INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Cholinephosphotransferase catalyzes the transfer of phosphocholine from CDP-choline to diacylglycerol (DAG)¹ thus forming phosphatidylcholine (PtdCho) and CMP. As the final step of the Kennedy pathway (1, 2) cholinephosphotransferase identifies both the site of de novo PtdCho synthesis as well as the site from which PtdCho is transported to other organelles within the cell or assembled with proteins and other lipids for export from the cell during the genesis of lung surfactant, bile, and lipoproteins (3-5). Two genes coding for cholinephosphotransferase activities exist in Saccharomyces cerevisiae, CPT1, which encodes a cholinephosphotransferase (6), and EPT1, which codes for a dual specificity choline/ethanolaminephosphotransferase (7). In vitro, Cpt1p and Ept1p contribute equally to measurable cholinephosphotransferase activity (8); however, in vivo metabolic labeling analysis revealed that Cpt1p is responsible for 95% of the PtdCho-synthesizing cholinephosphotransferase activity with Ept1p contributing the final 5% (9). Both CPT1 and EPT1 predict proteins containing seven membrane spanning domains. The integral membrane bound nature of cholinephosphotransferase has prevented its purification from any source and has precluded the use of many standard structure/function approaches for analyzing S. cerevisiae Cpt1p and Ept1p enzymes. However, the uninterrupted similarity in predicted secondary structures and corresponding nucleic acid sequences allowed for the construction of a series of CPT1/EPT1-encoded chimeric enzymes (10, 11). The difference in CDP-alcohol specificity between the parental Cpt1p and Ept1p enzymes was exploited to delineate the CDP-choline binding site. A region encompassing the first soluble loop, residues 79-186 of Cpt1p, was pinpointed. Data base searches for known proteins with homology to the CDP-choline binding region of Cpt1p identified sequences within: S. cerevisiae ethanolaminephosphotransferase (EPT1) (7), phosphatidylinositol (PtdIns) synthase (PIS1) (12), and phosphatidylserine (PtdSer) synthase (PSS1/CHO1) (13); prokaryotic phosphatidylglycerol (PtdGro) phosphate and PtdSer (Gram-positive only) synthases (14, 15); soybean aminoalcoholphosphotransferase (16); and rat PtdIns synthase (17). Each of these enzymes catalyzes the synthesis of a phospholipid by the displacement of CMP from a CDP-alcohol by a second alcohol to form a phosphoester bond. Alignment of the sequences within each of the above enzymes revealed a completely conserved motif, Asp-Gly-(X)₂-Ala-Arg-(X)₈-Gly-(X)₃-Asp-(X)₃-Asp, termed the CDP-alcohol phosphotransferase motif. Data base searches revealed this motif was specific to the above enzymes and hence is predicted to be diagnostic for the reaction type catalyzed by each. To lend credence to this prediction, and to provide insight into catalytic mechanism, scanning alanine mutagenesis of the conserved residues within the CDP-alcohol phosphotransferase motif of S. cerevisiae cholinephosphotransferase (CPT1) was performed.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- [-³²P]dATP (3000 Ci/mmol) and [methyl-¹⁴C]choline were products of NEN Life Science Products. [methyl-¹⁴C]CDP-choline was purchased from American Radiolabeled Chemicals. Custom oligonucleotides were purchased from Life Technologies, Inc. Dideoxy sequencing was performed utilizing the T7 sequencing kit (Amersham Pharmacia Biotech). Lipids were purchased from Avanti Polar Lipids. Triton X-100 was purchased from Pierce. Hemagglutinin (HA) monoclonal antibody (mAb) was purchased from Boehringer Mannheim. All other reagents were of the highest quality commercially available.

Site-directed Mutagenesis-- Plasmid pRH150 (6) contains the CPT1 gene in the high copy plasmid YEp352 (18) and was used as the template for all mutagenesis reactions. The T4 DNA polymerase-directed MORPH plasmid DNA mutagenesis protocol (5'  3') was used with the appropriate mutagenic oligonucleotides (Table I) as directed by the manufacturer. All mutations were confirmed by DNA sequencing. Plasmids were routinely propagated in DH5 Escherichia coli (endA1 recA1 hsdR17 (r_kmk⁺) supE44 thi-1 gyrA(NaI^r) relA1 deoR (80lacZM15) (lacZYA-argF)U169.

