The contribution of individual and pairwise combinations of SNPs in the APOA1 and APOC3 genes to interindividual HDL-C variability.

Apolipoproteins (apo) A-I and C-III are components of high-density lipoprotein-cholesterol (HDL-C), a quantitative trait negatively correlated with risk of cardiovascular disease (CVD). We analyzed the contribution of individual and pairwise combinations of single nucleotide polymorphisms (SNPs) in the APOA1/APOC3 genes to HDL-C variability to evaluate (1) consistency of published single-SNP studies with our single-SNP analyses; (2) consistency of single-SNP and two-SNP phenotype-genotype relationships across race-, gender-, and geographical location-dependent contexts; and (3) the contribution of single SNPs and pairs of SNPs to variability beyond that explained by plasma apo A-I concentration. We analyzed 45 SNPs in 3,831 young African-American (N=1,858) and European-American (N=1,973) females and males ascertained by the Coronary Artery Risk Development in Young Adults (CARDIA) study. We found three SNPs that significantly impact HDL-C variability in both the literature and the CARDIA sample. Single-SNP analyses identified only one of five significant HDL-C SNP genotype relationships in the CARDIA study that was consistent across all race-, gender-, and geographical location-dependent contexts. The other four were consistent across geographical locations for a particular race-gender context. The portion of total phenotypic variance explained by single-SNP genotypes and genotypes defined by pairs of SNPs was less than 3%, an amount that is miniscule compared to the contribution explained by variability in plasma apo A-I concentration. Our findings illustrate the impact of context-dependence on SNP selection for prediction of CVD risk factor variability.

Expression of the mRNA encoding truncated PPAR alpha does not correlate with hepatic insensitivity to peroxisome proliferators.

Two alternatively spliced forms of human PPAR alpha mRNA, PPAR alpha1 and PPAR alpha2, have been identified. PPAR alpha1 mRNA gives rise to an active PPAR alpha protein while PPAR alpha2 mRNA gives rise to a form of PPAR which lacks the ligand-binding domain. PPAR alpha2 is unable to activate a peroxisome proliferator response element (PPRE) reporter gene construct in transient transfection assays. Both PPAR alpha1 and PPAR alpha2 mRNA are present in human liver, kidney, testes, heart, small intestine, and smooth muscle. In human liver, PPAR alpha2 mRNA abundance is approximately half that of PPAR alpha1 mRNA; a correlation analysis of PPAR alpha1 and PPAR alpha2 mRNA mass revealed an r-value of 0.75 (n = 18). Additional studies with intact liver from various species, showed that the PPAR alpha2/PPAR alpha1 mRNA ratios in rat, rabbit, and mouse liver were less than 0.10; significantly lower than the 0.3 and 0.5 ratios observed in monkey and human livers, respectively. To determine if a high PPAR alpha2/PPAR alpha1 mRNA ratio was associated with insensitivity to peroxisome proliferators, we treated human, rat, and rabbit hepatocytes with WY14643, a potent PPAR alpha activator, and measured acyl CoA oxidase (ACO) mRNA levels. Rat ACO mRNA levels increased markedly in response to WY14643 while human and rabbit hepatocytes were unresponsive. Thus, although the PPAR alpha2/PPAR alpha1 mRNA ratio is low in rabbits, this species is not responsive to peroxisome proliferators. Further studies with male and female rats, which vary significantly in their response to peroxisome proliferators, showed little difference in the ratio of PPAR alpha2/PPAR alpha1 mRNA. These data suggest that selective PPAR alpha2 mRNA expression is not the basis for differential species or gender responses to peroxisome proliferators.

Elevated hepatic apolipoprotein A-I transcription is associated with diet-induced hyperalphalipoproteinemia in rabbits.

