Effect of live yeast supplementation on energy partitioning and ruminal fermentation characteristics of steers fed a grower-type diet in heat-stress conditions

Abstract

The objective of this trial was to determine the influence of live yeast supplementation (LY), environmental condition (ENV), and their interaction (TRT) on energy partitioning, nitrogen metabolism, and ruminal fermentation dynamics of steers receiving a grower-type diet. The effects of LY and ENV were investigated using a 2 × 2 crossover design that spanned five periods. Eight Angus-crossbred steers were randomly split into pairs and housed in four outdoor pens outfitted with an individualized feeding system. Animals were limit-fed a grower diet (DIET) at 1.2% shrunk body weight (SBW) with no live yeast supplementation (NOY) or a grower diet top-dressed with 10 g LY/d for 14 d (1.2 × 10¹² CFU/d). On days 13 and 14, animals were subjected to one of two ENV conditions, thermoneutral (TN; 18.4 ± 1.1 °C, 57.6 ± 2.8% relative humidity [RH]) or heat stress (HS; 33.8 ± 0.6 °C, 55.7 ± 2.7% RH), in two side-by-side, single-stall open-circuit, indirect respiration calorimetry chambers. Data were analyzed using a random coefficients model. Carryover effects were examined and removed from the model if not significant. Gross (GE), digestible, metabolizable, heat, and retained energies were not influenced by DIET, ENV, or TRT (P ≥ 0.202). Gaseous energy, as a percentage of GE, tended to increase during HS (P = 0.097). The only carryover effect in the study was for oxygen consumption (P = 0.031), which could be attributed to the tendency of NOY (P = 0.068) to have greater oxygen consumption. DIET, ENV, or TRT (P ≥ 0.154) had no effects on total animal methane or carbon dioxide emissions. Similarly, DIET, ENV, or TRT (P ≥ 0.157) did not affect ruminal pH, redox, protozoa enumeration, ruminal ammonia concentrations, and acetate-to-propionate ratio. Propionate concentrations were the greatest in animals in TN conditions receiving LY (P = 0.034) compared to the other TRT. This effect is mirrored by TN-LY tending to have greater acetate concentrations (P = 0.076) and total VFA concentrations (P = 0.065). Butyrate concentrations tended to be greater for animals fed LY (P = 0.09). There was a tendency for LY to have elevated numbers of Fusobacterium necrophorum (P = 0.053). Although this study lacked effects of LY on energy partitioning, nitrogen metabolism, and some ruminal parameters during HS, further research should be completed to understand if LY is a plausible mitigation technique to enhance beef animals’ performance in tropical and sub-tropical regions of the world.

This study investigates the efficacy of live yeast supplementation as a mitigation technique for heat stress in beef cattle fed grower-type diets, and its effects on energy partitioning, nitrogen metabolism, and ruminal fermentation dynamics. Additionally, this study utilizes an in vitro model to simulate the performance of rumen inoculum after heat stress or live yeast supplementation.

Introduction

With an exponentially growing global population, the demand for meat consumption is expected to increase by 30% by 2050 (FAO, 2009). As approximately 70% of global beef production is located in tropical and sub-tropical regions of the world, such as Brazil, Australia, Mexico, and the southeastern region of the United States, the identification of plausible mitigation techniques to heat stress is necessary in order to continue to increase the efficiency of rearing beef cattle in these regions (Cooke et al., 2020). Elevated temperatures paired with humid environments can have detrimental effects on ruminant production, such as decreased feed intake and poor energy metabolism (Fuquay, 1981; Beede and Collier, 1986; Rhoads et al., 2010; NASEM, 2016; Tedeschi and Fox, 2020). The inimical energy losses due to increased tissue metabolism and thermoregulation from heat stress lead to an increase in maintenance energy requirements which, in turn, requires an increase in tissue metabolism and thermoregulation. As a lowered dry matter intake (DMI) cannot support this feedback loop, animal performance suffers and production is significantly reduced. As the animal’s core temperature increases, panting also increases to dissipate heat and cool the body faster, which subsequently results in an increase in the amount of carbon dioxide (CO₂) released from the animal (Bernabucci, 2019; Lacetera, 2019). Although biogenic CO₂ is typically not considered a greenhouse gas, its quantification is still required to account for carbon flux from beef cattle production (IPCC 2001, 2007; Stackhouse-Lawson et al., 2012). Furthermore, when comparing treatments, a changing in biogenic CO₂ emission should be discussed, mainly when control animals emit more than treated animals. While ruminal fermentation continues as rumination and rumen motility slow during heat stress events, there are no a concurrent increases in enteric methane (CH₄) emissions (Ngwabie et al., 2011; Barnett et al., 2015; Yadav et al., 2016). Therefore, heat stress events for ruminant animals are not considered significant contributors to greenhouse gasses and climate change.

