2.2. Experimental Dietary Treatments and Feed Management
The four roughage-based basal dietary treatments supplemented with formulated concentrate in TMR bases were: Control (natural pasture hay (NPH)); treated teff straw silage (TTS); Napier grass hay (NGH); and
Brachiaria hybrid grass hay (BhH). The natural pasture hay was purchased from a private dairy farm and was composed predominantly of grass species (
Andropogon,
Cynodon,
Digitaria,
Hyparrhenia, and
Panicum spp.) and legumes (
Trifolium quartinianum,
T. polystachyum, and
Indigofera atriceps) as characterized by Denekew et al. [
29]. The improved forages were planted in accordance with the recommended agronomic practices for Napier grass (accession number 1574) [
13,
30] and
Brachiaria grass (accession number CIAT 36087) [
15] in 2.65 ha of irrigated land at the Andassa Livestock Research Center. The forages were irrigated to support continuous growth for harvesting sufficient amounts of hay. The harvested forage was air-dried for 2–3 days in the field and mixed before storage for the feeding trial. The hay was chopped into 2–5 cm lengths and mixed with the concentrate in the TMRs. For chemical analysis, samples of each forage were taken and oven-dried to a constant dry weight, ground to pass through a 2-mm screen, and then stored in airtight plastic bags. For teff straw treated silage preparation, 1 L of effective microbes, 1 kg of molasses, 2.5 kg of urea, and 18 L of chlorine-free water were mixed and used to treat 50 kg of teff straw as described previously by Dejene et al. [
31]. Treated teff straw silage was put into airtight plastic bags and well packed to avoid trapping air. To facilitate anaerobic fermentation, filled bags were kept at room temperature (25 °C) for 21 days, as recommended for environmental conditions such as those in an Ethiopian climate [
26]. The formulated concentrate mixture was purchased from a private animal-feed factory company and consisted of maize (
Zea mays) (40%), noug (
Guizotia abyssinica) seed cake (49%), wheat bran (
Triticum Aestivum) (8%), iodized salt (1%), and ruminant premix (2%). The ruminant premix (produced by INTRACO Ltd., Antwerp, Belgium) contained the following additives (per kg): Ca, 1310.5 g; Na, 192.2 g; Mg, 520.8 g; Fe, 5000 mg; Mn, 5000 mg; Zn, 10,000 mg; I, 150 mg; Se, 40 mg; Co, 15 mg; vitamin A, 999,750 IU; vitamin D3, 199,950 IU; vitamin E, 800 mg; butylated hydroxytoluene, 50 mg; and ethoxyquin, 55 mg.
Cows were fed TMR diets consisting of 70% roughage and 30% concentrate on a dry matter (DM) basis. To achieve the targeted milk yield (4 kg milk/d), the TMR diets were formulated based on National Research Council [
32] for the maintenance and lactation requirement of the lactating dairy cow (average body weight, 238.3 kg) for each treatment group. The TMR diets were offered in the morning (8:00 h), at noon (12:00 h), and during the afternoon (16:00 h). Cows had individual and free access to drinking water throughout the entire experiment. The chemical composition of feed ingredients and the TMR diets are presented in
Table 1. The acid detergent lignin (ADL)/neutral detergent fiber (NDF) ratio of the dietary treatments for NPH, TTS, NGH and BhH is 0.13, 0.17, 0.22, and 0.25, respectively.
2.3. Measurements and Sample Collection
The TMR feed offered and the orts were recorded daily at each feeding time over the whole experimental period. From day 15 to 21 of each experimental period, samples of the TMR offered and ort were collected and stored for laboratory analysis. Daily milk yields for all 8 cows were recorded at each milking time. Cows were hand-milked twice daily, and the milking time was in the morning (08:00 h) and afternoon (16:00 h) throughout the experiment. During each period (from 15 to 21 days), milk samples were collected at each milking time (08:00 h and 16:00 h), pooled by cow, and then transported in an ice box for milk composition analysis. Another pooled milk sample for each period was stored immediately at −20 °C for determination of milk urea nitrogen (MUN) content.
Spot urine samples from all cows were collected in plastic containers between days 15 and 21 of each experimental period three times per day (07:00, 13:00, and 17:00 h) and immediately stored at –20 °C for determination of N and creatinine contents. A sample of 100 mL urine was collected from every cow and 8 mL of an aqueous solution of sulfuric acid 10% (v/v) added, in which outflowing air was led to trap aerial ammonia and reduce urine pH to below 3, as described previously [
33]. Fecal samples were collected on days 15 through 21 of each experimental period by direct sampling from the rectum of each cow in the morning (07:00), afternoon (13:00), and evening (17:00 h). Composite fecal samples from each cow in each period were maintained at −20 °C until laboratory analysis.
After feeding the cows on the twenty-first day of each experimental period, blood samples were collected from the jugular veins of individual animals into 10 mL sodium heparin and potassium EDTA vacuum tubes, as described previously by Nichols et al. [
28]. The blood samples were pooled over sampling time points according to cow and period; plasma was prepared by centrifuging blood at 1000×
g for 5 min at 23 °C; the supernatant was transferred to identified plastic tubes and stored at –20 °C until laboratory analysis. Rumen fluid samples were collected from each cow two hours after the morning feeding on the twenty-first day of each experimental period by inserting a rumen-fluid collector through an esophageal gavage with a manual sucker [
34]. The samples were temporarily placed on ice and then processed for ruminal ammonia nitrogen analysis. The body weight (BW) of each dairy cow was weighed at the start, middle, and end of each period for the whole experiment time by using a ground weight balance; the weight of cows was obtained by averaging the weights taken before feeding and after milking on two successive days.