Past studies have shown that a high saturated fatty acid diet containing coconut oil elevates plasma HDL cholesterol and apolipoprotein A-I (apoA-1) in rabbits through a mechanism involving increased synthesis. We have extended those studies by investigating expression of the hepatic apolipoprotein A-I gene and other lipid related genes in that model. Rabbits fed a diet containing 14% coconut oil for 4 weeks showed HDL-C elevations of 170% to 250% over chow-fed controls with peak differences occurring at 1 week. Plasma apoA-I levels were also increased over this time frame (160% to 180%) reflecting the HDL-C changes. After 4 weeks, there were no differences in plasma VLDL-C or LDL-C levels in chow versus coconut oil-fed rabbits. Hepatic levels of apoA-I mRNA in coconut oil-fed animals were elevated 150% after 4 weeks compared to chow-fed controls; hepatic mRNA levels for ten other genes either decreased slightly (apoB, LCAT, hepatic lipase, albumin, ACAT, and HMG CoA reductase) or were unchanged (CETP, apoE, LDL-receptor, and acyl CoA oxidase). Nuclear run-on transcription assays revealed that coconut oil feeding for 4 weeks caused a 220% increase in hepatic apoA-I transcription rate compared to controls; no change was observed for CETP and apoE. Treatment of cultured rabbit liver cells with various saturated fatty acids and sera from chow-fed and coconut oil-fed rabbits did not alter apoA-I mRNA levels as observed in vivo. These data demonstrate that coconut oil elevates plasma HDL-C and apoA-I by increasing hepatic apoA-I transcription while expression of other genes involved in lipid metabolism are reduced or unchanged in response to coconut oil feeding.

A novel compound that elevates high density lipoprotein and activates the peroxisome proliferator activated receptor.

In the current studies we describe the effects of PD 72953 and related compounds on lipoprotein levels in chow-fed male rats. After 2 weeks, 10 mg/kg of PD 72953 daily was as effective as 100 mg/kg gemfibrozil for elevating HDL-cholesterol. At 100 mg/kg, PD 72953 further elevated HDL-cholesterol to 232% of control levels, and was associated with increased HDL size and plasma apoE (169% of control), despite no change in hepatic apoE mRNA. ApoA-I rose transiently (at 1 week), but by 2 weeks only apoE remained elevated. PD 72953 dose-dependently reduced plasma apoB, VLDL-cholesterol, LDL-cholesterol, and triglyceride. Hepatic apoC-III mRNA reduction parallelled triglyceride lowering. After 1 week, 30 and 100 mg/kg per day PD 72953 reduced plasma apo-CIII levels by 30 and 34%, and triglycerides by 60 and 83%, respectively. PD 72953 treatment had no effect on triglyceride production rates; however, 125I-labeled VLDL apoB disappearance was enhanced. We compared PD 72953 to a structurally similar diacid, PD 69405, that also reduced VLDL and LDL, but had no effect on HDL elevation. Compared to PD 72953, PD 69405 further accelerated 125I-labeled VLDL apoB disappearance, decreased triglyceride production, and elevated the ratio of post-heparin hepatic to lipoprotein lipase activity. Whole animal studies, transient transfection studies in HepG2 cells, and chimeric receptor studies in kidney 293 cells suggest that PD 72953 is a ligand for the peroxisomal proliferation activated receptor alpha (PPARalpha), and PPARgamma. Overall, PD 72953 may act through a peroxisomal proliferation activated receptor and result in plasma triglycerides and apoB-containing lipoprotein reduction, while also raising HDL cholesterol. Reduced apoC-III may allow triglyceride-rich remnants to more efficiently bind and present substrate to peripheral tissue lipoprotein lipase, and therefore allow enhanced shedding of remnant phospholipid surface for HDL production.

A cDNA-dependent scintillation proximity assay for quantifying apolipoprotein A-I.

