3-Nitrooxypropanol supplementation of a forage diet decreased enteric methane emissions from beef cattle without affecting feed intake and apparent total-tract digestibility

Abstract

Supplementation of ruminant diets with the methane (CH₄) inhibitor 3-nitrooxypropanol (3-NOP; DSM Nutritional Products, Switzerland) is a promising greenhouse gas mitigation strategy. However, most studies have used high grain or mixed forage-concentrate diets. The objective of this study was to evaluate the effects of supplementing a high-forage diet (90% forage DM basis) with 3-NOP on dry matter (DM) intake, rumen fermentation and microbial community, salivary secretion, enteric gas emissions, and apparent total-tract nutrient digestibility. Eight ruminally cannulated beef heifers (average initial body weight (BW) ± SD, 515 ± 40.5 kg) were randomly allocated to two treatments in a crossover design with 49-d periods. Dietary treatments were: 1) control (no 3-NOP supplementation); and 2) 3-NOP (control + 150 mg 3-NOP/kg DM). After a 16-d diet adaption, DM intake was recorded daily. Rumen contents were collected on days 17 and 28 for volatile fatty acid (VFA) analysis, whereas ruminal pH was continuously monitored from days 20 to 28. Eating and resting saliva production were measured on days 20 and 31, respectively. Diet digestibility was measured on days 38–42 by the total collection of feces, while enteric gas emissions were measured in chambers on days 46–49. Data were analyzed using the mixed procedure of SAS. Dry matter intake and apparent total-tract digestibility of nutrients (DM, neutral and acid detergent fiber, starch, and crude protein) were similar between treatments (P ≥ 0.15). No effect was observed on eating and resting saliva production. Relative abundance of the predominant bacterial taxa and rumen methanogen community was not affected by 3-NOP supplementation but rather by rumen digesta phase and sampling hour (P ≤ 0.01). Total VFA concentration was lower (P = 0.004) following 3-NOP supplementation. Furthermore, the reduction in acetate and increase in propionate molar proportions for 3-NOP lowered (P < 0.001) the acetate to propionate ratio by 18.9% as compared with control (4.1). Mean pH was 0.21 units lower (P < 0.001) for control than 3-NOP (6.43). Furthermore, CH₄ emission (g/d) and yield (g/kg DMI) were 22.4 and 22.0% smaller (P < 0.001), respectively, for 3-NOP relative to control. Overall, the results indicate that enteric CH₄ emissions were decreased by more than 20% with 3-NOP supplementation of a forage diet without affecting DM intake, predominant rumen microbial community, and apparent total-tract nutrients digestibility.

Providing beef producers with effective enteric methane mitigation options is crucial in reducing the environmental impact of beef production. The present study showed that supplementation of high-forage diets with 3-nitrooxypropanol has tremendous potential for enteric methane mitigation.

Introduction

Methane (CH₄) is produced by ruminant animals during the normal enteric fermentation of ingested feed and represents 2–12% loss of dietary energy to the animal (Johnson and Johnson, 1995). The negative effects of enteric methanogenesis on both the environment (potent greenhouse gas) and feed utilization efficiency have motivated scientists over the past decades to find CH₄ mitigation strategies that are effective and economical (Beauchemin et al., 2020).

In an effort to minimize the environmental impacts of ruminants by reducing enteric CH₄ emissions, several feeding strategies, including dietary supplementation of CH₄ ­inhibitors, have been investigated (Beauchemin et al., 2020). One such inhibitor is 3-nitrooxypropanol (3-NOP). 3-Nitrooxypropanol is a structural analogue of methyl-coenzyme M, and therefore binds to the active site of methyl-coenzyme M reductase and temporarily oxidizes the nickel ion from oxidation state (+1) to (+2) in the active site, leading to an inhibition of the last steps of methanogenesis (Duin et al., 2016). Furthermore, 3-NOP is degraded rapidly in the rumen yielding 1,3 propanediol and very low concentrations of nitrate and nitrite and hence poses minimal health risk (Duin et al., 2016). Recently, the product received approval for commercial use in dairy cattle in the EU and dairy and beef cattle in Brazil and Chile under the name Bovaer (https://www.dsm.com/anh/news/press-releases.html).

Various studies indicated that supplementation of 3-NOP in the diet of ruminants reduced CH₄ emissions in a dose-dependent manner, with no sign of animal toxicity (Martinez-Fernandez et al., 2014; Lopes et al., 2016; Vyas et al., 2018a, 2018b; McGinn et al., 2019; Alemu et al., 2021a, 2021b). A recent meta-analysis of 14 studies (Kim et al., 2020) revealed a linear decrease in CH₄ yield (g/kg dry matter (DM) intake) with 3-NOP supplementation regardless of animal type (beef, dairy) and length of experimental period without adverse effect on animal performance. Similarly, a meta-analysis of 11 experiments (Dijkstra et al., 2018) indicated that the impact of 3-NOP on enteric CH₄ emissions was positively associated with 3-NOP dose and negatively associated with dietary fiber concentration. It has been reported that supplementation of 3-NOP impacts DM and apparent total-tract nutrient digestibility (Romero-Perez et al., 2014; Haisan et al., 2017; van Gastelen et al., 2020).

To date, studies evaluating 3-NOP have focused on growing and finishing beef cattle (Vyas et al., 2018a, 2018b; Kim et al., 2019; McGinn et al., 2019; Alemu et al., 2021a, 2021b) and dairy cows (Hristov et al., 2015; van Gastelen et al., 2020; Schilde et al., 2021; Pitta et al., 2022) consuming mixed forage and concentrate diets containing a relatively low amount of fiber. Few studies (e.g., Zhang et al., 2021; Gruninger et al., 2022) have investigated the impacts of 3-NOP in animals consuming high-fiber diets, such as those fed to beef cows. While the mechanism underpinning its CH₄ mitigation is known (Duin et al., 2016), other effects of 3-NOP on DM intake, rumen fermentation, microbial communities, and nutrient digestibility are not yet fully understood. Therefore, the objective of this study was to determine the effects of supplementing a high-fiber diet (90% forage DM basis) with 3-NOP (150 mg/kg DM) on DM intake, rumen fermentation, rumen microbial community, enteric gas emissions, salivary secretion, and apparent total-tract nutrient digestibility. We hypothesized that supplementation of a high-fiber diet with 3-NOP would decrease enteric CH₄ emissions without negatively affecting fiber digestibility.

Materials and Methods

Animal use for this experiment was approved by the Institutional Animal Care and Use Committee of the Lethbridge Research and Development Centre (Lethbridge, Alberta, Canada) and was in accordance with the guidelines of the Canadian Council on Animal Care (2009). Animals were housed in individual tie stalls on rubber mats bedded with wood shavings for the duration of the study (except the periods of total fecal and urine collection and enteric CH₄ measurement). Animals exercised daily except when in the chambers for CH₄ measurement or during the total collection period.

Experimental design and dietary treatments

Eight ruminally cannulated beef heifers (Angus cross; mean body weight (BW) ± SD, 556 ± 57.5 kg) were used in a replicated crossover design with two groups (block, four cattle/block), two periods, and two treatments. Body weight was measured (without fasting) on two consecutive days at the beginning and end of each period. The first 16 d of the 49-d periods was used for adaptation and the following 33 d for measurement and sample collection. Animals were blocked by BW and randomly assigned to the two treatments: 1) control (no 3-NOP supplementation) and ii) 3-NOP (control + 150 mg 3-NOP/kg DM). The level of 3-NOP was determined based on CH₄ mitigation effectiveness (previous study in our lab that used a similar dose and reported a 21% and > 30% reduction in CH₄ yield for beef cattle on high-forage and high-grain diets, respectively; Vyas et al., 2018a), and economic application rates in practice (manufacturer’s information). The basal diet was formulated to meet the nutrient requirements of growing beef cattle following the National Academies of Sciences, Engineering, and Medicine (NASEM, 2016), and contained 450 g/kg DM barley silage, 450 g/kg DM chopped grass hay, and 100 g/kg DM concentrate containing protein, minerals, and vitamin supplements (Table 1). Diets were prepared daily as a total mixed ration (TMR) and the required amount of 3-NOP was measured and mixed into the TMR daily to reach the target concentration. Ingredients were mixed daily in a Calan Super Data Ranger (American Calan, Inc., Northwood, NH, USA) equipped with a scale. The treatment supplement contained 11.1% 3-NOP on a carrier of SiO₂ (53.3%) and propylene glycol (35.6%), and the placebo supplement contained the carrier SiO₂ (60%) and propylene glycol (40%) only. Total mixed ration was offered to individual animals ad libitum (5% orts) at 1000 hours (90% of the feed) with a top-up (10%) at 1300 hours due to limited feed trough capacity and they had free access to water.

Table 1.

