Differences between Yaks and Qaidam Cattle in Digestibilities of Nutrients and Ruminal Concentration of Volatile Fatty Acids Are not Dependent on Feed Level

Abstract

The Qinghai–Tibetan Plateau (QTP) is characterized by highly fluctuating seasonal pastures. Yaks (Bos grunniens) graze at higher altitudes than Qaidam cattle (Bos taurus), but the two bovine species co-graze in their overlapping ranges. We hypothesized that yaks would digest nutrients to a greater extent and utilize energy more efficiently than cattle at low dietary intakes, but the difference between bovine species would not be apparent at high intakes. To test this hypothesis, six yaks (203 ± 6.0 kg) and six Qaidam cattle (214 ± 9.0 kg), all 3.5-year-old castrated males, were used in two concurrent 4 × 4 Latin square designs with two extra steers of each species in each period. The digestibilities of dry matter, organic matter, crude protein, ether extract, neutral and acid detergent fiber were greater (p < 0.05) in yaks than in cattle and decreased linearly (p < 0.05) when feed level (FL) increased. The average daily gain (ADG), the ratios of digestible energy (DE) to gross energy and metabolizable energy (ME) to DE, and ruminal total volatile fatty acids and ammonia-N concentrations were greater (p < 0.001) in yaks than in cattle and increased linearly (p < 0.001) when FL increased. Based on the regression equations of ADG on ME intake, the daily ME requirement for maintenance in yaks was 0.53 MJ BW^−0.75 d⁻¹, which was lesser (p < 0.05) than the 0.62 MJ BW^−0.75 d⁻¹ in cattle. We concluded that: (1) when differences between breeds emerged, the differences existed for all FLs; (2) maintenance energy requirement was lesser and ADG was greater in yaks than in cattle; (3) the digestibilities of nutrients were greater in yaks than in cattle when consuming only oat hay pellets. These findings indicate that yaks adapt to fluctuating dietary intakes in harsh environments by having a low energy requirement and high digestibility of nutrients, independent of the FL.

1. Introduction

Yaks (Bos grunniens), raised at 3000 to 6000 m above sea level (a. s. l.), are indigenous to the Asian highlands and are well adapted to the extreme, harsh climate. Today, there are approximately 17.6 million yaks worldwide, of which 95% are raised in China. Yaks are vital to the livelihood of Tibetan herdsmen residing in the Qinghai–Tibetan Plateau (QTP), providing milk, meat, fibers, fuel (dung), and transport. In addition, yaks are important to the culture and religion of the herders. Qaidam cattle (Bos taurus) were introduced to the QTP from the lowlands before the third century AD, and, today, there are approximately 10,000 cattle [1]. It was originally used as a work animal and was typically raised in agro-pastoral transition zones at an altitude between 2600 and 3600 m a. s. l [1]. Yaks and Qaidam cattle co-graze over their overlapping ranges on alpine meadows. Traditionally, yaks graze all year round without supplementary feed and are known for their low forage intake during the winter; however, cattle require dietary supplements and are provided with shelter at night during the winter [1].

The plant growing period on the QTP is short, 100 to 150 days per year, and the seasonal biomass and nutrient availability fluctuate greatly. For example, biomass was reported to be 0.25-ton dry matter (DM)/ha in winter and 4.58 ton DM/ha in summer, while crude protein (CP) content was 2.96% DM in winter and 15.4% DM in summer [2,3]. Consequently, energy and CP intakes of grazing yaks are often below the maintenance requirements during the cold winter, and, as a result, yaks can lose substantial body weight (BW).

Previous studies reported that nitrogen utilization and fiber degradation were greater in yaks than in cattle [4,5]. In addition, the rumen bacterial communities differed between these two bovine species, even under the same feeding conditions [6]. However, to our knowledge, the utilization of nutrients by yaks and cattle when dietary intakes are below energy and CP maintenance requirements is still unknown. The present study aimed to fill this gap. The concentrations of serum metabolites are important measurements as they reflect the energy and health statuses of the animals. Because of differences between these two bovine species in husbandry, nutrients metabolism, and rumen microbiota [7], we hypothesized that yaks would digest nutrients to a greater extent and utilize energy more efficiently than Qaidam cattle at low intakes, but the difference between bovine species would not be apparent at high intakes. Therefore, in comparing the two species when consuming different levels of energy, the interaction between bovine species (S) × feed level (FL) would be more significant than the effect of either S or FL. To test this hypothesis, we compared apparent digestibilities of nutrients, average daily gain (ADG), rumen fermentation parameters, and serum concentrations of metabolites and hormones between yaks and cattle. The results of this study could provide an insight into the adaptive mechanisms that allow the two bovine species to cope with the severe, harsh environment of the QTP.

