larger version (50K): [in this window] [in a new window]   Figure 5. Domain structure of proteins related to the mammalian PLD proteins. The relative positioning of the PH, CLS, and transphosphatidyl signature sequences are indicated. The scale and noted insertions are based on the numbers of amino acid residues. eCLS represents the aligned family of E. coli cardiolipin synthase, YLP2, and YWIE as described in the text. Accession numbers: (see Fig. 1); spo14, P36126; yag2, Q09706; eCLS P31071; YLP2, P31048; YWIE, P45860.

Tissue distribution of human PLD isoforms
With the cDNA sequences of both human PLD isoforms available, a determination of the human tissues expressing these genes was possible. Multiple tissue Northern blots were probed with cDNA corresponding to hPLD1 and hPLD2 ( Fig. 6). With the possible exceptions of spleen, ovary, pancreas, and spinal cord, there are no tissues particularly enriched with PLD1 message, which appears as two distinct sizes of 4.2 and 7.0 kb. The differences between these messages are not known. Placenta, thymus, prostate, ovary, thyroid, spinal cord, and trachea were the tissues identified as having higher levels of message for the PLD2 isoform, which has a single message of approximately 6.0 kb.

View larger version (52K): [in this window] [in a new window]   Figure 6. Tissue Northern analysis of human PLD isoforms. A) Human multiple tissue Northern blots probed with hPLD1. B) Human multiple tissue Northern blots probed with hPLD2. C) Human multiple tissue Northern blots probed with actin control primer. Hybridizations and visualization were performed as described in Materials and Methods.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
We have described the cloning and characterization of human PLD2 and the analysis of splice variants and domain structures for both human PLD isoforms. In our attempts to clone hPLD2 via RACE and hPLD1 by rtPCR, we uncovered putative alternate splice forms of both messages. The hPLD1b splice form agrees with the splice variation of PLD1 reported by Hammond et al. (10) during the preparation of this manuscript; however, our study revealed an additional alternate splice form of hPLD1: hPLD1c. Although the human PLDs are highly homologous (51% identical), human PLD2 exhibits a splice pattern distinct from that reported for hPLD1 although the chromosomal structure is similar (29). A shorter splice variant is found but in equal abundance with PLD2a; however, the spliced region is shorter than that of PLD1 (11 amino acids) and is located in a different part of the gene, toward the 3' end of the coding region. Since the deleted sequence is located in the second transphosphatidyl signature of the protein that is required for the activity of the enzyme (13), this splice variant is catalytically inactive, but may play a regulatory role due to its high abundance. It is not known whether both abundant splice variants of the two PLD isoforms are actually translated because they are indistinguishable by Western analyses (ref 10; P. Steed and W. Boyar, unpublished observations).

Since we were unable to obtain clones of the longer PLD1a splice variant in cDNA derived from several sources, we evaluated the relative abundance of PLD1 splice variants in numerous tissues to determine whether the PLD1a form is present in meaningful amounts. The shorter variant of PLD1, PLD1b, is the most abundant in all tissues tested, with the possible exception of HeLa cells from which PLD1 was originally cloned. Alignment of the known PLD family sequences ( Fig. 1) indicates that the region of insertion in PLD1a is not present in any other homologue sequence, consistent with the observation that this region appears to be extraneous since it has no effect on PLD1 enzyme activity (10).

In addition to these variants, we detected messages for both PLD1 and PLD2 that encode truncated proteins, PLD1c and PLD2c. These splice variants are likely inactive due to their lack of two transphosphatidyl domains (26). These variants are biologically relevant because we identified the alternate splicing forms by sequencing multiple cDNA clones, evaluated these messages by rtPCR, and verified all potential PLD fragments by sequence analysis of multiple clones. Further, the alternate splice forms described here correspond to intron/exon junctions of murine PLD2 (29), and the putative splice variants of PLD2 are found in relatively equal abundance with the full-length clone; therefore, it is unlikely that they represent products of PCR skipping or the amplification of pseudogenes or unprocessed message. These results allow for a determination of the physiologically relevant splice form(s) of the PLDs; hPLD2a and hPLD2b are most abundant and found in relative equal amounts. Since the longer form of PLD1, PLD1a, is not found in appreciable amounts in any nontransformed tissues tested, we conclude that the PLD1b form is more appropriate for biochemical analysis.

Determination of the cDNA sequences for both human isoforms of PLD allows for a study of the tissue distribution of these messages. Northern hybridizations indicate that the mRNA messages for both PLD isoforms are present, at least in small amounts, in all tissues tested. However, PLD1 is not detected in particular abundance in any tissue whereas there is some heterogeneity to the expression of PLD2. In addition, we detect two distinct messages for human PLD1 migrating to sizes of 4.2 and 7.0 kb, respectively. We detect only a single message of 6.0 kb for human PLD2. In contrast to the reported enrichment for mouse PLD2 in brain (14), we do not see high expression of human PLD2 message in brain.

