The ingestion of plant proteins compared with animal proteins generally results in lower plasma cholesterol concentrations in various animal species, including humans (Carroll 1995, Kritchevsky 1993, Sirtori et al. 1993, Sugano and Koba 1993). Two explanations have so far been provided for the different effects of plant and animal proteins; one is based on a difference in the physicochemical properties of dietary proteins or luminal digestion products and the other is based on differences in the amino acid composition of dietary proteins. The mechanism of the former is that plant proteins suppress the intestinal absorption or reabsorption of neutral and acidic steroids, and thereby decrease the pool size of cholesterol within the body. This resembles one of the mechanisms responsible for the hypocholesterolemic action of certain types of dietary fibers. In contrast, the manner in which amino acid composition exerts effects is not fully understood, although specific amino acids of dietary proteins are thought to affect, either directly or indirectly, the metabolism of cholesterol.

Recently we provided evidence that the alteration of liver microsomal phospholipid profile, as represented by the decrease in the ratio of phosphatidylcholine (PC)³ to phosphatidylethanolamine (PE), might be associated with the hypocholesterolemic action of some dietary proteins, including soybean protein, in rats fed a cholesterol-free diet (Sugiyama et al. 1996). There was a significant positive correlation between plasma cholesterol concentration and the PC:PE ratio (Sugiyama et al. 1996). The decrease in the PC:PE ratio of liver microsomes likely arose from a depression of PE N-methylation due to lower methionine contents of these proteins. However, it is still unclear how the decreased PC:PE ratio of liver microsomes is associated with a reduction in plasma cholesterol. Several groups have reported that the fatty acid composition of plasma PC, a major phospholipid class of plasma lipoproteins, can be significantly affected by the type of dietary protein (Huang et al. 1986, Sugano et al. 1988). In this regard, it is interesting that the uptake rate of reconstituted HDL cholesterol by perfused rat livers was most stimulated by 16:0-18:2 PC of the five PC molecular species used (16:0-18:2, 16:1-16:1, 18:0-18:2, 18:1-16:0 or 20:1-20:1) for the reconstitution of HDL (Kadowaki et al. 1993). These findings indicate that certain types of hypocholesterolemic substances, including some dietary proteins, may modify the plasma PC molecular species profile and thereby accelerate the uptake of plasma lipoprotein cholesterol by tissues. In support of this, we recently reported that the molecular species composition of plasma PC can be markedly influenced in rats fed a cholesterol-free diet by dietary eritadenine, a hypocholesterolemic factor of Lentinus edodes mushroom (Sugiyama and Yamakawa 1996). However, except for a report by Koba et al. (1994), little information is available concerning the effect of dietary protein types on plasma phospholipid molecular species composition. Although methionine has been postulated to be one of the amino acids which participate in the differential effects of dietary proteins on plasma cholesterol concentration, it is unclear whether the methionine content of dietary proteins also affects fatty acid metabolism or the phospholipid molecular species composition of plasma lipoproteins.

In this study, the effects of dietary casein, dietary soybean protein and methionine supplementation were investigated to determine whether the type and methionine content of dietary proteins can modify the molecular species profile of plasma PC, as well as the liver microsomal PC:PE ratio, and to obtain insights into the relationship between the plasma PC molecular species composition and the hypo- or hypercholesterolemic action of dietary proteins in rats fed a cholesterol-free diet.

MATERIALS AND METHODS

Three different diets were fed in this study: casein diet, soybean protein diet and methionine-supplemented soybean protein diet (Table 1). The soybean protein used was relatively pure according to the supplier (Fuji Oil, Osaka, Japan), having a concentration of crude fats of 3 g/kg or less. Based on information from the protein suppliers, casein (Nacalai Tesque, Kyoto, Japan) and soybean protein diets were estimated to contain 6.4 and 3.0 g methionine/kg, respectively. Therefore, the third diet was supplemented with L-methionine at a concentration of 3.4 g/kg diet to adjust the methionine concentration to equal that of the casein diet. In Experiment 1, 24 rats were divided into three groups of 8 rats (each with similar mean body wt) and given free access to the three experimental diets and water for 2 wk. In Experiment 2, the time-dependent effect of switching rats from the casein diet to the soybean protein diet was investigated. Forty-eight rats were given free access to the casein diet for 2 wk, and 8 of them were killed on the final day of 2-wk period (d 0). Of the remaining 40 rats, 8 continued to be given free access to the same diet and were killed after 5 d. The other 32 rats were given free access to the soybean protein diet and were killed in groups of 8 on the 1st, 2nd, 3rd and 5th d. The body weight and food consumption of rats were measured daily. The experimental design was approved by the Laboratory Animal Care Committee of the Faculty of Agriculture, Shizuoka University.