2. Materials and Methods

2.1. Study Site, Animals, Diets, and Experiment Design

This study was conducted between November 2020 and March 2021 at the Wushaoling Yak Research Facility of Lanzhou University, Wuwei, China (37°12.4′ N, 102°51.7′ E, 3 154 m a. s. l. in northeastern QTP).

Six Tianzhu white yaks (203 ± 6.0 kg) and 6 Qaidam cattle (214 ± 9.0 kg), all 3.5-year-old castrated males, were kept individually in metabolic cages (1.0 m × 2.2 m) that allowed for total feces and urine collection. The animals were offered ad libitum oat hay pellets (

Table 1. The chemical composition of oat hay pellets (on DM basis).

Items	Concentration
DM, g/kg	917
OM, g/kg	868
CP, g/kg	101
EE, g/kg	45.1
NDF, g/kg	622
ADF, g/kg	364
Calcium, g/kg	4.14
Phosphorus, g/kg	1.85
GE, MJ/kg	17.6

DM = dry matter; OM = organic matter; CP = crude protein; EE = Ether extract, NDF = neutral detergent fiber; ADF = acid detergent fiber; GE = gross energy.

, Baisheng Modern Agriculture and Animal Husbandry Development, Zhangye City, China) for 14 days to measure their voluntary intake (VI). The lowest VI, 4.4 kg DM per day, was used for the study. The 4 dietary feed levels (FLs) of oat hay pellets were 0.45, 0.60, 0.75, and 0.90 VI, which were approximately 0.66, 0.88, 1.10 and 1.32 times metabolizable energy (ME) requirements for yaks, respectively [8]. Oat hay pellets were offered in two equal portions at 08:00 am and 06:00 pm daily, and drinking water was available freely. Oat (Avena sativa) is cultivated widely as a supplementary feed and covers over 70% of the cropland used for planting forages on the QTP. Energy yield and CP of oat hay are comparable to natural grass in autumn.

The animals were allocated randomly according to a 2 (species) × 4 (FLs) factorial arrangement in two concurrent 4 × 4 Latin square designs balanced for carry-over effects with 2 additional yaks and cattle; treatment sequences for the additional animals were selected randomly from columns of a separate Latin square [9]. This study included four periods, each 28 days, which consisted of 22 days (days 1 to 22) for dietary adaptation to the new FL and 6 days (days 23 to 28) for sampling and data collection. In each period, one of the FLs was offered to three yaks and cattle, and the other three FLs were offered to one yak and cattle each [10].

2.2. Experimental Procedures and Collection of Samples

The yaks and cattle were weighed prior to 08:00 on days 1 and 29, and the ADG was calculated over the 28 days of each period. In each period, the digestibilities of nutrients were measured for 5 consecutive days. The total feces were mixed each day, weighed, and a representative sample of 5% of the total daily feces (wet weight, kg) was stored in a plastic bag at −20 °C for each animal for each period. The total daily urine was collected using a funnel-shaped latex bag (International Centre for Tibetan Plateau Ecosystem Management, Lanzhou University, China) that was connected by a tube for the urine to flow to a plastic pail that contained sufficient sulfuric acid (9 mmol/L) to maintain the urine pH < 3.0. The total volume of urine was measured, and 4% of the total volume (L) was stored at −20°C for later analysis. Approximately 80 g/d of oat hay pellets were collected daily during the digestibility trials for analyses.

One hundred mL of rumen fluid were collected from each animal at 08:00, 10:00, and 14:00 on day 28 using an oral stomach tube (length, 2.6 m, outer diameter, 19 mm; Anscitech, Wuhan, China). The first 35 mL were discarded to avoid saliva contamination, and the remaining 65 mL were used for measurements. The pH was determined immediately using a pH meter (PB-10, Sartorius, Göttingen, Germany), and then the rumen fluid was passed through 4 layers of cheesecloth and was stored at −80°C in plastic tubes for volatile fatty acids (VFA) and ammonia-N analyses.

Jugular vein blood samples of the two bovines were collected in evacuated tubes on day 28 prior to morning feeding, centrifuged at 3000× g (4 °C) for 15 min, and the serum was stored at −80 °C.

