Cloning and Characterization of a Second Human CTP:Phosphocholine Cytidylyltransferase*

CTP:phosphocholine cytidylyltransferase (CCT) is a key regulator of phosphatidylcholine biosynthesis, and only a single isoform of this enzyme, CCTα, is known. We identified and sequenced a human cDNA that encoded a distinct CCT isoform, called CCTβ, that is derived from a gene different from that encoding CCTα. CCTβ transcripts were detected in human adult and fetal tissues, and very high transcript levels were found in placenta and testis. CCTβ and CCTα proteins share highly related, but not identical, catalytic domains followed by three amphipathic helical repeats. Like CCTα, CCTβ required the presence of lipid regulators for maximum catalytic activity. The amino terminus of CCTβ bears no resemblance to the amino terminus of CCTα, and CCTβ protein was localized to the cytoplasm as detected by indirect immunofluorescent microscopy. Whereas CCTα activity is regulated by reversible phosphorylation, CCTβ lacks most of the corresponding carboxyl-terminal domain and contained only 3 potential phosphorylation sites of the 16 identified in CCTα. Transfection of COS-7 cells with a CCTβ expression construct led to the overexpression of CCT activity, the accumulation of cellular CDP-choline, and enhanced radiolabeling of phosphatidylcholine. CCTβ protein was posttranslationally modified in COS-7 cells, resulting in slower migration during polyacrylamide gel electrophoresis. Expression of CCTβ/CCTα chimeric proteins showed that the amino-terminal portion of CCTβ was required for posttranslational modification. These data demonstrate that a second, distinct CCT enzyme is expressed in human tissues and provides another mechanism by which cells regulate phosphatidylcholine production.

CCT␣ proteins have been identified and sequenced in rat (3), hamster (4), mouse (5), and human (6), and there are only minor differences among these mammalian cDNAs (see Ref. 6 for a comparison). Their catalytic properties are thought to be essentially identical, and CCT␣ can be divided into four distinct functional domains (see Fig. 1). The amino-terminal domain between residues 1 and 71 contains a sequence that specifies the nuclear localization of the protein between residues 2 and 28 (7,8). The catalytic core extends from residues 72 to 233. This region of the protein is conserved from yeast to mammals and is responsible for substrate binding and catalysis. In particular, the conserved HXGH motif is essential for cytidylyltransferase activity (9,10). The third domain, located between residues 256 and 288, contains three 11-residue amphipathic repeats that form ␣-helices following association with lipid regulators and contribute to the reversible membrane association of the enzyme (11)(12)(13)(14)(15). The binding of stimulatory lipids to this region greatly enhances catalytic activity by lowering the K m of the enzyme for CTP into the range corresponding to cellular concentrations of the nucleotide (16). CCT␣ is also negatively regulated by lipids and is potently inhibited by sphingosine (17), lysophosphatidylcholine (18), and antineoplastic phospholipids (18,19). The fourth domain of CCT␣ is the carboxyl-terminal phosphorylation domain between residues 315 and 367. CCT␣ membrane association and activity are modulated by reversible phosphorylation (20,21), and all of the phosphorylation sites are located in the carboxylterminal region (22). Phosphorylation attenuates CCT␣ biochemical activity by interfering with lipid stimulation (23), and unphosphorylated CCT␣ exhibits a greater degree of membrane association in cells (20).
CCT␣ has been localized using cellular in situ methods to the nucleus in Chinese hamster ovary cells (7,8), but in rat hepatocytes, the protein has been detected in both the nuclear and cytoplasmic compartments (24). CCT␣ has also been identified in association with Golgi membranes (25,26), endoplasmic reticulum, and transport vesicles (27) using biochemical methods. There is only one isoform of CCT expressed in yeast (28), and only one isoform of CCT (CCT␣) has been identified, purified, or cloned from mammalian sources (2). The existence of a conditionally lethal Chinese hamster ovary cell mutant with a temperature-sensitive defect in CCT␣ activity (29) also suggested that there was only a single CCT isoform in mammalian cells. A single genetic locus for CCT␣ was identified on mouse chromosome 16 (5), and the murine CCT␣ gene has been cloned (30). In this work, we identify a unique, second human CCT isoform, called CCT␤. CCT␤ catalyzes the same enzymatic reaction as CCT␣ and requires the presence of lipids for full activity. However, CCT␤ lacks the nuclear targeting sequence and the phosphorylation domain of CCT␣, suggesting that CCT␤ is distinct from CCT␣ with regard to its subcellular localization and regulation.
