Studies on persisting effects of soy-based compared with amino acid-supplemented casein-based diet on protein metabolism and oxidative stress in juvenile pigs

Juvenile growing pigs were studied to explore whether a soy-based diet can induce persistent physiological alterations, especially in protein and energy metabolism, nutrient oxidation and redox homeostasis. In former studies we have shown that in juvenile pigs chronically fed protein diets based on either casein (CAS) or soy protein isolate (SPI), the SPI diet significantly decreases growth rate and increases oxidative stress responsiveness as compared to CAS. In addition, here we show that chronic feeding of SPI vs. CAS diet decreases whole body protein synthesis (WBPS) (p = 0.007) and hepatic gene expression associated with protein synthesis. To study persistent SPI effects, a three-period feeding experiment was designed: In the test group 18 pigs received the CAS diet for 24 days (period 1), followed by 31 days on the SPI diet (period 2) and further 31 days on the CAS diet (period 3). In the control group 18 pigs were fed the CAS diet throughout the three periods (86 days). Temporary consumption of SPI diet results in persistent changes of protein metabolism and oxidative stress responsiveness. After switching back from SPI to CAS diet the decrease of WBPS of the test group vs. control group was of borderline significance (p = 0.061), transcript levels of hepatic gene expressions of leucine aminopeptidase, endopeptidase 24.16, glutathione-S-transferase and peptide methionine sulfoxide reductase were increased. In liver tissue, total glutathione was increased and thiobarbituric acid reactive substances were decreased in the test vs. control group. In conclusion, results suggest that SPI-induced changes in protein and amino acid metabolism as well as in redox homeostasis and antioxidative potential in growing pigs persist 4 weeks after the cessation of SPI feeding.

Keywords: Hepatic gene expression; oxidative stress; pigs; soy protein isolate; whole body protein synthesis

Soy protein-based diets have been demonstrated to possess health-promoting potential (Cordle [9]; Badger et al. [4]; Colacurci et al. [8]), which are believed to be mainly caused by non-nutritive components, in particular, isoflavones (Vitolins et al. [42]). Studies in growing pigs (Beyer et al. [5]; Junghans et al. [18]) and rats (Klein et al. [24]) fed highly purified soy protein isolate (SPI) as compared to casein (CAS) diet have shown that the amino acid imbalance in SPI leads to decreased protein retention and growth retardation. Previously we have shown that SPI-based diet modifies hepatic gene and protein expression associated with protein biosynthesis and oxidative stress responsiveness in growing pigs (Schwerin et al. [40]; Junghans et al. [19]). Thus, exposure of infants to soy protein-based diet might also have adverse long-term effects due to methionine and lysine deficiency (Friedman & Brandon [13]; Mendez et al. [33]). In order to assess the benefits and risks of soy-based diets it is important to study metabolic long-term effects. Therefore, the purpose of the present work was to examine whether feeding of SPI vs. CAS-based diet to growing juvenile pigs would induce persistent physiological alterations, such as whole body protein synthesis, nutrient oxidation and redox homeostasis.

The experimental protocol was approved by the Ethics Committee of the Ministry of Nutrition, Agriculture, Forestry, and Fishery, Schwerin, State Mecklenburg-Vorpommern, Germany (No. VI 522a-7221.31-1-018/99).

Fifty-four male, castrated pigs (German Landrace) were weaned at 28 days (Exp. 1) or at 21 days of life (Exp. 2) and reared on a starter diet "Start 2" (crude protein 21.1%, crude fat 5.3%, crude fibre 4%, metabolizable energy 15.6 MJ/kg) (Trede and von Pein, Itzehoe, Germany) for 4 weeks (Exp. 1) or 3 weeks (Exp. 2). In Exp. 1, 18 pigs (mean body weight: 16 kg) consumed a casein-based control diet (CAS, Deutsches Milchkontor GmbH, Hamburg, Germany) for 24 days. Nine pigs were switched to a soy protein isolate-based diet (SPI, SUPRO® 1610, Interfood, Bad Homburg, Germany), which was fed for another 31 days while the other nine pigs continued on the CAS diet as the control group (Table I).

