Animals, diet and design
Twenty-four 12 weeks old female crossbred pigs [(German Landrace × Large White) x Pietrain] with an initial body weight of 34.9 ± 1.9 kg (mean ± SD) were randomly assigned to 2 groups of 12 pigs each. The animals received a hypercholesterolemic basal diet, one group with 230 g/kg lupin protein isolate from L. angustifolius and the other group (control) with 230 g/kg casein (Meggle, Wasserburg, Germany). The diets were fed for 4 weeks. Besides the experimental proteins, the hypercholesterolemic basal diet contained (in g/kg diet) wheat (400), corn starch (casein diet: 197; lupin protein diet: 192), coconut fat (120), cholesterol (20), vitamin and mineral mixture (20) and monocalcium phosphate (13). The lupin protein diet was supplemented additionally with 5 g/kg of DL-methionine at the expense of corn starch to meet the requirement for methionine for growing pigs . Minerals and vitamins were added according to the recommendations of the NRC  as precast mix (Mineral feed, Basu, Bad Sulza, Germany). The diets were calculated on the basis of GfE recommendations  to contain 15 MJ/kg. The pigs were individually kept in pens in an environmentally controlled facility with a temperature of 20°C, relative humidity between 55-60%, and light from 06:00 am to 06:00 pm. All pigs had free access to food. Water was available ad libitum from a nipple drinker system during the whole experiment. The pigs were weighed once a week. During the last 5 days of the experiment, six pigs of each group were kept in metabolism cages which allowed a quantitative collection of faeces. The experimental procedure was performed according to the established guidelines for care and handling of laboratory animals and was approved by the council of Saxony-Anhalt, Germany (No. 42502-2-1040MLU).
Preparation and characterisation of the experimental proteins
Casein that was used as control protein was not further processed. Defatted total protein isolate of L. angustifolius was provided by the Fraunhofer Institute (IVV, Freising, Germany). The crude protein content of the diet proteins was determined by official methods , and was 948 g/kg dry matter (DM) for casein and 915 g/kg DM for lupin protein isolate. The amino acid composition of the experimental proteins is shown in Additional file 1. Casein contained (in g/kg DM) 0.58 calcium, 2.7 phosphorus, 0.10 magnesium, and 0.056 zinc. The lupin protein isolate contained (in g/kg DM) 1.02 calcium, 9.41 phosphorus, 0.59 magnesium, and 0.032 zinc. The calculated amount of phytic acid based on the analysed phytate phosphate in the lupin protein isolate was 25.4 g/kg DM. No detectable phytate phosphate was found in casein.
Lupin conglutin γ was isolated from lupin seed flour according to Lovati et al. . The procedure consisted of various chromatographic steps, including metal affinity chromatography on NiNTA-agarose. The homogeneity of purified conglutin γ was assessed by SDS-PAGE .
Amino acid analysis of the experimental proteins
Samples were oxidised and then hydrolysed with 6 M HCl  to determine the amino acid concentrations. Separation and quantification of the amino acids were performed by ion-exchange chromatography following post-column derivatisation in an amino acid analyzer (Biotronic LC 3000; Eppendorf, Hamburg, Germany). After digesting the diets with barium hydroxide  tryptophan was quantified by reversed-phase HPLC .
Analysis of phytate phosphate and phytic acid of the experimental proteins
The concentration of phytate phosphate in the dietary proteins was analysed by a method of Harland and Oberleas . Briefly, phytic acid was extracted from 2 g protein with 40 ml 2.4% HCl. Afterwards the samples were filtered and immobilised onto an anion exchange column (AG1-X4, 100–200 mesh, chloride form, Bio-Rad Laboratories, Hercules, California, USA). After washings the retained inositol phosphates were eluted from the column with 0.7 M NaCl. A 50 μl aliquot of the eluent was hydrolysed with 50 μl of a 5 M HClO4 and 1 M H2SO4 mixture, and was ashed at 250°C. The inorganic phosphate concentration of the samples was quantified colourimetrically by use of standards. The phytic acid content of the sample was calculated to be 28.2% × phosphorus .
Blood samples that were taken from each animal at the beginning and at the end of the experiment were collected in heparinised tubes and centrifuged at 4°C and 1,100 g for 10 min to obtain plasma. Faeces were collected twice a day over a period of 5 days. Pooled faeces from each pig were frozen at −20°C pending analysis. At the end of the experimental period, the pigs were anaesthetised and killed by exsanguination 12 h after their last meal. The liver was harvested and weighed. Thirty cm of small intestine (10 cm before and 20 cm behind the papilla duodeni major) were removed, rinsed with physiological NaCl solution and cut lengthwise. The intestinal mucosa was scraped and mixed. Samples of liver and intestinal mucosa were snap-frozen in liquid nitrogen and stored at −80°C pending further analysis.