Enzyme Assays-- S. cerevisiae strain HJ091 (a ura3-52 his3-1 leu2-3, 112 trp1-289 cpt1::LEU2 ept1) was utilized in all studies. HJ091 is devoid of detectable cholinephosphotransferase activity (10, 11). Microsomal membranes were prepared from HJ091 grown at 30 °C to mid-log phase in synthetic dextrose media containing the appropriate nutritional supplements to ensure plasmid maintenance (19). Cholinephosphotransferase activity was assessed using 20 mM MgCl₂ as cation cofactor, 10 mol % PtdCho (egg) as phospholipid cofactor, and 10 mol % dipalmitoleoylglycerol (di16:1DAG) and 500 µM [¹⁴C]CDP-choline (500-4000 dpm/nmol as dictated by the sensitivity required) as substrates in a Triton X-100 mixed micelle assay as described (9).

PtdCho Biosynthesis-- Yeast strain HJ091 ± parental and mutagenized plasmids were grown to mid-log phase in synthetic dextrose media containing the appropriate nutritional supplements. [¹⁴C]Choline (20 µM, 1 × 10⁵ dpm/nmol) was added to the cultures for 30 min. Yeast cells were concentrated by centrifugation, washed with ice-cold water, and resuspended in 1 ml of CHCl₃/CH₃OH (1/1, v/v). Cells were disrupted using a BioSpec multibead beater containing 0.5 g of 0.5-mm acid washed glass beads on the homogenize setting for 1 min at 4 °C. The beads were washed with CHCl₃/CH₃OH (2/1, v/v). A sample was removed from the total homogenate for determination of total choline uptake. To facilitate phase separation, water and CHCl₃ were added to the remaining homogenate. Phospholipids in the organic phase were analyzed by thin layer chromatography on Whatman silica gel 60A plates using the solvent system CHCl₃/CH₃OH/NH₄OH/H₂0 (70/30/4/2, v/v). PtdCho was the only radiolabeled lipid detected. Aqueous metabolites were concentrated under vacuum, resuspended in H₂O, and separated by thin layer chromatography on Whatman silica gel 60A plates in a solvent system consisting of CH₃OH/0.6% NaCl/NH₄OH (50/50/5, v/v). Choline, phosphocholine, and CDP-choline were the only radioactive metabolites detected. Radiolabel was detected using a Bioscan System 200 imaging scanner and the radioactive bands were scraped into vials and subjected to scintillation counting.

Immunodetection-- The CPT1 gene and site-directed mutants were subcloned from YEp352 (URA3, 2 µ ori] to pRS426 [URA3, 2 µ ori] to facilitate the elimination of a HindIII site within the multicloning site. A 3× repeat of the HA epitope was amplified by polymerase chain reaction from the plasmid pGTEP (gift from Stephen Garrett, Duke University Medical Center) with HindIII sites added to the ends of each primer. The amplified 3× HA epitope was subcloned into a unique HindIII site within the CPT1 coding region corresponding to amino acids Tyr²⁸ and Leu²⁹. The insert was sequenced to ensure polymerase frame and fidelity. HJ091 cells were transformed with the constructs and microsomal membranes were isolated. Identical (10 µg) amounts of microsomal protein were incubated with equal volumes of 2 × SDS sample buffer (50 mM Tris-HCl (pH 6.8), 10% glycerol (v/v), 2% SDS (w/v), 5% 2-mercaptoethanol (v/v), 0.05% bromphenol blue (w/v)) at 37 °C for 20 min. Proteins were separated by SDS-PAGE (4% stacking gel, 12.5% resolving gel), and transferred to polyvinylidene fluoride membrane utilizing the Bio-Rad minigel transfer apparatus at 30 volts for 18 h in 48 mM Tris, 30 mM glycine, 0.1% SDS, 20% methanol (v/v) transfer buffer. The membrane was: blocked with phosphate-buffered saline (PBS) containing 5% skim milk powder and 0.05% Tween 20 for 1 h; washed three times with PBS, 0.05% Tween 20; incubated for 1 h in PBS, 0.05% Tween 20 with HA mAb (1:2,000); washed three times with PBS, 0.05% Tween 20; incubated for 1 h in PBS, 0.05% Tween 20 with goat anti-mouse Ab coupled to horseradish peroxidase (1:10,000); washed three times with PBS, 0.05% Tween 20; and signal was detected using the ECL (Amersham) method as per the manufacturer's instructions.