Anti-CCT␣ rabbit polyclonal antiserum was raised against a synthetic peptide (MDAQSSAKVNSRKRRKE) corresponding to the first 17 amino acids of CCT␣. AntiCCT␤ rabbit polyclonal antiserum was raised against a synthetic peptide (MEEIEHTCPQPRL) corresponding to amino acids 27-39 of CCT␤. Peptides and peptide antigens were prepared by the Molecular Resource Center of St. Jude Children's Research Hospital. To prepare antigen, each peptide was coupled to keyhole limpet hemocyanin via an additional cysteine at the carboxyl terminus of the peptide. Immunization of rabbits and collection of antisera was performed by Rockland, Inc., according to their standard schedule. Antisera were purified by affinity chromatography on Affi-Gel 10 cross-linked to the peptide.
Isolation of the CCT␤ cDNA-Human Genome Systems identified and provided a clone that exhibited significant sequence similarity to CCT␣. The protein expressed from this cDNA, however, did not exhibit significant CCT catalytic activity. Sequence information from this clone was used to search the public expressed sequence tagged data base. We identified a clone (GenBank TM accession no. AA382871) that contained 40 bp of related sequence at the 5Ј end. We purchased this clone from American Type Culture Collection and sequenced the cDNA on both strands using primers that flanked the multiple cloning sites and internal primers that were synthesized to ensure a complete read on both strands. The cDNA contained a single open reading frame we called CCT␤. A 1.3-kb BamHI-XhoI fragment was excised and subcloned into the mammalian expression vector, pcDNA3, to generate plasmid pPJ34, which expressed CCT␤ from the constitutive cytomegalovirus promoter.
Plasmid DNA was isolated, transcribed, translated, and labeled with [ 35 S]methionine using the Promega T7-coupled transcription/translation kit according to the manufacturer's instructions. The labeled proteins were analyzed by SDS-gel electrophoresis and visualized by autoradiography.
Construction of Plasmids for Expression of CCT Chimeras and CCT␤ Amino-terminal Truncation-Rodent CCT␣ (pWYCT) and human CCT␤ (pPJ34) cDNAs cloned into pcDNA3 were digested with SspI. The pcDNA3 vector has an SspI site distanced approximately 1 kb from the 5Ј-end of the T7 promoter. The CCT␣ cDNA has an SspI site at nucleotide 260, whereas CCT␤ has an SspI site at nucleotide 311. The fragments that contained either CCT␣ or CCT␤ sequence plus vector sequence were purified. The 1.3-kb fragment generated from CCT␤ cDNA that encoded the amino terminus was ligated to the 5-kb fragment of pWYCT to generate pCCT␤/CCT␣, and the 1.3-kb fragment from the CCT␣ cDNA encoding the amino terminus was ligated to the 5-kb fragment of pPJ34 to generate pCCT␣/CCT␤. The resulting plasmids were checked for correct orientation with the polymerase chain reaction using the T7 and SP6 primers of pcDNA3.
Plasmid pPJ34 was digested with NcoI, and the resulting 2185-bp fragment was purified. The DNA overhanging sequences were filled in with Klenow fragment and purified. The fragment was next digested with XhoI to yield a 1180-bp fragment, and after purification, this fragment was ligated into pcDNA3 that had been previously digested with EcoRV and XhoI, resulting in plasmid pPJ35. Inserts in pcDNA3 were screened by polymerase chain reaction using T7 and Sp6 primers. DNA sequencing confirmed the truncation of the first 26 amino acids of CCT␤.
CCT Assay-CCT activity was determined essentially as described previously (31). The standard assay contained 64 M lipid activator (PtdCho:oleic acid, 1/1), 4 mM CTP, 10 mM MgCl 2 , 150 mM bis-Tris-HCl, pH 6.5, 1 mM phospho[ 14 C]choline (specific activity, 4.5 mCi/mmol), in a final volume of 50 l. The reaction mixtures were incubated at 37°C for 10 min. The reaction was stopped by the addition of 5 l of 0.5 M Na 3 EDTA, and the tubes were vortexed and placed on ice. Next, 40 l of each sample was spotted on preadsorbent Silica Gel G thin layer plates, which were developed in 2% ammonium hydroxide/95% ethanol (1:1, v/v). CDP-[ 14 C]choline was identified by co-migration with a stand-ard, scraped from the plate, and quantitated by liquid scintillation counting. Protein was determined according to the Bradford method (32).
