Peptide-based targeting of fluorophores to peroxisomes in living cells.

Peptides carrying different fluorophores can be designed to incorporate spontaneously into living cells when added to the medium. By incorporating the peroxisome-targeting sequence PTS1, the peptide is recognized by the protein-import machinery of peroxisomes and, as a result, can accumulate in these organelles. Depending on the cell type, an inhibitor of the multidrug-resistance protein might be required to ensure strong accumulation. In this update, we discuss the potential of these peptide-linked fluorophores in solving issues related to organelle function and dynamics.

Effect of nonspecific phospholipid transfer protein on cholesterol esterification in microsomes from Morris hepatomas.

The content of cholesterol and cholesterol ester as well as the levels of acyl coenzyme A:cholesterol acyltransferase activity were determined in the microsomes from Morris hepatomas 7777, 5123D, and 7787. The free cholesterol content, expressed per mg microsomal protein, was significantly increased only in the microsomes from Morris hepatoma 7777 [47.8 +/- 0.4 micrograms (S.D.); p less than 0.001] and hepatoma 7787 (37.6 +/- 6.2 micrograms; p less than 0.01) as compared to normal liver (28.8 +/- 2.4 micrograms). The cholesterol ester content in the microsomes of the three different tumors did not significantly differ from that of normal liver (2.1 +/- 1.2 micrograms cholesterol per mg microsomal protein). The microsomal acyl coenzyme A:cholesterol acyltransferase activity was decreased in Morris hepatoma 7777 (8.6 +/- 2.3 pmol/min/mg protein; p less than 0.01) and in hepatoma 5123D (7.5 +/- 1.7 pmol/min/mg protein; p less than 0.02), and was normal in the hepatoma 7787 (16.5 +/- 7.8 pmol/min/mg protein) as compared to rat liver (16.0 +/- 2.9 pmol/min/mg protein). In a previous study (B. J. H. M. Poorthuis and K. W. A. Wirtz, Biochim. Biophys. Acta, 710: 99-105, 1982), this acyltransferase activity was shown to be stimulated by preincubation of rat liver microsomes with cholesterol-containing vesicles and the nonspecific phospholipid transfer protein. In this paper, a similar 4-fold stimulation of activity was observed for the microsomes of the various hepatomas investigated. The possible role of the nonspecific phospholipid transfer protein in intracellular cholesterol esterification is discussed.

Increased cholesterol esterification in rat liver microsomes in purified non-specific phospholipid transfer protein.

The effect of the non-specific phospholipid transfer protein purified from rat liver on the activity of acyl-CoA:cholesterol acyltransferase (EC 2.3.1.26) in rat liver microsomes was studied. The activity of cholesterol acyltransferase was measured from the rate of incorporation of [1-14C] oleoyl-CoA into cholesteryl oleate. Activity was stimulated by preincubation by the microsomes with the non-specific phospholipid transfer protein alone, but most effectively when vesicles consisting of phosphatidylcholine/cholesterol (molar ratio 1:1) also were present in the preincubation mixture. Preincubation with vesicles consisting of only phosphatidylcholine or phosphatidylcholine/phosphatidylethanolamine (molar ratio 1:1) had no effect. The stimulating effect is dependent on transfer protein and vesicle concentration and on the length of preincubation. Treatment of the transfer protein with N-ethylmaleimide abolished its effect on cholesterol ester formation. Preincubation of the microsomes with transfer protein and phosphatidylcholine/cholesterol vesicles containing radioactively labeled cholesterol shows that exogenous cholesterol is converted readily to cholesterol ester. The data are explained by the ability of non-specific phospholipid transfer protein to effect net transfer of cholesterol to those microsomes that contain cholesterol acyltransferase. Enlargement of the cholesterol substrate pool would then give rise to stimulation of the cholesterol acyltransferase activity. This study suggests a role for the transfer protein in modulating cholesterol metabolism by its ability to transport cholesterol between membranes.

The low-affinity lipid binding site of the non-specific lipid transfer protein. Implications for its mode of action.

