Rat intestinal apolipoprotein B gene expression. Evidence for integrated regulation by bile salt, fatty acid, and phospholipid flux.

We previously reported that intestinal apo B48 synthesis in the rat was unaltered by dietary triglyceride intake but demonstrated regulation in response to biliary lipid availability. Studies are now presented in which the mechanisms underlying biliary lipid dependent expression of intestinal apo B48 synthesis have been investigated further. Bile salt replacement was effective in a dose- and structure-dependent manner in reexpressing intestinal apo B48 synthesis after prolonged bile diversion. Further experiments suggested that this effect of bile salt may be related to facilitated uptake of fatty acid. A role for mucosal phospholipid flux was suggested by studies in which infusion of lysolecithin, with or without Na taurocholate, produced complete reexpression of apo B48 synthesis in jejunal enterocytes. Over a four- to sixfold range of apo B48 synthesis rates in both jejunum and ileum, there was no change in apo B mRNA size or abundance as determined by RNA blot hybridization. Analysis of both intestinal mucosa and microsome lipid content in a variety of settings revealed that apo B48 synthesis rates were correlated with microsome triglyceride fatty acid content (r = 0.65, P less than 0.005) but not free fatty acid or phospholipid content. These studies demonstrate a physiologic role for elements of biliary lipid flux in the regulation of apo B gene expression. The data suggest that an integrated mechanism may exist whereby apo B48 synthesis is related to microsome triglyceride flux, particularly at low levels of lumenal substrate availability.

Insulin regulates apolipoprotein B turnover and phosphorylation in rat hepatocytes.

Our laboratory has previously shown that insulin inhibits the secretion of newly-synthesized and immunoreactive apo B from rat hepatocytes. We have also shown that apo B is secreted as a phosphoprotein and that phosphorylation is increased in hypoinsulinemic nonketotic diabetes. The present studies were conducted to determine whether the ability of insulin to inhibit apo B secretion is related to alterations in apo B turnover and whether insulin itself affects apo B phosphorylation. Pulse-chase studies with [35S]methionine in primary cultures of hepatocytes from normal rats in the absence and presence of insulin show that the secretion of apo B100 and apo B48 are inhibited by insulin and that this inhibition may be due in part to enhanced intracellular degradation. In addition, there is a second intracellular apo B48 pool which is not insulin regulated or degraded. In experiments in which hepatocytes were incubated with [32P]orthophosphate, insulin decreased 32P incorporation into apo B100 (42%) with only small effects on apo B48 (11%). The small insulin effect on apo B48 may relate to an insulin-insensitive apo B48 intracellular pool. These studies show that insulin can affect the intracellular turnover, secretion, degradation, and phosphorylation of apo B and emphasize the differential regulation of apo B100 and apo B48 with regard to these parameters in rat liver.

Subfractionation of liver membrane preparations by specific ligand-induced density perturbation.

Immunofluorescence and immunoferritin staining with monospecific antibodies to dipeptidyl peptidase IV purified from rat liver plasma membrane showed that the antigenic sites of this glycoprotein was exposed only on the outer surface of the liver cell. In a vesiculated plasma membrane preparation the peptidase was located exclusively on right-side-out elements, which differed in their degrees of ferritin staining, and could be separated into subfractions of different buoyant densities corresponding to their concentration of dipeptidyl peptidase IV. The concomitant density perturbation of nucleotide pyrophosphatase was similar, but not identical, to that of the peptidase itself, indicating that these two marker enzymes are somewhat differently distributed in the plane of the liver plasma membrane. Since essentially all the galactosyl transferase in plasma membrane and none of that in Golgi membrane could be density-perturbed with the antipeptidase, the activity in the plasma membrane preparation could not be ascribed to contamination with discrete Golgi elements. On the other hand, the small amount of dipeptidyl peptidase IV found in the Golgi preparations was itself perturbed by the antipeptidase, indicating that it represented contaminating right-side-out plasma membrane vesicles. In preliminary experiments similar separations were also obtained with wheat germ agglutinin as the plasma membrane ligand. Density perturbation, mediated by the recognition of specific surface markers, should be a useful adjunct in the separation and characterization of subcellular components in other systems.

Improved method for the isolation of rat liver plasma membrane.

