Influence of Anti-Coccidial Compounds and Phytogenic Saponin Extracts on In Vitro and In Vivo Ruminal Fermentation and Methane Production of Cattle

Simple Summary There is rising interest globally in reducing the use of antibiotics in livestock feeding regimens. Phytogenic feed additives have been evaluated for their potential use to replace compounds such as ionophores in finishing cattle diets. The effects of saponin-containing extracts from Yucca schidigera (YSE) on ruminal fermentation, average daily gain, and feed efficiency appear to be similar to monensin; however, YSE is not approved for the control or prevention of coccidiosis in cattle. Therefore, the objectives of this study were to evaluate sources and levels of anti-coccidial compounds and saponins, individually or in combination, on in vitro and in vivo ruminal fermentation and CH4 production. In summary, the combination of decoquinate with YSE did not influence in vivo CH4 production but tended to increase ruminal propionate proportion. Monensin inclusion decreased in vitro CH4 production and the acetate:propionate. Increasing saponin inclusion increased the propionate proportion in vitro but was not accompanied by a reduction in CH4 production. Abstract Four experiments were conducted to evaluate sources of anti-coccidial compounds and phytogenic saponin extracts on in vitro and in vivo ruminal fermentation and CH4 production at multiple inclusion levels. In experiment 1, eight steers were fed either a finishing diet or a finishing diet supplemented with 0.5 mg/kg BW decoquinate (DCQ) and 3.33 mg/kg BW Yucca schidigera extract (YSE), and respiratory gas exchange was measured. In experiment 2, four ruminally-cannulated steers were fed the same treatments as experiment 1, and ruminal fermentation was evaluated. Anti-coccidial sources (experiment 3; monensin, DCQ, amprolium) and saponin sources (experiment 4; YSE, Quillaja saponaria extract) and levels were evaluated for effects on in vitro ruminal fermentation and CH4 production. DCQ + YSE supplementation did not influence (p ≥ 0.24) in vivo respiratory gas consumption/production, in situ DM degradation, or liquid passage kinetics. Ruminal propionate proportion tended to increase (p = 0.09) with DCQ + YSE. Monensin decreased (p ≤ 0.04) in vitro acetate:propionate and CH4 production; saponin supplementation linearly increased (p < 0.01) propionate proportion but did not influence (p ≥ 0.38) in vitro CH4 production. Saponins and non-antibiotic anti-coccidials did not influence in vitro or in vivo CH4 production with finishing diets.


Introduction
Ruminal methanogenesis is a potential pathway for the elimination of H 2 and the regeneration of NAD + for microbial glycolysis [1]. Methane mitigation strategies can largely be classified into three main groups: animal and feed management, diet formulation, and rumen manipulation [2]. High-concentrate diets containing moderate to large proportions of starch are fed to increase the energy density of the diet to allow for more efficient growth and improved product quality. When high-concentrate diets are fed to beef cattle, the metabolizable energy to digestible energy ratio increases compared to feeding highroughage diets [3]. The net result is proportionally lower CH 4 energy losses when feeding high-concentrate diets [3]. Although feeding high-concentrate diets results in less CH 4 production per unit of fermentable organic matter compared with high-roughage diets, the challenge remains to develop strategies to further decrease enteric CH 4 emissions from feedlot cattle. Thus, combining enteric CH 4 mitigation strategies that target animal and feed management, diet formulation, and rumen manipulation simultaneously could potentially be effective for achieving additive reductions in enteric CH 4 emissions [2,4].
Ionophores, such as monensin, have historically been supplemented in finishing cattle diets to improve average daily gain, feed efficiency, and reduce CH 4 emissions [5]. Dietary monensin inclusion leads to energetic advantages by modifying ruminal microbial populations, which results in increased propionate proportion and decreased CH 4 production [6]. However, there is rising global interest in removing antibiotics, such as ionophores, from livestock diets, and consequently, phytogenic feed additives have been explored for their potential use in ruminant livestock diets [7][8][9][10]. Recent studies have demonstrated that Yucca schidigera extract (YSE) supplementation increased the average daily gain and feed efficiency of finishing cattle when included in the diet at up to 4 g/d [11,12]. In ruminants, isolated saponins or saponins in phytogenic extracts, such as YSE, have been shown to decrease CH 4 production both in vitro [13,14] and in vivo [15,16]. Potential mechanisms by which YSE decreases CH 4 production include altering ruminal microbial populations, inhibiting ruminal H 2 production, or decreasing feed intake or ruminal digestibility [14]. Lila, et al. [15] found that feeding sarsaponin to steers at 0.5% or 1% of the diet (11.2 or 22.4 g/d) on a dry matter (DM) basis decreased in vivo CH 4 production by up to 12.7% and was associated with increased ruminal propionate proportion and decreased ruminal NH 3 and protozoa concentrations. Those authors also found that dietary sarsaponin inclusion decreased total-tract DM and neutral detergent fiber digestibility [15]. In general, the effects of YSE on ruminal fermentation, growth performance, and feed efficiency appear to be similar to the effects of ionophores in cattle diets [17].
Monensin functions as an anti-coccidial compound [18,19], and although YSE has some anti-coccidial activity [17], its potential application in the prevention or control of coccidiosis in finishing cattle diets has not yet been determined [20]. Decoquinate (DCQ) is a nonantibiotic feed additive approved for use in the control of coccidiosis in cattle in multiple countries worldwide [21]. Increasing DCQ inclusion in high-forage and high-concentrate diets did not influence diet digestibility or ruminal fermentation characteristics [22], but it is thought that increased average daily gain and feed efficiency of beef cattle supplemented with DCQ is because of its function as a cocciodiostat [21][22][23][24]. The objectives of the current study were to evaluate the combination of YSE with DCQ on CH 4 production of steers using indirect calorimetry and ruminal fermentation characteristics. We hypothesized that feeding YSE in combination with DCQ would decrease in vivo CH 4 production, and increase ruminal propionate proportion, without influencing the rate or extent of in situ ruminal DM degradability. We further evaluated sources of anti-coccidial compounds (monensin vs. non-antibiotic) and phytogenic saponin extracts (steroidal vs. triterpenoidal) individually for their effects on in vitro ruminal fermentation and CH 4 production at inclusion levels resembling practical feeding conditions or supranutritional levels.