Table I. Feeding regimen and study design of growing pigs fed soy protein isolate (SPI) and casein (CAS)-based diets.

Exp. 2 was designed to investigate potential persisting dietary effects of the SPI diet after pigs had been switched back to the CAS diet. In this experiment, 36 pigs (mean body weight: 11 kg) were fed a CAS diet in period 1 for 24 days. In period 2, pigs were randomly assigned to either the CAS or the SPI diet which were fed for the next 31 days. In the third period, 18 pigs in each group were either continued on the CAS diet or switched back from SPI to CAS diet (Table I). CAS and SPI were the sole protein sources of the diets. Both diets were isoenergetic and isonitrogenous. Semi-synthetic experimental diets were used, which contained wheat starch, protein source (CAS or SPI), sucrose, fat, cellulose, minerals and vitamins (Table II). Protein and ME of the rations amounted to 9% w/w and 1800 kJ · kg BW−0.62 · d−1 in Exp. 1, and 9% w/w and 1430 kJ · kg BW−0.62 · d−1 in Exp. 2, respectively. In the experiments dry matter intake was adapted depending on the live weight. The pigs were pair-fed to consider the different weight gain in both feeding groups. Content of nutrients, minerals and vitamins were already reported earlier (Junghans et al. [18]). The protein components of the diets are characterized by different amino acid patterns (Table III).

Table II. Composition of the experimental diets and energy content.

Table III. Amino acid contents [g/16 g N] of proteins in the experimental diets.

The CAS diet was supplemented by methionine, threonine and tryptophan. Protein supply was set to 50% of the recommendations of the Gesellschaft für Ernährungsphysiologie der Haustiere (GEH) (Anonymous [2]). As shown in previous studies, this protein level was appropriate to induce significant metabolic differences between SPI and CAS (Junghans et al. [21]). During the experimental feeding pigs were housed in metabolic cages. Pigs were fed once a day (at 07:00 h). On the days of tracer administration pigs were only supplied with water.

Four open-circuit respiration chambers as described by Nehring et al. ([34]) were used for the determination of heat production from measurement of CO2 production, O2 consumption and urinary nitrogen loss. The airflow through the chamber (1.8 m3) was adjusted to about 1500 l/h. Thus, in respiration experiments the CO2 concentration reached about 0.6 to 0.8% (v/v). The temperature in the chambers was kept at 25 ± 2°C for pigs up to 20 kg and at 24 ± 2°C for pigs weighing more than 20 kg. The relative humidity was 60 – 70%. The light:dark cycle was set to 12:12 h (from 07:00 – 19:00 h).

Six days prior to the tracer administration jugular vein and carotid artery catheters (prepared from Silastic® tubing 602 – 285, Dow Corning, Aromando Medizintechnik GmbH, Düsseldorf, Germany; length 80 cm, inner×outer diameter 1.57×3.18 mm) were placed for tracer infusion and blood sampling, respectively. During the 4-week experimental period catheters were maintained patent by a constant infusion (0.7 ml/h) of physiological saline containing heparin (10 000 IU/l; heparin-sodium, B. Braun Melsungen AG, Melsungen, Germany) and sulfonamide (2 ml/l; Sulfadimidin pro injection as sodium salt, Serumwerk Bernburg AG, Bernburg, Germany) using syringe pumps (Model 22, Harvard Apparatus, South Natick, Massachusetts, USA).

During the adaptation period (CAS feeding) pigs were adapted to the respiration chambers for 1 week. Tracer studies were performed on day 51 in Exp. 1 and day 79 in Exp. 2. The pigs were slaughtered on day 58 and day 86 in Exp. 1 and Exp. 2, respectively. Twenty-four hours after feeding, at 7:00 h, pigs received an intravenous bolus injection of [1-13C]glycine (0.22 mmol/kg BW, 99 atom %13C) and [15N]glycine (0.11 mmol/kg BW, 99 atom %15N) via the jugular vein. For the correction of the incomplete 13C release in breath CO2, on day 53 in Exp. 1 and day 81 in Exp. 2 the same pigs received an intravenous bolus injection of NaH13CO3 (0.06 mmol/kg BW, 99 atom %13C) 24 h after feeding (Table I).