The ingredients and nutrient composition of the basal diet (mean ± SD)

Item	Control	3-NOP	SEM	P-value6
Ingredient, % DM
Barley silage1	45.0	45.0
Chopped grass hay2	45.0	45.0
Supplement mix3	9.9	9.9
Placebo/3-NOP	0.1	0.1
Nutrient composition
DM, % as fed	56.1 ± 1.94	55.8 ± 2.01
OM, % of DM	93.1 ± 0.19	93.0 ± 0.23
CP, % of DM	9.88 ± 0.41	9.86 ± 0.46
NDF, % of DM	53.0 ± 2.66	52.8 ± 1.94
ADF, % of DM	29.8 ± 1.82	29.7 ± 1.41
Starch, % of DM	12.8 ± 2.20	12.8 ± 2.03
GE, Mcal/kg DM	4.32 ± 0.08	4.32 ± 0.11
NEm, Mcal/kg DM4	1.53	1.52
NEg, Mcal/kg DM4	0.89	0.88
TMR particle size distribution5, % as-is retained on sieves
19 mm	33.3	34.5	3.76	0.75
8 mm	23.5	22.9	1.44	0.67
1.18 mm	29.5	28.5	2.08	0.65
Pan	13.7	14.0	1.98	0.88
pef8.0	0.57	0.57	0.033	0.86
pef1.18	0.86	0.86	0.020	0.88
peNDF8.0, % of DM	30.3	30.1	1.78	0.89
peNDF1.18, % of DM	46.2	45.1	1.32	0.44

¹ Barley silage: DM, 36.3 ± 2.47% as fed; OM, 93.0 ± 0.17 % DM; CP, 9.7 ± 0.26 % DM; NDF, 47.4 ± 1.84 % DM; ADF, 25.3 ± 1.13 % DM; starch, 23.9 ± 1.92 % DM; GE, 4.5 ± 0.07 Mcal/kg DM.

² Chopped grass hay: DM, 91.4 ± 2.33% as fed; OM, 93.8 ± 0.37 % DM; CP, 6.2 ± 0.71 % DM; NDF, 64.9 ± 3.90 % DM; ADF, 38.0 ± 3.04 % DM; starch, 2.7 ± 1.05 % DM; GE, 4.3 ± 0.11 Mcal/kg DM.

³ Supplement included barley grain (dry rolled, 4.90%), canola meal (4.26%), urea (0.08%), salt (0.30%), molasses (0.21%), vitamin and mineral premix (0.04%), and canola oil (0.11%). Supplement nutrient composition: DM, 93.2 ± 0.77% as fed; OM, 91.3 ± 0.48 % DM; CP, 25.6 ± 1.64 % DM; NDF, 27.3 ± 1.89 % DM; ADF, 11.9 ± 0.70 % DM; starch, 27.5 ± 1.95 % DM; GE, 4.2 ± 0.13 Mcal/kg DM. Vitamin and mineral premix in supplement contained (g/kg of DM): 348.3 calcium carbonate (Pestell Minerals and Ingredients, New Hamburg, ON, Canada), 283.7 zinc sulfate monohydrate (Pestell Minerals and Ingredients, New Hamburg, ON, Canada), 103.1 cupric sulfate pentahydrate (Freeport-McMoRan Inc., Phoenix, AZ, USA), 146.1 manganese sulfate monohydrate (Pestell Minerals and Ingredients, New Hamburg, ON, Canada), 1.50 ethylenediamine dihydroiode (EDDI) (800 g EDDI/kg) (Pestell Minerals and Ingredients, New Hamburg, ON, Canada), 50.4 selenium (10 g selenium/kg; Selenium Premix 1%, Pestell Minerals and Ingredients, New Hamburg, ON, Canada), 0.80 cobalt carbonate (Pestell Minerals and Ingredients, New Hamburg, ON, Canada), 17.7 vitamin A (1,000,000 IU/g) (Zhejiang Nvb Company Ltd., Zhejiang, China), 1.70 vitamin D3 (500,000 IU/g) (Zhejiang Nvb Company Ltd., Zhejiang, China), and 47.3 vitamin E (all-rac-alpha tocopherol acetate) (500,000 IU/kg) (Zhejiang Nvb Company Ltd., Zhejiang, China).

⁴ Based on NASEM (2016) using measured values of animal BW, dry matter intake, and diet information.

⁵ Particle size distribution was determined using the Penn State Particle Separator (Kononoff et al., 2003). pef_8.0 and pef_1.18 = physical effectiveness factor determined as the proportion of particles retained on two sieves (19 and 8 mm; Lammers et al., 1996) and three sieves (19, 8, and 1.18 mm; Kononoff et al., 2003), respectively; peNDF_8.0 and peNDF_1.18 = physically effective NDF determined as NDF content of TMR multiplied by pef_8.0 and pef_1.18, respectively.

⁶Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

Data and sample collection

Feed intake and consumption rates

Feed offered and orts were measured and recorded daily to calculate DM intake. Average ad libitum intake was determined using intake data collected from days 17 to 35. Starting two days prior to the total collection of feces (days 38–42) till the end of enteric gas measurement (days 46–49), the amount of feed offered was reduced such that animals were fed 95% of the average ad libitum intake calculated based on individual animal intake collected from days 17 to 35. The approach was chosen to ensure similar feed intake between treatments, thus avoiding confounding effect of DM intake on digestibility and enteric gas production. From previous studies (Romero-Perez et al., 2014; Vyas et al., 2018b), we anticipated a reduction in intake of the animals when confined in metabolic chambers for enteric gas measurement. Samples of forage and concentrate feeds were collected once weekly, and samples of TMR and orts were obtained daily and pooled weekly. Composited samples were subsampled using a riffle splitter, dried at 55 °C for 72 h in a forced-air oven, ground through a 1-mm sieve (standard model 4, Arthur Thomas Co., Philadelphia, PA, USA), and retained for chemical analysis. Furthermore, composite fresh TMR and orts samples were processed for particle size distribution analysis using the Penn State Particle Separator, according to Kononoff et al. (2003). Physical effectiveness (pef) factor was determined as the proportion of particles retained on two sieves (19 and 8 mm) for pef_8.0 (Lammers et al., 1996) and three sieves (19, 8, and 1.18 mm) for pef_1.18 (Kononoff et al., 2003). Neutral detergent fiber (NDF) concentration of the TMR (DM basis) was multiplied by pef_8.0 and pef_1.18 to calculate physically effective NDF (peNDF) content, peNDF_8.0, and peNDF_1.18, respectively. Eating rate was measured on two days (days 24 and 26) by weighing the feed bunk to determine feed disappearance at 3, 6, 9, 12, and 24 h (refusal next day) post-feeding (Kim et al., 2019).

Eating and resting salivation and rumen liquid volume

To estimate salivary secretion during eating, samples of swallowed, and ingested feed were collected from each animal on day 20 as described by Cassida and Stokes (1986) and Maekawa et al. (2002). Briefly, the collections were made at the cardia via the rumen cannula. Approximately 20 min prior to sampling, the cannula plug was removed and digesta was evacuated from the anterior sac of the rumen to expose the cardia. The removed digesta was placed in pre-warmed, insulated, covered plastic tubs to maintain temperature, ensure microbial survival, and minimize temperature shock to the animal when digesta was replaced. Animals were allowed to eat undisturbed for 5–10 min before sample collection started using a collection bag (50 cm long and 18 cm wide sewn to a wire-hoop with a 9-cm circumference; Cassida and Stokes, 1986; Chibisa et al., 2016). All the swallowed feed boluses (ingested feed and saliva) in a 2-min period were collected, and collection was repeated every 5-min for a total of six collections or until the animal stopped eating for more than 20 min. Samples contaminated with rumen contents were discarded. The wet weight of collected masticate samples was recorded and samples were dried in a forced-air oven at 55 °C for 48 h to determine DM content. The amount of saliva added to masticates was calculated as the difference in moisture of the collected masticate and the TMR (as-fed basis) offered. The DM of saliva was assumed to be 1% (Bailey and Balch, 1961); therefore, its contribution to DM analysis was negligible. Salivation rate (mL/min) and ensalivation of the feed were calculated as described by Maekawa et al. (2002). To determine ensalivation rate, saliva production was expressed as a proportion of the quantity of TMR ingested.

Resting saliva production was collected on day 31 using a similar apparatus to that used for collecting masticates. Rumen contents were totally removed through the cannula 2 h before feeding time and kept in pre-warmed, insulated, and covered plastic tubs. To prevent eating and drinking, feed was removed and water flow was stopped, respectively. Swallowed saliva was collected for approximately 5 min for a total of four collections with a 2 min interval between consecutive samplings (Maekawa et al., 2002). Individual saliva sample volume was measured, and resting salivation rate was calculated as the amount produced divided by the duration of the collection period. Evacuated rumen contents were measured and two 1-kg samples were taken after thorough mixing. Collected samples were oven-dried at 55 °C for 72 h to determine ruminal liquid volume.

Apparent total-tract nutrient digestibility

To determine apparent total-tract digestibility, total excretion of feces was collected on days 38–42 into pans placed behind the animals. The daily collected feces samples were weighed, thoroughly mixed, and sampled (5% of total wet weight). Samples were composited by period for individual animals, dried at 55 °C in an oven for 72 h, ground through a 1-mm sieve (standard model 4, Arthur Thomas Co., Philadelphia, PA, USA), and retained for chemical analysis. To avoid contamination of feces, urine was collected separately into a closed collection container (20 L capacity) via tubing attached to an indwelling catheter (26 French, 75-cc balloon; C. R. Bard, Inc., Covington, GA, USA) inserted into the bladder of the animal.