Materials and Methods
All animal procedures were approved by the University of Kentucky Animal Care and Use Committee (Protocol 2020-3546).

Experiment 1
Eight Holstein steers (initial body weight (BW) = 309 ± 28.0 kg) were used to determine the effects of DCQ + YSE supplementation on in vivo CH 4 production. Steers were housed in individual pens (3 m × 3 m) in the Intensive Research Building of the University of Kentucky C. Oran Little Research Center in Versailles, KY, USA. The experimental design was a randomized complete block crossover design with two periods. Steers were fed either the basal diet (control) or the basal diet supplemented with 0.5 mg of DCQ per kg of BW (Deccox; Zoetis, Parsippany, NJ, USA) + 3.33 mg of YSE per kg of BW (30% solids; Micro-Aid Feed Grade Concentrate; DPI Global, Porterville, CA, USA) (DCQ + YSE). The amount of DCQ fed was for the prevention of coccidiosis in ruminating and non-ruminating calves (including veal calves) and cattle caused by Eimeria bovis and E. zuernii, according to the manufacturer (Zoetis, Parsippany, NJ, USA). The amount of YSE fed was based upon the manufacturer's recommendation for beef cattle to supply 1 g/d YSE (DPI Global, Porterville, CA, USA).
The basal high-concentrate diet ( Table 1) was formulated to supply two times the net energy required for maintenance (NE m ) and to exceed requirements for ruminally degradable protein, metabolizable protein, vitamins, and minerals [25]. The basal diet was mixed in one batch, vacuum-sealed in Cryovac barrier bags (0.559 m × 0.914 m; Sealed Air Corporation, Charlotte, NC, USA), and frozen at −4 • C until use. Rations were fed (7.09 ± 0.492 kg DM) once daily at 0800. Fine-ground corn (454 g) was mixed with treatments and fed 15 min before the morning feeding each day to ensure complete consumption of the offered dose of DCQ + YSE. Steers had ad libitum access to water throughout the experiment. Periods were 10 d in length, including a 7d adaptation to treatments, followed by a 3 d collection period. Because YSE decreased in vitro CH 4 production after 24 h of fermentation [14], it was assumed that a 7 d adaptation period would be sufficient to observe a response in in vivo CH 4 production. Previous research has shown the short-term effects of feed additive supplementation on CH 4 production [26], with decreases in CH 4 production occurring within 3 to 4 d [27,28]. Respiratory gas exchange was measured over three consecutive 24 h periods from day 8 to day 10. After completion of the collection period, steers were switched to the opposite treatment to begin the next adaptation period. , and mineral concentrations (Ca and P) were determined using inductively coupled plasma-optical emission spectroscopy (iCAP PRO XP ICP-OES; Thermo Fisher Scientific Inc., Beverly, MA, USA).

Respiratory Gas Exchange
Steers were fed treatments for 7 d in individual pens before being transferred to metabolism stalls (1.52 m × 2.13 m) for the measurement of respiratory gas exchange using indirect calorimetry. The design of the head-box-style respiration chambers was previously described by Koontz, et al. [29]. Each respiration chamber was fitted with a waterer, feeder, and air-conditioning unit to maintain consistent temperature (21 • C) and relative humidity (35%). Before use, the zero point of each gas analyzer was calibrated with pure N 2 gas (American Welding & Gas, Lexington, KY, USA) and the span point of each gas analyzer was calibrated with a custom analytical standard (American Welding & Gas, Lexington, KY, USA; 19.900% O 2 , 0.700% CO 2 , 0.0650% CH 4 ). Recovery of O 2 (105 ± 7.1%) and CO 2 (102 ± 6.8%) for each respiration chamber was determined by combusting a known amount of propane (119 ± 8.2 g) over a 120 min period.
The flow system and equipment arrangement were similar to those described by Hellwing, et al. [30]. Air flow was maintained at 600 L/min during measures of respiratory gas exchange via a fan motor control (Flow Max XL; Columbus Instruments, Columbus, OH, USA). The airflow from each respiration chamber was measured using mass flow meters (HFM-200; Teledyne Hastings Instruments, Hampton, VA, USA) with laminar flow elements (LS-4F; Teledyne Hastings Instruments, Hampton, VA, USA). Inlet and exhaust airflow from the chambers passed through a 10-channel expansion interface, system sample pump (0.5 L/min), and sample drier before analysis. Sampled air was analyzed for O 2 concentration by paramagnetic detection (Columbus Instruments, Columbus, OH, USA), and CO 2 (Columbus Instruments, Columbus, OH, USA) and CH 4 (VIA-510; Horiba Ltd., Kyoto, Japan) concentrations were measured using infrared gas analyzers. Data from the flow meter and gas analyzers were integrated through a CI-Bus Serial Interface (Columbus Instruments, Columbus, OH, USA), and respiratory gas measurements were recorded in 9 min intervals using Oxymax software (version 4.8.5; Columbus Instruments, Columbus, OH, USA). Oxygen consumption, CO 2 production, and CH 4 production were calculated as the total volume of gas consumed/produced after 24 h of measurements. The respiratory quotient was calculated as the liters of CO 2 produced divided by the liters of O 2 consumed.