The 13C and 15N tracer compounds were purchased from Campro Scientific Ltd. (Berlin, Germany). For determination of the 13C abundance in breath CO2 gas samples were drawn from the respiration chamber by means of a membrane pump at −10, 10, 30, 60, 80, 100, 120, 160, 190, 250, 360, 480, 720 and 1440 min after injection of the 13C tracer as reported previously (Junghans et al. [20]). Gas samples were transferred into breath bags (Tegobag®, Tesseraux, Bürstadt, Germany) for the measurement of 13C/12C ratios in breath CO2 by means of isotope ratio mass spectrometry (DELTA Plus XL, Thermo Quest, Bremen, Germany) coupled with the Gas Bench II (Finnigan, Bremen, Germany). Twenty-four hours urine samples were collected from the respiration chamber in 5-l plastic vessels containing 10 ml hydrochloric acid (6 N) at −1, 1, 2 and 3 days after injection of the 15N tracer (Junghans et al. [20]) and stored at −20°C. The 15N/14N isotope ratio was measured in urinary nitrogen by isotope ratio mass spectrometry (Delta S, Finnigan MAT, Bremen, Germany) after combustion of urine samples by an elemental analyzer (Carlo Erba, Milan, Italy). For the combustion, about 50 µl urine were weighed in tin capsules.

Twenty four hours fasting blood samples were taken on day 52 in Exp. 1 and on day 80 in Exp. 2. Blood was collected on ice before centrifugation at 4°C at 1450 g for 15 min for obtaining plasma which was stored at −20°C.

The concentrations of free amino acids in blood plasma were measured after deproteinization with sulfosalicylic acid by ion-exchange chromatography Biochrom 20 (Pharmacia LKB, Cambridge, UK) using lithium buffer and physiological amino acid standard solutions (Sigma, no. A9906) and with an amino acid analyzer. Standard methods were used for the analysis of plasma urea and glucose (Sigma Diagnostics, catalogue no. 535 and 315, respectively).

The insulin and glucagon concentrations were determined by RIA from LINCO (St Charles, MO, USA). The insulin test (PI-12K) applied on 100 µl duplicates of plasma samples shows 100% cross-reactivity with porcine insulin. The limit of detection was 9.3 pmol/l (human standard). The glucagon test (GL-32K) is specific for pancreatic glucagon. The limit of detection is 5.7 pmol/l (100 µl sample size). Plasma IGF-I concentrations were determined using a competitive electrochemiluminescence (ECL) immunoassay modified by the use of receptor grade recombinant hIGF-I (GPB, Mediagnost, Reutlingen, Germany) (Rehfeldt et al. [37]).

For measurements of thiobarbituric acid reactive substances (TBARS), reduced glutathione (GSH) and mRNA expressions hepatic tissue was taken after slaughtering of 24-h fasted pigs. Pigs were killed by means of a bolt shooting device, tissue samples were taken, immediately frozen and stored at −80°C until measurements.

TBARS were used to evaluate the stability of the liver lipids against stimulated lipid peroxidation (Buege & Aust [7]). TBARS were expressed in terms of malondialdehyde (MDA). To stimulate lipid peroxidation, 2 ml of liver homogenate were incubated in 0.1 mM ascorbate and 5 µM FeSO4 at 37°C. Aliquots (0.5 ml) were withdrawn and pipetted into 0.25 ml of 20% trichloroacetic acid in 100 mM KCl immediately (0 min) and after 60 min incubation. These samples were centrifuged at 10 000 g for 10 min and 0.5 ml of the supernatants were mixed with 0.5 ml thiobarbituric acid (0.67%) and boiled for 15 min in a water bath. The absorbance at 535 nm was determined immediately after cooling. Standard MDA solution was prepared by hydrolysis of 1,1,3,3-tetraethoxypropane and the results were expressed as nmol MDA/mg homogenate protein.