2.3. Laboratory Analysis

The oat hay pellets and fecal samples were dried at 60 °C in a forced air oven (DHG-9123A, Jiecheng Experimental Apparatus, Shanghai, China) for 72 h, ground, passed through a 1-mm sieve (JFSO-100, Topu Yunnong Instrument, Hangzhou, China), and stored at room temperature. The DM (method 925.45) of the oat hay pellets and feces were determined by drying samples at 105 °C for 24 h in a forced air oven (DHG-9123A, Jiecheng Experimental Apparatus, Shanghai, China). The content of ash in the oat hay pellets and feces was determined by the complete combustion of a sample at 550 °C for 6 h in a muffle furnace (HT-12-17A1700, Shanghai Heheng Instrument Equipment, Shanghai, China), and the organic matter content (OM; method 942.05) was calculated as DM minus ash [11]. The nitrogen (N) content was measured using the micro-Kjeldahl method (K1100, Hanon instruments, Jinan, China), and the CP content was estimated as N content × 6.25. Ether extract (EE; method 920.29) was measured using a reflux system (Ankom XT 15, Fairport, NY, USA) with petroleum ether at 90 °C for 1 h [11]. Neutral detergent fiber (NDF; exclusion of heat-stable amylase and with residual ash) and acid detergent fiber (ADF) were determined by an automatic fiber analyzer (Ankom Technology, Fairport, NY, USA) [12,13]. The gross energy (GE) of oat hay pellets and feces were measured by bomb calorimetry (standard, benzoic acid, 26,470 J/g; 6400 Calorimeter, Parr Instrument Company, Moline, IL, USA).

Concentrations of ruminal VFAs were measured by gas chromatography (2010 plus system, Shimadzu Corporation, Kyoto, Japan) following Liu et al. [14] and of ammonia-N by the method of Hristov et al. [15].

The concentrations of serum albumin, β-hydroxybutyrate (BHBA), glucose, lactic acid, non-esterified fatty acid (NEFA), and total protein were determined by an automatic biochemistry analyzer (Hitachi 7160, Hitachi High-Technologies Corporation, Tokyo, Japan), following the protocols of commercial kits (Shanghai Bangyi Biological Technology Co., Ltd., Shanghai, China). The concentration of globulin was calculated as the total protein minus albumin concentration. The serum concentrations of glucagon, growth hormone (GH), insulin, insulin-like growth factor-1 (IGF-1), leptin, norepinephrine, thyroxine, and triiodothyronine were measured by commercial enzyme-linked immune sorbent assay (ELISA) kits (Shanghai Bangyi Biological Technology Co., Ltd., Shanghai, China; product No: BYE98314, BYE98273, BYE98150, BYE98149, BYE98165, BYE99016, BYE98124, and BYE98188, respectively), according to the manufacturer’s instructions.

2.4. Calculations and Statistical Analysis

The energy loss due to enteric methane (CH₄) emission of these two bovine species was estimated from regression equations that used gross energy intake (GEI) as the independent variable. In cattle, CH₄ energy emission was estimated as [16]:

CH₄ energy (MJ/d) = 0.065 × GEI (MJ/d).

(1)

In yaks, CH₄ energy loss was estimated as [17]:

CH₄ energy (MJ/d) = 0.040 × GEI (MJ/d) + 0.10.

(2)

Digestible energy intake (DEI) was calculated as:

DEI (MJ/d) = GEI (MJ/d) − fecal energy (MJ/d).

(3)

Metabolizable energy intake (MEI) was calculated as:

MEI (MJ/d) = DEI (MJ/d) − (urinary energy (MJ/d) + CH₄ energy (MJ/d)).

(4)

The ME for maintenance (ME_m) was estimated by regressing ADG on daily MEI. The regression equation took the form:

ADG (g/d) = α MEI (MJ BW^−0.75 d⁻¹) + β.
ME_m = MEI, when ADG equals to zero.

(5)

The study employed a 2 × 4 factorial arrangement of treatments, with 2 animal species (yaks and Qaidam cattle) and 4 FLs (0.45, 0.60, 0.75, and 0.90 VI), in two concurrent 4 × 4 Latin square designs. The design was not orthogonal, as there were 4 rows, 6 columns, and 4 treatments, with one treatment measured three times per period (row). However, the effect of the period was not significant and, therefore, could be omitted [9]. Data for ADG, nutrient digestibilities, rumen fermentation parameters, and serum metabolites and hormones (each replicate served as an experimental unit) were analyzed by the mixed model procedure (SAS version 9.4, SAS Inst. Inc., Cary, NC, USA). Bovine species (S) and FL were fixed effects, while experimental animals were random effects. The parameters in rumen fluid, which were collected serially, were analyzed as repeated measures. When there was a significant interaction between S × FL, means of yaks and Qaidam cattle at the same FL were compared using a t-test. Polynomial contrasts were used to determine whether the effects of FL were linear or quadratic. Significance was accepted at p-values < 0.05 and a tendency at 0.05 ≤ p-values < 0.10.