Rumen sampling and rumen pH

Rumen contents were collected on days 17 and 28 of each period at 0 h (prior to feeding) and 3, 6, 9, 12, and 24 h post-feeding. For each animal, about 250 mL rumen content was sampled at each sampling point from four different locations in the rumen and then mixed together by sampling time for each animal to obtain a representative sample for volatile fatty acid (VFA), ammonia-nitrogen (NH₃–N), dissolved hydrogen (H₂), and protozoal and microbiome analysis. The rumen content sample was filtered through two layers of polyester monofilament fabric (355 µm mesh opening) to separate the liquid and solid fractions. A 5-mL sample of rumen fluid was acidified with 1 mL 25% w/v metaphosphoric acid (5:1) and frozen at −20 °C until analysis of VFA and NH₃–N. For protozoa identification and counting, a 5-mL rumen fluid sample was preserved with methylgreen–formalin–saline solution and stored in the dark at room temperature (23 ± 2 °C) until protozoa were identified and counted (Ogimoto and Imai, 1981). For rumen microbial community analysis, 50-mL subsamples of filtered rumen fluid (liquid phase) and samples of rumen digesta (solid phase) were snap frozen in liquid nitrogen and stored at −80 °C until DNA extraction. Dissolved H₂ was measured using a sensor (H₂-500; Unisense, Aarhus, Denmark) attached to a glass flow-through cell (2 mm internal diameter and 6 mm external diameter). The H₂ sensor was connected to a microsensor multimeter (Unisense, Aarhus, Denmark), which was calibrated as described by Guyader et al. (2017).

On day 20, indwelling pH meters (LRCpH Data Loggers, Dascor, Inc., Escondido, CA, USA) were inserted through the rumen fistula of each animal to record ruminal pH at 1-min intervals for a total of 7 d. The system was calibrated (in pH 4 and 7 solutions at 35–40 °C) prior to insertion on the first day and then upon removal on the last day. On day 28, the indwelling pH meters were removed and the data were downloaded and analyzed (model M5-version 755 Dascor Data Logger Software, Dascor Inc., Escondido, CA, USA). Continuous ruminal pH data were summarized for daily mean, minimum, and maximum.

Enteric gas measurements

Following total collection of feces, enteric gas (CH₄, CO₂, and H₂) measurements were conducted using four open-circuit calorimetry chambers (4.4 m wide × 3.7 m deep × 3.9 m tall, Conviron Inc., Winnipeg, MB, Canada) from days 46 to 49 according to the methods described in Beauchemin and McGinn (2006). The chambers were calibrated before and after the commencement of the study by releasing a known amount of CH₄ and CO₂ and calculating the recovered amount. Calibration factors were then used to correct the gas emissions. Variability in slopes across chambers was less than 5% and recovery rates ranged from 98% to 104%. Animals were trained for entry into the chambers prior to starting the trial using a training protocol to minimize stress and enteric gases were measured from animals within each group (four animals) simultaneously using the four chambers (72 h of continuous measurements), with the measurements for the additional square staggered by 1 wk. Gas concentrations were measured using infrared gas analyzers (CH₄; model Ultramat 5E; Siemens Inc., Karlsruhe, Germany; CO₂/H₂O; model LI-7000; LICOR Environmental, Lincoln, NE). Gas concentrations in the inlet and outlet air ducts of each chamber were monitored sequentially for approximately three consecutive min (total 6–7 min/chamber). All chambers were sampled within 27 min, with an additional 3 min used to measure the zero reference gas (pure nitrogen). The gas sampling procedure was repeated every 30 min, for a total of 48 times per chamber per day. Daily CH₄ and CO₂ emissions were calculated following McGinn et al. (2004). Daily CH₄ flux was determined for each animal and expressed as a portion of GE intake on the same day, assuming the energy content of CH₄ is 13.3 Mcal/kg.

The concentration of H₂ was monitored using a H₂ breath tester (BreathTracker Digital Microlyzer, QuinTron Instrument Company, Inc., Milwaukee, WI, USA) from gas samples taken in vacutainers from exhaust air ducts every 3-h post-feeding for 48 h. Gas samples were analyzed using a gas chromatograph equipped with a thermal conductivity detector within approximately 30 min after the sample was taken. Hydrogen flux was calculated by assuming background concentration to be zero. Hydrogen production was subsequently calculated by multiplication of the H₂ concentrations by the measured airflow rate through the chambers.

Laboratory analysis

Analytical DM was determined and used to correct chemical analysis results on a DM basis, by drying ground samples in a forced-oven at 135 °C for 2 h (AOAC, 2016; method 930.15). Organic matter (OM) was calculated as the difference between DM and ash (AOAC, 2016; method 924.05). Neutral detergent fiber (AOAC, 2016; method 2002.04) and acid detergent fiber (ADF; AOAC, 2016; method 973.18) were analyzed sequentially with amylase and sodium sulfite used during NDF determination. Diet gross energy (GE) content was determined using a bomb calorimeter (model E2k, CAL2k, Johannesburg, South Africa). Subsamples were further ground to ~5 μm (MM 400, Retsch Inc., Haan, Germany) and analyzed for crude protein (CP) using flash combustion with gas chromatography and thermal conductivity detection (Carlo Erba Instrumentals, Milan, Italy). Starch analysis was conducted by enzymatic hydrolysis of α-linked glucose polymers according to the methods by Hall (2015).

Concentrations of VFA were measured using an automated gas–liquid chromatography (Model 5890, Hewlett-Packard, Palo Alto, CA, USA) with splitless injection capability, a capillary column (30 m by 0.32 mm by 1 μm; ZB-FFAP, Phenomenex Inc., Torrance, CA, USA), and flame ionization detection (Vyas et al., 2018b). Ruminal NH₃–N concentration was determined by the salicylate–nitroprusside–hypochlorite method using a flow injection analyzer (Sims et al., 1995).

Rumen bacteria and archaeal diversity

Rumen samples collected for microbial analysis (solid, liquid) were freeze-dried (Labconco FreeZone 1, Labconco Corporation, Kansas City) and ground to a fine powder using a coffee grinder. Details on microbial sample analysis are reported by Zhang et al. (2020) and Gruninger et al. (2022). Briefly, DNA was extracted from approximately 0.1 g of the freeze dried, ground material using the Zymobiomics DNA extraction kit (Zymo Research, Irvine, CA, USA). The concentration and purity of the extracted metagenomic DNA were tested using a Nanodrop one (ThermoFischer, Mississauga, ON).

Library generation and sequencing were performed at McGill University and Genome Quebec Innovation Center, Montreal, Canada using the Illumina MiSeq Reagent Kit V2 (500 cycles) following the manufacturer’s guidelines. The V4 region of the 16S rRNA gene was amplified using the 515F (5ʹ-GTGCCAGCMGCCGCGGTAA-3ʹ) and 806R (5ʹ-GGACTACHVGGGTWTCTAAT-3ʹ) primer set, and 2 × 250bp libraries were generated using the Illumina MiSeq Reagent Kit (V2; 500 cycles). A second library was generated for each sample using the 915F (5ʹ-AGGAATTGGCGGGGGAGCAC-3ʹ) and 1386R (5ʹ-GCGGTGTGTGCAAGGAGC-3ʹ) primer set, which targets the V6-V8 region and is optimized for the analysis of rumen methanogens (Kittelmann et al., 2013). For this longer amplicon, 2 × 300bp libraries were generated using the Illumina MiSeq Reagent kit V3 to provide sufficient read overlap. All PCR reactions were performed with 1 μL of gDNA using the following conditions on a Fast Start High Fidelity PCR System (Roche, Montreal, PQ): denaturation at 94 °C for 2 min, followed by 33 cycles of 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 30 s, with a final elongation step at 72 °C for 7 min. Fluidigm Corporation (San Francisco, CA) barcodes were incorporated in a second PCR reaction using the FastStart High Fidelity PCR System under the following conditions: 95 °C for 10 min, followed by 15 cycles of 95 °C for 15 s, 60 °C for 30 s, and 72 °C for 1 min, followed by a final elongation step at 72 °C for 3 min. After amplification, PCR products were assessed in a 2% agarose gel to determine the success of amplification. All samples were quantified using the Quant-iT PicoGreen dsDNA Assay Kit (Life Technologies, Carlsbad, CA, USA) and were pooled in equal proportions. Pooled samples were then purified using calibrated Ampure XP beads (Beckman Coulter, Mississauga, ON). The pooled samples (library) were quantified using the Quant-iT PicoGreen dsDNA Assay Kit (Life Technologies, Carlsbad, CA, USA) and the Kapa Illumina GA with Revised Primers-SYBR Fast Universal kit (Kapa Biosystems, Wilmington, MA, USA). Average fragment size was determined using a LabChip GX (PerkinElmer, Waltham, MA, USA) instrument.

Bioinformatic analysis

Amplicon sequencing data were analyzed using QIIME2 (Bolyen et al., 2019), with V4 and V6-V8 data analyzed separately. FASTQC (Andrews, 2010) was used to inspect the quality of the read pairs. Following the removal of primer and adapter sequences using Bbduk (Bushnell, 2014), sequences were denoised into amplicon sequencing variants (ASVs) using DADA2 (Callahan et al., 2016), with quality control, phiX read filtering, and chimera removal procedures also implemented in this step. MAFFT was used to align sequences against a reference alignment, highly variable regions were then masked, and FastTree was used to generate a phylogenetic tree (Price et al., 2010). A Naïve-Bayes classifier trained to either the V4 or V6-V8 region of the Silva 138.1 reference database was used to assign taxonomy to sequences with the feature-classifier plugin within QIIME2 (Bokulich et al., 2018). Sequences were subsampled (rarefied) to ensure that α- and β-diversity analysis used the same number of sequences per sample. Microbial α-diversity indices (Chao1, Shannon) were calculated in R using the Phyloseq package (McMurdie and Holmes, 2013). Principal coordinate analysis (PCoA) plots were generated in R using the Vegan and Phyloseq packages, based on Bray Curtis and Weighted UniFrac distance matrices (Lozupone et al., 2011). All microbiome figures were generated using ggplot2 (Wickham, 2009). All sequencing reads have been deposited to the Small Reads Archive (NCBI) under accession numbers PRJNA558986 (V4) and PRJNA558986 (V6-V8).

