Animals, adaptations and grouping
The experimental procedure was evaluated by the ethical committee of and approved by the State Office for Agriculture, Food Safety and Fisheries Mecklenburg-Vorpommern, Germany (LALLF 7221.3-1-052/17). All experiments were performed in accordance with relevant regulations and the reporting in the manuscript followed the ARRIVE guidelines. Non-pregnant, lactating Holstein-Friesian cows from two commercial farms were screened for milk yield and MUC, considering the last five monthly milk testing. Animals (N = 36) in second to fourth lactation with comparable milk yield (32.5 ± 0.9 kg/d) were bought in pairs of two with a high (HMU: 276 ± 4 mg/L; n = 18) and low (LMU: 186 ± 4 mg/L; n = 18) MUC and transported to the institutional barn at FBN in 18 blocks. Farmers were informed about and agreed to the use of cows in the present study. Cow pairs were assigned alternating to an isocaloric ration (10.1 ± 0.1 MJ ME/kg DM) with normal (NP: 15.9 ± 0.1%) or low (LP: 13.8 ± 0.2%) crude protein content (Table 1), resulting in 4 groups (HMU-NP, HMU-LP, LMU-NP, and LMU-LP) of n = 9 cows in each group. Cows were adapted to the ration and the loose housing system for 2 weeks. They were fed at 0500 h and 1700 h for ad libitum intake and milked at 0430h and 1630h. The body condition score (BCS) and dorsal back fat thickness were determined once a week as previously described14. At the end of the second week, animals were 329 ± 13 days in milk, had a BCS of 3.5 ± 0.5 on a five-point scale, and a milk yield of 23.4 ± 0.8 kg/d.
Adaptation, nitrogen balance study, 13C urea tracer administration, and tissue sampling
The N balance study was conducted according to the animal trait of ontology ATOL_0005338 and the respectively described method15. Animals were transferred to a climate-controlled room at 15°C and adapted to a tie-stall on a 2.43 x 1.56 m stanchion and for one week. Cows were continuously fed ad libitum from one batch, which was aliquoted in vacuumed 40 kg-plastic bags, and stored at +4°C. A feed sample was taken and frozen at -20°C. On the last day of the adaptation period, cows were prepared for a 4-d N balance study by fitting the cows with a urinal consisting of a customized synthetic leather model which had a cutout around the vulva area. The urinal was glued to the cow below the anus around the vulva and the cutout was covered with a sewn-on piece of synthetic leather.
On the first to fourth day of the balance trial, animals were continuously fed from the same batch but only 97% of the daily ad libitum intake to ensure homogenous intake. Between 0600 h and 2000 h, feed was provided in 7 equally-sized portions in 2-hour intervals amounting to 58% of the daily ad libitum intake. The remaining 39% were given at 2000 h in one portion. On the first day of the balance study, a flexible plastic tube (4.5 diameter) was connected to the urinal and a 30 L-container, and urine was collected for 24 h to measure the weight and volume and from these to calculate the density of the urine. Samples were taken and stored at -20°C until analysis. On the second day of the balance study and onwards, the container was prefilled with 400 mL (564 g) of 50% sulfuric acid and placed on a shaker or magnetic stirrer to minimize volatile N losses. Feces was collected several times a day, combined and stored in a container at 4°C. The 24-h collections were weighed and homogenized, and subsamples were taken and stored at -20°C. Animals were milked twice daily (0630 h and 1830 h), milk yield was recorded, and two 15-mL aliquots from each milking were pooled on a daily basis and stored at -20°C until analysis.
On the first day of the N balance study, cows were equipped with a jugular vein catheter, which was flushed with 20 mL 0.9% saline and 10 mL 3.5% sodium citrate. On the third day, a 13C urea-tracer (13C urea, ≥99 atom%; Sigma-Aldrich, St. Louis, USA) was injected as a bolus (2 g per 650 kg body weight (BW), dissolved in saline), and the catheter was flushed with 20 mL saline. Blood samples were collected at several time points after bolus injection in EDTA-containing tubes, immediately placed on ice, centrifuged (1345 ×g, 20 min, 4°C), and stored at -20°C as described previously6.