Experiment 2
Four ruminally-cannulated Holstein steers (initial body weight = 469 ± 22.1 kg) were used to determine the effects of DCQ + YSE supplementation on ruminal fermentation, in situ ruminal degradability, and liquid passage rate. The experimental design was a randomized crossover design with two periods. Steers were fed either the basal diet (control) or the basal diet supplemented with 0.5 mg of DCQ per kg of BW + 3.33 mg of YSE per kg of BW (DCQ + YSE). Steers were housed in individual pens (3 m × 3 m), fed the same diet (9.70 ± 0.341 kg DM) once daily, as described in experiment 1, and had ad libitum access to water. Treatments were mixed with finely ground corn, as described in experiment 1. Periods were 14 d in length, including 7 d for adaptation to treatments and 7 d for sample collection. The in situ degradability experiment was conducted from day 8 to day 12, followed by the collection of ruminal fluid on day 14. After completion of the collection period, steers were switched to the opposite treatment to begin the next adaptation period.

In Situ Ruminal Degradability
The in situ ruminal DM degradability of the basal diet was measured using methods previously described [31]. Twenty grams of the basal diet, ground to pass a 2 mm screen, was weighed into nylon bags (10 cm × 20 cm; 50 µm pore size; R1020 Forage Bag; ANKOM Technology, Macedon, NY, USA). The nylon bags were incubated in the rumen for 0, 3,6,9,12,24,36,48,72, and 96 h. One nylon bag for each incubation timepoint was placed into a zipped wash bag (25 cm × 31 cm; k2107; HomeAide Delicate Wash Bag) that was suspended in the ventral rumen of each steer. Two stainless steel magnets (1.27 cm × 7.62 cm; Silver Star AlniMAX II, Sundown Industries Co., Plainview, NY, USA) were added to each wash bag to ensure immersion in the ventral rumen. Wash bags were attached to a steel chain with a breeching snap clip (2710231; Koch Industries, Inc., Minneapolis, MN, USA). The steel chain was secured to the rumen cannula cap by connecting the steel chain to an inverted U-bolt on the inner portion of the cannula cap (#1 Eazy-out Stopper; Bar Diamond, Inc., Parma, ID, USA) with a breeching snap clip. Wash bags were inserted in reverse order so that all bags were removed from the rumen and rinsed simultaneously. At removal, the wash bags were removed from the steel chain and placed into an ice-water bath to stop fermentation. The nylon bags were removed from the wash bag and rinsed 5 times in a washing machine with 1 min rinse and 2 min spin cycles [32]. The nylon bags were then dried in a 100 • C forced-air oven (1680; Sheldon Manufacturing, Inc., Cornelius, OH, USA) for 48 h to determine in situ DM disappearance.
The potential rate and extent of in situ DM degradation were determined using the first-order asymptotic model [33]: where y is the degradation after t hours, a is the soluble fraction, b is the potentially degradable fraction, k d is the fractional rate of degradation of b, t is the incubation time (h), and Lt is the lag time (h). Dry matter degradability data from times 0, 3, 6, 9, 12, 24, 36, 48, 72, and 96 h were fitted to the above nonlinear model using SAS according to the procedures described by Fadel [34] to generate the parameters described. The in situ ruminal degradability was determined by using the parameters generated for the rate and extent of degradation, as previously described, and modeled with the rate of passage [35]: where isRD is the in situ ruminal degradability, a is the soluble fraction, b is the potentially degradable fraction, k d is the fractional rate of degradation of b, and k p is the fractional rate of the liquid passage. The fractional rate of liquid passage measured on day 14 was used for k p . Degradation coefficients from the generated parameters were converted to percentages by multiplying coefficients by 100.