3. Results

3.1. Average Daily Gain and Apparent Digestibilities of Nutrients

The interactions between S × FL were not significant for ADG or nutrient digestibilities (p ≥ 0.10). The ADG was greater (p < 0.001) in yaks than in cattle and increased linearly (p < 0.001) with increasing FL (Figure 1). Based on the regression equations between ADG and MEI, ME_m in yaks was lesser (p < 0.05) than in cattle (0.53 vs. 0.62 MJ BW^−0.75 d⁻¹; Figure 2). The digestibilities of DM, OM, CP, EE, NDF, and ADF were greater (p < 0.05) in yaks than in cattle and decreased linearly (p < 0.05) with increasing FL (

Table 2. Apparent digestibilities of dry matter and dietary nutrients in yaks and cattle offered oat hay pellets at different feeding levels.

Items	S	Oat Hay Pellets FL				SEM	p-Values
		0.45 VI n = 6	0.60 VI n = 6	0.75 VI n = 6	0.90 VI n = 6		S	FL	S × FL	FL-L	FL-Q
DM, g/kg	Yak	656	647	644	626	15.6	0.042	0.038	0.820	0.042	0.410
	Cattle	630	627	621	596
OM, g/kg	Yak	675	672	662	642	11.2	0.046	<0.01	0.951	0.030	0.190
	Cattle	657	652	642	618
CP, g/kg	Yak	711	701	692	677	11.9	<0.001	<0.001	0.234	0.015	0.463
	Cattle	682	676	672	656
EE, g/kg	Yak	707	666	638	637	13.5	<0.001	<0.01	0.893	<0.001	0.410
	Cattle	643	626	619	602
NDF, g/kg	Yak	592	581	573	546	10.8	0.027	0.016	0.908	<0.01	0.431
	Cattle	565	555	557	533
ADF, g/kg	Yak	556	549	536	515	11.8	0.021	<0.001	0.927	<0.001	0.167
	Cattle	542	536	521	494

S = species; FL = feeding level; SEM = standard error of means; VI = voluntary intake; DM = dry matter; OM = organic matter; CP = crude protein; EE = ether extract; NDF = neutral detergent fiber; ADF = acid detergent fiber. FL-L = Linear effect of oat hay pellets FL; FL-Q = Quadratic effect of oat hay pellets FL.

).

As designed, GEI increased linearly (p < 0.001) with increasing FL, with no difference between species (

Table 3. Gross, digestible, and metabolizable energy intakes and energy output in yaks and in cattle offered oat hay pellets at different feeding levels.

Items	S	Oat Hay Pellets FL				SEM	p-Values
		0.45 VI n = 6	0.60 VI n = 6	0.75 VI n = 6	0.90 VI n = 6		S	FL	S × FL	FL-L	FL-Q
GE intake, MJ/d	Yak	35.4	47.3	59.1	71.1	0.17	0.885	<0.001	0.964	<0.001	0.255
	Cattle	35.4	47.3	59.1	71.1
Fecal energy, MJ/d	Yak	13.4	17.1	20.7	25.1	0.48	0.014	<0.001	0.497	<0.001	0.617
	Cattle	13.9	17.7	22.2	26.0
DE intake, MJ/d	Yak	22.0	30.1	38.4	46.0	0.51	0.041	0.036	0.792	<0.001	0.813
	Cattle	21.5	29.6	36.9	45.1
DE/GE	Yak	0.622	0.637	0.649	0.647	0.0079	0.013	0.013	0.838	<0.01	0.551
	Cattle	0.608	0.626	0.625	0.634
Urinary energy, MJ/d	Yak	0.34	0.36	0.41	0.45	0.013	<0.01	<0.001	0.353	<0.001	0.622
	Cattle	0.29	0.34	0.39	0.42
Methane energy, MJ/d	Yak	1.52 A	1.99 A	2.47 A	2.94 A	0.179	<0.001	<0.001	<0.001	<0.001	0.208
	Cattle	2.30 B	3.07 B	3.84 B	4.62 B
ME intake, MJ/d	Yak	20.2	27.8	35.5	42.6	0.44	<0.001	<0.001	0.089	<0.001	0.901
	Cattle	18.9	26.2	32.7	40.1
ME/DE	Yak	0.916	0.922	0.925	0.926	0.0033	<0.001	0.018	0.957	<0.01	0.302
	Cattle	0.879	0.886	0.885	0.888

S = species; FL = feeding level; VI = voluntary intake; GE = gross energy; DE = digestible energy; ME = metabolizable energy. ^A,B Means with different superscript letters in the same column within an item are significantly different from each other (p < 0.05). FL-L = Linear effect of oat hay pellets FL; FL-Q = Quadratic effect of oat hay pellets FL.