Statistical analysis

Data were analyzed across treatments using the mixed procedure of SAS (SAS Inst., Inc., Cary, NC, USA). For all traits of interest animal was the experimental unit. The statistical model included treatment (control, 3-NOP), sampling time (h, d) and their interaction as fixed effects, and group, period nested within group and animal nested within group as random effects. Sampling time was considered as repeated measure in the model with animal within group as the subject. For particle size distribution, salivation, rumen digesta, and digestibility, day and the repeated measure were removed from the model. Kenward–Roger’s option was used in the model statement to estimate denominator degrees of freedom. Residual plots were used to check the validity of the underlying statistical assumptions of homogeneity of variances and normality. Different time-series covariance structures were evaluated and the best one (autoregressive of order-one) was selected based on the lowest Akaike Bayesian information criteria.

Statistically significant differences in microbial community structure (based on Weighted Unifrac and Bray-Curtis distances) were identified using a PERMANOVA test, with a model that included treatment, period, sampling time (h, d), phase (solid, liquid), and their interactions. Differences in microbial composition across treatments, phases, periods, and sampling time were identified using PROC MIXED in SAS. Prior to analysis, ASV prevalence was calculated, and all ASVs present in less than 10% of the samples were discarded as recommended by Nearing et al. (2022) for robust statistical analysis of microbiomic data. Data were then normalized by converting from raw to proportional abundances, with the proportion of each individual genus/phyla treated as a trait. The statistical model used to detect differences in microbial composition and diversity included the fixed effects of treatment, phase (solid, liquid), sampling time (h), and their interactions. Sampling time was considered as a repeated measure. All P-values from tests of α-diversity and taxa abundance were corrected for a false discovery rate of 0.05 using the Benjamini–Hochberg method (Benjamini and Hochberg, 2000). Data were presented as least squares means ± SEM. Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

Results

Supplementation of 3-NOP had no effect on ad libitum DM, and GE intake (Table 2). Similarly, apparent total-tract digestibility of DM, OM, CP, NDF, ADF, and starch was not affected by 3-NOP supplementation (Table 2). Furthermore, although there was a tendency (P = 0.09) of higher feed consumption rate for the 3-NOP treatment within the first 3-h after feeding (32.3 vs 35.4% of feed), consumption rate for 3 to 6, 6 to 9, 9 to 12, and 12 to 24 h after feeding was not affected by 3-NOP supplementation (P ≥ 0.36, Table 3). Animals consumed 60% and 89% of the offered feed within the first 6 and 12 h, respectively. Particle size distribution for TMR and its associated peNDF were not affected by treatment (Table 1). However, for orts, the proportion of particles retained on the 19 mm sieves was significantly greater (P = 0.02) with a lower quantity of particles retained on the 1.18 mm sieve (P = 0.02) for 3-NOP supplementation as compared to the control treatment (Table 3). Furthermore, the proportion of ort particles retained on the 8 mm sieve and bottom pan tended to be lower for 3-NOP (P ≤ 0.07).

Table 2:

Dry matter (DM) intake and nutrient digestibility of beef heifers (n = 8) fed a high forage diet (45% finely chopped grass hay and 45% barley silage) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (3-NOP, 150 mg/kg DM). Animals were fed at 1000 hours and the small amount left due to the size of the feed bunk was topped-up at1300 hours.

	Control	3-NOP	SEM	P-value3
Ad lib DM intake, kg/d1	10.2	10.1	0.12	0.48
GE intake, Mcal/d	43.8	43.4	0.54	0.43
Nutrient digestibility, %2
DM	66.4	67.2	1.31	0.58
OM	67.8	68.6	1.26	0.50
CP	65.1	65.5	1.66	0.81
NDF	57.9	59.7	1.67	0.32
ADF	43.5	43.4	2.45	0.96
Starch	97.6	97.3	0.26	0.15

¹Ad libitum intake from days 17 to 35.

²To calculate the apparent total-tract digestibility (percent of nutrient digested), the following equation was used: % ND = (NC − NF)/NC × 100, where ND is nutrient digestion (%), NC is nutrient consumed (kg), and NF is nutrient in feces (kg) (Schneider and Flatt, 1975; Cochran and Galyean, 1994).

³Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

Table 3:

Feeding frequency (n = 4) and particle size distribution of orts for beef heifers (n = 8) fed a high forage diet (45% finely chopped grass hay and 45% barley silage) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (3-NOP, 150 mg/kg DM).

Items	Control	3-NOP	SEM	P-value3
Feed intake, kg/d as-is1	18.3	18.3	1.25	0.99
% of feed consumed
0–3 h	32.3	35.4	2.53	0.09
3–6 h	27.2	25.9	1.57	0.38
6–9 h	18.3	17.8	1.85	0.76
9–12 h	10.9	8.8	2.57	0.36
12–24 h	11.2	11.5	3.03	0.92
Orts particle size distribution, % as-is retained on sieves2
19 mm	60.3	70.5	3.31	0.02
8 mm	23.0	18.7	1.66	0.07
1.18 mm	13.2	8.3	1.83	0.02
Pan	4.52	2.43	1.021	0.06

¹Intake values were for the days that feeding rate measurement was taken (days 24 and 26). Animals were provided with 90% of the TMR at 1000 hours and feed bunks were topped-up with the remaining feed at 1300 hours.

²Particle size distribution was determined using the Penn State Particle Separator (Kononoff et al., 2003).

³Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

No effect was observed on the rate (mL/min) of eating and resting saliva production as well as the weight of fresh ruminal contents due to supplementation of 3-NOP (Table 4). Dry matter content of masticates was less (P < 0.01) for the control treatment, which may relate to the greater ensalivation of feed (amount of saliva added per unit of feed ingested) for the control treatment (7.04 mL/g DM) relative to 3-NOP (5.49 mL/g DM). Furthermore, ruminal liquid fraction tended to be higher (P = 0.07) for the control treatment (45.8 kg) as compared to 3-NOP (43.0 kg).

Table 4:

Eating and resting saliva production from beef heifers (n = 8) fed a high forage diet containing finely chopped grass hay (45%) and barley silage (45%) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP)

Items	Control	3-NOP	SEM	P-value4
Eating salivation
DM content of masticates, %	11.2	13.7	0.45	<0.01
Eating rate
g masticate/min	192	214	32.7	0.51
g DM/min	22.1	30.3	5.15	0.15
Salivation rate, mL/min1	148	157	22.9	0.72
Ensalivation of feed, mL/g DM2	7.04	5.49	0.396	0.01
Resting salivation
Salivation rate, mL/min	477	499	86.4	0.81
Ruminal content3
Digesta weight, kg	51.5	48.7	1.44	0.10
Digesta DM, %	11.1	11.5	0.39	0.31
Digesta DM, kg	5.68	5.63	0.332	0.87
Liquid fraction, kg	45.8	43.0	1.21	0.07

Saliva (mL) = weight of masticate (g)—weight (g) of feed (as-fed) consumed. Saliva DM is considered to be 1% and assumed to contribute no weight to the DM analysis (Bailey and Balch, 1961).

¹Salivation rate (mL/min) = total amount of saliva obtained (mL) divided by duration of collection (min).

²Ensalivation of feed = amount of saliva added (mL) per g of feed ingested.

³Determined following total evacuation of the ruminal contents.

⁴Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

Feeding 3-NOP increased minimum and maximum pH (P ≤ 0.01), which coincides with higher mean rumen pH (+0.21 units, P < 0.001) relative to control treatment (6.43, Table 5). No interaction effect between treatment and sampling time was observed for rumen pH. Diurnal pattern of mean rumen pH (Figure 1) indicated a consistent drop in rumen pH post-feeding for both treatments with mean pH reaching the lowest point 5–6 h after the morning feeding (6.41) for 3-NOP and 11–12 h (6.20) for control. Subsequently, pH increased but rumen pH was consistently greater for 3-NOP throughout the day. Relative to the control treatment, total VFA concentration was significantly lower (P < 0.01) for 3-NOP treatment. The molar percentages of propionate and isobutyrate (P ≤ 0.01) were increased while the molar percentage of acetate and the acetate to propionate ratio were decreased following 3-NOP supplementation (P ≤ 0.001, Table 5). Compared to control, supplementation of 3-NOP increased propionate percentage by 11.7% while the percentage of acetate was reduced by 7.9% which caused an 18.9% reduction in acetate to propionate ratio. Furthermore, interaction effect between treatment and sampling time (P ≤ 0.001) was observed for the molar percentages of butyrate, valerate, isovalerate, and caproate where the lowest percentages were observed at 0 and 24 h sampling while the largest percentages were observed at 6 and 9 h after feeding. No effect was observed on rumen NH₃–N concentration with supplementation of 3-NOP.

Table 5.

