MOLECULAR CLONING, SPLICE VARIANTS, EXPRESSION, AND PURIFICATION OF PHOSPHOLIPASE C-DELTA 4

Abstract Complementary DNAs encoding a previously unidentified phosphoinositide-specific phospholipase C (PLC) isozyme were cloned from a rat brain cDNA library by the polymerase chain reaction with degenerate oligonucleotide primers based on sequences common to three known -type PLC isozymes. The encoded polypeptide contains 772 amino acids (calculated molecular mass, 88,966 daltons) and is similar in primary structure to -type PLC isozymes, with overall sequence identities of 45% to PLC-1, 72% to PLC-2, and 47% to PLC-3. Thus, the new PLC isozyme was named PLC-4. Recombinant PLC-4 was purified from extracts of HeLa cells that had been infected with vaccinia virus containing the corresponding cDNA. The purified protein exhibited an apparent molecular mass of 90 kDa on SDS-polyacrylamide gels. The specific activity of PLC-4 and its dependence on Ca were similar to those of PLC-1. The distribution of PLC-4 in 16 different rat tissues was studied by immunoblot analysis with PLC-4-specific antibodies of fractions obtained after an enzyme-enrichment procedure. The 90-kDa immunoreactive protein was detected unambiguously in only eight tissues and was present at concentrations that were low compared to those of other major PLC isozymes. A 93-kDa immunoreactive protein was also prominent in testis but was not detected in the other seven positive tissues. The 93-kDa enzyme appears to be derived from a splice variant of the mRNA that encodes the 90-kDa PLC-4 and contains an additional 32 amino acids between the X and Y catalytic domains. Splice variants have not previously been detected for -type PLC isozymes.

The amino acid sequences of PLC isozymes are relatively variable with the exception of two well conserved regions, identified as the X (ϳ170 amino acid residues) and Y (ϳ260 residues) domains, that appear to constitute the catalytic site. The amino acid sequence similarity in the X and Y domains is ϳ60 and 40%, respectively, among the nine mammalian enzymes (3,4); the similarity is greater when members of the same type of PLC are compared. The sequence between the X and Y domains is short (40 -110 residues) in the ␤and ␦-type isozymes. However, in ␥-type isozymes, this region is much longer (ϳ400 residues) and contains two Src homology 2 (SH2) domains, which bind phosphotyrosine-containing sequences in other proteins, and one SH3 domain, which interacts with proline-rich sequences in cytoskeletal proteins (5,6). Furthermore, unlike ␥and ␦-type enzymes, ␤ type isozymes have a long carboxylterminal sequence (ϳ450 residues) downstream of the Y domain.
All mammalian and Drosophila PLCs possess an aminoterminal region of ϳ300 residues that precedes the X domain and contains a pleckstrin homology (PH) domain. The PH domain is a loosely conserved protein module of ϳ100 amino acids and targets various proteins to the membrane surface by interacting with either the ␤␥ subunits of G proteins or PIP 2 (7,8).
The various PLC isoforms appear to be activated by different receptors through different mechanisms. Activation of ␥-type enzymes is achieved by phosphorylation by autophosphorylated tyrosine kinases as a result of binding to these kinases via the SH2 domains. Isozymes of the ␤ type are activated as a result of binding either to the ␣ subunits of G q class G proteins, via the long carboxyl-terminal region, or to G␤␥ subunits, probably through the PH domain (4,9). Regulation of ␦-type enzymes is not yet understood. Despite the presence of PH domains in ␥and ␦-type enzymes, there is no evidence that these isozymes are modulated by G␤␥ subunits.
As part of our continuing effort to detect previously unidentified PLC isoforms and gather clues pointing to a regulatory mechanism for ␦-type PLC enzymes, we screened a rat brain cDNA library by the polymerase chain reaction (PCR) with oligonucleotide primers based on the amino acid sequences conserved in the X and Y domains of ␦-type isozymes. Here, we now describe the molecular cloning of a cDNA corresponding to a new ␦-type PLC, named PLC-␦4. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The  The underlined sequences at the 5Ј-end of each oligonucleotide indicate restriction enzyme cleavage sites (EcoRI and SphI sites for the forward and reverse primers, respectively) and two nucleotides (CG) added to facilitate cloning of PCR products. Primers SS and VA were based on amino acid sequences common to all three types of PLC, whereas the remaining three primers were based on sequences specific to ␦-type isozymes.
