Sphingosine kinase 1 expression is downregulated during differentiation of Friend cells due to decreased c-MYB.

Sphingosine kinase 1 (SPHK1) overexpression in malignant cells has been reported. Mouse Friend cells showed higher SPHK1 but not SPHK2 expression compared with other mouse cell lines. A Sphk1 promoter analysis demonstrated the region between -53bp and the first exon as the minimal promoter. Further promoter truncation revealed the importance of a MYB-binding site. EMSA using this region as the probe demonstrated one band containing c-MYB protein, and its intensity decreased during erythroid differentiation with hexamethylane bisacetamide (HMBA), a potent inducer of erythroid differentiation of Friend cells. ChIP assay also revealed in vivo binding of c-MYB. c-MYB overexpression and siRNA for c-Myb affected SPHK1 expression, confirming the important regulatory role of c-MYB in SPHK1 expression. HMBA reduced c-MYB expression rapidly. Induced differentiation by HMBA caused a marked and rapid reduction of SPHK1 mRNA, protein and enzyme activity leading to the rapid decrease of cellular sphingosine 1-phosphate level. Moreover, terminally differentiated cells did not resume SPHK1 expression. Compared with original Friend cells, stable overexpression of wild-type SPHK1 showed higher cell proliferation, resistance to cell death by serum depletion. Interestingly, HMBA-induced differentiation of these cells was delayed but not completely suppressed. In contrast, SPHK inhibitor and its siRNA inhibited cell growth and enhanced HMBA-induced differentiation significantly, suggesting that SPHK1 delayed HMBA-induced differentiation by its cell proliferation-promoting activity. Effects of pertussis toxin, a G-protein-coupled receptor inhibitor, and S1P receptor antagonist on Friend cell growth and differentiation were negligible, suggesting the importance of the intracellular SPHK1/S1P signaling in Friend cells.

Transcription factor specificity protein 1 (Sp1) is the main regulator of nerve growth factor-induced sphingosine kinase 1 gene expression of the rat pheochromocytoma cell line, PC12.

Sphingosine kinase (SPHK) is known to exert an anti-apoptic role in various cells and cell lines. We previously reported that human brain is rich in SPHK1 (Murate et al. 2001). After showing a high expression of SPHK1 in rat brain, we examined the gene expression mechanism using nerve growth factor (NGF)-stimulated rat PC12 cells. With RT-PCR, we found that both rat brain and PC12 utilized exon 1d mostly out of eight untranslated first exons. NGF induced an increase in SPHK enzyme activity and protein about double those in PC12 cells, and NGF-induced SPHK1 mRNA was three times higher than in the control. The minimal 5' promoter was determined, and TrkA specific inhibitor K252a inhibited the NGF-induced promoter activity of SPHK1. The truncation or mutation of putative transcription factor-binding motifs revealed that one specificity protein 1 (Sp1) binding motif of the 5' region of exon 1d is prerequisite. Electrophoresis mobility shift assay confirmed the promoter analysis, indicating increased Sp1 protein binding to this motif after NGF treatment. Chromatin immunoprecipitation assay also showed the binding of Sp1 and the promoter region in vivo. These results suggest the signal transduction pathway from NGF receptor TrkA to transcription factor Sp1 protein binding to the promoter Sp1-like motif in NGF-induced rat SPHK1 gene expression.

Nuclear localization of neutral sphingomyelinase 1: biochemical and immunocytochemical analyses.

To examine the intracellular localization of neutral sphingomyelinase 1 (nSMase 1), a rabbit polyclonal antibody was raised against a recombinant form of the enzyme expressed in E. coli. It has been reported that, in rat liver or in ascites hepatoma AH7974, high activity of neutral sphingomyelinase (SMase) is found at the plasma membrane, with a lesser but significant amount in nucleus and cytoplasm. The biochemical properties, dithiothreitol requirement and high salt concentration dependency, of cloned and expressed nSMase 1 resemble those of previously described nuclear neutral SMase of AH7974. The present study was therefore focused on the nuclear localization of this enzyme. Western blotting of subcellular fractions using anti-rat nSMase 1 antibody revealed most nSMase 1 to be associated with the nuclei and some with microsomes, but not with plasma membranes. Consistently, neutral SMase activity in nuclear extract was immunoprecipitated by the antibody, while that of plasma membranes was not. The results indicate that nSMase 1 mainly resides in the nucleus and may thus differ from neutral SMase in plasma membrane. On gel-filtration column chromatography of nuclear extract, the profile of neutral SMase activity corresponded well with immunoreactive protein bands on western blotting, suggesting that a large part of nuclear neutral SMase may be nSMase 1. Removal of the nuclear envelope by treatment with Triton X-100 did not significantly decrease the amount of nuclear nSMase 1, and western blotting of subnuclear fractions (i.e. nuclear envelope, chromatin, and nuclear matrix) revealed nSMase 1 signal exclusively in the nuclear matrix. Immunocytochemistry with AH7974, as well as rat fibroblast cell line 3Y1, demonstrated nSMase 1 to be localized mainly in the nucleus, with some in the cytoplasm. Moreover, immuno-electron microscopy clearly showed the signal of nSMase 1 to be more dense in the nucleus than in the cytoplasm of AH7974.