Protein and Lipid Determinations-- Protein was determined by the method of Lowry et al. (20) using bovine serum albumin as standard. DAG was prepared from PtdCho by phospholipase C digestion, and yield was estimated using the method of Stern and Shapiro (21). Phospholipid phosphorus was determined by the method of Ames and Dubin (22).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Enzyme Activity of the Cpt1p CDP-alcohol Phosphotransferase Mutants-- Scanning alanine site-directed mutagenesis of the CDP-alcohol phosphotransferase motif (Asp¹¹³-Gly¹¹⁴-(X)₂-Ala¹¹⁷-Arg¹¹⁸-(X)₈-Gly¹²⁷-(X)₃-Asp¹³¹-(X)₃-Asp¹³⁵, the one Ala residue was converted to Gly) of S. cerevisiae Cpt1p was performed to lend credence to the prediction that this motif is diagnostic for the reaction type catalyzed by this class of enzymes, which also includes: S. cerevisiae ethanolaminephosphotransferase (EPT1) (7), PtdIns synthase (PIS1) (12), and PtdSer synthase (PSS1/CHO1) (13), prokaryotic PtdSer synthase and PtdGro phosphate synthase (14, 15), soybean aminoalcoholphosphotransferase (16), and rat PtdIns synthase (17) (Fig. 1). The CDP-alcohol phosphotransferase motif of Cpt1p was chosen as the target motif for several reasons: (i) a well established mixed micelle assay exists enabling kinetic analysis of enzyme activity (8, 10, 23); (ii) Cpt1p is a well characterized member of the known CDP-alcohol phosphotransferase enzymes and coupled with S. cerevisiae mutants devoid of cholinephosphotransferase activity result in an effective and established expression system (8, 9, 24); (iii) the metabolic fate of radiolabeled choline for PtdCho synthesis has been rigorously characterized in S. cerevisiae with defects in cholinephosphotransferase activity allowing for in vivo corroboration of in vitro results (9, 24). The CPT1 gene was carried on the high copy plasmid YEp352 and was used for all mutagenesis and expression experiments. CPT1 is utilizing its own promoter in all analyses. Mutagenesis of each Ala residue was represented by the preferred GCT codon to ensure against codon bias during mRNA translation (25).

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Known enzymes possessing a CDP-alcohol phosphotransferase motif. Footnote a, other prokaryotic homologues catalyzing the same reaction also contain the motif.

In Vivo Analysis of the Cpt1p CDP-alcohol Phosphotransferase Mutants-- To corroborate the above in vitro assessment of the ability of the various CDP-alcohol phosphotransferase motif site-directed mutants to confer cholinephosphotransferase activity, each mutant was expressed in S. cerevisiae cells devoid of cholinephosphotransferase and their capacity to incorporate radiolabeled choline into PtdCho was determined (Fig. 2). Since the rate-limiting step for PtdCho synthesis is normally at the level of phosphocholine cytidylyltransferase, alterations in cholinephosphotransferase activity do not generally affect the ability of cells to incorporate radiolabeled choline into PtdCho (9, 24, 26, 27). In addition, the metabolic labeling protocol is much more sensitive than the in vitro enzyme assay and does not rely on the ability of the various mutant enzymes to survive cell disruption and subcellular fractionation and hence provides a second level of confidence for determining if mutants with undetectable cholinephosphotransferase activity in vitro are indeed devoid of enzyme activity. The enzymatically active Asp¹¹³ to Ala, Ala¹¹⁷ to Gly, and Arg¹¹⁸ to Ala mutants incorporated labeled choline into PtdCho at a level similar to that of cells carrying the parental Cpt1p protein. The Gly¹¹⁴ to Ala, Gly¹²⁷ to Ala, Asp¹³¹ to Ala, and Asp¹³⁵ to Ala mutants, each of which was inactive in vitro, were unable to incorporate radiolabeled choline into PtdCho (Fig. 2B). In agreement with a metabolic block at the level of cholinephosphotransferase in the S. cerevisiae cells used in this study (cpt1::LEU2 ept1), this yeast strain (and each of the inactive Cpt1p mutants) accumulated CDP-choline (Fig. 2C), while the wild type cells (and each of the mutants with detectable cholinephosphotransferase activity in vitro) did not accumulate CDP-choline due to its successful conversion to PtdCho (Fig. 2, B and C).

