All experimental procedures were approved by the Northwest A&F University Animal Care and Use Committee.
Animals, diets, and experimental design
Twenty-four primiparous Guanzhong dairy goats (113 ± 9 days in milk (DIM), 39 ± 3.8 kg of body weight (BW) at the start of the experiment) were chosen from a dairy goat farm (Shaanxi, China) and blocked into six blocks by DIM, BW, and daily milk production (DMP). Animals within each block were randomly assigned to 1 of 4 dietary treatments: control (CON), a basal diet without any additives; basal diet supplemented with FUM (Aladdin®, Shanghai, China) at 34 g/d; basal diet supplemented with NPD (J&K Scientific®, Beijing, China) at 0.5 g/d; and the basal diet supplemented with both FUM and NPD. The ration was fed as total mixed ration (TMR, Table 1) twice daily at 0730 and 1730 h and was provided individually at 105% of the expected feed intake (as-fed basis) based on the amounts of feed offered and refused from the previous day. The FUM or NPD was top-dressed on one-quarter of the TMR offered for each meal and was fed firstly to ensure complete intake. All goats were individually housed in 24 tie-stalls in a barn and had free access to water. The goats were milked twice daily at feeding time. The milk produced by the goats receiving NPD was discarded.
Table 1
Ingredients and chemical composition of the experimental diet
Item | Dry matter % |
Ingredients | |
Corn silage | 21.3 |
Alfalfa hay | 30.8 |
Ground corn | 22.9 |
Soybean meal | 6.6 |
Cottonseed meal | 5.0 |
Corn germ meal | 3.2 |
Wheat bran | 8.2 |
CaHPO4 | 0.5 |
CaCO3 | 0.5 |
NaHCO3 | 0.3 |
Salt | 0.5 |
Vitamin-mineral premixa | 0.2 |
Chemical composition | |
DM | 47.0 |
EE | 4.1 |
CP | 18.6 |
NDF | 36.1 |
ADF | 20.4 |
aVitamin-mineral premix (per kg): 600 mg of Mn, 950 mg of Zn, 430 mg of Fe, 650 mg of Cu, 30 mg of Se, 45 mg of I, 20 mg of Co, 450 mg of nicotinic acid, 800 mg of vitamin E, 45 KIU of vitamin D, and 120 KIU of vitamin A. |
The feeding experiment lasted 12 weeks, and all samples were collected or measured at wk 3, 6, 9, and 12. The six blocks of goats were divided into 3 groups by DIM, and the feeding experiment started in a staggered manner for the 3 groups with a 7-d interval so that gas emissions from each group could be measured in turn using the four indoor environmental chambers (each 7.4 m × 4.2 m × 2.7 m) available. Two goats within the same treatment were placed in one chamber and were separated by placing each in a metabolic cage (1.5 m ×1.0 m × 1.5 m). The goats were moved from barn to chambers one day before sample collection and measurements, and no stress responses were observed because they had already adapted to the chambers before the feeding experiment. On d 1–4 during each sample collecting week, total-tract digestibility of dietary nutrients, milk composition, and CH4 emissions were measured simultaneously, and the samples of blood and rumen content were collected on d 4–5 and d 5–6, respectively.
Measuring CH4 emissions and milk performance
Gas emissions in the environmental chambers were measured as previously described [17, 20] with minor changes. Briefly, the daily (22 h; 08:30 to 17:30 and 18:30 to 07:30) gas emissions from each chamber were measured in 3 consecutive days. During the gas measurement, the internal temperature of the chambers was maintained to be the same as the ambient temperature outside the building. The air inside each chamber was mixed for 30 s every 10 min by 4 draft fans. The gases from the four chambers and external environment were continuously and constantly pumped at a rate of 4 L/min by 5 exhaust fans. The pumped gases were analyzed sequentially by an FID sensor (Thermo Scientific 55i, USA), 12 min for each in every 60 min.
The daily CH4 production was calculated as follows:
CH4 production (L/d) = Σ [(Ci – Ci−1) × Vc + Vf × (Ci – COi)]/1000
Where (Ci – Ci−1) = the difference of CH4 concentration (mL/m3) every 60 min; COi = the CH4 concentration of external environment; Vc = the chamber volume (83.9 m3); and Vf = the gas volume pumped from each chamber over each 60-min measurement (0.24 m3).
During each of the two one-hour no-measurement periods, the chamber doors were opened, and the fresh-air exchange fans were running to exchange fresh air. Meanwhile, the goats were milked and fed, and the samples of milk and orts of individual goats were collected. During these 3 consecutive days, the morning and evening milk production of each goat were recorded and mixed, and 50 mL was subsampled and stored at 4°C until analysis for milk composition. Milk samples were analyzed for fat, protein, lactose, and milk urea nitrogen (MUN) using an infrared milk analyzer (MilkoScan FT 120, FOSS, Hillerød, Denmark) within 24 h. Fat corrected milk (FCM) was calculated according to NRC (2001) [21]: milk fat yield (kg/d) × 16.216 + milk yield (kg/d) × 0.4324, and net energy for lactation (NEL, Mcal/d) = milk yield, kg/d×((0.0929×percent fat)+(0.0563 ×percent true protein)+(0.0395 ×percent lactose)).