The non-specific lipid transfer protein (nsL-TP) from bovine liver was studied by using the following fluorescent lipid analogs: phosphatidylcholine species with a sn-2-pyrenylacyl-chain of different length [Pyr(x)PC], sn-2-pyrenyldecanoyl-labelled phosphatidylinositol [Pyr(10)PI], -phosphatidylinositol 4-phosphate [Pyr(10)PIP], -phosphatidylinositol 4,5-bisphosphate [Pyr(10)PIP2] and dehydroergosterol. These analogs provided information on the effect of hydrophobicity and charge on lipid binding and transfer by nsL-TP. Binding of the Pyr(x)PC species decreased with increasing sn-2 acyl-chain length. Under equilibrium conditions, the fraction of nsL-TP that carried a PC molecule did not exceed 8%, which is consistent with a low affinity binding site. Also nsL-TP-mediated transfer of the Pyr(x)PC species decreased with increasing sn-2 acyl-chain length and was highly correlated with spontaneous transfer. Binding of the phosphoinositides increased in the order Pyr(10)PI less than Pyr(10)PIP less than Pyr(10)PIP2, indicating that an increase in lipid negative charge stimulates binding. The transfer of the phosphoinositides, however, decreased in the same order, which suggests that a high negative charge impairs the dissociation of the phospholipid from nsL-TP. Cholesterol, at concentrations up to 50 mol% in the donor membrane, hardly affected binding and transfer of Pyr(6)PC, strongly suggesting that nsL-TP has no high binding affinity for cholesterol. In agreement with this, binding of dehydroergosterol to nsL-TP was not detectable. Despite this apparently negligible affinity, nsL-TP-mediated transfer of dehydroergosterol was in the same order as that of Pyr(6)PC. The results are interpreted to indicate that transfer of lipids by nsL-TP involves the formation of a putative low-affinity lipid-protein complex. This formation is enhanced when lipid hydrophobicity decreases or lipid negative charge increases. Based on the binding and transfer data, the mode of action of nsL-TP is discussed in terms of change in free energy.

Induction of a relatively fast transbilayer movement of phosphatidylcholine in vesicles. A 13CNMR study.

[N-13CH3]Phosphatidylcholines are introduced into the outer monolayer of phosphatidylcholine vesicles with the phosphatidylcholine exchange protein from bovine liver. The transbilayer distribution of the [N-13CH3]phosphatidyl-choline is measured with 13C NMR. The transbilayer movements of [N-13CH3]-dioleoyl phosphatidylcholine and [N-13CH3]dimyristoyl phosphatidylcholine at 30 degrees C in vesicles composed of these phosphatidylcholines are extremely slow processes with estimated half-times of days. [N-13CH3]Dioleoyl phosphatidyl-choline introduced into dimyristoyl phosphatidylcholine vesicles migrates from the outer to the inner monolayer with a half-time of less than 12 h. The data suggest that differential changes in the lateral packing of the two monolayers might be a driving force for transbilayer transport of phospholipids.

Phosphatidylinositol transfer protein from bovine brain. Substrate specificity and membrane binding properties.

A recently developed fluorimetric transfer assay (Somerharju, P., Brockerhoff, H. and Wirtz, K.W.A. (1981) Biochim. Biophys. Acta 649, 521-528) has been applied to study the substrate specificity and membrane binding of the phosphatidylinositol-transfer protein from bovine brain. The substrate specificity was investigated by measuring the rate of transfer, either directly or indirectly, for a series of phosphatidylinositol analogues which included phosphatidic acid, phosphatidylglycerol as well as three lipids obtained from yeast phosphatidylinositol by partial periodate oxidation and subsequent borohydride reduction. Phosphatidylglycerol and the oxidation products of phosphatidylinositol were transferred at about one tenth of the rate observed for phosphatidylinositol while phosphatidic acid was not transferred. It is concluded that an intact inositol moiety favours the formation of the putative transfer protein-phosphatidylinositol complex. In addition to phosphatidylinositol, the transfer protein also transfers phosphatidylcholine. In order to obtain information on the possible occurrence of two sites of interaction, vesicles consisting of either pure 1-acyl-2-parinaroylphosphatidylinositol or 1-acyl-2-parinaroylphosphatidylcholine were titrated with the protein. Binding of labeled phospholipid to the protein was represented by an increase of lipid fluorescence and found to be much more efficient for phosphatidylinositol than for phosphatidylcholine. This is interpreted to indicate that the protein contains an endogenous phosphatidylinositol molecule which can be easily replaced by exogenous phosphatidylinositol but not by phosphatidylcholine, a lipid with a lower affinity for this protein. Thus the binding sites for the two phospholipids are mutually exclusive, i.e. phosphatidylinositol and phosphatidylcholine cannot be bound to the protein simultaneously. Finally, the effect of acidic phospholipids on the transfer protein activity was studied either by varying the content of phosphatidic acid in the acceptor vesicles or by adding vesicles of pure acidic phospholipids to the normal assay system. The latter vesicles consisted of either phosphatidic acid, phosphatidylglycerol, phosphatidylserine, phosphatidylinositol or cardiolipin. In both instances the transfer protein activity was inhibited, obviously through the enhanced association of the protein with the negatively charged vesicles. These findings strongly suggest that relatively nonspecific ionic forces rather than specific protein-phospholipid headgroup interactions contribute to the association of the phosphatidylinositol-transfer protein with membranes.