Ruminal Fermentation and Liquid Passage
On day 14, steers were administered 500 mL of CrEDTA (2.3 g Cr) solution [36] through the rumen cannula 2 h after the morning feeding to evaluate the ruminal liquid passage rate. Approximately 150 mL of ruminal contents were collected from the midventral region of the rumen immediately before administration of the dose (0 h) and at 0.5, 1, 2, 4, 6, 8, 10, 12, 16, 20, and 24 h after dosing. The ruminal contents were squeezed through four layers of cheesecloth to separate the ruminal fluid for analysis of NH 3 and L(+)-lactate by UV-VIS spectrophotometry and VFA by gas chromatography (Hewlett-Packard 6890 Plus GC; Agilent Technologies Inc., Santa Clara, CA, USA). Squeezed ruminal fluid (5 mL) was combined with 0.5 mL of 500 g/L metaphosphoric acid as a deproteinizing agent [37] and 0.5 mL of 85 mM 2-ethylbutyrate as an internal standard. Acidified samples were frozen at -20 • C to facilitate protein precipitation. Samples were thawed, centrifuged at 20,000× g for 15 min at 4 • C, and the supernatant fractions were transferred to autosampler crimp-top vials. Each aqueous sample (0.2 µL) was injected into the inlet with an automatic liquid sampler (7693A; Agilent Technologies Inc., Santa Clara, CA, USA) and vaporization occurred at 260 • C. The sample was carried to the fused silica capillary column (25326; Nukol Capillary GC Column; Supelco Inc., Bellefonte, PA, USA) with He at a 2:1 split ratio. The initial oven temperature was 110 • C for 1 min, ramped to 125 • C at 5 • C per min, ramped to 195 • C at 65 • C per min, and cooled to 110 • C post-detection. Upon column exit, separated SCFA were detected with a flame ionization detector and quantified using electronic integration. Total VFA concentration was considered as the sum of acetate, propionate, isobutyrate, butyrate, isovalerate, and valerate concentration. Molar VFA proportions were calculated as the individual VFA concentration divided by the total VFA concentration and multiplied by 100. Ruminal NH 3 concentration was analyzed using the glutamate dehydrogenase procedure [38] adapted to a Konelab 20XTi Clinical Analyzer (ThermoFisher Scientific Inc., Beverly, MA, USA). Ruminal L(+)-lactate concentration was measured using the L(+)-lactate dehydrogenase procedure [39,40] adapted to a multi-mode plate reader (BioTek Synergy HTX; Agilent Technologies Inc., Santa Clara, CA, USA).
Chromium concentrations for each sample were determined using atomic absorption spectroscopy (Aanalyst 200; PerkinElmer Inc., Waltham, MA, USA) at a wavelength of 357.87 nm. Baseline concentrations of Cr (0 h) were used to correct the concentrations measured at each time point. The concentration of Cr after dosing and fractional clearance rate of Cr was determined by calculation of the exponential decay rate for Cr using the NLIN procedure of SAS (version 9.4; SAS Institute Inc., Cary, NC, USA) and the following equation: where Cr t represents the Cr concentration at a given time, Cr 0 represents the Cr concentration at time 0-h, k represents the fractional rate of Cr clearance which is assumed to be equivalent to the fractional rate of liquid passage (k p ), and t represents the time in hours [41,42]. The liquid retention time in the rumen was calculated as the absolute value of 1/k p . The rumen liquid volume (L) was determined by dividing the amount of Cr dosed by the amount of Cr present at time zero (C 0 ). The liquid flow rate (L/h) was calculated as k p × rumen liquid volume.

Experiments 3 and 4 2.3.1. Experiment 3
The objectives of this experiment were to evaluate the effects of antibiotic or nonantibiotic sources of feed additives containing anti-coccidial activity on in vitro CH 4 production and ruminal fermentation. Additionally, sources of the anti-coccidal compounds were evaluated at multiple inclusion levels that resemble practical feeding conditions or supranutritional inclusion. The experimental design was a randomized complete block design with 7 treatments in a 3 × 2 + 1 factorial arrangement. Three sources of anti-coccidial compounds were tested: monensin (Rumensin 90; Elanco Animal Health Incorporated, Greenfield, IN, USA), DCQ (Deccox; Zoetis Inc., Parsippany, NJ, USA), and amprolium (Corid 9.6% Oral Solution; Huvepharma Inc., Peachtree City, GA, USA). There were two levels for anticoccidial sources: 1X (based on current feeding recommendations) and 10X (10-fold greater than the 1X dose). The 1X and 10X treatment levels were 30 and 300 mg/kg substrate for monensin, respectively. The 1X and 10X treatment levels were 25 and 250 mg/kg substrate for DCQ, respectively. The 1X and 10X treatment levels were 500 and 5000 mg/kg substrate for amprolium, respectively. Treatment concentrations were corrected for active ingredient percentages and DM content. The basal substrate without any feed additives (0X) was included as a negative control.
Treatments were added to the fermentation vessels according to the proportions described in Table 2. To aid in the accuracy of treatment dispersal, treatment premixes (Rumensin 90 and Deccox) were mixed with the basal substrate. Monensin was prepared as a 10% (w/w) mixture of Rumensin 90 with the basal substrate. Decoquinate was prepared as 10% and 50% (w/w) mixtures of Deccox with the basal substrate for the 1X and 10X treatments, respectively. Amprolium was prepared as a 6.4% or 64% (v/v) solution of Corid 9.6% Oral Solution diluted in water for the 1X and 10X treatments, respectively. All vessels received 500 mg of the basal substrate or basal substrate mixed with treatment so that all flasks contained the same substrate volume. All vessels received 2 mL of water or added treatment diluted in water so that all flasks contained the same liquid volume. There were two replicate fermentation vessels for each treatment, and the experiment was replicated on four separate days.  1 Rumensin 90 premix was mixed with the basal substrate on a 10% (w/w) basis. 2 Deccox premix was mixed with the basal substrate on a 10% (w/w) basis for the 1X treatment. 3 Deccox premix was mixed with the basal substrate on a 50% (w/w) basis for the 10X treatment. 4 Two milliliters of each solution were added to their respective fermentation vessels. Corid 9.6% Oral Solution was diluted with 6.4 mL/100 mL and 64 mL/100 mL for the 1X and 10X treatments, respectively. Liquid saponin treatments were mixed with water on a v/v basis.