Apparent total tract digestibility and energy balance
The apparent total tract digestibility and energy balance of each goat were estimated by daily total collection of feces and urine from d 1–4 during experimental wk 3 and 9. All refusals and feces of individual goats were dried at 55℃ for 72 h in forced air ovens to a constant weight and subsample (about 100 g, wool removed) was ground through a 1-mm screen for further analysis. Urine was collected through a funnel into buckets and acidified by adding 100 ml of 10% (vol/vol) sulfuric acid to prevent microbial degradation and the loss of volatile ammonia-N. These samples were determined for DM (Method 934.01, AOAC, 2005) and CP (Method 954.01, AOAC, 2005) [22]. NDF and ADF contents were measured using the filter bag method with sodium sulfite and heat-stable α-amylase (Ankom® A200I fiber analyser, ANKOM Technology, Macedon, NY, USA). The BW was recorded twice daily after milking. The gross energy (GE) content of the samples was analyzed in an automatic adiabatic bomb calorimeter (model 1600 Parr Instrument Co., Moline, IL, USA). Digestible energy (DE) was calculated as the difference between energy intake and fecal energy; the energy lost as CH4 was calculated as the CH4 emitted in L/day × 39.54 kJ/L; metabolizable energy (ME) was the difference between DE and the sum of the energy in urine and CH4.
Collection and analysis of blood samples
Blood samples were collected from an external jugular vein into two 10-mL blood tubes before the morning feeding on two consecutive days in each sample collection week. The sample in the tube was allowed to clot at room temperature for 30 min and centrifuged (3,000 × g, 15 min) thereafter to obtain serum, which were stored at − 80°C for later analysis. Serum malondialdehyde (MDA) concentration, total antioxidant capacity (T-AOC), and the activities of serum glutathione peroxidase (GSH-Px) and superoxide dismutase (SOD) were analyzed using respective commercial kits (Jiancheng Bioengineering Institute, Nanjing, China).
Collection and analysis of ruminal samples
Ruminal content samples were collected using an oral tube and a hand vacuum pump at 6 h after the morning feeding in 2 consecutive days in each sample collection week. To minimize saliva contamination, approximately 50 mL of ruminal fluid was discarded before sample collection. Ruminal pH was measured immediately after sampling. Rumen fluid was subsampled for analysis of volatile fatty acids (VFA, 5 mL with 1 mL of 25% metaphosphoric acid added), organic acids (5 mL), and microbiota (45 mL), and then stored at -80°C until analysis.
Ruminal VFA concentration was determined using gas chromatography (Agilent Technologies 7820A GC system, Palo Alto, CA, USA) as described by Li et al. [23]. Ruminal organic acid (fumarate, succinate, and lactate) concentration was determined using an Agilent 1260 high-performance liquid chromatography system as done in previous studies [24, 25].
Bacterial community analysis
Rumen content samples of each goat from each week were freeze-dried and mixed. Microbial genomic DNA was extracted using a QIAamp DNA Stool Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instruction. The concentration and purifity of the DNA samples were analyzed using a Nanodrop spectrophotometer (Thermo Fisher Scientific, Inc., Madison, WI, USA). The V4-V5 hypervariable region (515F-926R) of the 16S rRNA gene was amplified using the primers: 5’-GTGYCAGCMGCCGCGGTAA-3’ and 5’-CCGYCAATTYMTTTRAGTTT-3’ [26] and paired-end sequenced (2×250) on the Illumina MiSeq platform.
The paired-end reads were quality-filtered, assembled, and trimmed as described previously [27]. The trimmed sequences were clustered into operational taxonomic units (OTUs) at ≥ 97% sequence similarity using Uclust in QIIME [28]. Subsequently, the OTUs were taxonomically assigned using the Silva 16S rRNA databases (SSU132; https://www.arbsilva.de/) at a confidence threshold of 80%.
Statistical analysis
The duplicate measurements (i.e. VFA and CH4) of individual goat within each sampling week were averaged as one replicate for the statistical analysis. All data were analyzed as a one-way repeated measures ANOVA using the PROC MIXED program in SAS 9.2 (SAS Institute Inc., Cary, NC, USA). The statistical model included NPD, FUM, wk, and NPD×FUM, NPD×wk, FUM×wk, and NPD×FUM×wk interactions as fixed factors, and goat and block as random effects. Sampling week was treated as a repeated measure and goat as a subject. The most desirable covariance structure (unstructured, compound symmetric, and first-order autoregressive) for analysis was determined according to the smallest Bayesian information criterion [23, 29]. When there was a treatment×wk interaction, differences among treatments at each sampling week were reanalyzed using MIXED procedure with NPD, FUM and NPD×FUM interactions as fixed factors, and block as a random effect. When there was an NPD×FUM interaction, Tukey’s multiple comparison tests were used to assess differences among treatment means.
The alpha diversity of the samples was estimated using the abundance-based coverage (ACE) estimators, Shannon diversity index, and observed OTUs. Beta diversity of the samples was computed using principal coordinates analysis (PCoA) based on Bray-Curtis dissimilarity [30] in R v.3.6.3 (http://www.R-project.org). Permutational multivariate analysis of variance was performed using the anosim() function in the R package vegan to compare the statistical difference in microbial composition across the experimental periods and between treatments.
Statistical significance was declared at P < 0.05, while tendency was declared at 0.05 ≤ P < 0.10.