Structural requirements of sphingosine molecules for inhibition of DNA primase: biochemical and computational analyses.

Using 28 chemically well-defined compounds containing D-erythro-sphingosine and its analogues, we analyzed structure-activity relationships for DNA primase inhibition. Biochemical studies demonstrated a positively charged amino group at C2 and a long aliphatic chain to be absolutely required for inhibition. Whereas C2-amino group is intact, sphingosine 1-phosphate was totally inactive. This result could be due to cancellation of positive charge of the amino group by the interaction with negatively charged C1-phosphate, since simulations with the software INSIGHT II showed these two groups to be close enough to interact. The hydroxyl group at C3 and trans-double bond at C4-C5 were also found to be important for the inhibition. Dehydroxylation of C3, as well as saturation or cis-conversion of the trans-double bond led to decrease of inhibitory activity. Despite saturation of the double bond, introduction of a hydroxyl group into C4 of dihydrosphingosine resulted in restoration of inhibition. Conversion of the double bond into a triple bond did not abolish but rather enhanced the inhibitory activity. Among sphingosine stereoisomers, the naturally occurring D-erythro-sphingosine proved to be the strongest inhibitor. To ascertain the contribution of the total conformation to the inhibition, especially of the long aliphatic chain, we constructed a 3D-quantitative structure-activity relationship model using the computer program Catalyst/HipHop on the basis of information described above. Analysis of the hypothesis model for active compounds revealed that the orientation of aliphatic chain, represented by the dihedral angle of C2-3-4-5, correlated well with the inhibition. Modifications such as deletion of the hydroxyl group at C3 or saturation of the C4-C5 double bond caused shifts in the dihedral angle of C2-3-4-5, with concomitant decrease in inhibitory activity.

Cloning and expression of rat neutral sphingomyelinase: enzymological characterization and identification of essential histidine residues.

Using cross-species sequence homology, we cloned a cDNA for rat neutral sphingomyelinase (nSMase) composed of 422 amino acids that shares 87.6 and 79.0% identity with the mouse and human forms respectively. The rat nSMase expressed in Escherichia coli catalyzed sphingomyelin hydrolysis at neutral pH in a Mg(2+)-dependent manner, and required Triton X-100, dithiothreitol, and KCl for its full activity. The cloned rat enzyme shares conserved sequences with nSMases from both eukaryotes and prokaryotes. Introduction of single mutations into either of the histidine residues at positions 136 and 272, putative active sites, entirely abolished the activity, supporting a common mechanism for the nSMase family independent of the species. However, mutation in histidine 151, conserved only in eukaryotes, also abolished the activity, suggesting eukaryote-specific control of nSMase linked to this histidine 151. This enzyme also catalyzed the hydrolysis of lyso-platelet activating factor to yield 1-alkylglycerol at a rate that is slightly lower than that with sphingomyelin.

A possible role of nuclear ceramide and sphingosine in hepatocyte apoptosis in rat liver.

BACKGROUND/AIMS: Portal vein branch ligation induces apoptosis of hepatocytes in the ligated lobes in rat liver. Sphingomyelin degradation was studied during the process to evaluate its possible involvement in apoptosis in vivo. METHODS: DNA scissions were detected by the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) and an agarose gel electrophoresis of DNA. Using both ligated and non-ligated lobes, we measured activities of sphingomyelin degradation enzymes and contents of their products in purified nuclei and plasma membrane. RESULTS: DNA fragmentation was detectable in the ligated lobes at 90 min after the portal vein branch ligation by gel electrophoresis. At 15 h after the ligation, 27% of hepatocytes became TUNEL-positive. Prior to the onset of apoptosis, the activity of neutral sphingomyelinase increased in the nuclei of hepatocytes in ligated lobes (30 min after the ligation). The increase in sphingomyelinase paralleled its reaction product, ceramide. This was followed by the elevation of ceramidase activity in nuclei (60 min after the ligation) in association with an increase of its reaction product, sphingosine. Activities of these two enzymes and their products increased for at least 90 min. These changes were not observed in nuclei of the non-ligated lobes, or in the plasma membranes from either ligated or non-ligated lobes. CONCLUSIONS: These results, specific to the liver where apoptosis is being generated, suggest that nuclear sphingomyelin breakdown with an accumulation of ceramide and/or sphingosine in nuclei may induce the apoptosis of hepatocytes in vivo.