Experiment 4
The objectives of this experiment were to evaluate the effects of steroidal or triterpenoidal sources of saponins from phytogenic extracts on in vitro CH 4 production and ruminal fermentation. Additionally, saponin sources were evaluated at multiple inclusion levels that resemble practical feeding conditions or supranutritional inclusion. The experimental design was a randomized complete block design with 7 treatments in a 2 × 3 + 1 factorial arrangement of treatments. Two liquid sources of saponins containing 50% solids were tested: YSE (Micro-Aid Liquid; DPI Global, Porterville, CA, USA) and Quillaja saponaria extract (QSE; Phytogenic Patch Plus Triple P; DPI Global, Porterville, CA, USA). There were three levels for each saponin source: 1X, 10X, and 20X. The 1X, 10X, and 20X treatments were 100 mg/kg substrate, 1000 mg/kg substrate, and 2000 mg/kg substrate, respectively. Treatments were corrected for active ingredient percentages of the premix. The basal substrate without any feed additive (0X) was included as a negative control.
Treatments were added to the fermentation vessels according to the proportions described in Table 2. Saponin treatments were prepared as 0.25%, 2.5%, or 5% (v/v) solutions diluted in water for the 1X, 10X, and 20X treatments, respectively. All vessels received 2 mL of water or added treatment diluted in water so that all flasks contained the same liquid volume. There were two replicate fermentation vessels for each treatment, and the experiment was replicated on four separate days.

Fermentation Preparation
Conditions of the in vitro gas production protocol were similar to those described previously [43][44][45][46][47]. The same four ruminally-cannulated Holstein steers (initial BW = 554 ± 30.9 kg) used in experiment 2 were housed outdoors in dry lot pens (2.4 m × 14.6 m). Steers were fed the same high-concentrate diet used in experiments 1 and 2, and the diet was fed ad libitum once per day. Ruminal contents (1000 g) were collected 4 h after feeding from each steer and combined into an insulated container (YETI Rambler One Gallon Water Jug; Yeti Holdings, Inc., Austin, TX, USA) for transport to the laboratory. Ruminal contents were blended under CO 2 headspace for 30 s and then squeezed through four layers of cheesecloth. Squeezed ruminal fluid (600 mL) was combined with buffer, macromineral, micromineral, and reducing solutions (total = 2700 mL) that were prepared as described by Goering and Van Soest [44]. Fermentation vessels used in the current study were 250 mL coated glass bottles (Cat. #7056; ANKOM Technology, Macedon, NY, USA). Samples of the basal diet were dried at 55 • C for 24 h and ground to pass a 2 mm screen for use as the substrate for the in vitro ruminal fermentation. Substrates (501 ± 0.281 mg) were pre-weighed into fermentation vessels and combined with 2 mL of water or liquid treatment diluted in water to aid in the dispersal of the buffered inoculum. The buffered inoculum (103.2 ± 2.52 g) was then added to each fermentation vessel. Fermentation vessels were gassed with CO 2 for 20 s, capped with RF1 gas production modules (ANKOM Technology, Macedon, NY, USA), and placed into a 39 • C circulating water bath (89 L; Precision 2868 Circulating Water Bath; ThermoFisher Scientific Inc., Beverly, MA, USA) to equilibrate. After all fermentation vessels were added to the water bath, valves of the gas production modules were opened simultaneously to release any accumulated pressure. Then, valves were closed, and cumulative gas pressure was measured in 5 min intervals over a 24 h incubation period using ANKOM Gas Pressure Monitor software (version 11.4; ANKOM Technology, Macedon, NY, USA).

Sample Collection and Analysis
At the completion of the 24 h incubation period, vessels were removed from the water bath and transferred to an ice bath to stop the fermentation. Valves of the RF1 gas production modules were opened and samples of gas from the headspace of each vessel were collected with a syringe and stored in serum tubes (BD Vacutainer; Beckton, Dickinson and Company, Franklin Lakes, NJ, USA). Gas samples were analyzed for CH 4 concentration by gas chromatography (Agilent 7890A GC; Agilent Technologies, Santa Clara, CA, USA). Each gas sample (50 µL) was manually injected into the inlet with a small hub removable needle (7784-06; Hamilton Company, Reno, NV, USA) in a 100 µL Teflon syringe (7656-01; Hamilton Company, Reno, NV, USA). The samples were carried to the fused silica capillary column (19091J-413; HP-5 GC Column; Agilent Technologies, Santa Clara, CA, USA) with He at a 20:1 split ratio. Inlet, oven, and detector temperatures were set to 200 • C, 150 • C, and 250 • C, respectively. Upon column exit, separated CH 4 was detected with a flame ionization detector and quantified using electronic integration. Methane concentrations were determined in reference to analytical standards composed of 1% (v/v) and 10% (v/v) CH 4 balanced with N 2 (American Welding & Gas, Lexington, KY, USA). The RF1 gas production modules were opened, pH of the fermentation media was measured using a combination electrode and meter (SevenCompact S220; Mettler-Toledo International Inc., Columbus, OH, USA), and a 1 mL aliquot of the fermentation media was prepared and analyzed for VFA and NH 3 concentrations as described in experiment 2.