). The fecal energy of yaks was lesser (p = 0.014) than cattle and increased linearly (p < 0.001) in both species when FL increased. The DEI and the ratio of DE:GE were greater (p < 0.05) in yaks than in cattle and increased linearly (p < 0.05) when FL increased. Urinary energy was greater (p < 0.01), while the estimated CH₄ emission was lesser (p < 0.001) in yaks than in cattle, and both increased linearly (p < 0.001) with increasing FL. The ME intake and the ratio of ME:DE were greater (p < 0.001) in yaks than in cattle, and both increased linearly (p < 0.01) when FL increased.

3.2. Rumen Fermentation

The S × FL interactions were not significant for the ruminal concentrations of the total volatile acids (TVFAs) and ammonia-N and molar proportions of individual VFAs (p ≥ 0.10). Ruminal pH decreased linearly (p < 0.001) when FL increased and did not differ (p = 0.076) between yaks and cattle (

Table 4. The rumen fermentation parameters in yaks and in cattle offered oat hay pellets at different feeding levels.

Items	S	Oat Hay Pellets FL				SEM	p-Values
		0.45 VI n = 18	0.60 VI n = 18	0.75 VI n = 18	0.90 VI n = 18		S	FL	S × FL	FL-L	FL-Q
pH	Yak	7.10	6.90	6.81	6.79	0.098	0.076	<0.001	0.311	<0.001	0.510
	Cattle	7.04	6.84	6.80	6.71
Ammonia-N, mg/100 mL	Yak	6.65	6.89	7.44	8.06	0.220	<0.001	<0.001	0.216	<0.001	0.779
	Cattle	6.02	6.47	6.99	7.10
Total VFAs, mmol/L	Yak	53.8	61.6	63.0	65.0	0.73	<0.001	<0.001	0.379	<0.001	0.231
	Cattle	50.5	59.2	60.3	64.3
VFAs, mol/100 mol
Acetate	Yak	73.3	72.5	71.1	71.0	0.46	0.463	<0.01	0.435	<0.001	0.545
	Cattle	73.4	72.6	72.1	71.7
Propionate	Yak	14.8	14.5	14.9	14.7	0.33	0.372	0.767	0.295	0.613	0.460
	Cattle	14.7	15.0	14.8	15.6
Butyrate	Yak	8.66	9.58	10.7	10.8	0.397	0.281	<0.001	0.813	<0.001	0.103
	Cattle	8.40	9.30	10.0	9.70
Iso-VFAs	Yak	3.25	3.34	3.24	3.46	0.211	0.543	0.283	0.153	0.067	0.645
	Cattle	3.52	3.09	3.10	2.95
Acetate: propionate	Yak	4.97	4.99	4.81	4.83	0.116	0.585	0.502	0.398	0.191	0.514
	Cattle	5.00	4.86	4.89	4.60

S = species; FL = feeding level; VI = voluntary intake; VFAs = volatile fatty acids; N = nitrogen. Total VFAs = acetate + propionate + butyrate + iso-VFAs. FL-L = Linear effect of oat hay pellets FL; FL-Q = Quadratic effect of oat hay pellets FL.

). The concentration of total VFAs in yaks was greater (p < 0.001) than in cattle, and both increased linearly (p < 0.001) when FL increased. No difference (p > 0.10) was observed between species in molar proportions of acetate, propionate, butyrate, and iso-VFAs, and the ratio of acetate:propionate, but acetate decreased linearly (p < 0.01), whereas butyrate increased linearly (p < 0.001) when FL increased. The ruminal concentration of ammonia-N was greater (p < 0.001) in yaks than in cattle and increased linearly (p < 0.001) when FL increased.

3.3. Serum Metabolites and Hormones

The S × FL interactions were not significant for serum concentrations of metabolites and hormones (p ≥ 0.10). The concentration of serum glucose was greater (p = 0.021) in yaks than in cattle and decreased linearly (p < 0.01) when FL increased, whereas BHBA did not differ (p > 0.10) between the species but decreased linearly (p = 0.044) with increasing FL (

Table 5. The concentrations of serum metabolites in yaks and in cattle offered oat hay pellets at different feeding levels.