Three sequential PCR amplifications were performed with the GeneAmp PCR system 9600 (Perkin Elmer Corp.). A rat brain cDNA library in the Uni-ZAP XR vector (Stratagene) was used as the template for the first amplification reaction. The reaction mixture contained 0.5 l of library (1 ϫ 10 7 plaque-forming units), 50 pmol of each of the primers SS and VA, 0.1 mM dNTPs, 1 unit of native Taq polymerase (Perkin Elmer Corp.), and the manufacturer's buffer in a final volume of 50 l. Amplification was performed for one cycle of 5 min at 80°C, 3 min at 95°C, 15 s at 45°C, and 30 s at 72°C, followed by 44 cycles of a 15-s denaturating step at 94°C, a 15-s annealing step at 45°C, and a 30-s extension step at 72°C (the final extension step was prolonged to 5 min). Although sequences corresponding to primers SS and VA are present in all three types of PLC isozymes, the efficiency of amplification of PLC-␥ sequences was expected to be low because of the long distance (ϳ1.7 kb) between the two primer sites. To eliminate products amplified from the cDNA sequences of PLC-␤1 and PLC-␦1, two abundant PLC isoforms in brain, the products of the first amplification were treated with HindIII and AflIII, which cleave a 908-bp product derived from the PLC-␤1 sequence and a 833-bp product derived from the PLC-␦1 sequence, respectively.
The second amplification reaction was performed with the more internal, ␦ type-specific primers EP and GW and with the restriction enzyme-treated products from the first amplification reaction as the template. The reaction mixture contained 0.2 l of template and the same reaction components as for the first PCR. DNA amplification was achieved by 35 cycles with steps similar to those described above, with the exception that annealing was performed at 50°C. The products of the second amplification were subjected to digestion with AflIII, XhoI (which cleaves PLC-␦2 sequence), and PstI (which cleaves PLC-␦3 sequence).
The products from the second amplification reaction were then further amplified with the innermost ␦-specific primer, IL, and GW. The reaction was performed for 30 cycles under the same conditions as for the second PCR. Amplified DNA was digested with EcoRI and SphI, and the products were separated on an 8% polyacrylamide gel. A 426-bp oligonucleotide was cut and electroeluted from the gel, purified, and ligated into pUC19 vector. DNA sequencing was performed according to the Taq dye primer cycle sequencing method and Taq dideoxy terminator cycle sequencing method on an automated DNA sequencer (Applied Biosystems model 373A). The 426-bp oligonucleotide revealed a PLC-like sequence that differed from the sequences corresponding to known PLC isozymes.
Hybridization Screening of cDNA Libraries-A 312-bp 32 P-labeled probe was prepared by PCR in the presence of [ 32 P]dCTP with the 426-bp fragment as the template. The size of the probe was reduced (from 426 to 312 bp) to remove sequences corresponding to the X and Y domains, thus preventing hybridization to a known PLC sequence through the residual X and Y sequences. The radiolabeled probe was used to screen the rat brain cDNA library. Standard methods (10) were used for hybridization, cloning of DNA fragments from positive plaques into pBluescript SK vector (Stratagene), transforming Escherichia coli with the recombinant plasmids, and determining the nucleotide sequence of the cloned inserts. These procedures yielded two clones with a 2.7-kb insert that contained a full-length open reading frame for a PLC-like enzyme, a potential translational termination codon, and a poly(A) tail.
Expression and Purification of PLC-␦4-cDNA encoding the entire rat PLC-␦4 sequence was subcloned into the pTM1 vaccinia virus expression vector (11), with the PLC-␦4 coding sequence downstream of the bacteriophage T7 promoter in the resulting PTM1-PLC-␦4 construct. Recombinant vaccinia viruses encoding PLC-␦4 were generated by transfecting PTM1-PLC-␦4 into CV-1 cells, which had been infected with wild type vaccinia virus (12), and were finally selected by propagation in human TK Ϫ cells.