Isolation of CCT␤ from Endogenous Lipids-CCT␤ was isolated from COS-7 cells 48 h after transfection with plasmid pPJ34. Cells were washed with phosphate-buffered saline and harvested by centrifugation, and the pellet was lysed by incubation in lysis buffer (10 mM NaCl, 1 mM EDTA, 2 mM dithiothreitol, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 50 mM NaF, 100 M Na 3 VO 4 , 10 mM HEPES, pH 7.4) for 1 h on ice. The cells were disrupted by sonication, and the particulate matter was removed by centrifugation. The supernatant was loaded onto a 0.5-ml DEAE-Sepharose column and the column was washed with 1.5 ml of each of the following in succession: lysis buffer, lysis buffer plus 1% Nonidet P-40, lysis buffer, lysis buffer plus 0.25 M NaCl, lysis buffer plus 0.5 M NaCl, and lysis buffer plus 1.0 M NaCl. The eluant was collected in 0.5-ml fractions, and CCT activity was located in the 0.25 M NaCl wash. This procedure is essentially the same as described in previous papers (16,23). CCT␤ activity that was eluted from the column could only be detected in the presence of added lipid activators.
Transfection Experiments-COS-7 cells were grown in 100-mm dishes to 80% confluency in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 1% glutamine. Transfections using LipofectAMINE reagent were performed according to the manufacturer's instructions. Briefly, 10 g of plasmid and 60 ml of LipofectAMINE reagent were separately diluted into 0.8 ml of serumfree medium. The two solutions were combined and incubated at 25°C for 45 min. Next, 6.4 ml of serum-free medium was added to each tube, and the diluted solution was overlaid onto the COS-7 cells that had been previously rinsed with serum-free medium. The cells and reagents were incubated at 37°C for 5 h, and then 8 ml of growth medium containing twice the normal amount of serum was added. The medium was replaced 24 h after the start of the transfection procedure, and the cells were incubated for an additional 24 h at 37°C and then harvested for analysis.
Metabolic Labeling-COS-7 cells were transfected with vectors expressing either CCT␣, CCT␤, or a control. The total plasmid amount in each of the transfections was 10 g. At 24 h after transfection, the medium was changed, and the cells were labeled for an additional 24 h with [methyl-3 H]choline (3 Ci/ml). Cells were washed three times with 10 ml of phosphate-buffered saline and harvested in 10 ml of the same buffer, and the cell pellets were extracted using 720 l of chloroform/ methanol/concentrated HCl (1:2:0.02, v/v). Next, 240 l of chloroform and 240 l of water were added, and following vortex mixing, the phases were separated by centrifugation. The radioactivity in the soluble and phospholipid phases was quantitated. Samples of the soluble phase were separated on Silica Gel G thin layers developed with 2% ammonium hydroxide/95% ethanol (1:1, v/v), and the organic phase was analyzed on Silica Gel 60 thin layers developed with chloroform/methanol/ammonium hydroxide (60:35:8, v/v). About 90% of the labeled material in the organic phase was PtdCho under these labeling conditions. Choline-derived metabolites were identified by co-migration with standards.
Northern Blots-Three multiple human tissue Northern blots were purchased from CLONTECH and were hybridized and washed according to the manufacturer's instructions. The blots were first hybridized with 32 P-labeled probe prepared from a 1.3-kb BamHI-XhoI fragment that covered the entire CCT␤ cDNA. The blots were then stripped and hybridized with a 32 P-labeled probe prepared from the 582-bp SacI fragment of the human phosphatidylinositol synthase (pis1) cDNA (33). The blots were stripped again and hybridized with a 32 P-labeled probe prepared from a 320-bp PstI-ApaI fragment representing the 3Ј region of the CCT␤ cDNA. This area of the CCT␤ cDNA did not share any sequences in common with CCT␣.
Immunoblots-Cell lysates (50 g of protein) were separated by SDS-gel electrophoresis on 12% polyacrylamide gels and transferred by electroblotting onto nitrocellulose membranes. Immunoblotting was performed by incubation of the membranes with either purified anti-CCT␣ (1:200 dilution) or purified anti-CCT␤ (1:200 dilution) as the primary antibody. The Amersham Pharmacia Biotech ECL Western blotting reagents and protocol were used to identify the immunoreactive proteins.
Immunofluorescence Microscopy-HeLa cells grown on coverslips were fixed with 3.7% paraformaldehyde, permeabilized with cold acetone, and processed as described (34). Affinity-purified anti-CCT␤ primary antibody was diluted in 0.15 M NaCl, 10 mM Tris-HCl, pH 8.0. The cells were incubated with anti-CCT␤ antibodies at increasing dilutions followed by fluorescein-conjugated secondary antibodies. The coverslips were mounted with p-phenyldiamine, the cells viewed in a Zeiss IM-35 microscope equipped with fluorescence optics, and photographs were made on Kodak Tri-X pan film. Controls from which the primary antibody was excluded showed no significant fluorescence. Preincubation of the primary antibody with the peptide did not yield significant fluorescence. Two other controls assured selective labeling of the nuclear and cytoplasmic compartments. Both a nucleolar marker (anti-p120 antibodies, Becton Dickinson) and a cytoplasmic (cytoskeletal) marker (anti-vimentin antibodies, Boehringer Mannheim) were used to label the cells to confirm appropriate staining of cellular compartments.