Items	S	Oat Hay Pellets FL				SEM	p-Values
		0.45VI n = 6	0.60VI n = 6	0.75VI n = 6	0.90VI n = 6		S	FL	S × FL	FL-L	FL-Q
Glucose, mmol/L	Yak	5.52	5.23	5.01	4.36	0.164	0.021	<0.01	0.225	<0.01	0.799
	Cattle	4.80	4.59	4.60	4.13
BHBA, umol/L	Yak	273	264	262	238	12.3	0.123	0.044	0.952	<0.01	0.430
	Cattle	260	250	239	225
NEFA, umol/L	Yak	727	666	653	641	42.8	0.049	0.802	0.954	0.451	0.673
	Cattle	631	585	578	562
Lactic acid, mmol/L	Yak	1.24	0.91	1.16	1.22	0.120	0.964	0.480	0.129	0.328	0.730
	Cattle	0.92	1.10	1.14	1.32
Total protein, g/L	Yak	64.3	60.0	62.9	66.7	3.23	0.274	0.759	0.502	0.347	0.663
	Cattle	61.9	64.4	69.8	67.8
Albumin, g/L	Yak	39.3	38.5	39.0	45.5	2.65	0.334	0.745	0.530	0.924	0.286
	Cattle	43.1	41.9	43.0	42.4
Globulin, g/L	Yak	25.0	21.5	23.8	21.2	4.06	0.883	0.819	0.574	0.413	0.757
	Cattle	21.9	22.5	26.9	25.4

S = species; FL = feeding level; VI = voluntary intake; BHBA = β-hydroxybutyrate; NEFA = non-esterified fatty acid. FL-L = Linear effect of oat hay pellets FL; FL-Q = Quadratic effect of oat hay pellets FL.

). The concentration of NEFA was greater (p = 0.049) in yaks than in cattle but did not differ among FLs. The serum concentrations of lactic acid, total protein, albumin, and globulin were not affected by either species, FL, or their interactions (p ≥ 0.10).

The serum concentration of glucagon was higher (p = 0.012) in yaks than in cattle and decreased linearly (p = 0.035) when FL increased, whereas insulin did not differ between yaks and cattle but increased linearly (p < 0.01) when FL increased (

Table 6. The serum hormones concentrations in yaks and in cattle offered oat hay pellets at different feeding levels.

Items	S	Oat Hay Pellets FL				SEM	p-Values
		0.45VI n = 6	0.60VI n = 6	0.75VI n = 6	0.90VI n = 6		S	FL	S × FL	FL-L	FL-Q
Insulin, mIU/L	Yak	19.9	20.4	21.0	21.8	0.65	0.485	0.064	0.997	<0.01	0.863
	Cattle	19.5	19.9	20.7	21.3
Glucagon, pg/mL	Yak	380	369	350	341	12.1	0.012	0.028	0.750	0.035	0.412
	Cattle	344	339	338	308
Leptin, ng/mL	Yak	17.1	17.3	17.3	17.1	0.31	0.047	0.442	0.989	0.213	0.731
	Cattle	16.7	16.9	16.8	16.9
Growth hormone, ng/mL	Yak	12.3	12.7	13.2	13.3	0.52	<0.01	<0.001	0.367	<0.001	0.914
	Cattle	12.1	12.3	12.6	13.1
IGF-1, ng/mL	Yak	304	283	280	277	16.0	0.512	0.646	0.995	0.276	0.870
	Cattle	291	281	274	272
Triiodothyronine, nmol/L	Yak	11.7	12.6	11.1	11.4	0.69	0.292	0.222	0.539	0.388	0.287
	Cattle	12.7	13.8	11.8	10.8
Thyroxine, nmol/L	Yak	46.6	48.6	47.8	48.5	2.78	0.974	0.891	0.848	0.614	0.889
	Cattle	49.3	48.7	47.0	46.8
Norepinephrine, ng/mL	Yak	5.64	5.29	5.34	5.77	0.362	0.666	0.266	0.725	0.618	0.186
	Cattle	5.49	4.73	5.54	5.85

S = species; FL = feeding level; VI = voluntary intake; IGF-1 = insulin-like growth factor-1. FL-L = Linear effect of oat hay pellets FL; FL-Q = Quadratic effect of oat hay pellets FL.

). The serum concentration of leptin was greater (p = 0.047) in yaks than in cattle and was not affected by FL, whereas serum concentration of GH was also greater (p < 0.01) in yaks than in cattle but increased linearly (p < 0.001) when FL increased. The concentrations of serum IGF-1, triiodothyronine, thyroxine, and norepinephrine were not affected by either species, FL, or their interactions (p > 0.10).

4. Discussion

The interactions between S × FL were not significant when comparing ADG, the digestibilities of dry matter (DM) and nutrients, ruminal concentrations of total volatile fatty acids (VFAs) ammonia-N, molar proportions of individual VFAs, and the serum concentrations of metabolites and hormones (p ≥ 0.10). Therefore, our hypothesis was rejected, and differences between bovine breeds held for all feed levels.