For the purification of PLC-␦4, HeLa cells were grown at 37°C to a density of 5 ϫ 10 5 cells/ml in MEM spinner medium supplemented with 5% horse serum, infected at a ratio of 10 viruses per cell with the recombinant virus and a vTF7-3 recombinant vaccinia virus that contained the bacterial T7 RNA polymerase gene, and harvested 2 days after infection (12). Cell pellets (80 ml) were washed three times with phosphate-buffered saline, suspended in 2 volumes of homogenation buffer (50 mM Tris-HCl (pH 7.4), 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride, leupeptin (10 g/ml), and aprotinin (10 g/ml)), and were disrupted by sonication. The homogenate was centrifuged at 100,000 ϫ g for 1 h, and the resulting supernatant (200 ml) was collected and divided into five 40-ml portions, each of which was applied to a preparative TSKgel DEAE-5PW HPLC column (21.5 ϫ 150 mm) that had been equilibrated with 50 mM Tris-HCl (pH 7.4) containing 1 mM EGTA and 0.1 mM DTT. Proteins were eluted at a flow rate of 5 ml/min with linear gradients of 0 -0.3 M NaCl for 35 min and 0.3-1 M NaCl for 5 min. Fractions (5 ml) were collected and assayed for PLC as described (13). The fractions (49 -51 min) corresponding to the second peak of phosphatidylinositol-hydrolyzing activity from each column were pooled and concentrated. The subsequent three chromatography steps on a preparative TSKgel phenyl-5PW column, a TSKgel heparin-5PW column, and a Mono Q column were performed according to procedures similar to those previously described (13).
Preparation of Antibodies to PLC-␦4-Peptides corresponding to PLC-␦4 amino acid residues 454 -464 (KDEGSDLDPAS) and residues 759 -772 (VYTCMQEDLDMDEP) were synthesized and conjugated to keyhole limpet hemocyanin with glutaraldehyde. Antisera to the peptides were generated separately in rabbits, and specific antibodies were purified by immunoaffinity chromatography with purified PLC-␦4 protein.
Tissue Distribution of PLC-␦4-Frozen rat tissues that had been harvested in liquid nitrogen were purchased from Pel-Freeze Biologicals. Regenerating liver was prepared from partially hepatectomized Sprague-Dawley rats 24 h after removal of approximately two-thirds of the liver mass (14). Each tissue (5-10 g) was thawed in 50 ml of homogenization buffer (10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, leupeptin (10 g/ml), aprotinin (10 g/ml), and calpain inhibitors I and II (each at 4 g/ml)) and homogenized in a glass homogenizer with a motor-driven Teflon pestle (10 strokes). The homogenate was centrifuged at 1000 ϫ g for 10 min. The supernatant was adjusted to 2 M KCl by adding solid KCl, stirred for 2 h at 4°C, and then centrifuged at 35,000 ϫ g for 30 min. The resulting supernatant was dialyzed overnight against 4 liters of homogenization buffer and again centrifuged. The supernatant (100 -200 mg of protein) was applied to a heparin-Sepharose CL-6B (Pharmacia Biotech Inc.) column (20 ml of gel packed in a 1.5 ϫ 15 cm Econo column) that had been equilibrated with 20 mM Hepes (pH 7.0) containing 1 mM EGTA and 0.1 mM DTT. Bound proteins were eluted at a flow rate of 4 ml/min with the equilibration buffer containing 1.2 M NaCl. Fractions (16 ml) were collected and assayed for PLC activity. Essentially all detectable PLC activity eluted in six fractions, which were pooled and concentrated in a stirred ultrafiltration cell fitted with a YM 30 membrane (Amicon). After the final NaCl concentration was adjusted to 50 mM, the concentrate was centrifuged at 100,000 ϫ g for 10 min. Proteins (20 mg) from the supernatant were injected into a TSKgel heparin-5PW HPLC column (7.5 ϫ 75 mm) and fractionated as described previously (13). Fractions (0.5 ml) were collected and assayed for PLC activity. For immunoblot analysis, fractions were concentrated with Centricon-30 (Amicon), and one-fourth of the pooled peak fractions (61-64 min) from each tissue was resolved by SDS-polyacrylamide gel electrophoresis on a 7% gel. Proteins were transferred to nitrocellulose and incubated with purified antibodies to PLC-␦4, and immune complexes were visualized with alkaline phosphatase-conjugated goat antirabbit IgG.