View larger version (47K): [in this window] [in a new window]   Fig. 2.   Radiolabeled choline uptake and incorporation into the metabolites of the CDP-choline pathway. Exponentially growing S. cerevisiae cells (cpt1::LEU2 ept1) containing CPT1 CDP-alcohol phosphotransferase motif site-directed mutants were radiolabeled with 20 µM choline for 30 min as described under "Experimental Procedures." A, the Kennedy pathway for PtdCho synthesis; B, total uptake of radiolabel and its incorporation into PtdCho; C, the accumulation of radiolabeled choline within the metabolites of the Kennedy pathway. Results are the mean of two experiments performed in duplicate; variation was less than 10% between experiments.

Kinetic Analysis of the Cpt1p CDP-alcohol Phosphotransferase Mutants-- As a first step in determining catalytic mechanism for cholinephosphotransferase, a kinetic analysis of each of the enzymatically active mutants was performed and compared with the parental enzyme. The mixed micelle assay employed revealed saturation kinetics for both substrates when the other was in excess (data not shown) implying surface dilution of the lipophilic substrate DAG occurs within the mixed micelle and hence the kinetic parameters reported here are reflective of true values. Assay buffer contained 500 µM CDP-choline and 10 mol % di16:1 DAG, and under these conditions wild type Cpt1p displayed an apparent K_m of 66 µM for CDP-choline and 1.75 mol % for DAG (Tables III and IV). Previous measurements of these same parameters for Cpt1p revealed apparent K_m values of 110 µM and 8.0 mol % for CDP-choline and DAG, respectively (8). However, the values obtained in the current study are predicted to be more accurate, since the DAG substrate used in the former study (8) was di18:1 DAG, and it has since been discovered that di16:1 DAG is the preferred substrate for Cpt1p (10). In addition, di18:1 DAG is insoluble at concentrations above 10 mol % in the mixed micelle assay employed in both studies and hence was previously used at suboptimal levels for the accurate measurement of kinetic parameters. Di16:1 DAG does not present this problem as it is soluble up to 20 mol %. The measured apparent K_m value of 1.75 mol % for di16:1 DAG is 5.7-fold lower than the concentration of di16:1 DAG (10 mol %) utilized in this study.

1

1

1

1

1

1

1

1

1

1

Immunodetection of Parental and Mutant Cholinephosphotransferases-- In this study, parental and mutant CPT1 genes were expressed from high copy (URA3, 2 µ ori) plasmids using the endogenous CPT1 promoter. Western blot analysis was performed to ensure that the various mutations created within Cpt1p (with activities significantly different from wild type) were not affecting Cpt1p levels. A 3× repeat of the HA epitope was inserted into the coding region of the parental and site-directed mutant genes at a site between amino acids Tyr²⁸ and Leu²⁹. Plasmids were transformed into HJ091 cells and microsomal membrane proteins were subjected to SDS-PAGE and Western blot analysis using mAb specific for the HA epitope (Fig. 3). The HA mAb recognized a protein of the expected size for Cpt1p+3× HA epitope (49 kDa) that was absent in cells expressing Cpt1p that did not contain the 3× HA epitope, demonstrating that the mAb was specific for epitope tagged Cpt1p proteins (the increased size of Ala¹¹⁷ is a function of the cloning strategy utilized that resulted in a 6× insertion of the HA epitope). Parental and mutant Cpt1p proteins Ala¹¹⁷ to Gly, Arg¹¹⁸ to Ala, Gly¹²⁷ to Ala, Asp¹³¹ to Ala, and Asp¹³⁵ to Ala were present in similar amounts (Fig. 3) confirming that the decreased cholinephosphotransferase enzyme activity associated with the amino acid substitutions of these residues was due to altered kinetic properties of the mutant proteins and not due to decreased protein levels. No Cpt1p protein was detected for the Gly¹¹⁴ to Ala mutant, indicating that the absence of cholinephosphotransferase activity associated with this amino acid substitution was due to increased protein lability.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Immunodetection of parental and site-directed mutants of Cpt1p. Microsomal membranes were isolated from HJ091 cells expressing parental and mutant forms of 3× HA epitope-tagged Cpt1p. Equivalent amounts of microsomal protein were subjected to SDS-PAGE and analyzed by Western blot utilizing mAb specific for the HA epitope as described under "Experimental Procedures."