The measurement of reduced thiol (GSH, cysteine) concentrations in liver tissue samples was performed using monobromobimane (Petzke et al. [35]). In brief, frozen liver tissue was powdered in liquid N2 and aliquots (0.05 – 0.1 g) were added to 1 ml of Tris-HCl buffer. The samples were vortexed and 100 µl of 1.97 mol/l sulfosalicylic acid was added. Liver tissue extract (10 µl) was mixed with 580 µl of 25 mmol/l Tris-HCl buffer (pH 8.0 at 20°C) and with 10 µl of 20 mmol/l monobromobimane in acetonitrile. The derivatization mixtures (20 µl) were measured by HPLC.

Total RNA was extracted from liver samples using the RNeasy Total RNA Kit (Quiagen, Hilden, Germany). Synthesis of first strand cDNA was performed with MMLV-RT (Promega, Madison, USA) and random hexamer primers using 2 μg total RNA. Quantitative analysis of PCR products was carried out in the LightCycler® (Roche, Mannheim, Germany) according to optimized PCR protocols and LightCycler DNA Master SYBR Green I® (Roche, Mannheim, Germany). Quantitative RT-PCR was carried out essentially as described by Schwerin et al. ([40]). Fluorescence signals, which were recorded on-line during amplification, were subsequently analysed using the second derivative maximum method of the Light Cycler Data analysis software. Copy numbers were calculated relative to the amount of total RNA. The 18S RNA abundance was measured to normalize for equal RNA amounts (Table IV).

Table IV. Genes studied by RT-PCR, PCR and LightCycler conditions.

The measured 15N/14N and 13C/12C ratios were converted to δ values per 103 relative to standards, where:

Graph

The δ values were then converted to isotope enrichments (APE, atom % excess) for calculation of 15N and 13C recoveries in urinary nitrogen and breath CO2, respectively (Klein [25]).

The isotope recoveries f as percentage of the isotope dose D (in mol) after administration of the labelled [15N]glycine, [1-13C]glycine or NaH13CO3 were obtained from the cumulated products of the isotope enrichment APE((ti+1 + ti)/2) in total nitrogen of urine or breath CO2 (in atom % excess) and the respective nitrogen or CO2 amount n(ti+1−ti) (in mol) excreted in consecutive time periods (ti+1−ti) over 12 h:

Graph

The 15N recovery in urinary N was used to estimate the whole body protein synthesis applying the end-product method described by Golden and Waterlow ([14]).

Following the single dose of [15N]glycine, the whole body protein synthesis (WBPS) [g · protein d−1] is

Graph

where fP(15N-glycine) is the "plateau value" of 15N recovery, which was reached between 48 and 72 h after tracer administration. EU is the urinary nitrogen excretion [mol N d−1]. The whole body protein breakdown WBPB was calculated as follows:

Graph

Similar to the 15N approach the plateau value of 13C recovery in breath CO2 at 12 h after administration of [1-13C]glycine fP(13C-glycine) was also used as the indicator for the whole body protein synthesis. The 13C technique needs a 12-h period of sampling for the assessment of the WBPS whereas the 15N technique measures WBPS over the 72-h period after tracer administration. Therefore, both tracer techniques can be considered as complementary, especially since their different end product pools breath CO2 and urinary nitrogen.

The plateau 13C recovery fP(13C-glycine) was corrected for incomplete 13C release in breath CO2 as determined with NaH13CO3fP(13C-bicarb).

Thus, the corrected 13C recovery fP(13C-glycine) was calculated as:

Graph

Heat production was calculated according to Brouwer ([6]), carbohydrate and fat oxidation according to Simonson and DeFronzo ([41]).

