The binding of phosphatidylcholine to the phosphatidylcholine transfer protein: affinity and role in folding.

Bovine liver phosphatidylcholine transfer protein (PC-TP) has been expressed in Escherichia coli and purified to homogeneity from the cytosol fraction at a yield of 0.45 mg PC-TP per 10 mg total cytosolic protein. In addition, active PC-TP was obtained from inclusion bodies. An essential factor in the activation of PC-TP was phosphatidylcholine (PC) present in the folding buffer. PC-TP from the cytosol contains phosphatidylethanolamine (PE) and phosphatidylglycerol (PG) with a preference for the di-monounsaturated species over the saturated species as determined by fast atom bombardment mass spectrometry (FAB-MS). By incubation with microsomal membranes the endogenous PE and PG were replaced by PC. Relative to the microsomal PC species composition, PC-TP bound preferentially C16:0/C20:4-PC and C16:0/C18:2-PC (twofold enriched) whereas the major microsomal species C18:0/C18:1-PC and C18:0/C18:2-PC were distinctly less bound. PC-TP is structurally homologous to the lipid-binding domain of the steroidogenic acute regulatory protein (Nat. Struct. Biol. 7 (2000) 408). Replacement of Lys(55) present in one of the beta-strands forming the lipid-binding site, with an isoleucine residue yielded an inactive protein. This suggests that Lys(55) be involved in the binding of the PC molecule.

The protein kinase C-dependent phosphorylation of serine 166 is controlled by the phospholipid species bound to the phosphatidylinositol transfer protein alpha.

The charge isomers of bovine brain PI-TPalpha (i.e. PI-TPalphaI containing a phosphatidylinositol (PI) molecule and PI-TPalphaII containing a phosphatidylcholine (PC) molecule) were phosphorylated in vitro by rat brain protein kinase C (PKC) at different rates. From the double-reciprocal plot, it was estimated that the V(max) values for PI-TPalphaI and II were 2.0 and 6.0 nmol/min, respectively; the K(m) values for both charge isomers were about equal, i.e. 0.7 micrometer. Phosphorylation of charge isomers of recombinant mouse PI-TPalpha confirmed that the PC-containing isomer was the better substrate. Phosphoamino acid analysis of in vitro and in vivo (32)P-labeled PI-TPalphas showed that serine was the major site of phosphorylation. Degradation of (32)P-labeled PI-TPalpha by cyanogen bromide followed by high pressure liquid chromatography and sequence analysis yielded one (32)P-labeled peptide (amino acids 104-190). This peptide contained Ser-148, Ser-152, and the consensus PKC phosphorylation site Ser-166. Replacement of Ser-166 with an alanine residue confirmed that indeed this residue was the site of phosphorylation. This mutation completely abolished PI and PC transfer activity. This was also observed when Ser-166 was replaced with Asp, implying that this is a key amino acid residue in regulating the function of PI-TPalpha. Stimulation of NIH3T3 fibroblasts by phorbol ester or platelet-derived growth factor induced the rapid relocalization of PI-TPalpha to perinuclear Golgi structures concomitant with a 2-3-fold increase in lysophosphatidylinositol levels. This relocalization was also observed for Myc-tagged wtPI-TPalpha expressed in NIH3T3 cells. In contrast, the distribution of Myc-tagged PI-TPalpha(S166A) and Myc-tagged PI-TPalpha(S166D) were not affected by phorbol ester, suggesting that phosphorylation of Ser-166 was a prerequisite for the relocalization to the Golgi. A model is proposed in which the PKC-dependent phosphorylation of PI-TPalpha is linked to the degradation of PI.

Rapid replenishment of sphingomyelin in the plasma membrane upon degradation by sphingomyelinase in NIH3T3 cells overexpressing the phosphatidylinositol transfer protein beta.

In order to study the in vivo function of the phosphatidylinositol transfer protein beta (PI-TPbeta), mouse NIH3T3 fibroblasts were transfected with cDNA encoding mouse PI-TPbeta. Two stable cell lines were isolated (SPIbeta2 and SPIbeta8) in which the levels of PI-TPbeta were increased 16- and 11-fold respectively. The doubling time of the SPIbeta cells was about 1.7 times that of the wild-type (wt) cells. Because PI-TPbeta expresses transfer activity towards sphingomyelin (SM) in vitro, the SM metabolism of the overexpressors was investigated. By measuring the incorporation of [methyl-(3)H]choline chloride in SM and phosphatidylcholine (PtdCho), it was shown that the rate of de novo SM and PtdCho synthesis was similar in transfected and wt cells. We also determined the ability of the cells to resynthesize SM from ceramide produced in the plasma membrane by the action of bacterial sphingomyelinase (bSMase). In these experiments the cells were labelled to equilibrium (60 h) with [(3)H]choline. At relatively low bSMase concentrations (50 munits/ml), 50% of [(3)H]SM in wt NIH3T3 cells was degraded, whereas the levels of [(3)H]SM in SPIbeta cells appeared to be unaffected. Since the release of [(3)H]choline phosphate into the medium was comparable for both wt NIH3T3 and SPIbeta cells, these results strongly suggest that breakdown of SM in SPIbeta cells was masked by rapid resynthesis of SM from the ceramide formed. By increasing the bSMase concentrations to 200 munits/ml, a 50% decrease in the level of [(3)H]SM in SPIbeta cells was attained. During a recovery period of 6 h (in the absence of bSMase) the resynthesis of SM was found to be much more pronounced in these SPIbeta cells than in 50% [(3)H]SM-depleted wt NIH3T3 cells. After 6 h of recovery about 50% of the resynthesized SM in the SPIbeta cells was available for a second hydrolysis by bSMase. When monensin was present during the recovery period, the resynthesis of SM in bSMase-treated SPIbeta cells was not affected. However, under these conditions 100% of the resynthesized SM was available for hydrolysis. On the basis of these results we propose that, under conditions where ceramide is formed in the plasma membrane, PI-TPbeta plays an important role in restoring the steady-state levels of SM.