Lipoprotein isolation and lipid analysis
Plasma lipoproteins were separated according to their density (very low density lipoprotein, VLDL, ρ < 1.006 g/ml; LDL, 1.006 g/ml < ρ < 1.040 g/ml; high density lipoprotein, HDL, 1.063 g/ml < ρ < 1.21 g/ml) by step wise ultracentrifugation. Lipids from liver and freeze dried faeces were extracted with a mixture of n-hexane and isopropanol (3:2, v/v) according to the method from Hara & Radin  modified by Eder & Kirchgessner . To measure lipid concentrations of liver and faeces, aliquots of the lipid extracts were dried and residues dissolved in a mixture of Triton X-100 and chloroform (1:1, w/w). The cholesterol and triglyceride concentrations of plasma, plasma lipoproteins, liver and faeces were determined using enzymatic reagent kits (Diagnostic Systems, Holzheim, Germany).
Analysis of faecal neutral sterols and bile acids by GC-FID and GC-MS
Lyophilised aliquots of homogenised faeces were analysed for neutral sterols (cholesterol, coprostanol, cholestanol, coprostanone and cholestanone) and for bile acids (iso-lithocholic acid, lithocholic acid, chenodeoxycholic acid, hyodeoxycholic acid, hyocholic acid) according to gas chromatographically methods, which have been previously described [19, 20]. In brief, 50 mg faeces were provided with 5α-cholestane (Sigma-Aldrich, Taufkirchen, Germany) as internal standard and underwent a mild alkaline hydrolysis with freshly prepared 1 M ethanolic NaOH. Free sterols were extracted with cyclohexane, extracts were dried under nitrogen stream and residues were resolved in decane. The analysis was performed using GC-FID (GC17A-AF Vers. 3, Shimadzu Corp., Kyoto, Japan). After extraction of the neutral sterols the samples underwent a strong alkaline hydrolysis with 10 M NaOH solution. The samples were than adjusted to pH 1 with HCl and free bile acids were extracted with diethyl ether. Norcholic acid (Sigma-Aldrich) was added as internal standard and extracts were methylated, silylated and finally dried under nitrogen stream. Residues were resolved in decane and injected in the GC-MS (GC17-QP5000, Shimadzu Corp.). Bile acid quantification was based on the multi-ion detection with m/z = 253.20 for norcholic acid, m/z = 215.25 for iso-lithocholic acid and lithocholic acid, m/z = 73.10 for chenodeoxycholic acid, m/z = 255.3 for hyodeoxycholic acid, and m/z = 458.50 for hyocholic acid.
RNA isolation and real-time RT-PCR
For the determination of mRNA abundances total RNA was isolated from intestinal mucosa using Trizol™ reagent (Invitrogen, Karlsruhe, Germany) according to the manufacturer’s protocol. The RNA concentration was estimated from the optical density at 260 nm. RNA purity was proofed by agarose gel electrophoresis. A total of 1.2 μg of total RNA was used for cDNA synthesis using the RevertAid™ M-MuLV Reverse transcriptase (Thermo Fisher Scientific Inc., Waltham, MA, USA). For the determination of relative mRNA concentrations real-time detection RT-PCR using the Rotorgene 6000 system (Corbett Research, Mortlake, Australia) was applied. A total of 1 μl cDNA templates were amplified in a total volume of 20 μl using 200 μM dNTPs (Genecraft, Cologne, Germany), 1.5 mM MgCl2, 0.5 U GoTaq DNA polymerase, 4 μl 5× buffer (all from Promega, Mannheim, Germany), 0.2 μl 10× SYBR Green (Sigma-Aldrich), and 5.4 pM of each primer. The PCR protocol provided an initial denaturation at 95°C for 3 min and 20–35 cycles of amplification comprising denaturation at 95°C for 25 s, annealing at primer-specific temperatures (57-60°C) for 30 s and elongation at 72°C for 55 s. Subsequently melting curve analysis was performed from 50°C to 99°C with continuous fluorescence measurement. The amplification of a single product of the expected size was confirmed using 2% agarose gel electrophoresis. Relative mRNA concentrations were calculated using the ∆∆Ct method . Ct-values of target genes and the reference genes were obtained using Rotorgene Software 5.0. Succinate dehydrogenase, subunit A (SDHA) and ribosomal protein S9 (RPS9) served as appropriate reference genes. Relative mRNA expression was expressed as fold change of mRNA abundance relative to the casein group. Sequences of the gene-specific primers are shown in Additional file 2.