Table 1. Composition of the experimental diets [View Table]

For the determination of phospholipid class composition, the phospholipids in the extract of liver microsomes were separated by TLC on silica gel 60 plates (Merck, Darmstadt, Germany) using chloroform:methanol:water (65:25:4, v/v/v) as a developing solvent. The bands of each phospholipid class were visualized with iodine vapor, scraped off the plate, and analyzed directly for inorganic phosphorus (Bartlett, 1959). For the determination of fatty acid and molecular species composition, plasma and liver microsomal PC were likewise separated by TLC, visualized with dichlorofluorescein, and extracted with chloroform:methanol (1:2, v/v). A portion of PC extract was methylated with BF₃:methanol reagent (14:86, wt/wt, Wako). The fatty acid methyl esters formed were extracted with hexane and analyzed by GLC (Model GC-17A; Shimadzu, Kyoto, Japan) using a TC-FFAP capillary column (0.25 mm × 30 m; GL Sciences, Tokyo, Japan). Another portion of PC was converted to diacylglycerol benzoates according to the method of Blank et al. (1984). The diacylglycerol benzoates were separated into each molecular species by HPLC (Model LC-6A; Shimadzu) using an ODS column (4.6 mm x 250 mm; Merck) essentially according to the method of Blank et al. (1984). Since some peaks consisted of two molecular species, the ratio of two molecular species was determined by GLC. A representative HPLC chromatogram of PC molecular species was shown previously (Sugiyama and Yamakawa 1996).

The hepatic concentrations of S-adenosylmethionine (SAM) and S-adenosylhomocysteine (SAH) were estimated essentially according to the method of Cook et al. (1988) with slight modifications as described previously (Sugiyama et al. 1995). The hepatic concentrations of free amino acids were measured with an amino acid autoanalyzer (Model 835; Hitachi, Tokyo, Japan) as described previously (Sugiyama et al. 1996).

RESULTS

Table 2. Effects of dietary protein types and methionine supplementation on plasma and liver lipid concentrations, on liver microsomal phospholipid composition, and on the concentrations of methionine and its metabolites in the liver of rats (Experiment 1)1 [View Table]

The proportion of linoleic acid in both plasma and liver microsomal PC was significantly higher in rats fed the soybean protein diet than in rats fed the casein diet, and this increase by soybean protein was significantly suppressed by methionine supplementation (Table 3). Conversely, the proportion of arachidonic acid in plasma and liver PC was significantly lower in rats fed the soybean protein diet than in rats fed the casein diet, and this decrease was also effectively suppressed by methionine supplementation. Consequently, the ratio of arachidonic acid to linoleic acid was significantly affected by dietary protein types and methionine supplementation. Of the minor fatty acids, 22:5(n-6) also clearly changed in a manner similar to that of arachidonic acid.

Table 3. Effects of dietary protein types and methionine supplementation on the fatty acid composition of plasma and liver microsomal phosphatidylcholine in rats (Experiment 1)1 [View Table]

In plasma PC, both 16:0-18:2 and 18:0-18:2 molecular species were significantly greater in rats fed the soybean protein diet compared with those fed the casein diet, and these increases were effectively suppressed by methionine supplementation (Table 4). In liver microsomal PC, the effect of soybean protein diet on 16:0-18:2 molecular species was greater than the effects on other molecular species containing 18:2 in the sn-2 position. In both plasma and liver microsomal PC, the effect of dietary protein types and methionine supplementation was greater on 18:0-20:4 molecular species than the effects on other molecular species containing 20:4 in the sn-2 position.