4.1. Body Weight Changes and Digestibilities of Nutrients in Yaks and in Cattle

As expected, the ADGs of yaks and cattle increased when FL increased, which was in accordance with previous studies in sheep [14,18]. Both species lost BW at the two low FLs and gained BW at the two high FLs. However, losses were lesser, and gains were greater in yaks than in cattle; consequently, the ADG in yaks was greater than in cattle. The estimated ME_m for yaks was 14.5% lesser than for cattle (0.53 vs. 0.62 MJ BW^−0.75 d⁻¹). The ME_m for yaks was similar to the value for grazing yaks (0.55 MJ BW^−0.75 d⁻¹) [19]; however, it was 15.2% greater than for feedlot yaks (0.46 MJ BW^−0.75 d⁻¹) [8]. If we assume that the difference in metabolic rates between the current study and the results of Han et al. [8] was due to the heat increment of feeding, then the efficiency of utilization of energy for maintenance (BMR/ME_m or k_m) equaled 0.83 (0.46 MJ BW^−0.75 d⁻¹/0.53 MJ BW^−0.75 d⁻¹).

The digestibilities of all nutrients were greater in yaks than in cattle. Ruminants are not able to degrade plant fibers without ruminal microorganisms, especially bacteria, which contain fibrolytic enzymes that convert structural carbohydrates to VFAs. Fibrolytic bacteria, such as Ruminococcus at the genus level, and Ruminococcaceae at the family level, are the most abundant bacteria in the gastrointestinal tract of yaks, which could explain, at least in part, the greater NDF and ADF digestibilities in yaks than in cattle [20,21]. Furthermore, Verrucomicrobia was also identified as a dominant bacterium in the gastrointestinal microbiota for yaks, and this phylum plays a vital role in plant fiber degradation [21,22].

Tibetan sheep, a high-altitude ruminant that co-grazes with yaks on the QTP, digested nutrients to a greater extent than fine-wool and Small-tailed Han sheep [14,23]. High digestibilites of energy and nutrients may have evolved in high-altitude grazing herbivores in order to better utilize the sparse vegetation available during long, cold winter on the QTP. In addition, the size of the rumen relative to the omasum is larger in yaks than in lowland cattle, which prolongs the rumen retention time in yaks [24].

As FL increased, the digestibilities of all nutrients decreased linearly, which is consistent with previous studies in yaks, goats, and Tan sheep [14,25,26]. With a low feed intake, mastication is greater, and the passage rate of digesta is slower, which increases the digesta retention time, and, ultimately, increases the digestibility of nutrients [27].

The GEI increased linearly with increasing FL, with no difference in intake between yaks and cattle. Fecal energy loss also increased when FL increased in these two species but was lesser in yaks than in cattle. Fecal energy was derived mainly from undigested dietary NDF, and the greater fiber digestibility in yaks than in cattle could explain the difference in fecal energy between species and the greater DEI and DE:GE ratio in yaks than in cattle [28]. Urinary energy excretion increased linearly when FL increased, which was consistent with a study on dairy cows [29]. In addition, urinary energy excretion was greater in yaks than in cattle. Urinary energy was correlated positively with urinary N concentration in ruminants, but this did not occur in the present study, as urinary N output was lower in yaks than in cattle (40.0 vs. 42.8 g/d, unpublished data) [30]. This suggests that the components in the urine, other than urea, had greater energy content in yaks than in cattle.

Methane emission, estimated from predictive equations, increased linearly when FL increased. The increase with increasing FL could be explained by the high correlation between enteric CH₄ emission and DMI or GEI in ruminants [28]. The estimated energy loss due to CH₄ emission ranged between 1.52 and 2.94 MJ/d for yaks, which was approximately 35% less than the loss for cattle, which ranged between 2.30 and 4.62 MJ/d. There is some support for this difference in CH₄ emission between species. An in vitro study demonstrated that rumen inoculum from yaks produced less CH₄ than from cattle when the same substrate (barley straw and concentrate at 1:1) was used, mainly because yaks produced less H₂ [31]. Moreover, yaks possess a “special rumen microbial ecosystem” that differs substantially from cattle, in which the metagenome has fewer methanogenesis pathways than cattle [32,33].

The MEI and the ratio of ME:DE increased when FL increased and were greater in yaks than in cattle. The lesser loss of estimated CH₄ energy in yaks than in cattle was the main reason for these differences between the species. Both ME and CH₄ energy increased with increased feed intake; however, as a proportion of MEI, CH₄ energy decreased to a greater extent in yaks than in cattle with increasing FL. It is apparent that yaks use a relatively low CH₄ emission as a means to increase MEI, a strategy that has also been noted for goats [27].

4.2. Rumen Fermentation

In the current study, ruminal pH fell within the normal range of 6.2 to 7.2 for yaks and cattle at all FLs, and no difference was observed between species. The pH decreased linearly when FL increased, which is consistent with a study in sheep [14]. The increase in the concentration of VFAs when FL increased in the present study could explain the decrease in pH, as ruminal pH is correlated negatively with VFAs concentration.