Analysis of bax protein in sphingosine-induced apoptosis in the human leukemic cell line TF1 and its bcl-2 transfectants.

Sphingosine, a sphingolipid breakdown product, has been proposed as an apoptosis-inducing agent. In this study, we examined the effect of sphingosine in bcl-2-overexpressing cells compared with cells that do not express the bcl-2 gene. The human erythroleukemic cell line TF1, which lacks bcl-2 expression, was easily induced to undergo apoptotic cell death by a variety of stimuli, including depletion of granulocyte-macrophage colony-stimulating factor (GM-CSF) or exposure to methylmethane sulfonate (MMS) (100 microg/mL), ultraviolet light (15 J/m2), X-ray irradiation (20 Gy), or sphingosine, a sphingolipid breakdown product (5 microM). In contrast, bcl-2 transfectants of TF1 (TF1-bcl2), which we established, were resistant to most of these treatments but remained sensitive to sphingosine. Neither C2- nor C6-ceramide (short-chain ceramide) induced apoptosis in TF1-mock and TF1-bcl2 cells. Sphingosine-induced apoptosis could not be inhibited by fumonisin B1, which can prevent conversion of sphingosine to ceramide, suggesting that sphingosine itself, not ceramide, possesses apoptosis-inducing capability. Western blotting, which revealed a 21-kDa bax protein in untreated cells, revealed the presence of an additional 18-kDa protein in GM-CSF-depleted and MMS- or sphingosine-treated TF1-mock cells. In TF1-bcl2 cells, this protein was not detected after GM-CSF depletion or MMS treatment, but was observed after sphingosine treatment. Immunoprecipitation with anti-bcl2 antibody, followed by immunoblotting with anti-bax antibody, showed that both the 21-kDa bax protein and the 18-kDa protein heterodimerized with bcl-2 protein. These results suggest that sphingosine is a unique reagent for apoptosis and that it can overcome bcl-2 gene expression. Furthermore, induction of 18-kDa bax-related protein may play an important role in apoptosis. Sphingosine, but not ceramide, may prove applicable as a reagent for future cytotoxic drugs used to treat intractable tumors overexpressing bcl-2.

Inhibition of DNA primase by sphingosine and its analogues parallels with their growth suppression of cultured human leukemic cells.

Sphingosine is a potent inhibitor of a mammalian DNA primase in vitro (Simbulan et al., Biochemistry 33, 9007-9012, 1994). Here we measured the inhibition of DNA primase in vitro by 9 sphingosine-analogues with respect to RNA primer synthesis and DNA primase-dependent DNA synthesis, and their potencies of inhibition in vitro were compared with their in vivo effects on human leukemic cells. Sphingosine, phytosphingosine and N, N-dimethylsphingosine strongly inhibited the activity of purified calf thymus DNA primase, and also inhibited the growth of human leukemic cell line HL-60, exerting strong cytotoxicity. Dihydrosphingosine and cis-sphingosine, which showed more subtle inhibition of DNA primase in vitro, moderately inhibited the cell growth in vivo and caused cell death. In contrast, N-acyl-, N-octyl-, and N-acetylsphingosine (ceramides) showing little inhibition of DNA primase suppressed cell growth only slightly. HL 60 cell was arrested at Go/G1 phase by exogenously added sphingosine. From these results, it is suggested that DNA primase is one of targets of sphingosine, an effector molecule in apoptosis.

Nuclear ADP-ribosylation factor (ARF)- and oleate-dependent phospholipase D (PLD) in rat liver cells. Increases of ARF-dependent PLD activity in regenerating liver cells.