RESULTS
Identification of the CCT␤ Clone-A BLAST search of the proprietary human expressed sequence tagged data base of Human Genome Sciences revealed the existence of a cDNA with considerable similarity to mammalian CCTs. However, the protein expressed by this cDNA was not catalytically active, indicating that it did not contain a complete CCT coding sequence. Sequence information from this clone was used to search the public expressed sequence tagged data base. We identified a second clone (GenBank TM accession no. AA382871) isolated from a human testis library that contained a 140-bp related sequence at the 5Ј end. This clone, called CCT␤, was purchased from American Type Culture Collection. Both cDNA strands were sequenced, and the clone contained the entire CCT␤ coding sequence. The cDNA sequence of human CCT␤ was compared with the cDNA sequence of human CCT␣ (see Fig. 2). The analysis of the sequence (see below) indicated that the CCT␤ cDNA encoded a new CCT isoform.
Similarities and Differences between the Predicted Protein Sequences and the cDNAs of CCT␣ and CCT␤-The predicted amino acid sequences of human CCT␣ and CCT␤ are compared in Fig. 1. The catalytic cores of human CCT␣ and CCT␤ are nearly identical and extend from amino acids 72 through 233. The catalytic core in CCT␤ has 64% identity with the equivalent yeast CCT domain that is located between amino acids 99 and 260 of the yeast protein (28). Three of the amino acids in CCT␤ that are different from CCT␣ (N120K, V136L, and R162K) are identical to the yeast CCT sequence. Three other amino acids that are different in CCT␤ compared with CCT␣ (E126D, D134E, and E160K) are identical to the residues found in the catalytic core of the yeast MUQ1 sequence, which has been identified as phosphoethanolamine cytidylyltransferase (36). Also, the catalytic domains of human phosphoethanolamine cytidylyltransferase are highly related to the analogous domains in CCT␣ and CCT␤ (37). These sequence similarities strongly suggested that the CCT␤ cDNA encoded a protein with cytidylyltransferase activity.
The amino-terminal domain of CCT␤ is distinct from the amino-terminal region of human CCT␣ (Fig. 1). CCT␤ lacks the human CCT␣ sequence 8 KVNaRKRRKEaPGPNGATEED 28 , which is postulated to mediate transit of the protein to the nuclear compartment. Amino acids 11 and 18 are designated in lowercase because they differ from the rodent CCT␣ sequence that is known to be the minimal protein sequence that is both necessary and sufficient for localization of rodent CCT␣ to the nucleus (8). The significant difference in the amino-terminal sequences of CCT␣ and CCT␤ and the lack of a discernible nuclear localization motif in CCT␤ suggested that CCT␤ would not be localized in the nucleus. On the other hand, the aminoterminal domain of CCT␤ has some identity to the yeast aminoterminal region (28) at amino acids 38 through 40 (RLT), and at Ala 48 , Thr 51 , and Asn 52 . These sequence similarities indicate that the amino terminus of CCT␤ is more related to the yeast CCT than mammalian CCT␣; however, the significance of this correlation is not obvious because a specific function for the amino terminus of yeast CCT has not been described.
The predicted CCT␤ protein sequence exhibits some significant similarities and distinct differences from the CCT␣ sequence in the amphipathic helical and phosphorylation regulatory domains (Fig. 1). The helical domain of human CCT␤ is highly related to the analogous domain in CCT␣, with 88% amino acid identity between residues 256 and 288 and with conservative substitutions at K259R, Q265N, and K266R. This domain in rodent CCT␣ is required for the lipid-dependent decrease in the CTP K m associated with stimulation of CCT activity (16) and mediates the reversible binding of CCT␣ to phospholipid bilayers (11)(12)(13)(14)(15). A similar amphipathic helical domain is absent in the yeast CCT (28), although this enzyme is activated by lipids (38). The similarity in the helical regions of CCT␣ and CCT␤ suggested that CCT␤ activity would be stimulated by the interaction with lipids in a manner similar to mammalian CCT␣ and that CCT␤ would undergo reversible association with cellular membranes. Lipid stimulation of CCT␣ activity is regulated by phosphorylation of the carboxylterminal domain (23). Sixteen serine residues are located in the CCT␣ domain, which extends from amino acid 315 through 367, and all of these serines are phosphorylated to some extent (22). In contrast, the shorter carboxyl-terminal domain of CCT␤ contains only three possible serine phosphorylation sites (Ser 315 , Ser 319 , and Ser 323 ). Ser 315 and Ser 319 are potential proline-directed phosphorylation sites analogous to Ser 315 and Ser 319 in CCT␣. Thus, the opportunities for the regulation of CCT␤ activity by reversible phosphorylation are more restricted in CCT␤ than in CCT␣.