An improved method for the isolation of plasma membrane from rat liver is presented. Gentle homogenization of perfused livers in buffered isotonic KCL, followed by direct flotation of a low-speed nuclear pellet through a discontinuous sucrose density gradient results in a 32% yield, and 25-fold enrichment for the plasma membrane marker, phosphodiesterase I, in a crude plasma membrane fraction. This fraction contains less than 1% of the mitochondria, and endoplasmic reticulum present in the original homogenate, but is more heavily contaminated with lysosomes and Golgi membrane. Vigorous mechanical disruption of this material, followed by a second discontinuous sucrose density gradient, gives a light plasma membrane fraction with an 80-fold purification and 20% yield of phosphodiesterase I over the original homogete (with further reduction of contaminants).

Effect of lovastatin on the secretion of very low density lipoprotein lipids and apolipoprotein B in the hypertriglyceridemic Zucker obese rat.

Recent studies demonstrated that inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase lower plasma triglyceride primarily by decreasing hepatic secretion of very low density lipoproteins (VLDL). A possible mechanism is that inhibition of cholesterol synthesis interferes with the assembly of VLDL particles. Since one molecule of apolipoprotein (apo) B is required for the proper assembly and secretion of each VLDL and secretion of apo B may be regulated by various lipid components of the lipoproteins, question arises whether HMG-CoA reductase inhibitors also decrease the secretion of apo B. To address this issue, we investigated the effect of lovastatin on the secretion rate of VLDL-apo B and on the composition of VLDL in the Zucker obese rat; a model for genetic hypertriglyceridemia. Lovastatin treatment (4 mg/kg day x 13 days), as compared with placebo, decreased the concentrations of fasting plasma triglyceride (1740 +/- 170 vs. 3130 +/- 790 micrograms/ml) and VLDL-triglyceride (1379 +/- 59 vs. 3082 +/- 715 micrograms/ml). There was a small but non-significant decrease in VLDL-apo B (19 +/- 2 micrograms/ml vs. 26 +/- 7 micrograms/ml). Thus, lovastatin significantly decreased the ratio of triglyceride to apo B in VLDL (76 in lovastatin vs. 124 in placebo group). Secretion rates of VLDL-lipids and VLDL-apo B were measured after intravenous injection of Triton WR-1339.(ABSTRACT TRUNCATED AT 250 WORDS)

Apolipoprotein B: structure, biosynthesis and role in the lipoprotein assembly process.

The complete amino acid sequence of the liver-synthesized apolipoprotein B (apoB) species, apoB 100, has been derived from cloned cDNA. The protein consists of 4536 amino acids (+ a 27 amino acid signal sequence). Cysteine is clustered in the N-terminal 1/10 of the protein, suggesting the presence of a stabilized tertiary structure in this part of the molecule. Three types of structure are suggested to be of importance for the binding of the protein to lipids; (i) hydrophobic sequences with a high probability for beta-sheet structure, (ii) strict amphipathic beta-sheets, and (iii) amphipathic alfa-helices. An apoB 100 molecule is completed within 10-14 min and secreted after approximately 30 min, 1/3 of which is due to the transfer through the endoplasmic reticulum (ER), while 2/3 is spent in the Golgi apparatus. ApoB 100 is co-translationally N-glycosylated and 25% of the oligosaccharide chains is processed in the Golgi compartment. Other posttranslational modifications that have been discussed include covalent acylation and phosphorylation. It has also been suggested that the lipid moiety of the apoB 100 lipoproteins are modified during the passage through the Golgi apparatus. The site of lipoprotein assembly is suggested to be separated from the site of apoB 100 synthesis, and apoB 100 appears to be co-translationally bound to the ER membrane and from this transferred to the ER lumen. Based on these observations a model for the assembly of apoB 100 lipoproteins is discussed in this paper. The intestinal derived apoB species, apoB 48, has a molecular mass of 210 kDa and appears to correspond to the N-terminal 48% of apoB 100. The mechanism by which apoB 48 is formed is still not known. Available data indicate that the protein is formed within the intestinal cells, these data also argue against the possibility that apoB 48 is formed by posttranslational proteolysis of apoB 100. The formation of a separate apoB 48 mRNA by alternative splicing has been suggested, based on the observation of a 7 kb mRNA which corresponds to the 5' portion of the apoB 100 mRNA. However, the most abundant apoB mRNA species found in the intestine have a size that corresponds to that of the apoB 100 mRNA, furthermore the observation that apoB 48 appears to terminate in a 7.5 kb exon that appears to lack alternative splice sites, does not favour the possibility of alternative splicing.