In Vitro Gas Production
The cumulative gas pressure was corrected for atmospheric pressure, converted to moles of gas produced using the ideal gas law, and then converted to milliliters of gas produced under standard conditions using Avogadro's law. Cumulative gas production was corrected for the gas volume in the headspace of each fermentation vessel. The gas volume in the headspace of each fermentation vessel was calculated as the total volume of the vessel (308.2 ± 5.22 mL) minus the sum of the volume of the buffered inoculum and substrate (103.8 ± 2.47 mL). Cumulative gas production was fitted to the exponential model described by Pitt, et al. [48] using GraphPad Prism 5 (Dotmatics, Boston, MA, USA): where F(t) is the cumulative gas production, r is the rate of gas production, t is the time in hours, and λ is the lag time in hours. The rate of gas production was converted to a percentage by multiplying by 100. Methane production was determined by multiplying the CH 4 concentration of gas samples by cumulative gas production after 24 h of fermentation.

Statistical Analysis
For experiment 1, variance-covariance structures for the statement of the repeated measure were assessed for fit using Bayesian information criterion for antedependence 1, autoregressive 1, compound symmetry, simple, and unstructured, including the steer as the subject. Whole-body O 2 consumption, CO 2 production, and CH 4 production were analyzed using the repeated measures statement of the MIXED procedure of SAS for fixed effects of replicate, day, treatment, period, and their interactions. The initial measurements of O 2 consumption, CO 2 production, and CH 4 production on day 8 were included in the model statements as covariates.
For experiment 2, ruminal Cr concentrations were fitted to the nonlinear equation previously described using the NLIN procedure of SAS. The in situ ruminal DM degradation data were analyzed using the NLIN procedure of SAS to estimate the parameters of the equation previously described. Degradation parameters and liquid passage characteristics were analyzed using the GLM procedure of SAS for fixed effects of period, treatment, and the period × treatment interaction. Variance-covariance structures were tested for ruminal VFA, NH 3 , and L(+)-lactate concentrations as described in experiment 1. Concentrations of ruminal fermentation end-products were analyzed using the repeated measures statement of the MIXED procedure of SAS for fixed effects of period, time, treatment, and the time × treatment interaction. The initial metabolite concentration (0 h) was included in the model statement as a covariate for ruminal metabolites.
For experiment 3, gas production kinetics and fermentation end-products were analyzed using the GLM procedure of SAS for fixed effects of the replicate and treatment. The IML procedure was used to generate orthogonal contrast coefficients to adjust for the unequal spacing between the treatment levels. Contrast statements were used to determine the differences between treatments. To determine the effect of anti-coccidial supplementation, a contrast was analyzed for basal substrate vs. others. To determine the effect of ionophore supplementation (antibiotic vs. non-antibiotic anti-coccidials), a contrast was analyzed for monensin vs. DCQ and amprolium. To determine the effect of the non-antibiotic anti-coccidial source, a contrast was analyzed for DCQ vs. amprolium. For each anti-coccidial source (monensin, DCQ, amprolium), a linear polynomial contrast (0X, 1X, 10X) was analyzed to determine the effects of supplementation level.
For experiment 4, gas production kinetics and fermentation end-products were analyzed using the GLM procedure of SAS for the fixed effects of the replicate and treatment. The IML procedure was used to generate orthogonal contrast coefficients to adjust for unequal spacing between treatment levels. Contrast statements were used to determine differences between treatments. To determine the effect of saponin supplementation, a contrast was analyzed for basal substrate vs. the others. To determine the effect of the saponin source, a contrast was analyzed for YSE vs. QSE. For each saponin source, linear and quadratic polynomial contrasts (0X, 1X, 10X, 20X) were analyzed to determine the effects of supplementation level.
All data were checked for normality using the Shapiro-Wilk test of the UNIVARIATE procedure of SAS. Blocks (replicates) were included in the model statement as fixed effects, as recommended by Dixon [49]. Pairwise differences of least squares means were sepa-rated using the Tukey-Kramer adjustment, protected by a significant F-test. Results were considered significant if p ≤ 0.05. Tendencies were declared when 0.05 < p ≤ 0.10.

Experiment 1
Supplementation of DCQ + YSE did not influence whole-body O 2 consumption and the CO 2 or CH 4 production of steers fed a high-concentrate diet ( Table 3). The respiratory quotient was not different between dietary treatment groups.

Experiment 2
Supplementation of DCQ + YSE did not influence the soluble or potentially degradable DM fractions of the basal diet ( Table 4). The potential rate and extent of DM degradation were not influenced by DCQ + YSE supplementation. The in situ ruminal DM degradability of the basal diet was not affected by dietary treatments. The fractional rate of liquid passage, ruminal liquid retention time, rumen liquid volume, and ruminal liquid outflow were not influenced by DCQ + YSE supplementation.