The cDNA sequences of CCT␣ and CCT␤ are distinct but there are regions that have significant similarity (Fig. 2). The similarities are most pronounced in the catalytic core and amphipathic helix domains that are the most conserved peptide regions between CCT␣ and CCT␤. However, there are many differences in the cDNA sequence in these regions, illustrating that CCT␣ and CCT␤ arise from the transcription of different genes and not by the alternative splicing of a single gene. The 5Ј and 3Ј regions of the cDNAs reflect the lack of similarity between the two proteins in the amino and carboxyl-terminal regions.
Pattern of CCT␤ mRNA Expression-The relative abundance of CCT␤ mRNA expression in a wide variety of human tissues was addressed by Northern blot analysis (Fig. 3). The blots were probed with the human phosphatidylinositol synthase (pis1) cDNA as a loading control. This enzyme in phosphatidylinositol biosynthesis is a "housekeeping" protein that is expressed at a relatively uniform level in human tissues (33). The blots were then probed with both the entire CCT␤ cDNA (Fig.  3) and with a 32 P-labeled fragment from the 3Ј untranslated region of the CCT␤ cDNA. The pattern of expression was the same with both probes (not shown). Two sizes of CCT␤ transcripts were detected. The largest CCT␤ mRNA (ϳ6.5 kb) was most abundant in brain, ovary, testis, and all fetal tissues examined. The second class of CCT␤ mRNAs were found between 1.1 and 1.9 kb. The 1.1-kb mRNA was detected in placenta, which was the most abundant source for CCT␤ mRNA in our survey. Testis also was an abundant source for CCT␤ transcripts, and two mRNAs of 1.6 and 1.9 kb were detected in this tissue. Although it is difficult to see in Fig. 3 due to the very high expression of CCT␤ in placenta and testis, CCT␤ mRNA species of either 1.1 or 1.9 kb were faintly detected in all tissues examined. Thus, CCT␤ mRNA is widely distributed in human tissues and expressed at very high levels in testis and placenta.
Substrate Specificity of CCT␤-The similarity of the catalytic core domains of CCT␤ and CCT␣ suggested that CCT␤ was a phosphocholine cytidylyltransferase. This prediction was confirmed by transfecting COS-7 cells with a CCT␤ expression construct (pPJ34) and measuring the CCT enzymatic activity in cell lysates (Fig. 4A). The introduction of the CCT␤ expression vector into the COS-7 cells led to a significant (8-fold) increase in the CCT specific activity from 4.8 to 38.4 nmol/ min/mg of protein in cell extracts (Fig. 4A). These data establish that the CCT␤ cDNA encodes an active CTP:phosphocholine cytidylyltransferase. Alternative substrates for CCT␤ (phosphoethanolamine, glycerol 3-phosphate, phosphatidic acid, and lysophosphatidic acid) were also screened. Substitution of these compounds for phosphocholine in the biochemical assay did not yield significant activity. Substitution of de-oxyCTP for the CTP in the assay at concentrations up to 10-fold higher also did not yield significant activity.
Effect of CCT␤ Expression on PtdCho Metabolism-COS-7 cells were transfected with the CCT␤ expression plasmid and were labeled with [ 3 H]choline for 24 h to determine whether CCT␤ functions as a CCT in vivo and whether the overexpression of this isozyme effects the PtdCho biosynthetic pathway (Fig. 4B). Although the cellular content of 3 H-labeled choline and phosphocholine were the same in control and CCT␤-transfected cells, there was a 3.4-fold increase in the CDP-choline pool from 1963 Ϯ 13 to 6706 Ϯ 847 cpm/mg. This finding was consistent with the identification of CCT␤ as a phosphocholine cytidylyltransferase and illustrated that overexpression of the enzyme leads to increased accumulation of its product, CDPcholine in vivo. There was also a 50% increase in the incorporation of [ 3 H]choline into PtdCho from 715,228 Ϯ 28,317 to 1.06 ϫ 10 6 Ϯ 15,959 cpm/mg in CCT␤-transfected cells indicating that CCT␤ overexpression accelerated PtdCho synthesis. The 5-fold increase in the amount of glycerophosphocholine, a breakdown product of PtdCho, from 7027 Ϯ 251 to 35,835 Ϯ 7083 cpm/mg in the CCT␤-expressing cells indicated an acceleration of PtdCho turnover similar to that previously observed with CCT␣ overexpression (39).