Results are presented as mean values with standard errors (SEM). Unpaired Student's t-test or Welsh test (SAS Procedure TTEST, SAS Institute Inc. (1999) SAS/STAT® User's Guide, Version 8, Cary, NC: SAS Institute Inc.) was used to examine differences between means. Statistical significance was accepted if the null hypothesis was rejected with p < 0.05.

Whole body protein metabolism was studied using 15N and 13C tracer kinetic techniques. The time course of the 15N recovery, i.e. the cumulative 15N amount (in % of 15N dose) excreted in urinary nitrogen, after a bolus dose of [15N]glycine is shown in Figure 1 (left panel).

Graph: Figure 1. Isotope kinetics in pigs fed SPI or CAS (9 animals each) after an intravenous bolus of [15N]glycine (left panel) and [1-13C]glycine (right panel) in Exp. 1. Values are means ± SEM. ×-× CAS, •-• SPI measured.

Urinary nitrogen excretion and 15N recovery over 72 h were used for calculation of whole body protein synthesis. Urinary nitrogen excretion, nitrogen retention and 15N recovery as well as whole body protein synthesis and protein breakdown were significantly different between pigs chronically fed SPI or CAS (Exp. 1, Table V). In pigs temporarily fed SPI (Exp. 2), there was a tendency for lower whole body protein synthesis (p = 0.061) and an increased nitrogen retention (p = 0.056) as compared to CAS-fed control pigs.

Table V. Parameters of protein and energy metabolism in pigs fed SPI or CAS diets☆ (Mean ± SEM).

The 13C recovery approaches a plateau value at about 12 h (Figure 1, right panel). In pigs chronically fed SPI-diet the 13C recovery was significantly higher than in the CAS-fed control. Interestingly, the 13C recovery increased from 21.7 ± 0.8% on day 2 of SPI feeding (corresponds to 81 days of age) to 29.7 ± 1.4% on day 22 of SPI feeding (p < 0.001) whereas CAS-fed animals remained at 16.7 ± 2.3%. Twenty eight days after replacement of SPI by CAS diet the 13C recovery was not different between control and test group (Table V).

Independently of the diet, the recovery of 13C in breath CO2 (% of 13C dose) after an intravenous bolus of NaH13CO3 amounted to 71.0 ± 4.0 in Exp. 1 and 2. Both chronically and temporarily SPI-fed pigs did not exhibit different heat production, carbohydrate and fat oxidation in comparison to corresponding CAS control. It became evident, however, that younger pigs oxidize less fat and more carbohydrates (Table V).

Molecules involved in the amino acid and protein metabolism (leucine aminopeptidase, endopeptidase 24.16) showed an increased mRNA expression with chronic and temporary SPI feeding (Table VI).

Table VI. Hepatic parameters in pigs fed SPI or CAS diets☆ (Means ± SEM).

Chronic SPI-feeding caused lower plasma concentration of several amino acids as compared with pigs on CAS feeding (Figure 2). Thus, the ratio (SPI:CAS) of concentrations of selected amino acids was as follows: methionine 0.57 ± 0.04; lysine 0.76 ± 0.08; leucine 0.62 ± 0.08; isoleucine 0.57 ± 0.03; valine 0.66 ± 0.09; tyrosine 0.59 ± 0.07; arginine 0.67 ± 0.08 (p < 0.05 each). In contrast, threonine showed an increased ratio (1.44 ± 0.17, p < 0.05).

Graph: Figure 2. Concentrations of plasma amino acids in pigs (9 animals each) fed SPI (open bars) or CAS (filled bars) after 24 h of fasting in Exp. 1. Values are means ± SEM. Asterisks (☆) denote significant differences between SPI and CAS fed pigs (p < 0.05).

Chronic SPI-feeding increased plasma urea concentrations (3.65 ± 0.20 mmol/l) as compared to CAS control (1.75 ± 0.16 mmol/l) (p < 0.05). The plasma glucose concentration was higher (p < 0.05) in pigs fed SPI (5.07 ± 0.12 mmol/l) as compared with CAS (4.51 ± 0.16 mmol/l). Feeding SPI did not alter the concentrations of plasma insulin, glucagon and IGF-1. Twenty eight days after replacement of SPI diet by CAS (Exp. 2) no differences remained in the blood parameters between the test and control group.