Yaks had a greater concentration of ruminal total VFAs than cattle, which is in agreement with an in vivo study when both species consumed the same diet, and with an in vitro study when the same substrate was used with rumen inocula for both species [32,33]. The greater concentration of VFAs in yaks than in cattle implied a greater substrate availability in yaks than in cattle, which resulted, at least in part, from the difference in the digestibility of nutrients between the species. In addition, there are 36 up-regulated genes in the ruminal epithelium of yaks linked with VFA production and absorption, and the rumen bacterial genes of yaks have more VFA-production pathways than in cattle [33]. The total VFA concentration increased with increasing FL, which is consistent with results for Tan sheep, and could be explained by the increased fermentable substrate with increasing FL [14].

The concentration of ruminal ammonia-N fell between 5 and 25 mg/100 mL for yaks and cattle at all FLs, which was reported to be the range for optimal microbial protein production [34]. As FL increased, the concentration of ammonia-N increased linearly, which was consistent with the results in Tibetan sheep and implied an increase in fermentable carbohydrates with increasing FL [35]. The concentration was greater in yaks than in cattle, which suggested a greater microbial protein yield in yaks than in cattle [4]. The relative abundances of fibrolytic and H₂-incorporating bacteria were greater, but those of amylolytic bacteria were lesser in yaks than in cattle [36].

4.3. Serum Metabolites and Hormones

Serum metabolites are often used to assess the physiological, nutritional, and health statuses of livestock. The concentration of serum glucose was greater in yaks than in cattle and decreased with an increase in FL in the two bovines. However, both species were well below the renal threshold of 5.56 to 7.78 mmol/L for cattle [37]. It was reported that serum GH is related positively to ADG, which occurred in the present study, as GH and ADG were greater in yaks than in cattle [38]. Furthermore, an increase in serum GH is generally associated with an increase in serum glucose, which could explain the greater concentration of serum glucose in yaks than in cattle. The yaks had a greater concentration of serum glucagon than cattle, while the concentration of serum insulin did not differ between the species, which could have also led to an increase in serum glucose. Glucagon is a catabolic hormone that promotes gluconeogenesis and lipolysis and increases the output of glucose from the liver. In contrast, insulin is an anabolic hormone that is responsible for the storage of glucose in peripheral tissues and decreases the concentration of serum glucose. The decrease in the concentration of serum glucose with an increase in FL was also reported in cattle and could be explained by the decrease in glucagon and increase in insulin with increasing FL [39]. Leptin increased the digestibility of nutrients in cattle [40], as was observed in the present study, with greater nutrient digestibilities in yaks than in cattle, and was correlated positively with ADG [41], which was also observed in this study.

Both species lost BW at the two low FLs and presumably were in negative energy balance. Body fat is mobilized, and the serum concentrations of NEFA and the ketone BHBA increase with a negative energy balance [42]. Consequently, the serum NEFA and BHBA concentrations were expected to decrease when FL increased. The serum BHBA concentration decreased linearly with increasing FL but did not differ between the species. The serum BHBA concentrations (0.225 to 0.273 mmol/L) were well below the subclinical level (between 1.0 and 1.4 mmol/L) for both species and for all FLs [43]. NEFA, the major component of triglycerides, is used as an energy source by many tissues and can reflect the nutritional status of the animal. They are generally derived from ingested dietary triglycerides; however, when in negative energy balance, NEFAs are produced by the hydrolysis of adipose tissue. However, the serum concentration of NEFAs did not change in both species with an increase in FL and BW. The serum NEFA concentrations were higher in yaks than in cattle, which was inconsistent with the greater BW loss in cattle than in yaks and, presumably, greater energy loss.

Proteins, the main and most abundant constituents of the blood serum, have many essential physiological functions, and are used as indicators of the health status of ruminants. Albumin is the most abundant protein, comprising 35 to 50% of total serum protein. It maintains homeostasis, contributing 75% of the osmotic pressure, transports substances, and scavenges free radicals. Globulins regulate inflammatory processes and defend against pathological damage. In the present study, the serum concentrations of the total protein, albumin, and globulins did not differ between species and among FLs and were generally in the normal range for healthy livestock. Therefore, neither species was affected by FL, which is in agreement with previous studies in cattle [44,45].

5. Conclusions

The digestibilities of nutrients and the concentration of total ruminal VFAs and ADG were greater in yaks than in cattle. The ratios of DE to GE and ME to DE were greater in yaks than in cattle, and the daily ME requirement for maintenance in yaks was lesser than in cattle. We concluded that yaks could utilize dietary nutrients more efficiently than cattle, independent of feed level. These findings provide insights into the adaptations that enable yaks to cope with the harsh QTP better than cattle.