Two forms of phospholipase D (PLD) have been found to be present in nuclei isolated from rat hepatocytes by measuring phosphatidylbutanol produced from exogenous radiolabeled phosphatidylcholine in the presence of butanol. In nuclear lysates from either rat liver or ascites hepatoma AH 7974 cells, the PLD activity was markedly stimulated by a recombinant ADP-ribosylation factor (rARF) in the presence of the guanosine 5'-O-(3-thiotriphosphate) (GTPgammaS) and phosphatidylinositol 4, 5-bisphosphate. ATP and phorbol-12-myristate 13-acetate had no synergistic effect on this PLD activity. On the other hand, the nuclear PLD was stimulated by unsaturated fatty acids, especially by oleic acid. The ARF-dependent nuclear PLD activity was increased in the S-phase of the regenerating rat liver after partial hepatectomy and also was much higher in AH 7974 cells than in the resting rat liver. In contrast, the levels of the oleate-dependent PLD activity remained constant throughout the cell cycle in liver regeneration. The intranuclear levels of the stimulating proteins of the nuclear PLD activity, e.g. ARF, RhoA, and protein kinase Cdelta increased in the S-phase of the regenerating liver. These results suggested that the nuclear ARF-dependent PLD activity may be associated with cell proliferation.

Replication protein-A mediates the association of calf thymus DNA polymerase alpha-DNA primase complex with guanine-rich DNA sequence.

We have shown that calf thymus DNA polymerase alpha-DNA primase complex (pol alpha-primase) preferentially binds to pyrimidine-rich sequences and initiates RNA primer synthesis [Suzuki, M. et al. (1993) Biochemistry 32, 12782-12792]. Here we tested the association of pol alpha-primase with a guanine-rich DNA fragment (SVG, 30-mer) containing in vivo initiation sites of simian virus 40 DNA replication. While pyrimidine-rich fragment (CTPPS 1, 30-mer), that is a preferred sequence for calf thymus DNA primase, was well co-precipitated with pol alpha-primase using anti-pol alpha antibody, SVG was hardly precipitated under the same conditions. Competition studies in either gel-retardation assay or during de novo DNA synthesis by pol alpha-primase demonstrated that the interaction of pol alpha-primase with SVG was much weaker than that with CTPPS-1. On the other hand, replication protein-A (RP-A) could bind SVG, although less efficiently than CTPPS 1. After preincubation with RP-A, SVG could bind pol alpha-primase that was immobilized on Sepharose beads. The simian virus 40 large T antigen also enhanced association of SVG to pol alpha-primase, while Escherichia coli single-stranded DNA-binding protein did not. However, pol alpha-primase, bound to SVG in the presence of RP-A, failed to synthesize RNA primers. When SVG was extended 10 nucleotides at its 5'-end, pol alpha-primase synthesized trace amounts of RNA primers, and this activity was stimulated more than 10-fold by adding RP-A. These results suggest a new role for RP-A, i.e., as a molecular tether that allows pol alpha-primase to bind guanine-rich regions of DNA in order to initiate RNA primer synthesis.

Purification and characterization of nuclear alkaline phospholipase A2 in rat ascites hepatoma cells.

The alkaline phospholipase A2 (PLA2) was purified from nuclei of rat ascites hepatoma cells (AH7974) by column chromatography with a Sephacryl S-300 column and an immunoadsorbent using anti-group II PLA2 monoclonal antibody. From these two columns, the alkaline PLA2 was eluted in parallel with a 17-kDa protein which is reactive to another anti-group II PLA2 polyclonal antibody. Approximately 80% of nuclear PLA2 was inhibited by this antibody. The alkaline PLA2 was found in association with the chromatin fraction among subnuclear fractions. By an immunocytochemical staining, the nuclei of AH7974 were stained more strongly than other parts of cells with anti-group II PLA2 antiserum.

Selective inhibition of DNA polymerase epsilon by phosphatidylinositol.

We studied the effects of various phospholipids on the DNA synthesizing reactions by calf thymus DNA polymerases alpha, delta, and epsilon. Of these three enzymes, DNA polymerase epsilon was most sensitive to acidic phospholipids, i.e., phosphatidylinositol (PI), phosphatidylinositol 4-monophosphate (PIP), phosphatidylserine (PS), phosphatidic acid (PA), and cardiolipin (CAR). Of these acidic phospholipids, PI (from bovine liver) is of special interest because it inhibited DNA polymerase epsilon much strongly than DNA polymerase alpha and delta. The inhibition of DNA polymerase epsilon by PI was competitive with the DNA template-primer and was noncompetitive with dTTP substrate. The Ki value was estimated to be 16 microM. These results indicate that PI from bovine liver can be used as a specific inhibitor for DNA polymerase epsilon to analyze its role in DNA replication. Interestingly, the PI isolated from soybean, which has a different fatty acid composition, inhibited not only DNA polymerase epsilon but also DNA polymerase alpha.