Table 4. Effects of dietary protein types and methionine supplementation on the molecular species composition of plasma and liver microsomal phosphatidylcholine in rats (Experiment 1)1 [View Table]

DISCUSSION

At present, the mechanism by which methionine exerts its hypercholesterolemic action is not fully understood. Previously we reported that the PC:PE ratio of liver microsomes could be influenced by the type of dietary proteins, and the ratio has significant positive correlations with both plasma cholesterol concentration and the methionine content of dietary proteins in rats fed a cholesterol-free diet (Sugiyama et al. 1996). These results suggest that dietary methionine might affect the plasma cholesterol concentration through altered phospholipid metabolism. It has also been shown that fatty acid composition of plasma and liver microsomal PC could be influenced by the type of dietary proteins (Huang et al. 1986, Sugano et al. 1988). The suppression of linoleic acid metabolism by soybean protein was attributable to the decrease in liver microsomal 6-desaturase activity, the rate-limiting step in the metabolism of linoleic acid (Choy et al. 1988). Although 6-desaturase activity was not measured in this study, the results obtained here can be explained essentially in terms of alterations in linoleic acid metabolism. The most striking finding here is that in addition to the liver microsomal PC:PE ratio, the fatty acid and molecular species compositions of PC were sensitive to dietary methionine level. These results suggest a correlation between the PC:PE ratio of liver microsomes and the metabolism of linoleic acid. In support of this, it has been observed that the decrease in liver microsomal PC:PE ratio was coupled with a decrease in 6-desaturase activity in rats fed various types of diets, such as PE-supplemented (Imaizumi et al. 1989), vitamin B-6 deficient (She et al. 1994) and eritadenine-supplemented (unpublished data). Furthermore, the PC:PE ratio of liver microsomes apparently decrease prior to the alteration of fatty acid and molecular species composition of plasma PC in response to diet switch (Figs. 2 and 3), indicating that the decrease in the PC:PE ratio was a cause, rather than an effect, of altered fatty acid metabolism. Thus, it is likely that dietary methionine concentration affects linoleic acid metabolism and plasma PC molecular species composition primarily through alteration of the liver microsomal PC:PE ratio, finally leading to an increase in the plasma cholesterol concentration.

Phosphatidylcholine is synthesized by both the CDP-choline pathway and the PE N-methylation pathway in the liver of rats (Tijburg et al. 1989), and PC synthesis by these two pathways can be stimulated by dietary choline and methionine, respectively. In this study, PC synthesis by the CDP-choline pathway likely was fully functional in all groups, since the diet contained an adequate amount of choline chloride (4 g/kg). On the contrary, PC synthesis by the PE N-methylation pathway is presumably reduced in rats fed the soybean protein diet because liver microsomal PE concentration was significantly enhanced in these rats. This assumption is based on the idea that microsomal PE concentration is regulated mainly by PE N-methylation, rather than PE synthesis, although the depression of PE N-methylation stimulates PE synthesis via the CDP-ethanolamine pathway. In support of this, other treatments to depress PE N-methylation (e.g., dietary suplementation with eritadenine) have been shown to increase the concentration of liver microsomal PE (Sugiyama et al. 1995). The depression of PE N-methylation by dietary soybean protein can be ascribed to the decrease in hepatic SAM concentration, since the activity of PE N-methyltransferase is regulated by both of the enzyme substrates, SAM and PE (Tijburg et al. 1989). The hepatic SAM concentration is thought to reflect the methionine level of diet. Consistent with this, the present study confirmed that the hepatic SAM concentration was sensitive to dietary protein type and methionine supplementation. In addition, the parallel time-dependent changes in hepatic SAM concentration and liver microsomal PE concentration or PC:PE ratio (Fig. 2) also support the direct association of these variables. Thus, the decreased PC:PE ratio of liver microsomes due to feeding the soybean protein diet may be attributable to the low methionine content of the protein. The results obtained here also support the concept that the liver microsomal PC:PE ratio is mainly regulated by the PE N-methylation pathway, rather than the CDP-choline pathway, when the diet contains adequate amounts of choline.