Several lines of evidence point to the mammalian PLDs as proteins likely to contain PH domains. In addition to the plasma membrane localization of mouse PLD2, both isoforms (in purified form) require inositol lipids for maximum activity in purified preparations; therefore, PLDs must interact directly with these lipid domains (15, 16). Further, PLD activity has been shown to be regulated by phosphatidylinositol-3' kinase (20). However, the original reports on the mammalian PLD sequences indicated that both were devoid of PH domains. In light of these apparent contradictions between the characteristics of recombinant PLD activity and the reported lack of PH domains, we tried to carefully evaluate our sequence of human PLD2. By analysis of the PLD primary sequences with the algorithms originally used to identify the motif, we have found that both PLD1 and PLD2 contain putative PH domains. Since the PLD family is recognized as a new and highly conserved gene family (15) and PH domains typically exhibit low sequence similarity (25), it is not surprising that the PH domains of the PLDs are relatively unique and difficult to detect. A recent report on the cloning of an Arabidopsis PLD indicates the potential presence of a PH domain (30). Though there is some correlative data for a PH domain in the Arabidopsis PLD isoforms, this putative motif is not present in the human PLDs, with the exception of hPLD1a (amino acids 549–556; within the inserted region). Since our data show that PLD1a is the minor splice form of PLD1 and this region is not present in PLD2 whereas both isoforms are regulated by PIP₂, this putative motif cannot be the primary PIP₂ binding site for the human PLDs. The domain structure of the human PLDs was further elucidated by identifying the potential extent of the catalytic domain based on comparisons with the cardiolipin synthase family. The putative catalytic domain encompasses both transphosphatidyl signature motifs and begins carboxyl-terminal to the PH domain. Analysis of the human PLD sequences determined that the previously identified transphosphatidyl signature motif extends beyond the ARF-dependent PLD family. This homology between the PLD genes and the cardiolipin synthase family points to common evolutionary history for the PLD signaling enzymes and lipid biosynthesis. These results extend the definition of the catalytic domain of the PLDs and the PLD family that has been postulated by mutagenesis experiments (26).

The stimulation of PLD activity has been correlated with a number of important cellular functions, including phagocyte activation, mitogenesis, and regulated secretion. Many of these functions are critical to disease states, including inflammation and cancer. Our study provides a molecular characterization of the two human PLD isoforms as well as several tools for the detailed study of the role of PLD in these cellular functions.

	FOOTNOTES

 
¹ Correspondence: Novartis Institute for Biomedical Research, 556 Morris Ave., Summit, NJ 07901, USA. Email: paul.steed{at}pharma.novartis.com

² Abbreviations: human PLD1, human phospholipase D1; ARF, ADP-ribosylation factor; PH, Pleckstrin homology; est, expressed sequence tags; rh, recombinant human; rtPCR, reverse transcriptase polymerase chain reaction.