Sphingosine inhibits the synthesis of RNA primers by primase in vitro.

We have previously shown the presence of sphingomyelin and sphingomyelinase in cell nuclei, suggesting that they may play a role in the intranuclear production of sphingosine, a potent bioactive molecule modulating diverse cellular functions. In the present study, the direct effects of sphingosine (C18:1) on the activity of DNA replication/repair polymerases were studied in vitro. Sphingosine had no effect on DNA polymerases alpha and beta and slightly inhibited DNA polymerases gamma, delta, and epsilon. In contrast, sphingosine strongly inhibited the activity of primase in a dose-dependent manner. On the other hand, dihydrosphingosine (C18:0), glycolipids, sphingomyelin, and ceramide had no effect on primase activity. Sphingosine equally inhibited the activity of primase complexed with DNA polymerase alpha, as well as its free form, with a Ki value of 4 microM. A gel-retardation analysis showed that the binding of primase with 32P-labeled template DNA was suppressed by sphingosine. Inhibition by sphingosine was competitive with the DNA template, but not with the substrate NTPs. After product analysis, a dose-dependent decrease in the amount of RNA primer products, consisting mainly of 10- and 11-mers, was observed in the presence of sphingosine, indicating that it inhibits the synthesis of RNA primers by primase. Sphingosine, however, had no effect on T7 RNA polymerase.

Purification and characterization of nuclear phospholipase C specific for phosphoinositides.

A phosphoinositide-specific phospholipase C (PLC) was solubilized from the isolated nuclei of rat ascites hepatoma AH7974 cells by ultrasonication in 2 M KCl. The extract was then subjected to five steps of column chromatographies in the order of Sephacryl S-300, phosphocellulose, Mono Q, Mono S, and Superose 6. Four forms of PLC (tentatively designated as N1, N2, N3, and N4) were purified 440-1400-fold. N1, N2, N3, and N4 showed apparent molecular masses of 85, 83, 80, and 88 kDa, respectively, on SDS-polyacrylamide gel electrophoresis. N1 cross-reacted with the antibody against the delta 1 isoform, while the other three forms did not cross-react with any of the antibodies against PLC-delta 1, -gamma 1, -gamma 2, and -beta 1. They hydrolyzed phosphatidylinositol (PI), phosphatidylinositol 4-monophosphate (PIP), and phosphatidylinositol 4,5-bisphosphate (PIP2) but did not show any activities against phosphatidylcholine and phosphatidylethanolamine. They showed the same optimal pH:pH 6.5 for PI hydrolysis and pH 7.0 for both PIP and PIP2 hydrolyses. They absolutely required Ca2+ for activity, with optimal concentrations of 10(-3)-10(-5) M for PIP and 10(-4)-10(-5) M for PIP2. For PI hydrolysis, N1, N2, and N3 required a Ca2+ concentration higher than 10(-2) M whereas N4 revealed significant activity even at 10(-5) M Ca2+ concentrations. Two forms of plasma membrane PLC and three forms of cytosolic PLC were purified from AH7974 cells by the same procedure as for nuclear PLC. Comparative study with these three groups revealed that all of the purified PLC isoforms shared similar enzymological properties except N4, which showed an exceptionally high affinity to Mono S column and was active at low concentrations of Ca2+ for PI as substrate. Furthermore, when PLC isoforms of nuclei from adult resting rat liver were compared with those from regenerating rat liver after partial hepatectomy, a PLC isoform corresponding to N4 of AH7974 cells was found only in regenerating liver nuclei. From these results, it was suggested that the nuclei of growing liver cells possessed a unique form of PLC (N4).

Sulfate- and sialic acid-containing glycolipids inhibit DNA polymerase alpha activity.