Rumen pH, volatile fatty acids (VFA), and ammonia nitrogen (NH₃–N) concentration from beef heifers (n = 8) fed a high forage diets containing finely chopped grass hay (45%) and barley silage (45%) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP). Animals were fed at 1000 hours with small top-up at 1300 hours

Item	Control	3-NOP	SEM	P-value2
				Treatment (treat)	Sampling time (St)	Treat × St
Rumen pH1
Minimum	5.84	5.97	0.040	<0.01	0.01	0.65
Mean	6.43	6.64	0.037	<0.001	<0.001	0.82
Maximum	6.86	7.01	0.030	<0.001	<0.01	0.92
Range	1.05	1.00	0.036	0.16	0.10	0.63
Total VFA, mM	112	104	2.5	<0.01	<0.001	0.80
VFA molar percentage, mol/100 mol
Acetate	68.6	63.2	0.32	<0.001	<0.001	0.21
Propionate	17.0	19.3	0.21	<0.001	<0.001	0.25
Butyrate	10.7	12.5	0.27	<0.001	<0.001	<0.001
Valerate	1.10	1.48	0.020	<0.001	<0.001	<0.001
Isobutyrate	0.81	0.88	0.019	<0.01	<0.001	0.16
Isovalerate	1.20	1.77	0.066	<0.001	<0.001	<0.001
Caproate	0.55	0.96	0.044	<0.001	<0.001	<0.001
Ac:Pr ratio	4.07	3.30	0.055	<0.001	<0.001	0.10
NH3–N, mM	3.15	2.95	0.190	0.31	<0.001	0.18

¹Rumen pH was measured at 1 min intervals over 7 d using continuous pH loggers. Range = maximum rumen pH—minimum rumen pH.

²Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

Figure 1.

Diurnal pattern of mean rumen pH from beef heifers (n = 8) fed a high forage diet containing finely chopped grass hay (45%) and barley silage (45%) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP). Error bars indicate SD.

While animals were in chambers for enteric gas measurements, no difference in DM and GE intake was observed between treatments (Table 6). Relative to control, feeding 3-NOP decreased (P < 0.001) enteric CH₄ emissions when expressed as g/d (−22.4%), g/kg DM intake (−22.0%), and % GE intake (−21.1%, Table 6). However, supplementation of 3-NOP increased H₂ gas yield (g/kg DM intake) by about 333-fold, and concentration of dissolved H₂ (µmol/L) in rumen fluid by about 10-fold (P < 0.001). Furthermore, CO₂ production and yield were not affected by treatment, and no interaction effect between treatment and sampling time was observed for feed intake and enteric gas production. The diurnal pattern of enteric CH₄ emissions (Figure 2) indicated that for the control treatment, enteric CH₄ production increased after feeding and reached its maximum of 9–10 h after feeding. However, immediately after feeding 3-NOP, there was a short 1-h negative effect on CH₄ production, with CH₄ production increasing thereafter but to a lesser extent than for control. At 12–13 h after feeding, CH₄ production for control and 3-NOP followed a similar pattern.

Table 6.

Methane (CH₄), hydrogen (H₂), and carbon dioxide (CO₂) emission (g/d) and yield (g/kg dry matter (DM) intake) from beef heifers (n = 8) fed a high forage diet containing finely chopped grass hay (45%) and barley silage (45%) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP)

Item	Control	3-NOP	SEM	P-value2
Body weight, kg	563	563	1.8	0.95
Chamber DM intake, kg/d1	8.37	8.53	0.187	0.42
Chamber GE intake, Mcal/d	35.6	36.5	1.08	0.40
CH4 gas emissions
g/d	209	162	4.5	<0.001
g/kg DM intake	25.3	19.7	0.63	<0.001
% GE intake1	7.83	6.18	0.149	< 0.001
g/kg BW/d	0.37	0.29	0.008	<0.001
H2 gas emissions
g/d	0.00	1.68	0.199	<0.001
g/kg DM intake, ×10−2	0.06	19.21	1.573	<0.001
Dissolved H2, µmol/L	33.4	372.3	29.00	<0.001
CO2 gas production
g/d	7082	7191	132.6	0.41
g/kg DM intake	872	876	23.0	0.85
g/kg BW/d	12.6	12.8	0.23	0.41

¹Dry matter intake and GE intake values were for the 3 d (days 46–49) that animals were maintained in the chamber system for methane measurement. Energy content of CH₄ was assumed as 13.3 Mcal/kg.

²Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

Figure 2.

Diurnal methane emissions pattern from beef heifers (n = 8) fed a high forage diet containing finely chopped grass hay (45%) and barley silage (45%) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP). Error bars indicate SD.

Rumen microbial sequencing

16S rRNA amplicon sequencing of V4 and V6-V8 hypervariable regions yielded a total of 47,516,642 reads across all samples, which were denoised into 26,563 unique ASVs using DADA2. Goods coverage was ≥0.99 in all of the samples indicating that sequencing depth was adequate to capture the diversity present in each sample. PCoA and accompanying PERMANOVA tests showed significant effects of treatment, rumen digesta phase (solid, liquid), and rumen sampling hour, though this varied with the choice of dissimilarity matrix (Figures 3 and 4). There was no ­interaction effect among treatment, digesta phase, and sampling hour (P ≥ 0.15). Moreover, the R² values accompanying the statistically significant results were generally low, with the exception of the rumen digesta phase comparison using the Weighted Unifrac matrix, which returned R² values of 0.35 and 0.38 for bacterial (V4) and archaeal (V6-V8) community, respectively (P = 0.001, Figures 3 and 4). This means that 35% and 38% of the variation in microbial composition was attributable to rumen digesta phase. This observation was reflected in the PCoA plots, which showed only rumen content clearly delineated the V4 dataset (Figure 3). Similarly, for the V6-V8 data, rumen content was the major discriminator between the samples, though a clear influence of treatment was also evident (Figure 4).

Figure 3.

Comparison of rumen microbiome based on sequencing V4 region of the 16s rRNA gene associated with rumen content (solid (SOL) and liquid (LIQ) phases) collected at different sampling times (0, 6, and 12 h) from beef heifers fed a high forage diet (45% finely chopped grass hay and 45% barley silage) supplemented without 3-nitrooxypropanol (3-NOP, control (CON)) and with 3-NOP (150 mg/kg DM, 3-NOP). PcoA analysis of rumen samples was calculated using weighted UniFrac (A, B, and C) and Bray Curtis (D, E, and F) distances.

Figure 4.

Comparison of rumen methanogen populations based on sequencing V6–V8 region of the 16s rRNA gene associated with rumen content (solid (SOL) and liquid (LIQ) phases) collected at different sampling times (0, 6, and 12 h) from beef heifers fed a high forage diet (45% finely chopped grass hay and 45% barley silage) supplemented without 3-nitrooxypropanol (3-NOP, control (CON)) and with 3-NOP (150 mg/kg DM, 3-NOP). PcoA analysis of rumen samples was calculated using weighted UniFrac (A, B, and C) and Bray Curtis (D, E, and F) distances.

The mean proportion of rumen bacterial and methanogens according to treatment group, rumen digesta phase and sampling time is presented in Tables 7 and 8, respectively. Overall, no interaction effect was observed among treatment, rumen digesta phase, and sampling time on the relative abundance of rumen bacterial and methanogen community (P > 0.05). Ampelicon sequence variants richness and evenness, measured by the Chao1 and Shannon indices, respectively, were higher for the control treatment relative to 3-NOP for both rumen bacterial as well as archaeal community (P ≤ 0.04). For rumen bacterial community, Chao1 and Shannon indices were higher in the solid phase compared to liquid samples (P < 0.01), but the archaeome was both richer and more even in the liquid phase (P < 0.01). Furthermore, Chao1 value for bacterial community was affected by sampling time (P = 0.02) where significant reduction was observed after feeding.

Table 7.

Relative abundance of bacterial taxa (phyla and genera level, sequenced targeting the V4 region of the 16s rRNA gene) associated rumen content (solid and liquid phase) collected at different sampling times (0, 6, and 12 h) from beef heifers fed a high forage diet (45% finely chopped grass hay and 45% barley silage) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP)