RT-PCR-Sprague-Dawley rats were sacrificed by cervical dislocation, and various tissues were collected in liquid nitrogen. Total RNA was isolated from these frozen tissues by the guanidinium isothiocyanate method, and poly(A) ϩ RNA was purified from total RNA with Oligotex-dT (Qiagen). RT-PCR was carried out with a StrataScript kit (Stratagene). First-strand cDNAs were synthesized with 100 ng of poly(A) ϩ RNA from each tissue and 300 ng of oligo(dT) primer in a 50-l reverse transcriptase reaction mixture. The first-strand products were amplified with two sets of PLC-␦4-specific primers (first set, nucleotides (nt) 1108 -1130 and 1640 -1662; second set, nt 1305-1325 and 1516 -1535 (nucleotides were numbered beginning with the first residue of the ATG initiation codon; those on the 5Ј-side of residue 1 are indicated by negative numbers)) by two consecutive rounds of PCR. The first amplification was performed for 30 cycles of 15 s at 94°C, 15 s at 60°C, and 30 s at 72°C with 3 l of the cDNA products and the first set of primers; the final extension step was increased to 5 min. The second-round amplification was achieved by 20 cycles with the second set of primers and 0.5 l of the first-round PCR products as template. The final products were separated on a 1.8% agarose gel. The expected 230-bp oligonucleotide was amplified specifically from PLC-␦4 cDNA. A measure of the efficiency of cDNA synthesis and the ability of the resulting cDNA to serve as a template for amplification was obtained for each of the RNA samples by amplifying the first-strand products with rat ␤-actin primers (Clontech); 30 cycles of PCR were performed under the same conditions as described above, and a 764-bp fragment was amplified.
Cloning and Sequencing of Two Alternatively Spliced Forms of PLC-␦4-RT-PCR of rat testis poly(A) ϩ RNA yielded 330-and 270-bp bands in addition to the 230-bp band. These two additional fragments were subcloned directly into the pGEM-T plasmid vector (Promega). Restriction enzyme analysis and sequencing of these two fragments revealed that an additional 96 and 42 nt, respectively, were inserted immediately before the Y region of the PLC-␦4 open reading frame. To determine whether these two additional fragments were amplified from alternatively spliced forms of PLC-␦4 mRNA in testis, the entire coding region of PLC-␦4 cDNA was amplified with two sets of primers (first set, nt Ϫ115-95 and 2421-2444; second set, nt Ϫ58 -35 and 2477-2497), all of which corresponded to the 5Ј-or 3Ј-untranslated regions of the cDNA. Two rounds of PCR were performed as described above but with a 1-min extension step. The second-round PCR products were digested with StuI and EcoRI for restriction enzyme analysis. PLC-␦4 cDNA contains only one site for each of these two restriction enzymes; thus, digestion of the second-round PCR products should yield three distinct fragments of 1493, 597, and 411 bp. The presence of additional two forms of PLC-␦4 cDNA in testis would be expected to result in the generation of additional 507-and 453-bp fragments.

PCR Amplification, Cloning, and Sequencing of PLC-␦4
cDNA-The existence of previously unidentified PLC-␦ isozymes was investigated by subjecting a rat brain cDNA library to three sequential PCR amplifications. Five blocks of conserved amino acid sequences, three (SS, EP, and IL) from the X domain and two (GW and VA) from the Y domain, were chosen for the synthesis of degenerate oligonucleotide primers. The outermost pair of primers (SS and VA) was used for the first PCR amplification and was based on sequences common to all three types of PLC isozymes, whereas the other three primers (EP, IL, and GW), used for the second (EP and GW) and third (IL and GW) PCR amplifications, were specific to ␦-type isoforms. The first-and second-round PCR products were treated with restriction enzymes known to cleave sequences corresponding to various PLC isozymes before serving as templates for the next round of PCR. The link between the X and Y domains of PLC isozymes varies both in length and sequence. Thus, whereas the presence of conserved residues in the X and Y domains identifies a PCR product as a member of the PLC family, a comparison of the linker sequence with known sequences further reveals whether it represents a newly discovered PLC-␦.