A new fluorimetric method to measure protein-catalyzed phospholipid transfer using 1-acyl-2-parinaroylphosphatidylcholine.

A new, simple and versatile method to measure phospholipid transfer has been developed, based on the use of a fluorescent phospholipid derivative, 1-acyl-2-parinaroylphosphatidylcholine. Vesicles prepared of this phospholipid show a low level of fluorescence due to interaction between the fluorescent groups. When phospholipid transfer protein and vesicles consisting of non-labeled phosphatidylcholine are added the protein catalyzes an exchange of phosphatidylcholine between the labeled donor and non-labeled acceptor vesicles. The insertion of labeled phosphatidylcholine into the non-labeled vesicles is accompanied by an increase in fluorescence due to abolishment of self-quenching. The initial rate of fluorescence enhancement was found to be proportional to the amount of transfer protein added. This assay was applied to determine the effect of membrane phospholipid composition on the activity of the phosphatidylcholine-, phosphatidylinositol- and non-specific phospholipid transfer proteins. Using acceptor vesicles of egg phosphatidylcholine and various amounts of phosphatidic acid it was observed that the rate of phosphatidylcholine transfer was either stimulated, inhibited or unaffected by increased negative charge depending on the donor to acceptor ratio and the protein used. In another set of experiments acceptor vesicles were prepared of phosphatidylcholine analogues in which the ester bonds were replaced with ether bonds or carbon-carbon bonds. Assuming that only a strictly coupled exchange between phosphatidylcholine and analogues gives rise to the observed fluorescence increase, orders of substrate preference should be established for the phosphatidylcholine- and phosphatidylinositol transfer proteins.

Binding of phospholipids to the phosphatidylinositol transfer protein from bovine brain as studied by steady-state and time-resolved fluorescence spectroscopy.

The phosphatidylinositol transfer protein isolated from brain, liver, heart and platelets was found to be present in two subforms which could be distinguished on the basis of the isoelectric points. In this study we have demonstrated that the two subforms isolated from bovine brain are due to the presence of either phosphatidylinositol or phosphatidylcholine in the lipid binding site of the protein. The transfer protein accommodates one phosphatidylinositol molecule in the binding site. The binding site for the sn-2 fatty acyl chain was investigated by incorporating in the transfer protein either phosphatidylinositol or phosphatidylcholine carrying a parinaroyl-chain attached at the sn-2 position. Time-resolved fluorescence spectroscopy revealed that the sn-2 fatty acyl chains for both phospholipids in the lipid-protein complex were completely immobilized (i.e., rotational correlation times of 17.4 ns for phosphatidylcholine and 16.3 ns for phosphatidylinositol). The similarity in correlation times suggests that the sn-2 fatty acyl chains of both phospholipids are accommodated in the same hydrophobic binding site of the protein.

The effect of polyphosphoinositides and phosphatidic acid on the phosphatidylinositol transfer protein from bovine brain: a kinetic study.