Item	Treatment (treat)			Sampling time (hour, H)4				Rumen content (RC)			P-value3
	Control	3-NOP	SEM	0	6	12	SEM	Liquid	Solid	SEM	Treat	H	RC
Observed amplicon sequence variants (ASV)	695	611	17.4	675a	654ab	630b	15.8	633	673	12.9	<0.01	0.02	<0.01
Chao1 value	696	612	17.5	676a	655ab	630b	15.9	634	674	13.0	<0.01	0.02	<0.01
Shannon index	5.56	5.32	0.089	5.47	5.44	5.41	0.029	5.36	5.52	0.024	0.04	0.09	<0.001
Relative abundances1, %
P Bacteroidates	50.8	50.9	3.31	49.2	50.4	53.0	1.65	58.4	43.3	1.51	0.98	0.07	<0.001
g Prevotella	39.0	35.3	3.20	34.5a	37.3ab	39.5b	1.45	43.7	30.5	1.18	0.32	<0.01	<0.001
g Rikenellaceae RC9 gut group	3.83	5.05	0.290	5.35a	4.32ab	3.65b	0.390	4.24	4.64	0.318	0.01	<0.001	0.20
g Prevotellaceae UCG-001	3.12	3.25	1.001	3.50	2.61	3.46	0.482	3.35	3.03	0.394	0.90	0.12	0.42
g Prevotellaceae UCG-003	1.08	1.68	0.355	1.11	1.52	1.50	0.360	1.89	0.87	0.296	0.12	0.43	<0.01
g Prevotellaceae NK3B31 group	0.78	0.96	0.346	1.18	0.70	0.73	0.285	0.79	0.95	0.227	0.62	0.16	0.48
g Prevotellaceae UCG-004	1.14	2.13	0.425	1.14	1.64	2.12	0.345	1.62	1.65	0.260	0.06	0.02	0.92
g Bacteroidales BS11 gut group	0.33	0.37	0.074	0.41	0.34	0.31	0.097	0.33	0.38	0.074	0.57	0.47	0.92
g Bacteroidates RF16 group	1.27	1.47	0.180	0.97b	1.45a	1.68a	0.176	1.58	1.16	0.141	0.32	<0.01	<0.01
g Muribaculaceae	0.53	0.63	0.102	0.76a	0.27b	0.70a	0.141	0.62	0.54	0.093	0.40	<0.01	0.40
g Paraprevotella	1.04	1.85	1.424	1.09	1.33	1.91	0.863	2.28	0.60	0.681	0.58	0.64	0.02
g Uncultured	0.73	0.86	0.070	0.86	0.76	0.77	0.111	0.86	0.74	0.089	0.11	0.62	0.17
P Firmicutes	32.6	31.1	1.49	33.1a	32.9a	29.6b	1.20	27.6	36.1	0.98	0.36	0.01	<0.001
g Succiniclasticum	15.6	18.0	1.34	17.9	16.6	15.8	1.13	16.0	17.6	0.92	0.10	0.18	0.10
g Christensenellaceae R-7 group	2.98	2.67	0.303	3.11	2.81	2.56	0.282	2.12	3.53	0.228	0.34	0.16	<0.001
g Ruminococcaceae	0.61	0.55	0.159	0.62	0.60	0.53	0.137	0.40	0.76	0.105	0.69	0.77	<0.01
g Lachnospiraceae NK3A20 group	1.60	1.80	0.284	1.60	1.94	1.52	0.307	1.14	2.26	0.250	0.50	0.38	<0.001
g Saccharofermentans	1.34	0.79	0.240	1.44a	1.09a	0.67b	0.146	0.75	1.38	0.116	0.04	<0.001	<0.001
g uncultured f Ruminococcaceae	2.14	0.26	0.636	0.88	1.44	1.27	0.685	1.20	1.19	0.484	0.02	0.60	0.98
g Lachnospiraceae XPB1014 group	1.29	0.67	0.246	0.90	1.19	0.84	0.201	0.65	1.30	0.160	0.06	0.16	<0.001
g Lachnospiraceae AC2044 group	0.81	0.52	0.115	0.77	0.70	0.51	0.132	0.41	0.91	0.108	0.04	0.14	<0.001
g NK4A214 group f Oscillospiraceae	1.50	1.04	0.163	1.35	1.39	1.07	0.194	1.11	1.43	0.158	0.02	0.21	0.05
g Pseudobutyrivibrio	0.94	0.78	0.135	1.02	0.79	0.77	0.126	0.47	1.25	0.098	0.29	0.09	<0.001
g Veillonellaceae UCG-001	0.79	0.71	0.093	0.86a	0.80a	0.60b	0.093	0.71	0.80	0.075	0.46	0.02	0.26
g Mogibacterium	1.08	0.58	0.066	0.70	0.95	0.84	0.098	0.58	1.08	0.082	0.01	0.05	<0.001
P Proteobacteria	2.42	2.83	0.963	0.32c	2.42b	5.13a	0.671	3.67	1.57	0.533	0.68	<0.001	<0.001
P Fibrobacteres	7.55	8.73	1.852	10.54a	7.97b	5.91b	1.260	4.81	11.5	1.02	0.54	<0.01	<0.001
P Spirochaetae	4.82	4.70	1.667	3.68b	4.66b	5.95a	0.517	3.81	5.72	0.422	0.95	<0.001	<0.001
g Treponema	4.59	4.60	1.593	3.45b	4.50ab	5.82a	0.524	3.65	5.54	0.427	0.99	<0.001	<0.001
P Verrucomicrobia	0.95	0.68	0.170	1.14a	0.71b	0.59b	0.165	1.04	0.59	0.133	0.15	<0.01	0.001
g WCHB1 41	0.90	0.53	0.087	1.03a	0.60b	0.52b	0.140	0.90	0.54	0.110	0.01	0.001	0.002
P Planctomycetota	0.43	0.49	0.111	0.65a	0.41b	0.38b	0.118	0.42	0.54	0.107	0.97	0.04	0.26
P Patescibacteria	1.67	1.39	0.692	2.08a	1.26b	1.25b	0.297	1.74	1.32	0.232	0.69	<0.01	0.08
Protozoa, 105 cell/mL2	4.71	5.73	0.612	6.52a	4.65b	4.53b	0.314	na	na	na	0.14	<0.001	na

¹ p = phylum, g = genus, f = family.

²Protozoa were identified and counted under microscope following Ogimoto and Imai (1981). Identified ruminal protozoa genera included: Isotricha spp., Dasytricha spp., Entodinium spp., Diplodinium spp., Ostracodinium spp., Metadinium spp., Polyplastron spp., and Osphyrscolex spp.

³Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

⁴Mean values with a different superscripted letter differ significantly among sampling times.

Table 8:

Relative abundance of rumen methanogens (phyla and genera level, sequenced targeting the V6-V8 region of the 16s rRNA gene) associated with rumen content (solid and liquid phases) collected at different sampling times (0, 6, and 12 h) from beef heifers fed high forage diet (45% finely chopped grass hay and 45% barley silage) supplemented without 3-nitrooxypropanol (3-NOP, control) and with 3-NOP (150 mg/kg DM, 3-NOP)

Item1	Treatment (Treat)			Sampling time (hour, H)3				Rumen content (RC)			P-value2
	Control	3-NOP	SEM	0	6	12	SEM	Liquid	Solid	SEM	Treat	H	RC
Observed amplicon sequence variants (ASV)	24.0	18.3	0.63	21.5	20.8	21.1	0.56	21.8	20.5	0.45	<0.01	0.45	0.01
Chao1 value	23.7	18.1	0.62	21.1	20.7	20.9	0.53	21.5	20.3	0.44	<0.01	0.71	0.01
Shannon index	1.99	1.63	0.053	1.79	1.83	1.80	0.027	1.90	1.71	0.022	<0.01	0.43	<0.001
Relative abundances1, %
P Euryarchaeota	68.1	69.6	1.31	73.4a	68.2ab	64.9b	1.82	60.9	76.8	1.49	0.31	<0.01	<0.01
g Methanobrevibacter	63.8	64.6	1.37	68.8a	63.7ab	60.2b	1.67	57.5	71.0	1.37	0.59	<0.01	<0.01
g Methanosphaera	4.26	4.98	0.270	4.61	4.55	4.69	0.295	3.40	5.84	0.241	0.05	0.89	<0.001
P Thermoplasmatota	30.2	30.4	1.70	25.6b	31.0ab	34.4a	1.75	37.8	22.8	1.43	0.93	<0.01	<0.01
g Candidatus Methanomethylophilus	16.8	25.4	1.79	16.8b	21.2ab	25.5a	1.39	25.7	16.6	1.13	0.01	<0.01	<0.01
g Uncultured f Methanomethylophilaceae	13.4	4.9	0.81	8.8	9.8	8.9	0.80	12.2	6.2	0.66	<0.01	0.40	<0.01
P Halobacterota	1.78	0.06	1.238	1.18	0.85	0.73	0.514	1.43	0.41	0.425	0.19	0.66	0.02
g Methanomicrobium	1.79	0.02	1.30	1.10	0.88	0.73	0.637	1.41	0.39	0.511	0.22	0.82	0.05
g Methanimicrococcus	0.15	0.00	0.001	0.20	0.00	0.10	0.088	0.19	0.00	0.001	0.37	0.12	0.04

¹ p = phylum, g = genus, f = family.

²Significance was declared at P ≤ 0.05 and a tendency to significance was declared at 0.05 < P ≤ 0.10.

³Mean values with a different superscripted letter differ significantly among sampling times.

The most abundant phyla in all samples were Bacteroidetes and Firmicutes, accounting for 82–83% of the sequences identified, and neither was affected by treatment (P ≥ 0.36). However, these predominant phyla varied according to rumen digesta phase, with higher proportions of Bacteroidetes observed in the liquid phase (58.4% vs. 43.3%, P < 0.001) and higher proportion of Firmicutes in the solid phase (36.1% vs. 27.6%, P < 0.001). Furthermore, the relative abundance of Firmicutes decreased with advancing sampling hour (P = 0.01), reaching its lowest abundance 12 h after feeding. The higher abundance of Bacteroidetes and Firmicutes was driven by the high relative abundance of Prevotella and Succiniclasticum generas, respectively, across all samples. Fibrobacter was the third most abundant phyla contributing 8–9% to the rumen microbial community. Mirroring Bacteroidetes and Firmicutes, Fibrobacter was not affected by treatment, and its abundance also decreased in the later sampling time points, reaching its lowest abundance 12 h after feeding (5.91%, P < 0.01). It was also more abundant in the solid phase samples compared to the liquid samples (P < 0.001). Furthermore, a targeted effect of 3-NOP supplementation on rumen microbial community was observed that resulted in increasing the relative abundance of Rikenellaceae RC9 gut group (P = 0.01) under Bacteroidetes phyla and decreasing the abundance of Saccharofermentans, Lachnospiraceae AC2044 group, Mogibacterium, NK4A214 group f Oscillospiraceae, and uncultured f Ruminococcaceae under Firmicutes phyla (P ≤ 0.04).