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Scanning alanine site-directed mutagenesis of the conserved amino acid residues within the CDP-alcohol phosphotransferase motif (Asp¹¹³-Gly¹¹⁴-(X)₂-Ala¹¹⁷-Arg¹¹⁸-(X)₈-Gly¹²⁷-(X)₃-Asp¹³¹-(X)₃-Asp¹³⁵) of S. cerevisiae cholinephosphotransferase was performed to lend credence to the prediction that this motif is diagnostic for the reaction catalyzed and to provide the first molecular investigation of the catalytic mechanism used by this class of enzymes which to date include: S. cerevisiae cholinephosphotransferase, ethanolaminephosphotransferase, PtdSer synthase, PtdIns synthase, prokaryotic PtdSer synthases, PtdGro phosphate synthases, and rat PtdIns synthase. Each of these enzymes catalyzes the formation of a phospholipid via the displacement of CMP from a CDP-alcohol by a second alcohol with concomitant formation of a phosphoester bond. The ability of each mutant to catalyze the cholinephosphotransferase reaction was assessed in vitro using a mixed micelle enzyme assay and in vivo by determining the capacity of each mutant to incorporate radiolabeled choline into PtdCho in a S. cerevisiae strain containing null mutations in its cholinephosphotransferase genes.

One mutant, Asp¹¹³ to Ala, displayed wild type characteristics both in vitro and in vivo. This is intriguing in that this, the first Asp residue within the CDP-alcohol phosphotransferase motif, is completely conserved over a wide range of species and hence has not drifted evolutionarily implying it is essential; however, its function is not apparent from the above analyses. Two mutants, A¹¹⁷ to G and R¹¹⁸ to A, displayed reduced in vitro catalytic activity due to 5.0-10.4-fold decreased V_max(app) and 2.5-3.8-fold increased K_m(app) values for both CDP-choline and DAG. These results imply Ala¹¹⁷ and Arg¹¹⁸ play a role in substrate binding or positioning with Arg¹¹⁸ predicted to be coordinated with one of the phosphate groups within the CDP-alcohol. Interestingly, in vivo the Ala¹¹⁷ to Gly and Arg¹¹⁸ to Ala mutants incorporated radiolabeled choline into PtdCho in cholinephosphotransferase null cells at levels similar to that of the parental enzyme. These data support two further conclusions: (i) the concentration of both CDP-choline and DAG at the intracellular site of cholinephosphotransferase must be sufficient to overcome the increase in K_m(app) demonstrated by the Ala¹¹⁷ to Gly and Arg¹¹⁸ to Ala mutants, and (ii) cholinephosphotransferase is not rate-limiting, since large decreases in V_max(app) did not affect the level with which labeled choline was incorporated into PtdCho. The latter conclusion is consistent with many other metabolic studies (24, 26). Mutations at Gly¹¹⁴, Gly¹²⁷, Asp¹³¹, and Asp¹³⁵, completely ablated detectable activity both in vitro and in vivo. The absence of cholinephosphotransferase activity in cells expressing Cpt1p carrying a Gly¹¹⁴ to Ala substitution correlated with an absence of detectable protein, indicating this residue is required for protein stability or folding. Since Gly residues do not possess a functional group, the lack of activity in the Gly¹²⁷ to Ala mutant suggests a steric role within the active site. The elimination of cholinephosphotransferase activity by mutating either Asp¹³¹ or Asp¹³⁵ (the last two residues within the motif) to Ala implies one of these is the catalytic residue.