Plasma concentrations of 25-hydroxy vitamin D3 (25(OH)D3) and tocopherols
The plasma concentration of 25(OH)D3 was determined by coupled liquid chromatography-mass spectrometry (LC-MS) using a MassChrom™ reagent kit (Chromsystems, Munich, Germany) as described recently . In brief, plasma samples were mixed with deuterated 25(OH)D3 (Chemaphor Inc., Ottawa, Canada) as internal standard. Precipitation reagent was added, and subsequent to centrifugation at 15,000 g for 5 min the supernatant was transferred into a HPLC vial and analysed by HPLC (Agilent 1100 HPLC, Agilent Technologies, Waldbronn, Germany), coupled to a MS system (API 2000, Applied Biosystems, Darmstadt, Germany).
The concentration of plasma tocopherol isomers was determined by HPLC analysis . Plasma samples were mixed with 1 ml of a 0.1 g/l pyrogallol solution (ethanol, absolute) and 150 μl saturated NaOH solution. This mixture was incubated at 70°C for 30 min, and tocopherols were extracted with n-hexane. Individual tocopherols of the extracts were separated isocratically by HPLC (Agilent 1100 HPLC, Agilent Technologies) using a mixture of n-hexane and 1,4 dioxane (94:6, v/v) as mobile phase and a LiChrospher Si-60 column (5 μm particle size, 250 mm length, 4 mm internal diameter; Agilent Technologies) and detected by fluorescence (excitation wavelength 295 nm, emission wavelength 325 nm).
Analysis of plasma and faeces minerals
For the determination of calcium, phosphorus, magnesium and zinc concentrations, plasma was diluted with distilled water and faeces were hydrolysed with 6 M HCl and 1.76 M HNO3 according to the official VDLUFA method . The minerals were analysed by inductively coupled plasma optical emission spectroscopy (ICP OES; Varian 715-ES; Agilent Technologies) using selected wavelengths (Ca: 318 nm, P: 215 nm, Mg: 280 nm, Zn: 214 nm). Analyses were run in duplicates.
Cell culture experiments
Caco-2 cells were cultured in minimal essential medium (MEM) with 10% fetal bovine serum (FBS) at 37°C with 5% CO2 and passaged at 90% of confluence . For cholesterol-uptake experiments, cells were seeded in common 6 well plates at a density of 0.8 × 106 cells per well and cultured for 20 to 23 days. Cells were used for experiments from passage 17 to 27. 24 h prior to the experiments cells received serum free medium.
For the cholesterol-uptake, cells were incubated for 2 h at 37°C with 2 mM radioisotopically labelled cholesterol ([4-14C]-cholesterol, 0.1 mCi/ml, American Radiolabeled Chemicals Inc., St. Louis, MO, USA) dissolved in MEM plus 2% FBS as previously described [25, 26] with or without the following test substances: 0.04 mg/ml ezetimibe (Santa Cruz Biotechnology Inc., Heidelberg, Germany), 5 mg/ml albumin (Sigma-Aldrich), 0.6 mg/ml lupin conglutin γ or 0.18 mg/ml sodium phytate (Sigma-Aldrich). The ezetimibe concentration which we used for the incubation study ranged in the upper concentration levels of ezetimibe used in other cell culture experiments [27, 28]. Prior to the incubation, the cells were washed twice with HEPES buffer (25 mM HEPES/Tris (pH 7.5), 140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl2, 0.8 mM MgSO4, 5 mM glucose, 37°C). After the incubation, the cells were quickly washed twice with phosphate buffered saline (37°C) containing 1% bovine serum albumin and four times with ice-cold HEPES buffer. Finally, cells were dissolved in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris/HCl (pH 8.0), 150 mM NaCl, 1% Triton X-100 (w/v), 0.5% sodium deoxycholate (w/v), 0.1% sodium dodecyl sulfate (w/v)) and the amount of absorbed radiolabelled cholesterol was determined by liquid scintillation spectrometry (Liquid Scintillation Analyser Tri-Carb 2800 TR, Perkin Elmer, Rodgau, Germany). Protein concentration was measured using BC assay (Uptima, Interchim, Montlucon, France). Data are expressed as pmol of cholesterol per mg of cell protein.
Values are expressed as means ± SD. Means of the two groups were compared by Student’s t-test using the statistic software MINITAB (Release 13, Minitab Ltd., Michigan, USA). Significances of differences between basal and final means were tested by the paired t-test. Means were considered significantly different at P < 0.05.