The predominant archaeal phyla were Euryarchaeota and Thermoplasmatota, neither of which were affected by treatment (Table 8). Methanobrevibacter (Euryarchaeota) and Candidatus Methanomethylophilus (Thermoplasmatota) were the major archaeal genera observed, contributing 64–65% and 17–25% of the total number of archaeal reads, respectively. Rumen sampling hour and digesta phase had a significant effect (P < 0.01) on the major archaeal taxa. Euryarchaeota was dominant in the solid phase of the rumen digesta and its abundance decreased with increasing rumen sampling time, reaching its lowest 12 h after feeding (P < 0.01). Moreover, supplementation of 3-NOP increased the relative abundance of Candidatus Methanomethylophilus and Methanosphaera (P ≤ 0.05) while decreasing the abundance of Uncultured f Methanomethylophilaceae (P < 0.01). Similar to the dominant rumen bacterial and methanogen community, feeding 3-NOP did not alter the total number of protozoa but sampling time had a significant effect where the lowest number was observed after feeding (P < 0.01, Table 7).

Discussion

Supplementation of 3-NOP has been reported as an effective feed additive for reducing enteric CH₄ emissions across ruminant species and diets (Dijkstra et al., 2018; Jayanegara et al., 2018; Kim et al., 2020; Yu et al., 2021). However, while the mechanism underpinning its CH₄ mitigation is known (Duin et al., 2016), other effects of 3-NOP on changes in DM intake, rumen fermentation (pH, VFA composition), microbial communities, and nutrient digestibility are not yet fully understood.

Feed intake and nutrient digestibility

Our study found that supplementing a high-fiber beef cattle diet with 3-NOP (150 mg/kg DM) had no impact on DM and GE intake. The effect of 3-NOP on DM intake differs across studies and may depend on dose, animal type, diet, and the duration of feeding (Kim et al., 2020; Yu et al., 2021). A meta-analysis of 14 studies (Kim et al., 2020) indicated that DM intake tended to decrease (slope = −0.0016, P = 0.06, and R² = 0.17) as the dose of 3-NOP supplemented increased in beef cattle. Conversely, using a dairy cattle database, the authors reported that 3-NOP supplementation had no significant linear relationship with DM intake (Kim et al., 2020). The observed difference in the response of DM intake to 3-NOP supplementation between beef and dairy could be related to the dose of 3-NOP used, with a lower dose in dairy studies (26.6–135.1 mg/kg DM intake) vs. a higher dose in beef studies (47.4–337.8 mg/kg DM intake, Kim et al., 2020). Furthermore, discrepancies in DM intake response to 3-NOP supplementation have been reported among short- and long-term studies that used different levels of supplementation. A long-term study with beef cattle using forage-based backgrounding diets (≥60% DM; Vyas et al., 2018a) reported no effect while Vyas et al. (2016b, 2018b) and Alemu et al. (2021a) reported a significant reduction in DM intake following 3-NOP supplementation at different doses. Similarly, for short-term studies i.e., Latin Square Design, Romero-Perez et al. (2014) and Zhang et al. (2021) reported a significant reduction in DM intake following supplementation with 3-NOP while Romero-Perez et al. (2015), Martinez-Fernandez et al. (2018), and Kim et al. (2019) observed no difference. The response of DM intake to 3-NOP supplementation can also be influenced by animal and feeding management (group size, bunk management and feed allocation frequency). For example, under commercial conditions, Alemu et al. (2021a) noticed a difference in DM intake and animal performance between groups of beef cattle receiving backgrounding diets (70% barley or corn silage) supplemented with 200 mg 3-NOP/kg DM, with different feeding management, larger groups in bigger pens (n = 253) vs smaller groups in smaller pens (n = 25). Using a high-fiber beef cattle diet (90% barley silage DM) supplemented with 3-NOP (without oil addition), Zhang et al. (2021) reported a 1.9% reduction in DM intake. However, in that study, 3-NOP was supplemented at 200 mg/kg DM and feed allocation was restricted to supply maintenance energy requirements, unlike in our study where 3-NOP was supplemented at 150 mg/kg DM and cattle were fed ad libitum.

The lack of impact of 3-NOP supplementation on DM intake in the current study suggests that supplementation of 3-NOP at 150 mg/kg DM had minimal or no effect on the organoleptic property of the diet. This is supported by the lack of differences in feed consumption rate between treatments, except for the increased tendency observed for the first 3 h after feeding (P = 0.09). A change in organoleptic properties of the diet usually affects feed consumption rate and causes sorting of feed by animals (Lee et al., 2015a, 2015b). Although feed consumption rate was not affected by treatment in this study, animals supplemented with 3-NOP preferentially sorted for small particles as indicated by the significantly higher proportion of ort particles retained on the 19 mm sieve (P = 0.02) and a lower proportion of particles retained on the 8 mm, 1.18 mm and bottom pan as compared to the control group. Preferential sorting for small particles of cattle fed 3-NOP is difficult to explain, as any negative effects of 3-NOP on organoleptic properties of the diet would be expected to cause preferential sorting against small particles, the fraction containing the 3-NOP supplement. Using a backgrounding diet with 65% corn silage, Kim et al. (2019) reported that supplementation of 3-NOP at 100 mg/kg DM had no effect on DM intake, feed consumption rate, or sorting of feed for animals receiving 3-NOP. Furthermore, no effect of 3-NOP on feeding behaviour was reported by Vyas et al. (2018b) for beef cattle fed backgrounding diets (65% DM barley silage) supplemented with 200 mg 3-NOP/kg DM.

The lack of negative impact on apparent total-tract nutrient digestibility due to 3-NOP supplementation in the current study is in line with previous studies that reported no or minimal negative effects of 3-NOP on apparent total-tract nutrient digestibility using beef cattle (Romero-Perez et al., 2014; Zhang et al., 2021). In a dose–response study using a backgrounding diet (60% DM barley silage), Romero-Perez et al. (2014) reported no effect on nutrient digestibility following 3-NOP supplementation, except a tendency for a quadratic response in DM digestibility. In that study, the amount of fed was restricted to 90% of ad lib intake as compared to the 95% in our study. Restriction of feed can influence total-tract nutrient digestibility (Klinger et al., 2007). Similarly, with a high-forage diet (90% DM) where animals were managed on a restricted level of feed to supply maintenance energy requirements, Zhang et al. (2021) reported an increase in CP and starch digestibility following 3-NOP supplementation (200 mg/kg DM). Using the in situ ruminal incubation technique, Zhang et al. (2020) reported that NDF degradability (288 h of ruminal incubation) of barley silage and grass hay was not affected by feeding 3-NOP (150 mg/kg DM). Barley silage and chopped grass hay were the main forage sources in our study. Thus, our study confirms the general lack of negative effects of 3-NOP on diet digestibility when supplemented to beef cattle.

Rumen fermentation

Reduction in total VFA concentration and changes in rumen fermentation profiles following 3-NOP supplementation have been reported previously. Multiple studies using beef cattle (Romero-Perez et al., 2015; Vyas et al., 2018b, Alemu et al., 2021a, 2021b; Zhang et al., 2021), sheep (Martínez-Fernández et al., 2014;) and dairy cows (Haisan et al., 2014, 2017; Melgar et al., 2020a, 2020b) reported that feeding 3-NOP shifted rumen fermentation pathways toward the production of propionate and butyrate rather than acetate. The observed decrease in the acetate to propionate ratio and increase in molar proportion of butyrate, valerate, and caproate in the present study following 3-NOP supplementation are consistent with the literature. McAllister and Newbold (2008) reported that the enhancement of propionate synthesis can be caused by the dissolved ruminal H₂ resulting from suppression of methanogenesis. As dissolved H₂ increases, as was observed for the 3-NOP treatment, propionate production becomes relatively more favorable because of the increasingly unfavorable thermodynamics of H₂ formation from electrons derived from fermentation (Wang et al., 2016).

The shift in rumen fermentation products following 3-NOP supplementation also affected rumen pH. The observed increase in mean, minimum, and maximum rumen pH is consistent with previous studies using beef cattle diets supplemented which 3-NOP (Romero-Perez et al., 2014, 2015; Kim et al., 2019). Increased rumen pH associated with reduced total VFA concentration has been reported with the use of 3-NOP (Melgar et al., 2020b; Zhang et al., 2021). Furthermore, the observed increase in rumen pH for the 3-NOP treatment could also be due to the increased (+17%) butyrate molar percentage. Penner et al. (2009) reported that increased rumen pH could be related to increased butyrate molar percentage and uptake from the rumen. Rumen pH can also be affected by saliva production, with significant buffering capacity (McDougall, 1948; Turner and Hodgetts, 1955). This is the first study that measured eating and resting salivation rates related to 3-NOP supplementation and the observed numerically higher eating and resting salivation rates for 3-NOP may also have contributed to the higher minimum, mean and maximum pH.