Several members of the CDP-alcohol phosphotransferase motif family of enzymes have been subjected to kinetic analyses; pure enzyme preparations of S. cerevisiae PtdSer synthase and E. coli PtdGro phosphate synthase, as well as microsomal mammalian cholinephosphotransferase, predicted sequential bi-bi reaction mechanisms (28-31). An in depth kinetic analysis of purified E. coli PtdGro phosphate synthase was consistent with a reaction mechanism that did not utilize an enzyme bound intermediate. A general acid-catalyzed reaction utilizing an Asp residue at its catalytic center would favor an enzyme bound intermediate, while a general base reaction would not, hence, CDP-alcohol phosphotransferases most likely utilize general base catalysis via nucleophilic attack of the hydroxyl group of the free alcohol directly on the phosphoester bond of the CDP-alcohol through one of the final two Asp residues within the motif without passing through an enzyme-bound intermediate.

From the above results it is clear that the CDP-alcohol phosphotransferase motif, Asp-Gly-(X)_2-Ala-Arg-(X)_8-Gly-(X)₃-Asp(X)₃-Asp, is diagnostic for the reaction catalyzed. However, it should be noted that this motif is not essential for this reaction type. E. coli PtdSer synthase catalyzes the formation of PtdSer via formation of a phosphoester bond utilizing CDP-DAG and serine as substrates with subsequent release of CMP in a manner similar to that of the Bacillus subtilus and S. cerevisiae PtdSer synthases (31-33). The latter two possess the CDP-alcohol phosphotransferase motif while the E. coli enzyme possesses a separate motif, HxK(U)₄D(X)₄UUGO, that also appears capable of the formation of the identical phosphoester bond (34, 35). This second motif is also found in prokaryotic cardiolipin synthase, as well as enzymes that catalyze the hydrolysis of a phosphoester bond including phospholipase D and some nucleases. Studies of E. coli and S. cerevisiae PtdSer synthases provide two further lines of evidence that are consistent with two distinct motifs catalyzing the same reaction via different mechanisms: (i) kinetic analysis of E. coli PtdSer synthase predicted a ping-pong reaction type that utilized an enzyme-bound intermediate (31-33), a mechanism distinct from the sequential bi-bi reaction of CDP-alcohol phosphotransferase motif enzymes (28-31), and (ii) an isotopic exchange NMR analysis comparing the PtdSer synthase activities from E. coli and S. cerevisiae observed a retention of chirality of the -phosphorus within the CDP-diacylglycerol substrate for the E. coli enzyme (consistent with a reaction mechanism proceeding via an enzyme-bound intermediate), while the S. cerevisiae enzyme displayed an inversion of chirality of the -phosphorus (implying a single displacement mechanism) (31).

This study is the first molecular characterization of a CDP-alcohol phosphotransferase motif and results indicate the motif plays an intimate role in catalysis. Several lines of evidence point to this conclusion: (i) the CDP-alcohol phosphotransferase motif is completely conserved in enzymes that catalyze the same reaction type; (ii) the conservation is observed across a wide range evolutionary range implying the preserved residues are essential for enzyme function; (iii) FASTAPAT and Motif searches of nonredundant data bases revealed only the above enzymes indicating its specificity; (iv) mutations within specific residues of the S. cerevisiae cholinephosphotransferase CDP-alcohol phosphotransferase motif abolish or reduce activity. The integral membrane-bound nature of all of the members of the CDP-alcohol phosphotransferase motif enzymes has precluded their purification from most sources, including any mammalian cell type, thus many of their respective cDNAs have yet to be isolated. The verification of a conserved motif diagnostic for the reaction type catalyzed by each of these enzymes will allow for the rapid identification of cDNAs coding for these proteins from expressed sequence tag data bases. These new molecular tools will allow for a precise dissection of the many biological and pathophysiological roles currently postulated for each of these enzymes.