Two processes are thought to participate in the hypocholesterolemic action of soybean protein: depressed secretion of lipoprotein cholesterol from the liver into the blood circulation (Sugano et al. 1982) and increased uptake of plasma lipoprotein cholesterol by tissues, including the liver (Khosla et al. 1991). A number of reports have shown that soybean protein enhances the excretion of neutral and acidic steroids into feces. This can contribute to both the processes. It has been shown by Yao and Vance (1988) that methionine stimulates the secretion of VLDL from cultured rat hepatocytes under the condition of choline deficiency. Based on these findings, they postulated that active synthesis of PC is required for the assmbly and secretion of VLDL. However, it seems unlikely that dietary methionine level also affects VLDL secretion even under the condition of adequate dietary choline level, since PC synthesis via the CDP-choline pathway is regulated through a feedback inhibition by PC. As for the latter process, Sirtori et al. (1984) have shown that soybean protein increased the activity of the lipoprotein receptor for -VLDL in liver cells of rats fed a cholesterol-enriched diet. It is not known whether the activity of lipoprotein receptors can be affected by dietary methionine level. However, the possibility that the methionine content of dietary proteins affects the phospholipid profile of liver cell membranes, where receptors exist, and thereby leads to an enhancement of lipoprotein receptor activities, cannot be excluded.

In addition to lipoprotein receptor activities of tissues, the nature of lipoproteins is thought to influence the uptake rate of plasma lipoprotein cholesterol by tissues. Kadowaki et al. (1993) have shown that the uptake rate of reconstituted HDL cholesterol by perfused rat livers was largely influenced by the difference in PC molecular species used for the reconstitution of HDL. They demonstrated that the uptake rate of cholesteryl oleate of reconstituted HDL by perfused livers was most stimulated by 16:0-18:2 PC of the five molecular species tested (16:0-18:2, 16:1-6:1, 18:0-18:2, 18:1-16:0 or 20:1-20:1), and that 16:0-18:2 PC was also the most hydrolyzed by hepatic lipase in vitro. These results were explained in terms of a critical role of hepatic lipase in the uptake of HDL cholesterol by the liver. Since hepatic lipase has phospholipase A₁ activity, it can hydrolyze the surface phospholipids of plasma lipoproteins. The hydrolysis of HDL phospholipids is thought to be necessary for the subsequent uptake of HDL constituents by the liver (Kadowaki et al. 1992). Furthermore, it has been suggested that hepatic lipase also participates in the uptake of other lipoproteins, such as remnants of chylomicrons and VLDL (Ji et al. 1994, Shafi et al. 1994). These facts suggest that the phospholipid molecular species composition of certain plasma lipoproteins might have an important role in the regulation of plasma cholesterol concentration through an influence on the process of lipoprotein cholesterol uptake by the liver. The work presented here demonstrate that the proportion of 16:0-18:2 molecular species in plasma PC was significantly enhanced by feeding the soybean protein diet, and that this enhancement was effectively suppressed by methionine supplementation. Therefore, the alteration of plasma PC molecular species composition might be associated with the methionine-dependent hypo- or hypercholesterolemic effect of soybean protein or casein, respectively, at least in rats fed a cholesterol-free diet.

Although plasma PC molecular species consist of approximately 20 molecular species, these can be divided into two groups based on fatty acid chain length in the sn-2 position. One group includes molecular species which contain 18:1 or 18:2 in the sn-2 position and another group includes molecular species which contain 20:4, 22:5 or 22:6. The proportion of each molecular species of the former group was enhanced or unaffected by soybean protein, and the proportion of each molecular species of the latter group was lowered or unaffected (Table 4). The changes over time in response to the diet switch of the sum of PC molecular species which contain 18:1 or 18:2 in the sn-2 position was found to be quite similar to that of plasma cholesterol concentration, although the direction of changes was reversed (Fig. 1, panel A and Fig. 3, panel F); there was a significant negative correlation between the two variables (r = 0.992, P < 0.001, n = 48). These results support the hypothesis that the plasma PC molecular species profile is associated with the regulation of plasma cholesterol concentration. Contrary to the results presented here, Koba et al. (1994) reported that in mice, dietary protein types (casein and soybean protein) had little or no effect on the liver microsomal PC:PE ratio and molecular species composition of PC, although the data for plasma cholesterol concentration were not provided. The response of phospholipid metabolism to dietary protein types may be different between rats and mice.