We have developed a cDNA-dependent scintillation proximity assay (SPA) for rabbit apolipoprotein A-I that follows a classic radioimmunoassay scheme, in that antiserum and radiolabeled ligand are used in a process to quantify a source containing unlabeled ligand. To synthesize radiolabeled ligand we isolated a full-length rabbit apolipoprotein A-I (apoA-I) cDNA, transcribed the corresponding RNA in vitro, and synthesized radiolabeled apoA-I by including tritiated leucine in an in vitro translation reaction. Assay conditions were established which allowed quantification of unlabeled apoA-I over a range of 0.2 to 4 nanograms with intra- and interassay coefficients of variation of 5% and 10%, respectively. Purified rabbit apoA-I, apoA-I in rabbit liver parenchymal cell conditioned media, and apoA-I contained in rabbit plasma all generated parallel titration curves. Quantification of rabbit plasma apoA-I was not affected when sheep anti-rabbit apoA-I serum was mixed with sheep anti-rabbit apoB or apoE serum; thus, the antibody need not be specific to quantify the ligand of interest. To show utility of the assay, apoA-I mass was quantified in in vitro and in vivo models displaying altered apoA-I levels. In each model apoA-I values from the cDNA-dependent SPA and the established methodologies of Western blotting and electroimmunodiffusion were highly correlated. The approach outlined in this report should permit rapid development of scintillation proximity assays for other proteins given the widespread availability of full-length cDNAs.

Large versus small unilamellar vesicles mediate reverse cholesterol transport in vivo into two distinct hepatic metabolic pools. Implications for the treatment of atherosclerosis.

Phospholipid liposomes are synthetic mediators of "reverse" cholesterol transport from peripheral tissue to liver in vivo and can shrink atherosclerotic lesions in animals. Hepatic disposal of this cholesterol, however, has not been examined. We compared hepatic effects of large (approximately equal to 120-nm) and small (approximately equal to 35-nm) unilamellar vesicles (LUVs and SUVs), both of which mediate reverse cholesterol transport in vivo but were previously shown to be targeted to different cell types within the liver. On days 1, 3, and 5, rabbits were intravenously injected with 300 mg phosphatidylcholine (LUVs or SUVs) per kilogram body weight or with the equivalent volume of saline. After each injection, LUV- and SUV-injected animals showed large increases in plasma concentrations of unesterified cholesterol, indicating mobilization of tissue stores. After hepatic uptake of this cholesterol, however, SUV-treated animals developed persistently elevated plasma LDL concentrations, which by day 6 had increased to more than four times the values in saline-treated controls. In contrast, LUV-treated animals showed normal LDL levels. By RNase protection assay, SUVs suppressed hepatic LDL receptor mRNA at day 6 (to 61 +/- 4% of control, mean +/- SEM), whereas LUVs caused a statistically insignificant stimulation. Hepatic HMG-CoA reductase message was also significantly suppressed with SUV, but not LUV treatment, and hepatic 7 alpha-hydroxylase message showed a similar trend. These data on hepatic mRNA levels indicate that SUVs, but not LUVs, substantially perturbed liver cholesterol homeostasis. We conclude that LUVs and SUVs mobilize peripheral tissue cholesterol and deliver it to the liver, but to distinct metabolic pools that exert different regulatory effects. The effects of one of these artificial particles, SUVs, suggest that reverse cholesterol transport may not always be benign. In contrast, LUVs may be a suitable therapeutic agent, because they mobilize peripheral cholesterol to the liver without suppressing hepatic LDL receptor mRNA and without provoking a subsequent rise in plasma LDL levels.

Lack of correlation between ACAT mRNA expression and cholesterol esterification in primary liver cells.