The third PCR amplification yielded a 426-bp fragment that revealed a PLC-like sequence distinct from those of known PLC isozymes. Removal of the sequences corresponding to the X and Y domains yielded a 312-bp fragment that was used as a probe to screen the rat brain library. Two clones with a 2.7-kb insert were obtained. Complete sequencing of these two clones revealed an open reading frame of 2,316 bp flanked by a 142-bp 5Ј-untranslated region and a 255-bp 3Ј-untranslated region including a poly(A) tail. The translational initiation site (ATG) was assigned to the first methionine codon at nt 143 because of the presence of an in-frame stop codon upstream of this methionine and flanking sequences that fulfill the Kozak criteria for initiation (15). An in-frame translational termination codon (TGA) was present after codon 772. Therefore, we concluded that this new PLC contained 772 amino acids, with a calculated molecular mass of 88,966 daltons (Fig. 1A).
A comparison of the deduced amino acid sequence with known PLC sequences revealed that the predicted protein was similar in primary structure and overall structural organization to PLC-␦ type isozymes, with overall sequence identities of 45, 72, and 47% to PLC-␦1, PLC-␦2, and PLC-␦3, respectively. Thus, the protein encoded by the cloned cDNA was named PLC-␦4 (Fig. 1B).
Purification and Characterization of PLC-␦4-To confirm that the isolated PLC-␦4 cDNA encodes a PLC enzyme, we subcloned the entire coding region of the cDNA into a vaccinia virus vector and expressed the encoded protein in HeLa cells. The expressed PLC-␦4 protein was purified by four successive column chromatographies (Fig. 2). The first chromatographic step, performed with a DEAE-HPLC column, separated PLC-␦4 from several endogenous PLC enzymes; immunoblot analysis with isozyme-specific antibodies revealed that the first activity peak ( Fig. 2A) contained PLC-␥1, PLC-␦1, PLC-␤1, and PLC-␤3 from HeLa cells (data not shown), whereas PLC-␦4 was detected in the second peak. HeLa cells infected with a control vaccinia virus vector not containing the PLC-␦4 cDNA yielded the first activity peak but not the second. The marked retention of PLC-␦4 by the heparin column (third step) facilitated its purification (Fig. 2C). SDS-polyacrylamide gel electrophoresis analysis of the peak fractions pooled from the last step on a Mono Q column revealed an apparently homogeneous (Ͼ90%) preparation of a 90-kDa protein (Fig. 2D), in good agreement with the calculated molecular mass of PLC-␦4.
Because of the low level of expression, only 200 g of purified PLC-␦4 could be obtained from 20 liters of cultured HeLa cells, despite good recovery yield (overall 25% relative to the DEAE column preparation). In contrast, ϳ1 mg of purified PLC-␦1 was obtained from 8 liters of cultured HeLa cells that had been transfected with the same virus vector harboring PLC-␦1 cDNA. 2 The catalytic activities of PLC-␦1 and PLC-␦4 expressed in HeLa cells were measured with [ 3 H]PIP 2 at various concentrations of free Ca 2ϩ ; both enzymes exhibited similar specific activities and dependence on Ca 2ϩ (data not shown).