Lipid Stimulation of CCT␤ Activity-The similarities in the amphipathic helical domains between CCT␣ and CCT␤ suggested that lipids would stimulate CCT␤ activity in a similar manner to CCT␣. The addition of the PtdCho:oleic acid lipid activator mixture to crude cell lysates did not enhance CCT␤ activity (not shown). However, the lack of lipid regulation in cell lysates could be attributed to the presence of endogenous lipid activators. This point was tested by removing the endogenous lipid activators from the transfected COS-7 cell lysates by ion-exchange chromatography (16,23) and determining the ability of PtdCho:oleic acid vesicles to activate the enzyme (Fig.   4C). CCT␤ activity was not detected following the removal of endogenous lipids, and it was potently stimulated by the addition of PtdCho:oleic acid vesicles to the sample. There remain many kinetic details related to the specificity of the lipid regulation of CCT␤ to be investigated with purified CCT␤, and there may be subtle differences between the two proteins because the amphipathic helical domains are not identical. Nonetheless, our experiments establish that CCT␤, like CCT␣, is critically dependent on the presence of stimulatory lipids for activity.
Expression and Modification of CCT␤ Protein-The predicted molecular size of the CCT␤ protein was confirmed by transcription and translation of the CCT␤ cDNA in vitro using a reticulocyte lysate (Fig. 5). The expressed proteins were ra- FIG. 3. Pattern of CCT␤ expression in human tissues. Three multiple human tissue Northern blots were purchased from CLON-TECH and were hybridized and washed according to the manufacturer's instructions. The blots were first hybridized with a 32 P-labeled probe prepared from a 1.3-kb BamHI-XhoI fragment that contained the entire human CCT␤ cDNA. The blots were then stripped and hybridized with a 32 P-labeled probe prepared from the 582-bp SacI fragment of the human phosphatidylinositol synthase (Pis1) cDNA (33).

FIG. 4. CCT␤ expression, and effect of CCT␤ overexpression on PtdCho metabolism in COS-7 cells, and lipid stimulation of CCT␤ activity.
A, CCT specific activity in COS-7 cells transfected with plasmids expressing CCT␤ (pPJ34) (q) or an empty vector control (pcDNA3) (E). The cells were harvested and assayed for CCT activity 48 h after transfections with the indicated expression construct as described under "Experimental Procedures." B, effect of CCT␤ overexpression on PtdCho metabolism. COS-7 cells were transfected with the CCT␤ or the control expression vectors, and 36 h later, the cells were labeled with [ 3 H]choline for 24 h. The cells were harvested, and the distribution of label among the soluble choline-derived metabolites and the amounts incorporated into PtdCho were determined by differential extraction and thin-layer chromatography as described under "Experimental Procedures." C, ability of PtdCho:oleic acid to stimulate the activity of CCT␤. COS-7 cells were transfected with the CCT␤ (pPJ34) expression construct, and 48 h later, the cells were harvested and extracted. Endogenous lipids activators were removed by ion-exchange chromatography, and the ability of added PtdCho:oleic acid (1/1) to stimulate CCT␤ activity was determined as described under "Experimental Procedures." Recovery of CCT␤ activity from the column was Ն90%. diolabeled with [ 35 S]methionine, and the products were separated by SDS-polyacrylamide gel electrophoresis on 12% gels. A protein of an apparent size of 35 kDa was identified in reactions using CCT␤ cDNA as template and was consistent with the predicted size of 36.3 kDa for CCT␤ protein. As a control, CCT␣ was expressed in the transcription/translation system, and the expected 42-kDa protein was detected (Fig. 5). The CCT␤ expression plasmid was transfected into COS-7 cells, and cell lysates were analyzed for expression of CCT␤ protein by immunoblotting with an antibody raised against amino acids 27-39 of the CCT␤ polypeptide sequence (Fig. 5). Two forms of CCT␤ were detected following expression in COS-7 cells. The less abundant species migrated at the same apparent size as the protein made in vitro (CCT␤), and there was a second, slower migrating form (CCT␤ M ). The location of both CCT␤ M and CCT␣ at approximately the same position on the gel was not due to cross-reactivity of the two affinity-purified antibodies. The specificities of the anti-CCT␣ and anti-CCT␤ aminoterminal antibodies were clearly demonstrated in the same experiment (Fig. 5). The larger apparent size of CCT␤ M suggests that a significant portion of the expressed protein is modified posttranslationally.