Received for publication February 23, 1998. Accepted for publication April 30, 1998.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Billah, M. M. (1993) Phospholipase D and cell signaling. Curr. Opin. Immunol. 5, 114–123[Medline]
2. Exton, J. H. (1994) Phosphatidylcholine breakdown and signal transduction. Biochem. Biophys. Acta 1212, 26–42[Medline]
3. Exton, J. H. (1997) New developments in phospholipase D. J. Biol. Chem. 272, 15579–15582[Free Full Text]
4. Kanoh, H., Imai, S., Yamada, K., and Sakane, F. (1992) Purification and properties of phosphatidic acid phosphatase from porcine thymus membranes. J. Biol. Chem. 267, 25309–25314[Abstract/Free Full Text]
5. Stutchfield, J., and Cockroft, S. (1993) Correlation between secretion and phospholipase D activation in differentiated HL60 cells. Biochem. J. 293, 649–655
6. Boarder, M. R. (1994) A role for phospholipase D in control of mitogenesis. Trends Pharmacol. Sci. 15, 57–62[Medline]
7. Dubyak, G. R., Schomisch, S. J., Kusner, D. J., and Xie, M. (1993) Phospholipase D activity in phagocytic leucocytes is synergistically regulated by G-protein and tyrosine kinase-based mechanisms. Biochem. J. 292, 121–128
8. McPhail, L. C., Qualliotine-Mann, D., Agwu, D. E., and McCall, C. E. (1993) Phospholipases and activation of the NADPH oxidase. Eur. J. Haematol. 51, 294–300[Medline]
9. Ha, K.-S., and Exton, J. H. (1994) Activation of actin polymerization by phosphatidic acid derived from phosphatidylcholine in IIC9 fibroblasts. J. Cell Biol. 123, 1789–1796[Abstract/Free Full Text]
10. Hammond, S. M., Jenco, J. M., Nakashima, S., Cadwallander, K., Gu, Q.-m., Cook, S., Nozawa, Y., Prestwich, G. D., Frohman, M. A., and Morris, A. J. (1997) Characterization of two alternately spliced forms of phospholipase D1. J. Biol. Chem. 272, 3860–3868[Abstract/Free Full Text]
11. Steed, P. M., Nagar, S. S., and Wennogle, L. P. (1996) Phospholipase D regulation by a physical interaction with the actin-binding protein gelsolin. Biochemistry 35, 5229–5237[Medline]
12. Gomez-Cambronero, J. (1995) Immunoprecipitation of a phospholipase D with antiphosphotyrosine antibodies. J. Interferon Cytokine Res. 15, 877–885[Medline]
13. Chung, J.-K, Sekiya, F., Kang, H.-S., Lee, C., Han, J.-S., Kim, S. R., Bae, Y. S., Morris, A. J., and Rhee, S. G. (1997) Synaptojanin inhibition of phospholipase D activity by hydrolysis of phosphatidylinositol 4,5-bisphosphate. J. Biol. Chem. 272, 15980–15985[Abstract/Free Full Text]
14. Lee, C., Kang, J.-S., Chung, J.-K., Sekiya, F., Kim, J.-R., Han, J.-S., Kim, S. R, Bae, Y. S., Morris, A. J., and Rhee, S. G. (1997) Inhibition of phospholipase D by clathrin assembly protein 3 (AP3). J. Biol. Chem. 272, 15986–15992[Abstract/Free Full Text]
15. Hammond, S. M., Altshuller, Y. M., Sung, T.-C., Rudge, S. A., Engebrecht, J., Morris, A. J., and Frohman, M. A. (1995) Human ADP-ribosylation factor-activated phosphatidylcholine-specific phospholipase D defines a new and highly conserved gene family. J. Biol. Chem. 270, 29640–29643[Abstract/Free Full Text]
16. Park, S.-K, Provost, J. J., Bae, C. D., Ho, W. T., and Exton, J. H. (1997) Cloning and characterization of phospholipase D from rat brain. J. Biol. Chem. 272, 29263–29271[Abstract/Free Full Text]
17. Colley, W. C., Sung, T.-C., Roll, R., Jenco, J., Hammond, S. M., Altshuller, Y., Bar-Sagi, D., Morris, A. J., and Frohman, M. (1997) Phospholipase D2, a distinct phospholipase D isoform with novel regulatory properties that provokes cytoskeletal reorganization. Curr. Biol. 7, 191–201[Medline]
18. Kodaki, T., and Yamashita, S. (1997) Cloning, expression, and characterization of a novel phospholipase D complementary DNA from rat brain. J. Biol. Chem. 272, 11408–11413[Abstract/Free Full Text]
19. Klarlund, J. K., Guilherme, A., Holik, J. J., Virbasius, J. V., Chawala, A., and Czech, M. P. (1997) Signaling by phosphoinositide-3, 4, 5-triphosphate through proteins containing Pleckstrin and Sec7 homology domains. Science 274, 1927–1930
20. Carrasco-Marin, E., Alvarez-Dominiquez, C., and Leyva-Cobain, F. (1994) Wortmannin, an inhibitor of phospholipase D activation, selectively blocks major histocompatibility complex class II-restricted antigen presentation. Eur. J. Immunol. 24, 2031–2039[Medline]
21. Mollinedo, F., Gajate, ., and Flores, I. (1994) Involvement of phospholipase D in the activation of transcription factor AP-1 in human T lymphoid Jurkat cells. J. Immunol. 153, 2457–2469[Abstract]
22. Natarajan, V., Vepa, S., al-Hassani, M., and Scribner, W. M. (1997) The enhancement by wortmannin of protein kinase C-dependent activation of phospholipase D in vascular endothelial cells. Chem. Phys. Lipids 86, 65–74[Medline]
23. Clark, K. (1991) Ph.D. dissertation. Kansas State University
24. Henikoff, S., and Henikoff, J. G. (1992) Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci. USA 89, 10915–10919[Abstract/Free Full Text]
25. Mayer, B. J., Ren, R., Clark, K. L., and Baltimore, D. (1993) A putative modular domain present in diverse signaling proteins. Cell 73, 629–630[Medline]
26. Sung, T.-C., Roper, R. L., Zhang, Y., Rudge, S. A., Temel, R., Hammond, S. M., Morris, A. J., Moss, B., Engebrecht, J., and Frohman, M. A. (1997) Mutagenesis of phospholipase D defines a superfamily including a trans-Golgi viral protein required for poxvirus pathogenicity. EMBO J. 16, 4519–4530[Medline]
27. Lemmon, M. A., Ferguson, K. M., O'Brien, R., Sigler, P. B., and Schlessinger, J. (1995) Specific and high-affinity binding of inositol phosphates to an isolated Pleckstrin homology domain. Proc. Natl. Acad. Sci. 92, 10472–10476[Abstract/Free Full Text]
28. Ferguson, K. M., Lemmon, M. A., Schlessinger, J., and Sigler, P. B. (1995) Structure of the high affinity complex of inositol triphosphate with a phospholipase C Pleckstrin homology domain. Cell 83, 1037–1046[Medline]
29. Redina, O. E., and Frohman, M. A. (1998) Organization and alternate splicing of the murine phospholipase 2 gene. Biochem. J. In Press
30. Qin, W., Pappan, K., and Wang, X. (1997) Molecular heterogeneity of phospholipase D (PLD). J. Biol. Chem. 272, 28267–28273[Abstract/Free Full Text]