A partial rabbit cDNA clone (14b) for ACAT has been characterized and used to demonstrate that hepatic and aortic ACAT mRNA14b abundance increased 2-3-fold in rabbits receiving a high fat/high cholesterol-diet compared to chow fed animals (Pape et al. (1995) J. Lipid Res. 36, 823-838). Because of those data we hypothesized that increased hepatic cholesteryl ester mass and synthesis rates in rabbit liver cells are associated with an increase in ACAT mRNA14b levels. To test this hypothesis we altered cellular cholesteryl ester mass and synthesis rates in primary parenchymal and nonparenchymal cells using various extracellular agents and measured the accumulated mass of ACAT mRNA14b. Parenchymal cells incubated with rabbit beta VLDL or mevalonolactone displayed a 6-10-fold increase in cellular cholesteryl ester mass over a three day treatment with no significant changes in cellular free cholesterol, triacylglycerols, or ACAT mRNA14b levels; HMG CoA reductase and LDL receptor mRNA mass decreased initially as a result of cholesteryl ester loading. Treatment of parenchymal cells with CI-976, an ACAT inhibitor, showed a marked reduction in cholesteryl ester synthetic rate compared to beta VLDL controls but displayed no change in ACAT mRNA14b levels. A mixed population of rabbit hepatic nonparenchymal cells was incubated with beta VLDL for 24 h in culture which resulted in a 6-fold increase in cellular cholesteryl ester mass; there was no change in ACAT mRNA14b levels. In an in vivo study, rabbits consuming a high fat/high cholesterol-diet for three weeks showed a 10-fold increase in hepatic cholesteryl ester with no significant changes in ACAT mRNA14b levels. Together these data indicate that rabbit liver cellular cholesteryl ester mass increases of up to 10-fold are not correlated with ACAT mRNA14b changes. Thus, hepatic ACAT mRNA14b expression and cellular cholesterol esterification do not appear to be coordinately regulated at this level of cholesteryl ester loading.

Tissue specific changes in acyl-CoA: cholesterol acyltransferase (ACAT) mRNA levels in rabbits.

A human cDNA clone (K1) was recently isolated that encodes functional acyl-CoA:cholesterol acyltransferase (ACAT) protein (Chang et al. J. Biol. Chem. 1993. 268: 20747-20755). We used the K1 clone to screen a rabbit liver cDNA library and isolated a 919 base pair partial rabbit cDNA (ACAT14b) that was greater than 90% homologous with the human nucleotide sequence. Northern blotting using the rabbit ACAT cDNA14b revealed the existence of at least six related mRNA species (ranging from 6.2 to 1.7 kb) in various rabbit tissues. Using an RNAse protection assay, ACAT mRNA14b was detected in twelve separate rabbit organs. Adrenal gland contained the highest concentrations of ACAT mRNA14b (per microgram of total RNA) being 20-, 30-, and 50-fold higher than small intestine, aorta, and liver, respectively. Additional studies with isolated liver cell populations revealed that rabbit hepatic nonparenchymal cells contained 30-fold more ACAT mRNA14b (per microgram of total RNA) than parenchymal cells. To determine whether ACAT mRNA14b levels are regulated in vivo, rabbits were fed for 4 weeks a high fat/high cholesterol diet (HFHC; 0.5% cholesterol, 3% coconut oil, 3% peanut oil) at which point they were either kept for an additional 4 weeks on the HFHC-diet or switched to the HFHC-diet plus CI-976 (50 mg/kg), a potent and specific ACAT inhibitor; another group of rabbits was fed a chow diet for the entire 8 weeks. The HFHC-diet caused a 2- and 3-fold increase in hepatic and aortic ACAT mRNA14b levels, respectively, in comparison to chow-fed animals; there was no change in adrenal or small intestine levels. CI-976 treatment lowered ACAT mRNA14b levels by 60% and 40% in liver and aorta, respectively, in comparison to the HFHC controls; again there was no change in adrenal or small intestine levels. These data indicate that ACAT mRNA14b levels increase in a tissue specific manner in response to dietary fat and cholesterol.

Rabbit liver apolipoprotein A-I synthesis is under nonparenchymal cell paracrine control.