Fatty acid synthesis in vivo and hepatic contribution to whole-body lipogenic rates in obese Zucker rats.

We have re-examined the claim by Godbole and York, based on the effect of surgical hepatectomy (Diabetologia 14:191, 1978), that liver contributed more than 90% of the newly synthesized FA found in adipose tissue of obese rats at the end of a 1-hr pulse of 3H2O. The amount of newly synthesized FA transported via plasma VLDL from liver to adipose tissue was estimated in lean and obese Zucker rats by determining the effects of Triton WR-1339, which blocks the uptake of VLDL-TGFA into tissues. Triton treatment was found not to cause any significant change in the amount of radioactive FA found in subcutaneous/perimetrial fat tissues, carcass or liver in either chow-fed or high-glucose, fed-refed lean or obese rats, although in the fed-refed dietary state the proportion found in the liver was increased over that in the chow-fed groups. Furthermore, the amounts of newly made FA which accumulated in the plasma of Triton-treated, chow-fed and glucose-fed refed animals during this period constituted only a few percentages of those found in the adipose tissue of these animals. Thus, in contrast to the claims of Godbole and York, no significant transfer of newly made FA from liver to adipose tissue occurs during a 1-hr experiment; it follows that the amount of these FA found in different tissues at the end of that period are valid measurements of their actual lipogenic activities in situ. It is suggested that the Godbole and York results are artifacts of their surgical hepatectomy procedure.

Biogenesis of plasma membrane glycoproteins. Tracer kinetic study of two rat liver plasma membrane glycoproteins in vivo.

Antibodies to purified nucleotide pyrophosphatase (NPPase) and dipeptidyl peptidase IV (DPP IV) were used to study the biogenesis of these rat liver plasma membrane glycoproteins in vivo. Following injection of tritiated leucine, the radioactivity in NPPase and DPP IV decayed at markedly different rates in the plasma membrane, with apparent half-lives of about 1 and 5 days, respectively. In short term experiments, labeling of total plasma membrane proteins was rapid and insensitive to colchicine, while labeling of both NPPase and DPP IV showed a lag of about 15 min, followed by colchcine-sensitive/cycloheximide-insensitive increases to half-maximal and maximal values at about 1 and 2 h, respectively. A peak of labeled DPP IV in rough microsomes at 15 min showed increased mobility on polyacrylamide gels and was largely inaccessible to antibodies in intact microsomes, consistent with its being an underglycosylated precursor, exposed on the cisternal side of the rough endoplasmic reticulum. In contrast, the behavior of unlabeled DPP IV in preparations of rough microsomes and Golgi was consistent with its being contributed by contaminating right-side-out plasma membrane vesicles. This conclusion was also necessary to fit the tracer kinetic data to a simple membrane-flow model, which gave precursor pools (1 microgram/g of liver) and fluxes (1 microgram/h/g of liver) for both DPP IV and NPPase which were about 3 orders of magnitude less than those for the synthesis of rat serum albumin. Thus, unlike hepatoma tissue culture cells (Doyle, D., Baumann, H., England, B., Friedman, E., Hou, E., and Tweto, J. (1978) J. Biol. Chem. 253, 967-973), normal rat liver does not contain large amounts of preformed intracellular plasma membrane precursors.

Biogenesis of plasma membrane glycoproteins. Purification and properties of two rat liver plasma membrane glycoproteins.