A second approach was used to verify that a posttranslational modification of the CCT␤ sequence was responsible for the retarded electrophoretic mobility of CCT␤ M . The cDNA sequences corresponding to the dissimilar amino-terminal regions of CCT␣ and CCT␤ were exchanged. Proteins were expressed from the chimeric cDNAs that represented the first 84 amino acids of CCT␤ followed by the remaining 283 amino acids of CCT␣ (CCT␤/CCT␣, Fig. 6) or the first 84 amino acids of CCT␣ followed by the latter 246 amino acids of CCT␤ (CCT␣/ CCT␤, Fig. 6). The anti-CCT␤ antibody recognized both the authentic CCT␤ and CCT␤ M proteins (Fig. 6, lane 8) expressed in COS-7 cells as well as the CCT␤/CCT␣ and CCT␤ M /CCT␣ chimeras (lane 6). The CCT␤/CCT␣ chimera exhibited a larger apparent molecular size than CCT␤ as anticipated due to the longer carboxyl terminus of CCT␣ (lane 6) and also displayed the previously reported gel shifts due to multiple phosphorylation of the CCT␣ carboxyl-terminal domain (40). In contrast, the anti-CCT␣ antibody only detected the gel shifts due to phosphorylation of the authentic CCT␣ protein (lane 4) and did not reveal any modification of the smaller CCT␣/CCT␤ chimeric protein (lane 7). The dissimilarities of the CCT␣ and CCT␤ NH 2 termini, the demonstrated specificities of the antibodies, and the reproduction of the gel shift with in vivo expression of a CCT␤/CCT␣ fusion protein strongly support the idea that CCT␤ is biochemically modified. This modification results in a protein that migrates with approximately the same apparent molecular weight as CCT␣ following denaturing gel electrophoresis.
To further localize the sequences in CCT␤ required for the posttranslational modification, the first 26 amino acids were deleted from the NH 2 terminus. The truncated protein, called CCT␤[⌬1 -26], was expressed in COS-7 cells and detected by immunoblotting with the anti-CCT␤ antibody (Fig. 6). The apparent molecular size of CCT␤[⌬1-26] was smaller than CCT␤, as expected; however, a CCT␤ M [⌬1-26] species was not observed, indicating that posttranslational modification of the truncated protein did not occur (Fig. 6, lane 11). These data suggested that the first 26 amino acids of CCT␤ were necessary for cellular processing of the protein.
Regulation of CCT␤ Activity-CCT␤ activity was examined in vitro to determine whether the NH 2 -terminal sequence and modification of the protein influenced expression and/or catalytic activity. COS-7 cells were transfected with the plasmids encoding CCT␤, CCT␣, the chimeric CCT proteins (CCT␤/  -26]. Cell lysates were prepared after 48 h, samples containing 50 g of protein were separated on 12% SDS gels, and expressed proteins were immunoblotted with either anti-CCT␣ or anti-CCT␤ affinity-purified antibodies as described under "Experimental Procedures." CCT␣ and CCT␣/CCT␤), and the NH 2 -terminal truncated CCT␤[⌬1 -26]. The CCT protein species were expressed at approximately equivalent levels based on the immunoblots (Fig.  6). The CCT protein species in the cell lysates were assayed in the presence of excess stimulatory lipids to evaluate their relative activities. Although cells overexpressing (CCT␤ plus CCT␤ M ) had higher activity (76 nmol/min/mg) than the endogenous control activity (8 nmol/min/mg), [CCT␤ plus CCT␤ M ] was less active than CCT␣ and its multiply phosphorylated species (1105 nmol/min/mg) (Fig. 7). Substitution of the NH 2terminal domain of CCT␣ onto CCT␤ enhanced biochemical activity (215 nmol/min/mg), whereas the amino-terminal domain of CCT␤ dramatically reduced the activity of CCT␣ (91 nmol/min/mg). These data suggested that the protein modification directed by the amino terminus of CCT␤ attenuates biochemical activity. In support of this hypothesis, truncation of the first 26 amino acids of CCT␤ elevated activity almost 10-fold (686 nmol/min/mg), supporting the idea that the NH 2 terminus plays a role in the cellular regulation of CCT␤ activity.