Tissue Distribution of PLC-␦4 as Determined by Immunoblot Analysis-The concentration of PLC-␦4 in most rat tissues is low, and it was not possible to detect the isozyme in crude homogenates by immunoblot analysis. It was thus necessary to prepare samples enriched in PLC-␦4 before such analysis. Tissue homogenates were extracted with 2 M KCl to solubilize PLC isozymes associated with particulate fractions. After dialysis, the extracts were fractionated on a conventional heparin column (mainly to remove turbid material) and then on a heparin HPLC column. Under the experimental conditions described in immunoblot analysis with affinity-purified monospecific antibodies to PLC-␦4 residues 454 -464 (Ab-454) or 759 -772 (Ab-759) (Fig. 3). The intensity of the 90-kDa band was strongest with testis and decreased in the order of brain Ͼ skeletal muscle Ͼ thyroid gland Ͼ stomach Ͼ thymus Ͼ aorta Ͼ heart.
No band was detected unambiguously with normal or regenerating liver, kidney, prostate, adrenal gland, intestine, pancreas, or lung. In addition to the 90-kDa band, testis showed a strong band of 93 kDa and a weak band of 86 kDa. Only the 86-kDa band was detected in spleen.
By comparing the immunoblot intensities of the samples with those of purified PLC-␦4, the amount of PLC-␦4 in various tissues was measured quantitatively. The complex formed between PLC-␦4 and antibodies was visualized with 125 I-labeled protein A, and the amount of radioactivity was determined with a PhosphorImager. From the amounts of protein loaded on the two heparin columns and the SDS gel, it was deduced that the amount of PLC-␦4 expressed in nanograms per milligram of protein in the KCl extracts was 7.2 for brain, 20 for testis (combining 90-and 93-kDa forms), 6.4 for skeletal muscle, and 5.6 for thyroid.
Tissue Distribution of PLC-␦4 mRNA as Determined by RT-PCR-We also examined the distribution of PLC-␦4 mRNA in rat tissues by RT-PCR with two pairs of PLC-␦4-specific primers. The relative intensities of the expected 230-bp band (Fig.  4A) match qualitatively those of the PLC-␦4 protein band on immunoblot analysis (Fig. 3) when the relative efficiency of PCR amplification is taken into account.
Alternatively Spliced Forms of PLC-␦4 mRNA-In addition to a 230-bp band of weak intensity, the RT-PCR products from testis poly(A) ϩ RNA contained two larger (330 and 270 bp) bands of stronger intensity. These larger bands were also present in the products from thyroid gland, normal liver, prostate, intestine, and pancreas (Fig. 4A). Cloning and sequencing analysis of the two larger RT-PCR products revealed 96-and 42-nt inserts positioned in-frame between nt 1461 and 1462 of the coding region of PLC-␦4 cDNA (Figs. 1 and 4B). These insertions add 32 and 14 amino acid residues and would result in putative PLC-␦4 proteins, designated ALT-I and ALT-II, with calculated molecular masses of 92,425 and 90,455 daltons, respectively, if there are no other differences.
To determine whether there are any modifications other than the insertions of 96 and 42 bp, we amplified fragments encompassing the entire coding region of PLC-␦4 cDNA by PCR. Two sets of primers based on sequences located in the 5Јand 3Ј-untranslated regions were used. The resulting products were then subjected to restriction enzyme digestion with StuI and EcoRI, each of which has only one recognition site in this region and together are expected to generate three fragments of 1493, 597, and 411 bp from the PLC-␦4 cDNA. Three fragments with these expected sizes were observed in the restriction reaction mixtures derived from brain and skeletal muscle   Fig.  2C. One-fourth of the pooled peak fractions (61-64) from each tissue was subjected to immunoblot analysis with rabbit antibodies to PLC-␦4, and immune complexes were detected with alkaline phosphatase-conjugated goat anti-rabbit IgG. Lanes at both sides contained prestained molecular size standards mixed with the indicated amounts of purified PLC-␦4 protein.
poly(A) ϩ RNA and PLC-␦4 cDNA (Fig. 4C). As predicted for fragments derived from the same molecule, the intensity of the three bands was proportional to their size. However, in the digestion mixture derived from testis poly(A) ϩ RNA, the 411-bp band was faint, whereas the 1493-and 597-bp bands were strong; two bands of medium intensity were apparent at positions corresponding to 507 and 453 bp, the sizes expected for fragments derived from ALT-I and ALT-II cDNA, respectively (Fig. 4C). This result suggests that ALT-I and ALT-II mRNAs are likely alternatively spliced forms of PLC-␦4 mRNA that differ from the latter only in the 96-and 42-bp inserts.