The effects of various glycolipids on the activity of immunoaffinity-purified calf thymus DNA polymerase alpha were studied in vitro. Preincubation with sialic acid-containing glycolipids, such as sialosylparagloboside (SPG), GM3, GM1, and GD1a, and sulfatide (cerebroside sulfate ester, CSE) dose-dependently inhibited the activity of DNA polymerase alpha, while other glycolipids, as well as free sphingosine and ceramide did not. About 50% inhibition was achieved by preincubating the enzyme with 2.5 microM of CSE, 50 microM of SPG or GM3, and 80 microM of GM1. Inhibition was noncompetitive with both the DNA template and the substrate dTTP, as well as with the other dNTPs. Since the inhibition was largely reversed by the addition of 0.05% Nonidet P40, these glycolipids may interact with the hydrophobic region of the enzyme protein. Apparently, the sulfate moiety in CSE and the sialic acid moiety in gangliosides were essential for the inhibition since neither neutral glycolipids (i.e., glucosylceramide, galactosylceramide, lactosylceramide) nor asialo-gangliosides (GA1 and GA2) showed any inhibitory effect. Furthermore, the ceramide backbone was also found to be necessary for maximal inhibition since the inhibition was largely abolished by substituting the lipid backbone with cholesterol. Increasing the number of sialic acid moieties per molecule further enhanced the inhibition, while elongating the sugar chain diminished it. It was clearly shown that the N-acetyl residue of the sialic acid moiety is particularly essential for inhibition by both SPG and GM3 because the loss of this residue or substitution with a glycolyl residue completely negated their inhibitory effect on DNA polymerase alpha activity.

RFP is a DNA binding protein associated with the nuclear matrix.

We reported that the RFP gene encodes a protein with putative zinc finger domains and was involved in the activation of the ret proto-oncogene. To further characterize the RFP protein, we developed a polyclonal antibody against the product synthesized from a fragment of the RFP cDNA expressed in Escherichia coli. Western blot analysis showed that RFP was identified as a 58 kDa protein in cell lysates from four human and rodent cell lines and from mouse testis. In addition, a unique 68 kDa protein was detected in the testis. Using AH7974 (rat ascites hepatoma) and Raji (human Burkitt lymphoma) cells, we demonstrated strong association of RFP with the nuclear matrix. Furthermore, RFP solubilized from the nuclear matrix had DNA-binding activity although it appears to bind more preferentially to double-stranded DNA than to single-stranded DNA. These results thus suggest that RFP may play a role in molecular processes which occur in the nuclear matrix.

Existence of phosphoinositide-specific phospholipase C in rat liver nuclei and its change during liver regeneration.

We found phosphoinositide-specific phospholipase C (PtdIns-PLC) activity in nuclei isolated from rat liver. The enzyme hydrolyzed phosphatidylinositol, phosphatidylinositol 4-monophosphate (PIP) and phosphatidylinositol 4,5-bisphosphate in a Ca(2+)-dependent manner, and produced inositol mono-, bis-, and triphosphate, respectively. Neither phosphatidylcholine, phosphatidylethanolamine, nor phosphatidylserine was utilized as a substrate. After partial hepatectomy, the PtdIns-PLC activity in isolated nuclei increased transiently in the S phase (20-22 h post-hepatectomy), to 2.5-fold higher than in the control, when measured with PIP. This result suggests a close relationship between the nuclear PtdIns-PLC, especially its PIP-hydrolyzing activity, and cell proliferation.

Growth-associated changes in fatty acid compositions of nuclear phospholipids of liver cells.

To know the possible relationships between nuclear phospholipids and cell proliferation, we have extensively analyzed phospholipids extracted from the nuclei of rat hepatic cells at various growth states. The content of phospholipid in nuclei as well as its composition was similar among liver cells tested, i.e., the regenerating rat livers (28 h, post-hepatectomy), sham-operated or non-treated control livers, and rat ascites hepatoma, AH7974 cells. In contrast, the fatty acid compositions of phospholipids differed from each other among these cells. At the 2-position of phospholipids in the regenerating liver nuclei at 28 h after partial hepatectomy, 18:1 (oleic acid) increased transiently at the expense of 20:4 (arachidonic acid) and 22:6 (docosahexaenoic acid), compared with those in the sham-operated control nuclei. This change in fatty acid composition was commonly observed throughout all phospholipids analyzed, i.e., phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI) and phosphatidylserine (PS). On the other hand, the change at 1-position was rather limited: in the regenerating liver nuclei (28 h), 18:1 increased only in PC at the expense of 18:0 (stearic acid). The similar and more marked deviation at the 2-position was observed with AH7974 nuclei it contained approximately 2-times more of 18:1 in PC, PE and PI than regenerating liver nuclei (28 h), and the decreased levels of 20:4 and/or 22:6. It should be noted that there were significant differences in the fatty acid compositions of PE and PS between sham-operated and non-treated controls. So, the sham-operated rat is the appropriate control for proliferation.(ABSTRACT TRUNCATED AT 250 WORDS)