Apolipoprotein A-I (apoA-I), the primary protein of high density lipoprotein, originates from intestine and liver of almost all mammalian species. In contrast to most species, intact rabbit liver is only capable of producing minute amounts of apoA-I mRNA and protein. In this report we demonstrate that purified rabbit hepatic parenchymal cells have apoA-I mRNA levels approximately 50-fold higher than intact liver after 48 h in monolayer culture. Investigations of the differential between in vivo and in vitro expression showed that conditioned media from nonparenchymal cells, a cell population essentially absent in parenchymal cell cultures, inhibited the elevation of apoA-I mRNA in a specific, concentration-dependent, and reversible fashion. Furthermore, at a concentration of nonparenchymal cell-conditioned media that inhibited apoA-I mRNA levels by > 80% compared to control, there were only slight changes in apoB, apoE, LDL receptor, LCAT, 7 alpha-hydroxylase, hepatic lipase, HMG-CoA reductase, and albumin mRNA levels. Metabolic labeling of parenchymal cell secreted proteins with [35S]methionine followed by apoA-I immunoprecipitation revealed that apoA-I synthesis and secretion corresponded to the changes observed for apoA-I mRNA. Initial biochemical characterization of the nonparenchymal cell media revealed that the inhibitory factor was > 30 kDa, heat-stable to 70 degrees C, and still active after urea denaturation and renaturation. These data suggest that, in rabbits, hepatic parenchymal-nonparenchymal communication in the form of a secreted factor may attenuate liver apoA-I expression in vivo.

Hepatic expression of genes regulating lipid metabolism in rabbits.

The liver plays a central role in lipid metabolism and plasma lipoprotein homeostasis. This dynamic process is regulated by a variety of liver-derived proteins. However, the specific liver cells that express these proteins are largely unknown. In the current study we measured mRNA levels for 13 genes encoding proteins involved in lipid metabolism in isolated rabbit hepatic parenchymal and nonparenchymal cells. For these analyses we cloned partial rabbit cDNAs for apolipoprotein A-I (apoA-I), apolipoprotein B (apoB), apolipoprotein E (apoE), cholesteryl ester transfer protein (CETP), hepatic lipase (HL), lipoprotein lipase (LPL), HMG-CoA reductase, LDL-receptor, 7 alpha-hydroxylase, albumin, bile salt-dependent cholesteryl ester hydrolase (CEH), lecithin:cholesterol acyl transferase (LCAT), and plasminogen activator inhibitor protein-1 (PAI-1). The cDNAs provided the basis for developing quantitative RNAse protection assays for each mRNA. These assays were used to determine whether differential patterns of mRNA expression existed between liver and other tissues and between hepatic parenchymal and nonparenchymal cells. The data demonstrate a diverse range in tissue distribution and mRNA abundance. Liver expressed all mRNAs except for LPL and CEH. Messenger RNA levels in isolated liver cell populations normalized to total RNA revealed a cell segregation pattern for hepatic gene expression: parenchymal cells showed higher levels of apoA-I, apoB, apoE, albumin, LCAT, HL, and 7 alpha-hydroxylase mRNAs compared to nonparenchymal cells while nonparenchymal cells showed higher levels of CETP, LDL-receptor, HMG-CoA reductase, and PAI-1 mRNAs compared to parenchymal cells. These data demonstrate the existence of differential mRNA expression patterns in rabbit liver cell populations for genes encoding proteins affecting lipid metabolism.

High-sulfur protein gene expression in a transgenic mouse.

We analyzed the effect of minoxidil on hair follicles isolated from transgenic mice. These transgenic animals synthesize the reporter enzyme CAT in their hair follicles only during the active phases of hair growth. The recombinant gene used to generate these mice contained the bacterial enzyme CAT under the control of the promoter from the gene of UHS protein. Studies using in situ hybridization showed that UHS proteins are expressed specifically in the matrix cells of the hair follicle during the terminal stages of hair differentiation. Hence the expression of the UHS proteins is a clear sign of active hair growth. With other in situ hybridization studies we demonstrated that CAT mRNA is expressed in differentiating matrix cells of the hair shaft in a location similar to that in which mRNA encodes UHS proteins. Thus we can use the levels of CAT activity as a measure of hair growth. We have confirmed that expression of the transgene is found in hair that is high in anagen and low in catagen follicles. The usefulness of our model was further demonstrated by showing that minoxidil, a drug that stimulates hair growth, increased the expression of CAT in cultured hair follicles. Thus we have demonstrated that expression of this reporter gene is sensitive, hair specific, and also useful for monitoring effects in cultured hair follicles. Hence these transgenic mice provide a model system for studying the biology of hair growth.