As a preliminary to a study of the biogenesis of individual plasma membrane glycoproteins, the marker enzyme nucleotide pyrophosphatase (NPPase) and a major rat liver plasma membrane sialoprotein, subsequently found to be identical with the enzyme dipeptidyl peptidase IV (DPP IV), were purified 10,000- and 2,000-fold, respectively, from rat liver. Both were amphipathic proteins which formed defined micellar complexes with detergents and aggregated in their absence. Gel filtration, sucrose density gradient centrifugation, and polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate showed the Triton X-100 complex of NPPase to contain a single 150,000-dalton peptide, while that of DPP IV was composed of two 120,000-dalton subunits; each complex also contained about 150,000-dalton Triton X-100. Trypsin cleaved the detergent complexes with release of major hydrophilic fragments which no longer bound detergent micelles; the accompanying change in peptide size was small for NPPase and undetectable for DPP IV, which also retained the dimer structure of its native form. DPP IV was the only major glycoprotein in rat liver plasma membrane which bound strongly to wheat germ agglutinin. Monospecific rabbit antibodies against NPPase and DPP IV precipitated the antigens without affecting their enzymatic activities.

Purification and properties of the fatty acids synthetase complex from Neurospora crassa, and the nature of the fas-mutation.

A procedure is described for the purification of the fatty acid synthetase complex (FAS) from Neurospora crassa. The enzyme complex has a molecular weight of 2.3 times 10(6), contains 6 mol of 4'-phosphopantetheine per mol, and on polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate gives a single band, or a closely spaced doublet, which comigrates with standard myosin (molecular weight, 2 times 10(5)). Since the slightly retarded component in the doublet accounts for all protein-bound 4'-phosphopantetheine, the complex appears to be made up of 11 to 12 equally sized subunits, 6 of which carry the acyl carrier protein function. In this unusual arrangement, notably the lack of the low-molecular-weight acyl carrier protein component seen in other FAS systems, as well as in its enzymatic properties, the Neurospora FAS complex is quite similar to the yeast enzyme. The FAS complex of a saturated fatty acid-requiring mutant, previously disignated cel-, contains less than 2% of the 4'-phosphopantetheine prosthetic groups found in the wild-type complex. The leaky phenotype of this mutant, here designated fas-, is accounted for by a residual fatty acid synthesizing activity in its FAS complex, which is several-fold higher than expected from its residual content of 4'-phosphopanthetheine.

A new class of lipids: chlorosulfolipids.

Acid hydrolysis of sulfolipids from Ochromonas danica liberates 13-chloro-docosan-1,14-diol; 11,15-dichloro-docosan-1,14-diol; and other docosan-diols with three to six chlorine atoms per molecule. In stationary cells grown on 1.47 mM Cl(-), the majority of sulfolipids are chlorinated, with the hexachloro species being the most abundant.

Apolipoprotein B is structurally and metabolically heterogeneous in the rat.

Electrophoresis of rat apolipoprotein B (apoB) on 5% polyacrylamide gels in the presence of NaDodSO4 separates three major components: PI, which comigrates with human low density lipoprotein (LDL) apoB; PII, a slightly faster-moving satellite band; and PIII, which migrates somewhat more slowly than myosin heavy chain. The proportion of PIII decreases with increasing density of the parent rat lipoprotein, from 90% an 70%, respectively, in chylomicrons and very low density lipoproteins (VLDL), to 7% in the major LDL2 (density 1.038-1.063 g/ml) fraction. A major component that comigrates with rat PIII is a marker for human chylomicron apoB, being absent from human VLDL, intermediate density lipoprotein (IDL), and LDL. Preliminary immunological and peptide mapping data show that rat apoB PI and PIII are closely related structurally, with the latter possibly being a large fragment of the former. Both peptides are synthesized in rat liver and found in Golgi secretory vesicles. Kinetic tracer experiments show that rat PI and PIII are present on separate VLDL particles, both of which are extensively removed from the circulation at the remnant stage, and that the declining PIII-to-PI/II ratios in IDL and LDL may be attributed to the more rapid turnover of PIII-containing lipoproteins at all levels, particularly within the LDL density range.

Nascent and remnant lipoprotein turnover in rats: experimental design and simulation.