The day after the balance study, animals were transferred to the institute's slaughterhouse. Animals were anesthetized by a captive bolt stunning and exsanguinated. A blood sample was collected in an EDTA-containing tube and stored on ice. The blood was centrifuged and the obtained plasma was frozen at -80°C. The liver was weighed and a tissue sample was taken from the middle of the left lobe. The tissue sample was immediately chilled on ice, cut into smaller pieces, shock-frozen in liquid N2 and stored at -80°C until further analysis.
Feed, milk, urine, feces, and plasma analyses
For the determination of feed dry matter, samples were air dried at 60°C, ground and dried again at 105°C for 4 h. For N measurement, frozen fresh feed samples were ground with the addition of dry ice. The drinking water was sampled in a glass bottle and analyzed for total-N content by the methods of the German Institute for Standardization (N concentration = 0.58 mg/L). According to the 24-hour excretion volume, a proportional pool sample of feces, milk, and acidified urine was prepared and sent to the accredited laboratory Landwirtschaftliche Untersuchungs- und Forschungsanstalt der LMS Agrarberatung GmbH (LUFA, Rostock, Germany) for analysis of total N using the Kjedahl method for feces and feed and a "vario max" element analyzer (Elementar; Langenselbold, Germany) for milk and acidified urine.
Frozen milk samples of the 4-day balance study were thawed and centrifuged for 10 min at 4°C and 50,000 × g to remove the milk fat. The resulting skim milk was subjected to urea concentration analysis using ABX Pentra C400 analyzer (HORIBA Europe GmbH, Oberursel, Germany) and the kit: LT-UR0010 (Labor + Technik Eberhard Lehmann GmbH, Germany). Urea concentration measured in skim milk was recalculated for whole-milk and a mean value of 4 days was calculated. The plasma sample obtained 10 min before tracer administration was subjected to the analysis of aspartate amino transferase (AST), GLDH, and gamma-glutamyltransferase (γ-GT) using the following kits: AST: A11A01629 (HORIBA Europe GmbH), GLDH: LT-GD 0010 (Labor + Technik Eberhard Lehmann GmbH), and γ-GT: AX0N00016 (Axon Lab AG, Germany). The analysis of plasma free amino acids was performed as described previously16 using a HPLC with fluorescence detection (HPLC 1200/1260 infinity II series; Agilent Technologies) with the following modifications. Amino acids were separated on a Gemini 250 × 4.6 mm 5 µm ODS (C18) 110 Å column using a phosphate buffer (pH 7.45): ACN/methanol/water (45:45:10) gradient from 7 to 100%. Urea concentrations in plasma samples obtained from slaughter were measured using again the LT-UR0010 kit, and urea concentrations from 50-fold diluted acidified urine samples were analysed by HPLC (1200/1260 infinity II Series; Agilent) with a 300 × 7.8 mm Rezex RCM-Monosaccharide column (Phenomenex Inc.) as described previously6. Non-urea N-metabolite concentrations in plasma obtained during slaughter and in non-esterified urine samples were analyzed by HPLC on a 250 × 4.6 mm Synergi 4 µm Hydro-RP 80 Å column protected by a corresponding 4 × 3 mm pre-column (both Phenomenex Inc., Aschaffenburg, Germany) as described before6.
13C urea analysis
The plasma 13C urea enrichment analysis was performed as previously described6. Briefly, the urea was derivatized and the t-butyldimethylsilyl derivate was analyzed on a gas chromatograph-mass spectrometer (GC-MS, QP 2010, coupled with GC 2010, AOC-20i; Shimadzu, Duisburg, Germany). Detection occurred at m/z 231 and 232. A two-exponential curve fitting to 13C urea enrichment was performed and mole % excess (MPE) was as: MPE (t) = a × e(-b × t) + c × e(-d × t) . The variables a, b, c and d were used to calculate the area under the curve (AUC) AUC= a / b + c / d and the mean residence time MRT (h) = (a / b2 + c / d2) / (a / b + c / d). Based on the tracer dosage (D) in mg, the urea pool size (Q) was calculated according to17: Q (mg/kg BW) = (D × MRT) / (AUC × BW).