Experiment 3
Increasing monensin inclusion linearly increased (p < 0.01) pH of the fermentation media after 24 h of incubation ( Table 6). The inclusion of anti-coccidial compounds increased (p = 0.02) gas production after 24 h, but gas production was greater (p < 0.01) for DCQ and amprolium compared with monensin. Gas production decreased (p < 0.01) linearly with increasing monensin inclusion, increased (p < 0.01) linearly with DCQ inclusion, and tended to increase (p = 0.07) linearly with amprolium inclusion. Monensin inclusion produced a faster (p < 0.01) rate of gas production compared with DCQ and amprolium. Increasing the monensin inclusion linearly increased the rate of gas production. Increasing DCQ and amprolium inclusion linearly decreased (p < 0.01) and tended to decrease (p = 0.07) the rate of gas production. Monensin inclusion decreased (p < 0.01) CH 4 percentage and production compared with DCQ and amprolium. Increasing monensin linearly decreased (p < 0.01) CH 4 percentage and production. Ammonia concentration linearly decreased (p = 0.04) with increasing DCQ inclusion.
Total VFA concentrations were lower (p = 0.05) for monensin compared with DCQ and amprolium. Acetate molar proportion linearly decreased (p = 0.02) with increasing levels of DCQ. Propionate proportion was greater (p < 0.01) for all treatments containing anticoccidial compounds compared to the basal substrate. However, the propionate proportion was greater (p < 0.01) when monensin was included in the in vitro fermentation compared with DCQ and amprolium. Increasing monensin and DCQ inclusion linearly increased (p < 0.01) the molar proportion of propionate. Isobutyrate and butyrate molar proportions were greater (p ≤ 0.04) for non-antibiotic anti-coccidial compounds compared with monensin. Increasing monensin inclusion linearly decreased (p < 0.01) butyrate proportion and tended to decrease (p = 0.08) isobutyrate proportion linearly. Isobutyrate proportion tended to increase (p = 0.08) linearly with increasing inclusion of DCQ. The isovalerate proportion was greater (p < 0.01) with monensin inclusion compared with DCQ and amprolium because increasing monensin inclusion linearly increased (p < 0.01) isovalerate proportion. Valerate proportion linearly decreased (p = 0.04) with monensin inclusion and linearly increased (p = 0.04) with amprolium inclusion. The anti-coccidial compound inclusion decreased (p < 0.01) the acetate:propionate, with the acetate:propionate being reduced (p < 0.01) to the greatest extent by monensin. Increasing levels of DCQ and monensin linearly decreased (p < 0.01) the acetate:propionate.

Experiment 4
The pH of the fermentation media after 24-h of incubation was greater (p < 0.01) with QSE inclusion compared with YSE (Table 7). Gas production after 24 h was greater (p < 0.01) for YSE compared with QSE. Increasing levels of YSE inclusion linearly increased (p = 0.02) gas production after 24 h. The rate of gas production was greater (p < 0.01) for QSE compared with YSE. Increasing levels of QSE inclusion linearly increased (p < 0.01) the rate of gas production. Methane percentage and production were not influenced by saponin inclusion. Ammonia and total VFA concentrations were not influenced by saponin inclusion.

Effects of DCQ + YSE on In Vivo Ruminal Fermentation and CH 4 Production
Previous research demonstrated that YSE supplementation decreased in vitro CH 4 production across substrates that contained low-, medium-, or high-proportions of roughages [14]. In the current study, supplementation of DCQ + YSE for up to 10 d did not influence in vivo CH 4 production of steers. Because YSE supplementation had previously resulted in reduced CH 4 production in vitro [14], we assumed that 7 d of adaptation would be adequate time to allow for ruminal turnover to observe effects on CH 4 production. The liquid retention time measured in experiment 2 (18.5-19.9 h) suggests that there was multiple ruminal turnovers within the 7 d adaptation period. However, it is possible that longer adaptation to dietary treatments was necessary to observe effects on CH 4 production. Also, feeding different sources and levels of DCQ and YSE and variation in the concentration of saponins and/or types of steroidal saponins present in YSE could potentially alter responses in CH 4 production.
Reductions in CH 4 production could potentially be associated with increases or decreases in economically important variables in cattle [2]. For example, decreased CH 4 production could be due to decreased intake, decreased digestibility, or decreased VFA production and, therefore, could negatively affect production outcomes [2]. In contrast, decreased CH 4 production associated with decreased protozoa, greater propionate production, and/or increased energy retention could be beneficial for both productive and environmental outcomes [2]. Feed intake was controlled by limiting energy intake to two times NE m in experiments 1 and 2. Results from the in situ degradability experiment demonstrated that the rate and extent of DM degradation and the in situ ruminal DM degradability of the basal diet were not affected by DCQ + YSE supplementation. Likewise, ruminal liquid passage kinetics were not affected by DCQ + YSE supplementation.
Changes in end-products of ruminal fermentation in experiment 2 may suggest that supplementation of DCQ + YSE modified ruminal fermentation similarly when compared to previous studies. In general, supplementation of YSE typically results in decreased ruminal NH 3 concentration, increased ruminal propionate proportion, and decreased ruminal protozoa concentration [50]. Although not statistically significant, ruminal NH 3 concentration was numerically reduced by 9.15% with DCQ + YSE supplementation. Ruminal propionate proportion tended to increase with DCQ + YSE supplementation. Although there was a tendency for a numerical increase in the propionate proportion, DCQ + YSE supplementation did not influence the acetate proportion or the acetate:propionate. It should be noted that the acetate:propionate was 1.43 and 1.41 for control and DCQ + YSE treatments, respectively. The decreased acetate proportion and increased propionate proportion result in a decreased acetate:propionate is closely correlated with decreased CH 4 production [51]. It is possible that the current diet composition and feeding level were propiogenic, limiting the opportunity for YSE to shift H 2 sinks. A recent review pointed out that the positive effects of YSE supplementation are not always observed when included in diets for cattle [50]. Sources of saponin-containing extracts, plant saponin composition, dietary inclusion levels, manufacturing processes, and interactions with dietary components are some factors that could potentially contribute to inconsistencies across studies [14].