To study diet-associated oxidative stress response hepatic glutathione and TBARS as well as mRNA abundance of glutathione-S-transferase α and peptide methionine sulfoxide reductase were analysed.

Chronic SPI-feeding (Exp. 1) resulted in a severe reduction of hepatic glutathione concentration (p < 0.001), whereas the TBARS were 5.6 times higher (p < 0.001). In contrast, after 28 days of replacement of SPI by CAS, hepatic glutathione was increased (p = 0.021) and the TBARS were 1.8 times lower (p = 0.012) (Table VI). The mRNA abundances of glutathione-S-transferase α and peptide methionine sulfoxide reductase linked with oxidative stress response were elevated in both experiments.

This study shows that feeding a diet containing SPI as the sole protein source causes long-term alterations in the protein metabolism and redox homeostasis in juvenile growing pigs. In particular, we observed that effects related to the oxidative stress response persisted for 3 weeks after cessation of SPI intake. The persistence of a reduced whole body protein synthesis rate and reduced N retention were of borderline significance (p = 0.062 and 0.056). The 13C recovery in breath CO2 at 12 h, parallel to the 15N recovery at 72 h, is related to the whole body protein synthesis because CO2 production is not different between test and control group (Table V). Therefore, whole body protein synthesis can be derived from the 13C recovery (Golden & Waterlow [14]; Yagi & Walser [44]). Differences in 15N and 13C recovery found in this experiment between the SPI and CAS diets confirm our previous results (Junghans et al. [21]).

Despite a decrease of the protein turnover in the test group compared with the control group by factor 2 in Exp. 1, heat production was not significantly changed (Table V).

Since we used highly purified SPI produced for infant formula, the level of total isoflavones was more than 200 times lower compared with other soy protein diets reported (Deutz et al. [11]). Also, effects of soy-associated constituents such as trypsin inhibitors and lectins could be largely excluded because of heat treatment under mild alkaline conditions.

In chronically SPI-fed-pigs (Exp. 1), differences in plasma amino acid concentrations, in part, reflect differences in the amino acid pattern of the protein diets. Thus, methionine, lysine, leucine, isoleucine valine and tyrosine concentrations are decreased in the diet and plasma as compared to CAS control (Table III and Figure 2). However, arginine is increased in the diet, but decreased in plasma (SPI vs. CAS fed pigs). This corresponds largely with our previous observations (Löhrke et al. [27]). The decrease of plasma arginine concentration in SPI-fed pigs vs. CAS control 24 h after feeding might be explained by increased utilization of arginine for urea genesis. Actually, the plasma urea concentration and daily urinary urea excretion were about 210% in SPI-fed pigs as compared to CAS control. The increased concentration of plasma glucose in SPI-fed pigs might indicate an increased gluconeogenesis and/or glycogenolysis. However, the 24-h fasting plasma insulin, glucagon and IGF-1 concentrations showed no differences between the test and control group in Exp. 1. In contrast to our findings, SPI-fed rats had lower fasting plasma glucose and insulin concentrations as compared with CAS-fed rats which might be due to different species and/or experimental design (Lavigne et al. [26]).

Several authors (Eggum et al. [12]; Mariotti et al. [31]) showed that the inflow of nutrients from the intestine is largely completed 24 h after feed intake. Therefore, measurements were performed in this fasting period. Thus, we could exclude the interference of diet-associated effects and direct effects caused by nutrient inflow. The present study is different from previous investigations of metabolic effects of SPI or CAS-containing diets, which were not focused on the prolonged fasting state (Deutz et al. [11]; Saggau et al. [38]).