The phosphatidylinositol transfer protein from bovine brain (PI-TP) has lipid transfer characteristics which make it well suited to maintain phosphatidylinositol (PI) levels in intracellular membranes (Van Paridon, P.A., Gadella, Jr., T.W.J., Somerharju, P.J. and Wirtz, K.W.A. (1987) Biochim. Biophys. Acta 903, 68-77). Using a continuous fluorimetric transfer assay we have investigated in what way phosphatidylinositol 4-phosphate (PIP), phosphatidylinositol 4,5-bisphosphate (PIP2) and phosphatidic acid (PA) affect the transfer activity of this protein in model systems. The effects were analysed by application of a kinetic model which yielded the association constant (K) and dissociation rate constant (k-) for the PI-TP/vesicle complex. Incorporation of PA, PIP and PIP2 into the phosphatidylcholine-containing vesicles increased the association constant solely by diminishing the dissociation rate constant. This effect could be completely accounted for by changes in the membrane surface charge density. In contrast to the inhibitory effect of PA, the inhibition caused by PIP2 was completely abolished by the addition of neomycin, in agreement with the observed preferential binding of this polyamine antibiotic to PIP2. A rise in pH from 5.5 to 8 drastically reduced the association constant for vesicles containing 16 mol% PA (e.g., from 38 to 2 mM-1), without affecting the Vmax. This effect could be mainly attributed to an increase in the negative charge on PI-TP (isoelectric point 5.5), resulting in an enhanced repulsion. Increasing the negative membrane surface charge at pH 7.4 had the opposite effect. This is interpreted to indicate that the membrane interaction site on PI-TP must be positively charged, overcoming the repulsive forces between PI-TP and the vesicle. Addition of PIP2 micelles as a third component in the transfer assay strongly inhibited PI-TP transfer activity. The extent of inhibition suggests a very high affinity of PI-TP for this lipid.

Polyphosphoinositides undergo charge neutralization in the physiological pH range: a 31P-NMR study.

The charge state of phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate was determined as a function of pH by way of 31P-NMR spectroscopy. The pK values for the first protonation of the phosphomonoester residues in phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate were found to be 6.2 and 6.6, respectively, for the 4-phosphate moiety, and 7.7 for the 5-phosphate moiety.

The phosphatidylcholine transfer protein from bovine liver discriminates between phosphatidylcholine isomers. A monolayer study.

The activity of the phosphatidylcholine transfer protein from bovine liver toward phosphatidylcholine isomers carrying a long and a short fatty acyl chain on either the sn-1- or sn-2-position was determined by way of the monolayer-vesicle assay. In this assay equimolar mixtures of the isomers were spread at the air/water interface and their transfer measured to the vesicles in the subphase initiated by addition of the transfer protein. The following isomers were tested: 1-decanoyl-2-[3H]oleoyl-sn-glycero-3-phosphocholine (C10:0/[3H]C18:1-PC) and 1-oleoyl-2-decanoyl-sn-glycero-3-phospho[14C]choline (C18:1/C10:0-[14C]PC); 1-lauroyl-2-[3H]oleoyl-sn-glycero-3-phosphocholine (C12:0/[3H]C18:1-PC) and 1-oleoyl-2-[14C]lauroyl-sn-glycero-3-phosphocholine (C18:1/[14C]C12:0-PC); 1-myristoyl-2-[3H]oleoyl-sn-glycero-3-phosphocholine (C14:0/[3H]C18:1-PC) and 1-oleoyl,2-myristoyl-sn-glycero-3-phospho[14C]choline (C18:1/C14:0-[14C]PC). It was found that the protein transferred C10:0/[3H]C18:1-PC twice as fast as C18:1/C10:0-[14C]PC. Similar differences in rate were observed for C12:0/[3H]C18:1-Pc and C18:1/[14C]C12:0-PC but not for the isomers carrying myristic acid. We propose that the transfer protein can discriminate between PC isomers due to the presence of distinct binding sites for the sn-1- and sn-2-acyl chain (Berkhout et al. (1984) Biochemistry, 23, 1505-1513).

Transfer of cholesterol and oxysterol derivatives by the nonspecific lipid transfer protein (sterol carrier protein 2): a study on its mode of action.

The nonspecific lipid transfer protein (nsLTP) facilitates the transfer of both phospholipids and cholesterol between membrane interfaces. In this study, we have investigated the transport of 14C-labelled cholesterol, 7-ketocholesterol, 7 alpha-hydroxycholesterol and 25-hydroxycholesterol from a mixed lipid monolayer at the air/water interface to acceptor vesicles in the subphase. In the absence of nsLTP the transport of cholesterol was virtually nil, whereas the spontaneous transport of the oxysterol derivatives increased in the order 7-ketosterol less than 7 alpha-hydroxycholesterol less than 25-hydroxycholesterol. In the presence of nsLTP, the transport of both cholesterol and the oxysterol derivatives was greatly enhanced; the highest rate of transport was observed for 25-hydroxycholesterol. In the absence of vesicles, binding of cholesterol and of 25-hydroxycholesterol from the monolayer to nsLTP was negligible. Similarly, nsLTP did not bind cholesterol from radiolabeled bovine heart mitochondria under conditions where it stimulated the transfer of cholesterol to vesicles. In agreement with this failure to bind, nsLTP was unable to carry cholesterol between two separate monolayers. From the monolayer experiments it became apparent that nsLTP is highly surface-active. Measurement of the transport of cholesterol and of oxysterol derivatives by the monolayer-vesicles assay and of a series of pyrene-labeled phosphatidylcholine species by the fluorescent transfer assay showed a high correlation between the spontaneous and the nsLTP-mediated lipid transport. This supports the notion that nsLTP lowers the energy barrier for the lipid monomer-membrane interface equilibration process. In view of the above observations, we propose that nsLTP may facilitate the transfer of lipids by being part of a transient collisional complex between donor and acceptor membrane.