Frozen liver tissue was ground to fine powder using a mortar and pestle under liquid N2. For protein extraction, 50 mg tissue powder and 200 µL lysis buffer consisting of Triton X-100 (1% v/v), sodium deoxycholate (DOC; 0.5% v/v), sodium dodecyl sulfate (SDS; 0.1% w/v), 8 mM NaCl, 2.7 mM KCl, 6.9 mM Na2HPO4, 1.5 mM KH2PO4, 100 µL PhosSTOP (Roche) and 10 µL Protease Inhibitor Cocktail (AppliChem) were placed in the shaker for 45 s. Samples were centrifuged for 20 min at 4 °C and 13,000 × g. The protein concentration was measured using the bicinchonin acid method and bovine serum albumin as standard. For each animal, two 2D-gels were prepared resulting in a total of 72 gels. For each gel, 300 µg protein extract was mixed with rehydration buffer (8 M urea, 2% CHAPS, 0.5% IPG strip buffer, 15 mM dithiothreitol (DTT) and 0.002% bromophenol blue to a final volume of 450 µL, and applied to 24 cm-IPG BlueStrip (Serva Electrophoresis) with a pH gradient of 3 - 10. The strips were covered with 1 mL of mineral oil and placed in the Ettan IPGphor3 (GE Health-care, Munich, Germany) for isoelectric focusing (IEF). Active rehydration and IEF was performed at 50 µA/strip at 50 V for 8 h, 500 V for 1 h, 1000 V for 2 h, and 8,000 V for 8 h 40 min. The IPGs were first equilibrated in a buffer of 75 mM Tris-HCl (pH 8.8), 29.3% glycerol, 2% SDS, 6 M urea, 0.002% bromophenol blue, and 1% DTT for 15 min and then agitated again for 15 min with the same buffer, but instead of DTT with 2.5% iodoacetamide (IAA). The IPG strips were transferred to 12.5% polyacrylamide gels (255 x 200 x 0.65 mm; 2D HPE Large Gel NF, Serva Electrophoresis) and the 2D-GE was performed on HPE BlueHorizon tower (Serva Electrophoresis) at 1 W for 30 min, 3 W for 30 min, 5 W for 10 min, 30 W for 3 h 50 min and 40 W for 50 min. The SDS-PAGE gels were stained overnight in colloidal coomassie and then de-stained in 25% methanol. The gels were scanned and the images saved as tiff format. Individual gels were warped to one master gel image to allow for image analysis using the software DELTA2D (version 4.6, DECODON, Greifswald, Germany; http://www.decodon.com). Spot volumes significantly different between groups or diets were stamped out of the gel with a spot cutter (Ø 2mm) and transferred to a 96-well micro titer plate. Protein identification was performed as previously described16. Briefly, proteins were digested using trypsin and the molecular masses of the tryptic digest were measured using a 5800 MALDI TOF/TOF Analyzer (Applied Biosystems, Forster City, CA, USA). The spectra were registered in a mass range of 900 to 3700 Da with a focus on 2000 Da. An automatic internal calibration was performed as a two-point calibration when the peptides with the monoisotopic (M+H)+ m/z at 1045,556 and 2211,104 reached a signal-to-noise (S/N) ratio of at least 20. Peak lists were generated using the "peak-to-maskot" script of the 4000 Series ExplorerTM software (V3.5). Peak lists were generated for a S/N ratio of 10, a peak density of 15 peaks per 200 Da, a minimum peak area of 100 and a maximum of 60 peaks per spot. The results of the MALDI-TOF-MS analysis were verified with TOF-TOF measurements and performed for the three highest peaks of a TOF spectrum. Internal calibration was performed as a one-point calibration using either the mono-isotope arginine (M+H)+ m/z at 175.119 or lysine (M+H)+ m/z at 147.107, if a S/N ratio of at least 5 was obtained. Peak lists were generated according to the above script with the following settings: Mass range 60 to precursor -20 Da, peak density of 15 peaks per 200 Da, minimum area of 100 and maximum 65 peaks per precursor. The peak list was created for a S/N ratio of 7. For the identification of the proteins, data base search with peptide mass fingerprint (PMF) was performed against the NCBInr (National Center for Biontechnology Information, http://www.ncbi.nlm.nih.gov/) database using the Mascot search engine version 2.4.1 (Matrix Science, London, UK). Search parameters were: taxonomy: “all entries”; variable modifications: “carbamidomethyl (C)” and “oxidation (M)”; precursor tolerance ± 50 ppm; peptide charge “1+”; MS/MS fragment tolerance “0.5 Da”; “monoisotopic”. The gene names of the identified proteins with a protein score greater than 64 (P < 0.05) were imported into ClueGO software18 for analysis of activated pathways. The results were visualized using Cytoscape software version 3.8.019.