Possible relations between dietary factors and persisting alterations of metabolism are of particular interest, but most of them are still poorly understood. Nevertheless, it has been considered that individual amino acids can regulate gene expression (Hesketh et al. [16]; Jousse et al. [17]). Thus, methionine is involved in the initiation of mRNA translation for protein biosynthesis and is the precursor for S-adenosyl methionine, the principal biological methyl donor (Lu et al. [28]). In this context, methylation of DNA affects gene expression, and it is believed that methionine plays a central role in epigenetic effects (Waterland & Jirtle [43]). As SPI vs. CAS is deficient in the content of methionine, persisting metabolic effects in temporarily SPI-fed pigs might result from this dietary factor. Further acute and, possibly, persisting effects might be related to methionine. Thus, the deficient dietary methionine supply in the chronically SPI-fed pigs (Exp. 1) obviously leads to a depletion of hepatic glutathione (Table VI) which indicates a reduced antioxidative capacity (Meister et al. [32]; Lyons et al. [30]). Higher concentrations of TBARS in the test group resulting from the experimentally Fe2+stimulated lipid peroxidation of liver lipids (Buege & Aust [7]) suggest higher oxidative stress responsiveness. In the temporarily SPI-fed pigs concentrations of hepatic glutathione are increased and TBARS are decreased. It indicates an 'over-regulating' process if SPI feeding is switched back to CAS. In contrast to these findings the examined hepatic mRNA expressions are increased in both chronically and temporarily SPI-fed pigs (Exp. 2, Table VI). The increased hepatic mRNA expression of glutathione-S-transferase and peptide methionine sulfoxide reductase in SPI-fed pigs imply a higher oxidative stress responsiveness (Schwerin et al. [40]; Junghans et al. [19]). The increase in the hepatic mRNA and protein expression of glutathione-S-transferase in SPI-fed pigs suggests the detoxification of products of oxidative stress by forming of thioethers. The increase in the hepatic mRNA and protein expression of peptide methionine sulfoxide reductase may be understood as a "repair" of oxidized functional proteins with methionine residues.

Leucine is also considered to regulate protein synthesis in skeletal muscle by enhancing both activity and synthesis of proteins involved in mRNA translation as shown in rats (Anthony et al. [1]). The lower plasma leucine concentration observed in SPI-fed pigs can be interpreted as a physiological response to alterations of muscle protein metabolism (Löhrke et al. [27]).

In our experiment the hepatic mRNA abundances that are linked with protein metabolism, especially proteolysis, such as leucine aminopeptidase and endopeptidase 24.16 showed a significant increase in pigs fed the SPI-diet compared with the CAS-diet. In line with our results, Rebolledo et al. ([36]) found an increased expression of intestinal and renal leucine aminopeptidase in rats fed a methionine deficient amino acid mixture. The increase of transcript levels of proteolytic enzymes reveals the adaptive potential of the organism to recover its amino acid homoeostasis.

In summary, our results suggest that SPI-induced changes in protein and amino acid metabolism as well as in redox homeostasis and antioxidative potential in growing pigs can persist 4 weeks after the cessation of SPI feeding. It is evident that the observed metabolic effects are amplified over SPI feeding time as found for the 13C recovery data. Our results on persisting dietary effects suggest links to findings on persisting protein turnover modifications induced by the protein feeding pattern (Arnal et al. [3]), the compensatory growth (e.g. Schadereit et al. [39]; Kamalzadeh et al. [22]) and the phenomenon of nutritional programming (e.g. Daenzer et al. [10]; Waterland & Jirtle [43]; Lucas [29]; Guzman et al. [15]). In order to elucidate these phenomena it remains to be investigated the regulatory mechanisms at the metabolic, transcript and protein levels as well as the interaction between these different levels (Kelleher [23]).

The study was funded by the core budget to the Research Institute for the Biology of Farm Animals. The technical assistance of I. Brüning, H. Schott, H. Pröhl, A. Jugert and J. Bittner is greatly appreciated. The authors thank Cornelia C. Metges for helpful comments and P. E. Rudolph for statistical advice and treatment of data. Furthermore, we thank W. B. Souffrant for amino acid analyses, F. Schneider and Charlotte Rehfeldt for support in hormone analyses.