Immunological comparison of phosphatidylinostiol and phosphatidylcholine exchange proteins in bovine brain, liver and heart.

The two phosphatidylinositol exchange proteins isolated from bovine cerebral cortex, I (isoelectric point pH 5.2) and II (isoelectric point pH 5.5), had essentially identical amino acid compositions. Rabbit antisera preparations specific to each of these brain proteins were equally effective in inhibiting the phosphatidylinositol transfer activity of both protein I and II. Judged by double diffusion on agar gels, immunoprecipitation was not observed between either of the brain phosphatidylinositol exchange proteins and anti-liver phosphatidylcholine exchange protein antibody or between liver phosphatidylcholine exchange protein and anti-brain phosphatidylinositol exchange protein antibody. Phosphatidylinositol and phosphatidylcholine transfer activity was measured in microsome-liposome assay systems. For membrane-free tissue preparations phosphatidylinositol activity increased in the order: brain greater than heart greater than liver, while phosphatidylinositol exchange proteins transferred phosphatidylinositol and phosphatidylcholine in the ratio 1.4: liver phosphatidylcholine exchange protein transferred exclusively phosphatidylcholine. Phosphatidylinositol transfer activity in brain, heart and liver was more than 80% inhibited by anti-brain phosphatidylinositol exchange protein antibody. The proportion of phosphatidylcholine transfer activity sensitive to anti-liver phosphatidylcholine exchange protein antibody was 15% for brain, 75% for liver and 20% for heart, while the proportion sensitive to anti-brain phosphatidylinositol exchange protein antibody was 65% for brain, 10% for liver and 60% for heart. Together these two classes of phospholipid exchange proteins accounted for approx. 80% of the phosphoatidylcholine transfer activity in selected bovine tissues. A protein which was chemically, immunologically, and catalytically similar to liver phosphatidylcholine exchange protein was identified in brain and contributed about 20% of the phosphatidylcholine transfer activity in that tissue.

Delipidation of the phosphatidylcholine exchange protein from beef liver by detergents.

1. A method is described to introduce [14C] phosphatidylcholine into the phosphatidylcholine exchange protein from beef liver. The effects of various detergents on this 14C-labelled phospholipid - protein complex are considered. 2. As shown by spectrophotometric and radioactivity analysis of polyacrylamide gels, sodium deoxycholate, Triton X-100, lysophosphatidylcholine and sodium dodecyl sulfate delipidate the exchange protein, while mixed phosphatidylcholine-detergents micelles are formed. 3. Protein delipidated by sodium deoxycholate, Triton X-100 and lysophosphatidylcholine retains its ability to catalyze the transfer of phosphatidylcholine between membranes. The immunological properties are similar to those of native protein as shown by double immunodiffusion in agar against an antiserum gamma-globulin fraction. 4. Sodium dodecyl sulfate and cetyltrimethylammonium bromide interact very strongly with the protein conferring their charge to the complex and destroying the antigenic determinants.

Identification of membrane proteins of human blood platelets with a hydrophobic photolabel.

A photoactivable glycolipid probe, 12-(4-azido-2-nitrophenoxy)stearoyl[1-14C]glucosamine, was used to label proteins and lipids of platelet membranes. The proteins were analyzed by two-dimensional high-resolution gelelectrophoresis. The labeling patterns showed that three membrane proteins were labeled which were not previously identified by ectolabeling (Sixma, J.J. and Schiphorst, M.E. (1980) Biochim. Biophys. Acta 603, 70-83). Analysis of the lipid fraction showed that phosphatidylcholine, phosphatidylethanolamine and phosphatidylserine were labeled by the probe. The distinct labeling of phosphatidylserine strongly suggests that the probe redistributes between the two halves of the bilayer.

Tissue distribution and subcellular localization of phosphatidylcholine transfer protein in rats as determined by radioimmunoassay.