We have attempted to predict the kinetic behavior of the complex very low-density lipoprotein (VLDL; d less than 1.006) fraction in blood plasma of rats in the steady state. Specifically we proposed a simple model with two different kinds of nascent VLDL particles derived from the liver, one containing apoprotein B (PI/II) [apoB(PI/II)], the high-molecular-weight apoB, and the other, apoprotein B (PIII) [apoB(PIII)], the low-molecular-weight apoB. Two other particles, the corresponding remnants derived from the nascent VLDL particles were also included. Then a number of feasible in vivo tracer experiments were considered in which VLDL labeled in the apoB and/or triglyceride (TG) moieties would be injected into recipient rats and the kinetic behavior of the various compartments predicted by simulation analysis. In addition the kinetic behavior of products such as free fatty acids formed during hydrolysis of labeled TG fatty acids and liver TG derived from labeled circulating remnants was considered. Both the relative sizes of nascent and remnant particles and the extent of average hydrolysis of nascent VLDL-TG (before formation of a remnant particle) were considered in our analysis. On the basis of these predictions we have suggested a number of experimental approaches that should be helpful in defining the relative pool sizes and the turnover rates of each kind of particle in vivo.

Radioimmunoassay of rat apolipoprotein B peptides in lipoproteins and tissues.

Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS-PAGE) separates rat apolipoprotein B (apoB) into one lower and two higher molecular weight components. Of the latter, peptide I (PI) corresponds to human B-100, while the slightly faster-migrating peptide II (PII) lacks a human counterpart; the smaller species peptide III (PIII) corresponds to human B-48. We describe here a competitive radioimmunoassay which separately measures the amounts of total (i.e., PI + PII + PIII) and larger (i.e., PI + PII) rat apoB peptides, with the amounts of PIII obtained by difference. Standard rat PIII and combined PI + PII (PI,II) were isolated by high-pressure gel filtration liquid chromatography in the presence of SDS, and the PI,II was used as an immunogen to raise rabbit antisera which were capable of binding all three forms of rat apoB. However, Scatchard analysis showed this binding to represent two distinct types of antibodies: one high-affinity class which bound only PI,II and a second class which bound all apoB peptides with equal but lower affinity. Thus, since 125I-labeled PIII was displaced equally effectively by PI,II and PIII, but 125I-labeled PI,II was displaced only by PI,II, the unabsorbed antiserum could be used to measure either total apoB or PI,II alone, depending on the choice of labeled ligand. The validity of the assay for apoB peptides in very-low-density and low-density lipoproteins and in liver microsomes was verified by comparison with peptide determinations by SDS-PAGE.

Rat McA-RH7777 cells efficiently assemble rat apolipoprotein B-48 or larger fragments into VLDL but not human apolipoprotein B of any size.

Studies of truncated apoB peptides in human subjects with familial hypobetalipoproteinemia, as well as of puromycin-generated spectra of nascent apoB peptides in rat and hamster liver, suggest that a minimum size is required for N-terminal fragments of apoB to be efficiently assembled into full-sized VLDL. We report here results of experiments undertaken to examine this phenomenon in greater detail by expressing individual carboxyl-truncated human apoB constructs in McArdle cells. Thus, apoB-29, -32, -37, -42, -47, -53, -70 and full length apoB-100 were transiently expressed in rat McA-RH7777 hepatoma cells, or human apoB-31 and apoB-53 were stably expressed in the same cells, and the secreted VLDL particles were characterized by kinetic gradient ultracentrifugal flotation. Calibration with rat plasma VLDL subfractions showed that about 90 and 50%, respectively, of lipoprotein particles containing endogenous rat B-100 and B-48 floated between fractions 2;-8 of the 11-fraction gradient. This corresponds to the normal VLDL diameter range of about 47 to 28 nm, with the remaining half of rat B-48 recovered as HDL particles in the 1.1 g/ml range. In contrast, regardless of their size, only 2;-5% of any of the truncated human apoB peptides expressed in these cells was recovered in the VLDL region of the gradient. The remaining 95+% of the lipoproteins were found as high density particles; as previously found in other systems the densities of the latter were inversely related to their peptide chain-length. Furthermore, transiently expressed full-length human apoB-100 was inefficiently secreted as VLDL by these cells, with the remainder appearing as LDL-sized particles. Thus, although we showed that McA-RH7777 cells secreted endogenous rat apoB as normal-sized VLDL, we found them unsuitable for our original purpose of using human apoB fragments to further define effects of apoB size on VLDL assembly. These cells appeared unable to efficiently use any size of human apoB for that process. Pulse-labeled untransfected McA-RH7777 cells chased in the presence of puromycin did, however, show a sharp decline in VLDL assembly efficiency for endogenous nascent rat apoB peptides shorter than B-48, similar to that originally found in normal rat liver.