The diurnal rumen pH pattern observed in the present study is similar to that reported by Romero-Perez et al. (2014, 2015), whereby a consistent reduction of mean pH occurred after feeding, reaching a nadir 7–9 h later with the lowest mean pH value ranging between 6.20 and 6.37 for the control and 3-NOP treatments, respectively. The longer period (11–12 h) for the mean pH to reach a nadir for the control treatment in our study could be related to the difference in diet composition: 65% forage (barley silage) in the previous study as compared to 90% forage (45:45 grass hay and barley silage) in the current study.

Microbial community response to 3-NOP supplementation

The unique ability of ruminant animals to digest fibrous feedstuffs is due to a diverse and sophisticated microbial consortium that functions together in a synergetic manner (Wang and McAllister, 2002). Our study revealed that supplementation of 3-NOP induced changes in ruminal VFA concentration without affecting feed digestibility or changes to the predominant rumen bacterial (Bacteroidates, Firmicutes) and archaeal taxa (Methanobrevibacter, Thermoplasmatota, Methanomicrobium). However, a targeted treatment effect was observed for some specific bacterial (e.g., Rikenellaceae RC9 gut group, Prevotellaceae UCG-004, Lachnospiraceae AC2044 group Saccharofermentans) and archaeal taxa (e.g., Candidatus Methanomethylophilus, Methanosphaera). Furthermore, a PCoA using UniFrac distances did not show a clear separation of control and 3-NOP treatment samples for the rumen microbiome while minor separation was observed for the rumen methanogens populations. The accompanying PERMANOVA tests also indicated a clear influence of treatment on the rumen microbial community. There was a significant effect of rumen digesta phase and sampling time on rumen bacteria and methanogens. Rumen microbial composition changes over time due to changes in substrate composition as feed degradation proceeds in the rumen (Cheng et al., 2017; Zhang et al., 2020). Ruminal Bacteroidetes are net H₂ utilizers (Ungerfeld, 2020) and the observed increase in the relative abundance over sampling hours could be associated with the increase in dissolved H₂ concentrations following feed degradation.

The effects of 3-NOP on the microbial community reported in the literature are inconsistent. Martínez-Fernández et al. (2014), Romero-Perez et al. (2014), and Gruninger et al. (2022) observed no changes in total copy number of 16S and 18S rRNA genes from different microbes (bacteria, methanogens, and/or protozoa) when providing different doses of 3-NOP to sheep or beef cattle. However, for beef heifers fed a high-forage diet (60% forage, DM basis) at 65% of ad libitum intake, Romero-Perez et al. (2015) reported that supplementation of 3-NOP had no effect on the total rumen bacterial population but reduced the abundance of methanogens. In our study, richness and evenness of ASVs measured by Chao1 and Shannon indices, respectively, were higher for the control treatment relative to 3-NOP for both bacteria and archaea. However, 3-NOP supplementation had minimal impact on the relative abundance of predominant bacteria and archaeal communities. Given that 3-NOP is a highly targeted inhibitor that specifically reduces the activity of the dominant rumen methanogens, its minor impacts on prominent rumen bacteria were not unexpected.

Specific effects of 3-NOP supplementation on the activities of certain rumen bacteria and methanogens have been reported (Haisan et al., 2014; Romero-Perez et al., 2015; Zhang et al., 2020; Gruninger et al., 2022). Feeding a high-fiber diet (90% forage, DM basis) at maintenance level supplemented with 3-NOP (200 mg/kg DM), Gruninger et al. (2022) reported no shift in the total composition of the rumen microbial community but observed a decrease in bacterial α-diversity in rumen fluid samples due to reduction in Firmicutes:Bacteroidetes ratio. Similarly, using an in situ ruminal incubation technique and beef heifers fed high forage diets supplemented with 3-NOP (150 mg/kg DM), Zhang et al. (2020) reported no effects on the overall composition of the microbial community colonizing the surface of feed in the rumen but a reduction in the relative abundance of genera Methanobrevibacter and Methanosphaera. Relatively few studies have examined the impacts of 3-NOP supplementation on the rumen microbial community in beef cattle and given the differences in findings among these studies, further research is warranted.

Enteric gas production

Among existing CH₄ mitigation practices, supplementation of 3-NOP can be highly effective in reducing CH₄ emissions from ruminants (Beauchemin et al., 2020; Yu et al., 2021). The CH₄ reduction potential of 3-NOP is attributed to the inhibition of methyl-coenzyme M reductase, the enzyme required for the last step of methanogenesis (Duin et al., 2016). The efficacy of 3-NOP in reducing enteric CH₄ emissions without apparent side effects has been reported earlier in beef (Vyas et al., 2018b; Kim et al., 2019; Alemu et al., 2021a, 2021b; Zhang et al., 2021) and dairy cows (Haisan et al., 2017; van Gastelen et al., 2020). Kim et al. (2020) conducted a meta-analysis of 14 studies and reported that supplementation of 3-NOP linearly decreased enteric CH₄ yield (g/kg DM intake) regardless of animal type (beef, dairy) and length of experimental period (slope = −0.041, P < 0.0001, R² = 0.744). Similarly, a meta-analysis of data from 11 experiments (Dijkstra et al., 2018) indicated that with a mean inclusion rate of 123 mg 3-NOP/kg DM, enteric CH₄ production (g/d) and yield (g/kg DM) were reduced by −22.2 and −17.1%, respectively, in beef cattle fed a range of diets. The authors also indicated that the effect of 3-NOP on enteric CH₄ production was positively associated with dose and inversely associated with diet NDF concentration. Based on the model proposed by Dijkstra et al. (2018), an increase of 10 mg 3-NOP/kg DM supplementation from the mean of 123 mg/kg DM resulted in a decline of 3-NOP effect on enteric CH₄ production by 2.56%. As such, the 150 mg/kg DM inclusion rate used in our study was expected to reduce CH₄ production by about 20.5%, which is comparable to the measured reduction value of 22.4%. Furthermore, the observed reduction in CH₄ production (g/d) and yield (g/kg DM intake), 22.4 and 22.0% respectively, in the current study (150 mg 3-NOP/kg DM) is comparable to a previous study by Alemu et al. (2021a) that reported a 20.2 to 25.5% reduction in CH₄ yield for beef cattle fed a high-forage diet (70% DM barley or corn silage and 29% steam-flaked barley grain) supplemented with 3-NOP at 150, 175 and 200 mg/kg DM.

Rumen H₂ is mainly consumed by archaea during methanogenesis to produce CH₄. As such, inhibition of CH₄ is often accompanied by an increase in H₂ accumulation or increase of products that serve as alternative H₂ sinks including formate, propionate, valerate, ethanol, lactate, unsaturated fatty acids, nitrate and sulfate reduction and microbial protein synthesis (Guyader et al., 2017). Furthermore, Pitta et al. (2022) reported that under inhibited methanogenesis by 3-NOP, fluctuations in H₂ concentrations were accompanied by changes in the expression of hydrogenases in H₂-producing bacteria to regulate the amount of H₂ production. In our study, 3-NOP supplementation greatly increased gaseous and dissolved H₂ in rumen fluid. Daily CH₄ production was reduced by 47 g/d which is equivalent to (e.g., in terms of reducing equivalents) releasing 24 g H₂/d. About 7% (1.7 g/d) of the produced H₂ was recovered in the gaseous form and the concentration of H₂ in the dissolved form was 11-fold higher. The large increase in gaseous and dissolved H₂ due to 3-NOP is in agreement with previous studies (Vyas et al., 2018b; van Gastelen et al., 2020; Melgar et al., 2020a; Zhang et al., 2020, 2021; Alemu et al., 2021a). In the present study, 3-NOP decreased the percentage of GE lost as CH₄ from 7.8 to 6.2% (21%). However, given that H₂ is energy-dense (142 kJ/g H₂; Afeefy et al., 2011), greater accumulation of H₂ partially offset the advantage of energy spared by CH₄ mitigation using 3-NOP.

Conclusions

This study investigated the impacts of 3-NOP supplementation using high-fiber diets on feed intake, apparent total-tract digestibility, enteric gas emissions, rumen fermentation, and microbial composition. The hypothesis that supplementation of a high-forage diet with 3-NOP would reduce enteric CH₄ emissions without affecting DM intake and apparent total-tract nutrient digestibility was verified. However, supplementation of 3-NOP (150 mg/kg DM) appeared to have affected the hydrogenotrophic methanogenesis pathway resulting in an increase in alternative H₂ sinks in the rumen as well as gaseous H₂ losses, without major impacts on the rumen microbial community composition.

Glossary

Abbreviations

3-NOP

3-nitrooxypropanol

ADF

acid detergent fiber

ASV

amplicon sequence variants

BW

body weight

CH₄

methane

CP

crude protein

DM

dry matter

GE

gross energy

H₂

hydrogen

NH₃–N

ammonia nitrogen

NDF

neutral detergent fiber

OM

organic matter

PCoA

principal coordinate analysis

Pef

physical effectiveness factor

peNDF

physically effective NDF

TMR

total mixed ration

VFA

volatile fatty acid

Funding

This project was provided by Agriculture and Agri-Food Canada, and DSM Nutritional Products supplied the 3-NOP.

Conflict of Interest Statement

Maik KINDERMANN is employed by DSM Nutritional products (Basel, Switzerland). The other authors have no conflicts of interest.