To determine whether the 93-kDa testis protein detected by immunoblot analysis (Fig. 3) is a product derived from such alternative splicing, we prepared rabbit antibodies (Ab-ALT-I and Ab-ALT-II, respectively) to the ALT-I-specific sequence KCPMSCLLICVHVLAQA and the ALT-II sequence KKAPN-SIPESILL. The PLC-␦4-enriched fractions from brain, skeletal muscle, thyroid gland, and testis were subjected to immunoblot analysis with a mixture of Ab-454 and Ab-759, with Ab-ALT-I, or with Ab-ALT-II (Fig. 5). The 90-kDa band was detected by the mixture of Ab-454 and Ab-759, but not by Ab-ALT-I or Ab-ALT-II, in all four tissues. The 93-kDa band was recognized by all antibodies and detected only in testis, suggesting that this protein corresponds to ALT-I. No band was detected between the 90-and 93-kDa proteins, where the putative ALT-II protein would be expected. These results indicate that ALT-II mRNA is not translated or that the translation product comigrates with ALT-I in the SDS gel. Because the majority (APN-SIPESILL) of the sequence against which Ab-ALT-II was prepared is also present in ALT-I, Ab-ALT-II also likely reacts with ALT-I. Therefore, the reactivity of Ab-ALT-II with the 93-kDa band does not necessarily suggest the presence of ALT-II. DISCUSSION Our results indicate that PLC-␦4 is expressed at low concentrations in a limited number of rat tissues. The 90-kDa form was unambiguously detected by immuoblot analysis in only 8 of 16 tissues examined, even after enrichment, and the 93-kDa form was detected only in testis. Brain is one of the rat tissues relatively rich in PLC-␦4, but the enzyme concentration of 7.2 ng per milligram of crude extract protein is significantly lower than those of PLC-␤1 (70 ng/mg), PLC-␥1 (140 ng/mg), and PLC-␦1 (180 ng/mg). Faint RT-PCR bands derived from PLC-␦4 mRNA were visible in a greater number of tissues, probably because of the greater sensitivity of this procedure.
Regenerating liver was examined for PLC-␦4 expression because a PLC-␦-like enzyme that is not recognized by antibodies to either PLC-␦1 or PLC-␦2 was shown to be expressed specifically during rat liver regeneration (14) and because our immunoblot analysis showed that PLC-␦3 is not the PLC isozyme specific to regenerating liver. 2 The present study suggests that PLC-␦4 also is not expressed in regenerating liver.
Like the other three ␦-type PLCs, PLC-␦4 contains a PH domain sequence at the amino terminus. However, none of the four PLC-␦ enzymes is significantly activated by G␤␥ subunits. 2 The PH domain of PLC-␦1 contains a 14-residue sequence (KVKSSSWRRERFYK) enriched in basic amino acids (net positive charge of 5) that was shown to form the core of the binding site for the negatively charged IP 3 and PIP 2 and was FIG. 4. Alternatively spliced forms of PLC-␦4 mRNA. A, RT-PCR analysis of PLC-␦4 mRNA from various rat tissues. First-strand cDNAs were generated with 100 ng of poly(A) ϩ RNA from each tissue and an oligo(dT) primer. PCR was performed with PLC-␦4-specific primers, as well as with ␤-actin-specific primers as a control, and the final products were separated on a 1.8% agarose gel. PLC-␦4 cDNA was also subjected to PCR amplification. Gel markers with a 100-bp ladder (Research Genetics and Life Technologies, Inc.) were loaded in the side lanes. B, schematic diagrams of PLC-␦4 splice variants. Two RT-PCR fragments of 330 bp (ALT-I) and 270 bp (ALT-II) were obtained in addition to the expected 230-bp fragment from testis poly(A) ϩ RNA. Sequencing of these two fragments indicated the presence of 96-and 42-bp inserts between codons 487 and 488 of PLC-␦4. The amino acid sequences deduced from these inserts are indicated. The regions against which antibodies Ab-ALT-I and Ab-ALT-II were generated are underlined. C, restriction enzyme digestion analysis of RT-PCR products containing the entire coding region of PLC-␦4. Amplification was achieved with primers corresponding to the 5Ј-and 3Ј-untranslated regions of PLC-␦4 cDNA. The resulting 2.5-kb products were digested with StuI and EcoRI and separated on a 1.8% agarose gel. Distinct fragments derived from ALT-I and ALT-II cDNAs in testis as well as that derived from PLC-␦4 cDNA are indicated.