2.4. Laboratory Analyses and Procedures
All feed and fecal samples were oven-dried at 60 °C for 48 h. The dried samples were ground to pass through a 2-mm screen and stored in plastic bags for subsequent determination of chemical components. The DM and organic matter (OM) contents of the diets and feces were determined according to AOAC [
35]. The crude protein (CP) concentration was estimated by multiplying the N concentrations by 6.25, and N concentrations were determined according to the Kjeldahl method [
36]. The neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) contents of the feed and feces were determined according to the procedures of Goering et al. [
37]. The metabolizable energy (ME) concentration (MJ/kg DM) was estimated on the basis of 24-h gas production (mL) from in vitro gas fermentation [
38];
where GP is the 24-h gas production volume (mL/0.2 g DM) and CP is the crude protein content (%) of the feed. Estimated net energy maintenance (NEm) and net energy lactation (NE
l) were calculated according to the following equation [
32]:
where the gross energy (GE) concentrations of feed were determined by using a bomb calorimeter (CA-4AJ, Shimadzu Corporation, Kyoto, Japan). For the TMR, these parameters were calculated according to the ratio composition of roughage and concentrate ingredients. Analysis of in vitro gas production for the feed samples was conducted at Shimane University (Japan) in accordance with the procedure described by Mekuriaw et al. [
39].
Feed intake was determined on days 16 through 21 of each period as the difference between the weight of the TMR feed offered to the animals and that left unconsumed:
whereas nutrient digestibility was calculated as ([nutrient intake (kg/d) – fecal nutrient output (kg/d)]/nutrient intake) × 100% [
40].
Milk fat, protein, and lactose contents were determined by using a LactoScan milk analyzer (Milkotronic Ltd., Nova Zagora, Bulgaria). Fat- and protein-corrected mik production (FPCM) (kg/d) was calculated according to De Koster et al. [
41]:
The milk yield efficiency was calculated according to Nichols et al. [
10]:
The milk urea nitrogen (MUN) concentrations were obtained through spectrophotometric enzymatic colorimetric methodology using commercially available kits (urease–GIDH method, λ = 525 nm; product code 410-55391 _ 418-55191, Fujifilm, Wako, Japan).
Urinary creatinine concentrations (mg/dL) were determined by using commercial enzymatic colorimetric assay kits (sarcosine oxidase method,
λ = 515 nm; product code 439-90901, Fujifilm, Wako, Japan) and UV–VIS spectrophotometry (DR6000, Hach, Dusseldorf, Germany). Total daily urine volume was estimated by dividing daily urinary creatinine excretions by the observed creatinine concentration of spot urine samples, assuming a daily creatinine excretion of 0.197 ± 0.047 mmol/kg BW [
42]. Urinary N was analyzed according to the Kjeldahl method [
35] by using the same equipment for feed, feces, and milk samples; hence, urinary N excretion was calculated by multiplying urinary N by urine volume. Nitrogen excreted in milk was calculated by using the equation [
43]:
The N excreted in feces was determined as [
44]:
Fecal output was estimated by using chromium oxide as an external indicator according to Kimura et al. [
44]. Chromium oxide was given daily (5 g for each cow) in the morning feeding on days 15 through 21 of each experimental period. The N balance was obtained by subtracting the values for N in urine, feces, and milk from the total N intake in grams [
43]:
Plasma underwent determination of glucose, non-esterified fatty acid (NEFA), β-hydroxybutyrate (BHBAA), and blood urea nitrogen (BUN) contents. These parameters were measured by using commercial enzymatic colorimetric assays (glucose: mutarotase-GOD method, product code 439-90901,
λ = 455 nm; NEFA: ACS-ACOD method, product code 279-75401,
λ = 550 nm; BHBAA, cyclic enzyme method, product code 279–75401,
λ = 405 nm; and BUN: urease-GIDH method, product code 410–55391 _ 418–55191,
λ = 340 nm; all from Fujifilm) and a UV–VIS spectrophotometer (DR6000, Hach). Immediately after collection, rumen fluid was tested for pH by using a handheld portable pH meter and then stored at −20 °C until analysis of ruminal ammonia nitrogen concentration. For ruminal fluid samples that were preserved with 1% H
2SO
4, the rumen fluid collected was centrifuged at 2000×
g for 15 min and the resulting supernatant was analyzed for ruminal ammonia nitrogen concentration, as described previously [
45].
2.6. Statistical Analysis
Data regarding feed intake, nutrient digestibility, nitrogen balance, plasma metabolites, milk yield, and milk composition were analyzed through analysis of variance (ANOVA) using the mixed model procedure of SAS (version 9.4; SAS Institute Inc., Cary, NC, USA). The model is:
where, Y
ijk represents the observation on cow
k given treatment
i at period
j;
µ is the overall mean;
Ti represents the fixed effect of the ith diet treatment, i = 1 to 4;
Pj represents the random effect of the jth period, j = 1 to 4;
Ck represents the random effect of the kth cow, k = 1 to 2;
εijk is the random residual error.
The model contained treatment as fixed effects, period and cow as random effects. Differences among the means were considered significant at the p ≤ 0.05 level according to Tukey’s test. To assess the relationships among plasma urea concentration, milk urea concentration, and ruminal ammonia N, as well among methane emission and milk yield, regressions were fitted by using the pooled data of all dietary treatments.