A radioimmunoassay for the phosphatidylcholine-transfer protein from rat liver was used to measure levels of PC-transfer protein in rat tissues. The assay as described before (Teerlink, T., Poorthuis, B.J.H.M., Van der Krift, T.P. and Wirtz, K.W.A., Biochim. Biophys. Acta 665 (1981) 74-80) was modified in order to measure PC-transfer protein in tissue homogenates and subcellular membrane fractions. To this end both a detergent (Triton X-100) and a proteolytic enzyme inhibitor (aprotinin) were added to the assay medium. The radioimmunoassay measured levels of PC-transfer protein in the range of 5-50 ng and was specific for PC-transfer protein from rat tissues. Subcellular distribution studies showed that in 10% (w/v) homogenates of liver approximately 60% of the PC-transfer protein was present in the 105000 X g supernatant fraction, the remainder being evenly distributed over the particulate fractions. PC-transfer protein associated with the particulate fractions was almost completely removed by a single washing step, suggesting a dynamic equilibrium between membrane-bound and soluble PC-transfer protein. Both 105000 X g supernatants and homogenates of various rat tissues were assayed. The highest levels of PC-transfer protein were measured in liver and intestinal mucosa. Lower values were found in kidney, spleen and lung, whereas heart and brain contained hardly any PC-transfer protein. PC-transfer protein levels in regenerating rat liver did not differ significantly from levels in normal liver. In fetal lung a change in PC-transfer protein content during development was observed, with a clear maximum 2 days before term, suggesting an involvement of PC-transfer protein in the secretion of lung surfactant.

Human erythrocyte membranes are fluid down to -5 degrees C.

This first observation of the deuterium nuclear magnetic resonance (2H-NMR) spectrum of phospholipid molecules incorporated into intact human erythrocyte ghosts shows that the liquid crystalline phase is stable down to a temperature of -5 degrees C. The quality of the 3H-NMR spectra indicate that it is now possible to carry out clinical studies of erythrocyte membranes using the techniques employed in this study.

A new high-yield procedure for the purification of the non-specific phospholipid transfer protein from rat liver.

Rat liver contains a non-specific phospholipid transfer protein that transfers phosphatidylethanolamine, phosphatidylcholine, phosphatidylinositol and sphingomyelin as well as cholesterol between membranes (Bloj, B. and Zilversmit, D.B. (1977) J. Biol. Chem. 252, 1613-1619). The present paper describes a new high-yield procedure for the purification of this protein which includes fractionation on DEAE-cellulose, Sephadex G-50 and hydroxyapatite. Starting from a pH 5.1 supernatant, a homogeneous protein was obtained after a 1 540-fold purification at a yield of 50%. The protein has a molecular weight of 14 800 as estimated by electrophoresis on polyacrylamide gels in the presence of SDS. It has a blocked N-terminal amino acid and a tryptophanyl fluorescence emission maximum at 335 nm. Its amino acid composition has been determined and compared to data published by others on similar proteins.

The occurrence of soluble and membrane-bound non-specific lipid transfer protein (sterol carrier protein 2) in rat tissues.

The occurrence and subcellular distribution of the non-specific lipid transfer protein (nsL-TP; sterol carrier protein 2) in rat tissues was investigated by the immunoblotting technique using the affinity purified antibody against rat liver nsL-TP. Highest levels of the protein were found in the homogenates of liver, lung and adrenals, whereas it could hardly be detected in brain. In other tissues (i.e., testis, kidney, heart and intestine) the protein was present at intermediate concentrations. Analysis of subcellular fractions obtained by differential centrifugation demonstrated that in all tissues except for the liver, nsL-TP was predominantly present in the particulate fractions. In the particulate fractions of all tissues, an immunoreactive 58 kDa-protein was detected. Density centrifugation of a nuclear-free homogenate from liver and testis indicated that the 58 kDa-protein did, and nsL-TP did not, cofractionate with catalase. This suggests that in these tissues the bulk of nsL-TP is extraperoxisomal. Membrane-bound nsL-TP in testis was sensitive to trypsin treatment, suggesting that it is exposed to the cytosol. Release of nsL-TP by washing the membranes with 0.1 M Na2CO3 (pH 11.5), indicated that nsL-TP is a periferal protein. It was shown by chromatofocussing that nsL-TP extracted from membrane fractions was more basic than nsL-TP present in the cytosol fraction from rat liver (isoelectric point of 8.7).