FIG. 5. Immunoblot analysis with antibodies specific to PLC-␦4 and its splice variants. Fractions enriched with PLC-␦4 were obtained from the indicated tissues as described in Fig. 3 and were subjected to immunoblot analysis with antibodies to PLC-␦4 (mixture of Ab-454 and Ab-759), ALT-I (Ab-ALT-I), or ALT-II (Ab-ALT-II).
suggested to tether the enzyme to membrane surfaces containing PIP 2 (18). The basic amino acid-rich sequence is well conserved in PLC-␦4 (KVRTKSWKKLRYFR); indeed, the ␦4 sequence has a net positive charge of 7 and thus would be predicted to interact better than PLC-␦1 with IP 3 and PIP 2 .
Two splice variants of PLC-␦4 mRNA, ALT-I and ALT-II mRNAs, were detected by PCR. ALT-I mRNA encodes the 93-kDa enzyme that contains an additional 32 amino acids located between the X and Y domains of PLC-␦4. It does not appear that ALT-II mRNA, which would produce an enzyme with an additional 14 amino acids in the same region between the X and Y domains, is translated in sufficient quantities to be detected by immunoblot analysis. PLC-␦4 is thus the first example a ␦-type PLC that exists in splice variants. Previously identified splice variants correspond to ␤-type PLC isozymes including rat PLC-␤1 (19), bovine PLC-␤4 (20), Drosophila PLC-p21 (21), and Drosophila PLC-norpA (22). Examination of the splicing differences in these PLC isozymes, including rat PLC-␦4, reveals that all occur outside of the X and Y domains (Fig. 6). Rat PLC-␤1 and Drosophila PLC-p21 are alternatively spliced in the carboxyl-terminal region following the Y domain, where the G␣ q -binding site is located (12,23). Splicing differences in Drosophila PLC-norpA variants occur in the PH domain, to which G␤␥ subunits and PIP 2 might bind. It is thus possible that the differences among splice variants result in differences in the ability to interact with signaling components.
Splice variants of rat PLC-␦4, like those of bovine PLC-␤4 (20), differ in the region separating the X and Y domains. This separating region of ␥-type PLC isozymes plays important roles by interacting with various signaling components through its SH2, SH3, and PH domains. Whether the region linking the X and Y domains of ␤and ␦-type PLC enzymes also interacts with signaling molecules is not known. However, this region of all mammalian ␤and ␦-type isozymes is rich in acidic amino acids; 20 of 70, 26 of 76, 29 of 137, and 24 of 100 residues are acidic in ␤1, ␤2, ␤3, and ␤4 isozymes, respectively, and 11 of 50, 19 of 53, 17 of 44, and 17 of 66 residues are acidic in ␦1, ␦2, ␦3, and ␦4 isozymes, respectively. Unlike the other three ␦-type PLC isozymes, PLC-␦4 contains a high density (12 of 66 residues) of serine and threonine in the region separating the X and Y domains. All mammalian ␤-type isoforms with the exception of PLC-␤4 also contain a high density of serine and threonine residues in this region. Furthermore, as with other splice variants (22), PLC-␦4 and ALT-I exhibit distinct tissue distributions. Together, these observations suggest that the 90and 93-kDa PLC-␦4 enzymes might be regulated differentially. FIG. 6. Schematic representations of the splice variants of various PLC isozymes. Differences between splice variants are illustrated by solid lines above or below each schematic. Numbers above each schematic represent amino acid positions at boundaries of the splicevariant domains, and those on the right indicate total number of amino acids.