Effects of Monensin on In Vitro Ruminal Fermentation and CH 4 Production
A recent meta-analysis reported that monensin supplementation decreased CH 4 production in beef steers and dairy cows [52]. Of the anti-coccidial and saponin sources tested, only monensin decreased in vitro CH 4 production in the current study. Monensin decreased total VFA concentrations and gas production in the current study, which is similar to the findings of others using in vitro gas production systems [53]. However, decreased total VFA concentration is not typically observed when cattle are fed monensin [54]. Rather, monensin can alter ruminal microbial populations by decreasing protozoa and Gram-positive bacteria, resulting in less acetate, butyrate, CH 4 , lactate, and NH 3 production [55,56]. Consistent with numerous previous in vitro and in vivo studies, monensin inclusion decreased the acetate:propionate and increased molar propionate proportion [53,[57][58][59][60]. Decreased isobutyrate and valerate proportions with monensin inclusion are likely due to the anti-microbial effects of monensin on proteolytic and/or amino acid-fermenting bacteria [61], suggesting reduced amino acid degradation [62]. Decreased butyrate proportion with monensin inclusion was also found previously when a concentrate substrate was used [53]. In the current study, monensin did not influence NH 3 concentration which contrasts with several studies reporting that monensin decreased NH 3 concentration [53,61,63]. The lack of effect of monensin on NH 3 concentration in the current study could be due to the excess N available from the in vitro buffer solutions [44], as well as, the basal substrate which exceeded requirements for ruminally degradable protein. Overall, monensin modified in vitro ruminal fermentation and CH 4 production consistent with several previous experiments [53,[64][65][66][67].

Effects of Non-Antibiotic Anti-Coccidial Compounds on In Vitro Ruminal Fermentation and CH 4 Production
In contrast to monensin, non-antibiotic anti-coccidial compounds had minimal impacts on in vitro CH 4 production and ruminal fermentation. Feeding increasing levels of DCQ did not influence the total-tract DM digestibility of steers fed a high-concentrate diet [22]. In contrast, results from the current experiment demonstrated that increasing levels of DCQ linearly increased in vitro gas production. However, this is likely due to changes in the fermentability of the Deccox premix, which replaced the basal substrate (2.1% and 20.1% inclusion rate for 1X and 10X treatments) in the current experiment. According to the manufacturer, the Deccox premix also contained corn meal, soybean oil, lecithin, and silicone dioxide. It is possible that some of the increase in propionate proportion in experiment 2 with DCQ + YSE supplementation could have been due to both DCQ and YSE, as increasing DCQ linearly increased in vitro propionate proportion in experiment 3. Whether or not those effects are due to DCQ or the premix itself remains to be determined. Similar to previous in vivo findings, DCQ had little or no influence on characteristics of ruminal fermentation, including VFA profiles or CH 4 production in the current study. Like DCQ, amprolium inclusion had minimal effects on in vitro ruminal fermentation in the current study.

Effects of YSE and QSE on In Vitro Ruminal Fermentation and CH 4 Production
Previous research had demonstrated that YSE inclusion decreased in vitro CH 4 production across 10%, 50%, and 100% forage-based diets [14]. However, increasing YSE or QSE did not influence in vitro CH 4 production in the current experiment. In the current study, increasing levels of saponins linearly decreased acetate proportion and increased propionate proportion, resulting in decreased acetate:propionate. Changes in molar propionate proportion with an absence of a change in in vitro CH 4 production in experiment 4 are similar to the results found in experiments 1 and 2, where the combination of DCQ + YSE tended to increase the molar proportion of propionate without influencing CH 4 production. Zúñiga-Serrano, et al. [50] proposed that YSE may modify ruminal fermentation through several mechanisms which can lead to downstream effects on enteric CH 4 production. Yucca schidigera extract or saponins from YSE can decrease ruminal cellulolytic bacteria and fungi [68], decrease methanogenic archaea [69], and decrease ruminal protozoa [13,15,70]. Reduced ruminal NH 3 , decreased acetate:propionate, and decreased fiber degradation are associated with decreased CH 4 production with YSE inclusion [50]. One study found that YSE and QSE decrease CH 4 production but at much greater concentrations than those used in the current experiment [13]. Sources of saponin-containing extracts, plant saponin composition, dietary inclusion levels, manufacturing processes, and interactions with dietary components are some factors that could potentially contribute to inconsistencies across studies [14].

Conclusions
Supplementation of DCQ + YSE for 7 to 10 days did not influence O 2 consumption, CO 2 production, or CH 4 production in steers consuming a high-concentrate diet at 2 × NE m . Supplementation of DCQ + YSE did not influence the rate or extent of ruminal DM degradation of the basal finishing diet or liquid passage kinetics. Supplementation of DCQ + YSE did not influence total VFA concentrations but tended to increase ruminal propionate proportion. Increasing levels of monensin decreased in vitro CH 4 production, and acetate:propionate, isovalerate, and valerate proportions. Decoquinate and amprolium had minimal effects on in vitro ruminal fermentation. Increasing YSE or QSE inclusion increased propionate proportion but was not accompanied by a reduction in in vitro CH 4 production. Further research is necessary to identify alternative non-antibiotic compounds for the simultaneous reduction of CH 4 emissions and control of coccidiosis in feedlot cattle diets.  Data Availability Statement: Data will be made available without reservation upon request to the corresponding author.