The extraction of RNA from 20 mg liver tissue powder was performed using the innuPREP RNA Mini Kit 2.0 and residual DNA was removed using the innuPREP Dnase I Digest Kit (both Analytik Jena AG, Jena, Germany). The RNA concentrations were determined on a NanoPhotometer (Implen GmbH, Munich, Germany). The RNA quality was evaluated by analyzing the RNA integrity number on a 2100 Bioanalyzer (Agilent), which resulted in values between 7.0 and 7.5. The cDNA synthesis (1000 ng total RNA) was performed using SensiFAST cDNA Synthesis Kit (Bioline, London, UK) and the obtained cDNA frozen at -80°C until use. Real-time qPCR was performed in duplicates using 2 µL diluted cDNA (10 ng/µL), 3 µL nuclease-free water, 0.5 µL of each primer (4 pmol/µL) listed in Supplemental Table 1 and 6 µL 2 × buffer SensiFAST SYBR No-ROX mix (Bioline) for one PCR reaction. The protocol was as follows: 60s at 95°C followed by 40 cycles of 5 s at 95°C, 10 s at 60°C and 5 s at 72°C using a LightCycler 2.0 (Roche). The efficiency of amplification was calculated using LinRegPCR software, version 2014.4 (Academic Medical Centre, Amsterdam, Netherlands). Amplicons were sequenced on an ABI 3130 Genetic Analyzer (Life Technologies GmbH, Darmstadt, Germany) to confirm sequence identity.
The experimental design bases on a 2-factoral variance analysis including MUC and CP content as factors. Based on this model, the minimum number of animals (n) in each group was calculated choosing type-I error (a= 0.05), type-II error (b = 0.2), and residual variance = 1. If aimed for a residual standard deviation equal to 1, n = 9 animals per group were required. Due to a technical problem with the climate control resulting in too low feed intake in one block, data from two animals fed NP diet were excluded from the statistical analysis. Thus, 8 HMU-NP, 9 HMU-LP, 8 LMU-NP, and 9 LMU-LP cows were considered for statistical analysis using SAS software (version 9.4, SAS Institute Inc., Cary, NC, USA). For the balance trial, means of the 4 daily collections were calculated. Data were analyzed with measurement analyses of variance using the MIXED procedure. The ANOVA model included the fixed factors MUC (levels: HMU/LMU), diet (levels: NP/LP), and the interaction of MUC × diet. Least-square means (LSM) and their standard error (SE) were computed for each fixed effect. The slice statement of the MIXED procedure was used to perform a partitioned analysis of the LSM for the interaction of group of milk urea concentration x diet, and pairwise differences were tested by using the Turkey-Kramer procedure. Results with a P-value < 0.05 were considered significant and 0.05 < P < 0.1 as trend.
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE20 partner repository with the with the dataset identifier PXD033412; username: [email protected]; password: 1qoR6BNM