Hair-specific keratins: characterization and expression of a mouse type I keratin gene.

A genomic clone for a member of the mouse type I hair keratin protein family has been isolated and analyzed in order to study the regulation of this keratin during the hair growth cycle. The coding sequence is divided into seven exons. The gene structure is typical of keratins in particular and intermediate filaments in general in that the intron-exon borders are not located at the domain borders of the protein. Comparison with a sheep wool keratin gene shows that the splice sites in the two hair keratin genes are found at identical locations relative to the amino acid sequence of the proteins. Similarly, comparison of the promoter areas of these genes shows several areas of nucleotide sequence conservation, including the area around the TATA box and an SV40 core enhancer sequence. In addition, a high degree of sequence identity exists in the fourth intron. In situ hybridization shows that transcripts of this gene are first found in the relatively undifferentiated proximal cortex area in the keratogenous zone of mouse vibrissae.

Evidence that the nonparenchymal cells of the liver are the principal source of cholesteryl ester transfer protein in primates.

Previous studies showed that 90% or more of the cholesteryl ester transfer protein (CETP) mRNA is contained in the liver of cynomolgus monkeys. The purpose of this study was to determine if the parenchymal cells (hepatocytes) were the hepatic cell type that contained that mRNA. The parenchymal and nonparenchymal cells were separated by standard methods, and the CETP, apoA-I, apoB, and apoE mRNA content of the preparation determined at each step in the purification process. ApoA-I and apoB are produced only in the parenchymal cells; apoE is produced by both cell types. The mRNA measurements showed that the CETP mRNA: apoA-I mRNA and the CETP mRNA: apoB mRNA ratios were more than 2500-fold greater in the nonparenchymal cell preparation than in the starting material, and that the purified parenchymal cell fraction was virtually devoid of CETP mRNA. In situ hybridization studies showed that, whereas the apoA-I mRNA signal was evenly distributed over the tissue section, the CETP mRNA signal was associated with the hepatic sinusoids, suggesting that it was the hepatic sinusoidal cells that were principally responsible for the high CETP mRNA levels in the liver. We conclude that the nonparenchymal cells are the principal source of CETP in the cynomolgus monkey.

Localization of TIMP in cycling mouse hair.

TIMP (tissue inhibitor of metalloproteinase) is a glycoprotein inhibitor of metalloproteinases that we hypothesize to be involved in the tissue remodeling that occurs during each hair growth cycle. We examined this hypothesis by studying the expression of TIMP at selected times during a single hair cycle using TIMP-lacZ transgenic mice to localize TIMP gene activity in the hair follicle. TIMP gene induction was visualized by staining mouse back skin for beta-galactosidase (beta-gal) activity. Paraffin sections were analyzed for the localization of TIMP expression. TIMP gene activation appears in hair follicles only during the mid-anagen (the growing stage of the hair cycle) primarily in Henle's layer of the inner root sheath. Some expression of TIMP is also seen in a few connective tissue cells, in the sebaceous gland and in cells at the proximity of the dermal papilla cells in catagen (regressing) and telogen (resting) follicles. These results are consistent with a role for TIMP in cyclic remodeling of connective tissue in hair follicles.

Hair-specific expression of chloramphenicol acetyltransferase in transgenic mice under the control of an ultra-high-sulfur keratin promoter.