Cellular Localization of CCT␤-The amino terminus of CCT␤ bears little resemblance to the amino terminus of CCT␣, which harbors a nuclear localization sequence. CCT␣ is reported to be predominately an intranuclear protein based on indirect immunofluorescence using an antibody raised against an amino-terminal peptide of CCT␣ (7,8). The lack of a nuclear localization motif in CCT␤ suggested that it would not be found in the cell nucleus. This hypothesis was tested by evaluating the distribution of CCT␤ in human HeLa cells using the anti-CCT␤ peptide antibody and indirect immunofluorescence microscopy (Fig. 8). The affinity-purified peptide antibody detected CCT␤ protein in the cytoplasm and did not detect the protein in the nucleus, even at the lowest antibody dilutions. The cytoplasmic staining appeared diffuse but was higher than the apparent background staining of the nucleus. The background signal associated with the cell nucleus/nucleolus was due to reaction with the fluorescein-conjugated secondary antibodies used in the assay (Fig. 8B). Monoclonal anti-p120 antibody positively identified the nucleolus in these cells (Fig.  8C), and anti-vimentin monoclonal antibody, which signaled the cytoskeleton, was used as a cytoplasmic marker (Fig. 8D). These data confirm that CCT␤ is localized primarily outside of the nucleus. DISCUSSION The existence of a second isoform of CCT opens the door to a series of experiments to determine the physiological function of CCT␤. Our data indicate that the biochemical properties of the two CCT isoforms are similar and the overexpression of either isoform is capable of perturbing PtdCho biosynthesis and metabolism. However, the distinct differences between the amino and carboxyl-terminal domains of CCT␣ and CCT␤ indicate that the two isoforms likely have unique regulatory properties. The two isoforms are clearly the products of different genes and are differentially expressed in tissues and perhaps also during development. Because CCT␣ is the only isoform detected in the large volume of work in this area (1,2), it is possible that CCT␤ is expressed at lower levels or in a specific developmental settings compared with than the more widely distributed CCT␣.
Previous data may have to be reinterpreted in light of the existence of CCT␤. For example, data on whole tissue CCT activity and distribution will need to be reevaluated. The Northern blots cannot be used to predict the relative levels of CCT␣ and CCT␤ proteins in particular tissues. Total CCT specific activity is relatively low in most tissues, illustrating that neither isoform is expressed at high levels. The properties of CCT␤ when expressed in vivo and assayed in vitro suggest that there are no significant differences in biochemical characteristics that could distinguish between the two isozymes in crude extracts. The general impression is that CCT␤ may be more restrictive in its expression compared with CCT␣ because CCT␣ has been the only mammalian isoform cloned to date. Nonetheless, we think that these data will be important to gather because some of the controversial issues in the CCT field may be explained by the presence of two isoforms.

FIG. 8. Immunofluorescence micrographs of HeLa cells labeled with anti-CCT␤ antibodies.
A, cells were incubated with anti-CCT␤ as the primary antibody. B, cells were labeled with fluoresceinconjugated secondary antibodies only. C, cells were labeled with anti-p120 antibodies as a marker for the nucleolus. D, cells were labeled with anti-vimentin antibodies to visualize the cytoskeleton in the cytoplasm. N designates the nucleus, and Nu designates the nucleolus. its nuclear localization (7,8) does not cross-react with CCT␤ (Figs. 5 and 6). Houweling et al. (24) report that CCT is both a nuclear and cytoplasmic protein in primary hepatocytes using a peptide antibody raised against amino acids 164 -176 of CCT␣. This sequence in the 164 -176 region (DFVAHD-DIPYSSA) is identical in human CCT␣ and CCT␤; therefore, antisera raised against this peptide would be predicted to react with both CCT isoforms. We detect both CCT␣ and CCT␤ transcripts in rodent liver, 2 and it will be very interesting to determine whether the cytoplasmic CCT detected in primary rodent hepatocytes can be attributed to the presence of CCT␤ in this tissue.
The co-migration of CCT␣ and CCT␤ M on denaturing gels and their similar reliance on lipid activators for biochemical activity makes it difficult to determine whether CCT␤ is a component of purified CCT preparations from mammalian sources. The regulation of CCT␣ is governed by phosphorylation of its carboxyl terminus at multiple sites, resulting in at least two species that migrate more slowly on denaturing gels (Figs. 5 and 6 and Ref. 40). Phosphorylation interferes with the stimulatory action of lipids on CCT␣ activity (23), and there is some correlation with membrane dissociation of CCT␣ in vivo (20), but the regulation by this mechanism is not an absolute on/off switch (23,35). On the other hand, the activity of CCT␤ M , a protein with approximately the same molecular weight as CCT␣, is lower (Fig. 7) due to posttranslational modification that is dependent on the amino terminus. The nature of the modification of CCT␤ protein and its role in the regulation of PtdCho biosynthesis is currently under investigation.