We have generated a transgenic mouse line by microinjection of a chimeric DNA fragment (KER-CAT) containing a hair-specific, murine ultra-high-sulfur keratin promoter (KER) fused to the coding region of the bacterial chloramphenicol acetyltransferase (CAT) gene. A 671-base pair (bp) stretch of the 5' promoter region was used to direct the expression of the CAT gene in this construct. Of the tissues tested for CAT activity in these transgenic animals only skin with growing hair, isolated hair follicles, and microdissected vibrissae showed substantial levels of activity. These are the same tissues where the endogenous ultra-high-sulfur keratin gene is expressed as shown by in situ hybridization. Furthermore, analysis of the CAT activity during the developmental stages of the hair growth cycle shows that the chimeric gene is expressed during the anagen phase of the hair growth cycle; this is the expected time during development for its expression. From these results we conclude that 671 bp of the promoter sequence from the ultra-high-sulfur keratin gene is sufficient to direct the correct development-specific and tissue-specific expression of the reporter gene construct in transgenic mice. The appropriate expression of the KER-CAT construct in transgenic mice is an important step in understanding the regulation of this gene during hair organogenesis.

Mapping and sequence of the gene for the pseudorabies virus glycoprotein which accumulates in the medium of infected cells.

RNA from pseudorabies virus (PRV)-infected cells was translated in a reticulocyte lysate with and without the addition of dog pancreas microsomes. Upon addition of the microsomes to the translation reaction, an additional prominent protein product was observed that was not present when microsomes were omitted. The gene coding for this processed protein and its lower-molecular-weight precursor was mapped within the small unique region of the genome by hybridization of mRNA to cloned fragments of PRV DNA and translation of the selected mRNAs. A fragment of the coding region of this gene was inserted into an open reading frame cloning vector to express part of this gene as a hybrid protein in Escherichia coli. This hybrid protein was injected into mice to raise an antiserum which was found to precipitate the glycoprotein which accumulates in the medium of PRV-infected cells. This allows us to conclude that the gene for the "excreted" glycoprotein (gX) maps to the small unique region of the genome, and that the precursor of this glycoprotein is readily processed by dog pancreas microsomes. The region of the PRV genome which codes for this glycoprotein was sequenced and found to include an open reading frame coding for 498 amino acids, flanked by sequences which contain features common to eucaryotic promoters and polyadenylation signals. The predicted protein sequence includes a hydrophobic sequence at the N-terminus which could be a signal sequence, and a hydrophobic sequence followed by a hydrophilic sequence at the C-terminus.

Three sizes of subunits in RNAs from feline sarcoma-leukaemia virus mixtures.

RNA from the Snyder-Theilen feline sarcoma-leukaemia virus complex (ST-FeSV-FeLV) sedimented in a double-peaked band between 50 and 70S, but Gardner-Arnstein (GA) FeSV-FeLV RNA sedimented in a single 70S peak. FeLV isolated from the ST virus mixture contained RNA which sedimented in a 70S band like GA-FeSV-FeLV RNA, but F422 FeLV RNA sedimented more slowly, at 50 to 60S. After thermal denaturation, resedimentation revealed three classes of RNA subunits in ST-FeSV-FeLV RNA: the first class, 35 to 37S, was also found in ST-FeLV and other FeLVs (except F422 FeLV), in the endogenous feline virus, RD114 and in GA-FeSV-FeLV; the second class, 32 to 34S, was similar to subunits in F422 FeLV and minor components of GA-FeSV-FeLV and ST-FeLV; the third class, 25S, was detected only in ST-FeSV-FeLV RNA. Electrophoresis of RNA species in buffered formamide provided evidence that the three classes of RNA subunits distinguishable on the basis of sedimentation rates actually represent three size classes of subunits. The ST virus mixture was shown to contain about equal titres of infectious FeLV and transforming FeSV whereas GA-FeSV-FeLV had at least a 10-fold excess fo FeLV over FeSV. These observations are discussed in terms of possible origins of the three sizes of FeSV-FeLV RNA subunits and their relationships to three species of FeSV-FeLV proviral DNA described recently (Sherr et al. 1979).