Yeast products, most commonly consisting of Saccharomyces cerevisiae variations, have been added to beef cattle rations to mitigate the adverse effects of multiple types of stress, such as shipping, transition, and heat (Batista et al., 2022). Yeast supplementation during heat stress is associated with increased DMI and digestibility, stabilization of rumen pH, and increased growth performance and rumen health (McAllister et al., 2011; Crossland et al., 2018, 2019). Additionally, when fed at a beneficial, targeted amount, thermostable live yeast products can lower redox in the rumen while alleviating drops in pH from a stimulated pool of non-fiber carbohydrate degrading bacteria (Cagle et al., 2019). Although there is a plethora of research surrounding the variable outcomes of yeast products in ruminant production, the direct effects of yeast products on energy partitioning and greenhouse gas emissions have been loosely examined (McGinn et al., 2004; Mutsvangwa et al., 2010; Crossland et al., 2018). Furthermore, ruminal performance after a heat stress event is weakly described in the literature (Kim et al., 2022).

This study aimed to examine the effects of live yeast supplementation in grower-type diets fed to steers undergoing heat stress conditions in indirect calorimetry chambers on energy partitioning, N balance, and ruminal performance. Additionally, this study utilized an in vitro fermentation technique to examine how previously-stressed rumen inoculum may perform once returned to favorable conditions.

Materials and Methods

Animals and experimental diet

Animals used in this study were registered and cared for according to guidelines approved by the Institutional Animal Care and Use Committee (AUP# 2018-0382) at Texas A&M University. This study originated in March and concluded in July. Eight Angus-crossbred steers (shrunk body weight [SBW] of 365 ± 32 kg), pre-adapted to analysis within respiration calorimetry chambers, were randomly split into pairs and housed in four partially-covered outdoor pens with a Calan gate feed system (American Calan, Northwood, NH). The Ruminant Nutrition System (https://www.nutritionmodels.com/rns.html, accessed February 2019, Tedeschi and Fox, 2020) was used to formulate a grower-type base ration reflective of beef grower rations fed in Texas. This diet and its nutritional densities are listed in Table 1. The concentrate portion of the diet, consisting of ground corn, cottonseed meal, cottonseed hulls, dried molasses, and a mineral-vitamin supplement, was pre-mixed with a reel mixer and subsequently hand-mixed with chopped bermudagrass (Cynodon dactylon) hay (5- to 10-cm particles) before offered to the animal. Animals were transitioned from ad-libitum bermudagrass hay to the base ration over a 21-d period. They were then further adapted to the base ration for 7 d before the origination of the study. In order to minimize exponential growth for calorimetric analysis, animals were limit-fed a total DMI (kg/d) of 1.2% SBW split into two feedings at 0700 and 1700 each day with unrestricted access to water. This feed restriction was implemented due to the size of the animals, the capcity of the calorimetry chamber, the energy density of the ration, and the maximal flow rate of the calorimetric equipment (Lighton, 2008). In order to monitor animal welfare during data collection, each animal was given a rumen temperature bolus (smaXtec Animal Care GmbH, Graz, Austria) which recorded in vivo rumen temperature every 10 min and uploaded readings to the cloud simultaneously (Alzahal et al., 2011). Rectal temperature was not recorded during the data collection periods as indwelling rectal temperature probes, while accurate and sensitive, can provide variable measurements due to probe stability, fecal temperature, and animal movement (Koltes et al., 2018). This provided automated monitoring of rumen temperature and a more accurate estimation of heat stress events as rumen temperature is 2 °C higher than rectal temperatures during high environmental conditions (Koltes et al., 2018).

Table 1.

Diet composition and nutrient densities of the grower-type base ration

Items	Amount
Feed Composition
Bermudagrass Hay, %DM7	20
Ground Corn Grain, %DM	46
Cottonseed Meal, %DM	14
Cottonseed Hulls, %DM	10
Molasses, %DM	8
Mineral-Vitamin Mix1, %DM	2
Nutrient Density2
DM, % of feed	86.2
CP7, %DM	12.6
SCP7, %DM	2.4
ADICP7, %DM	1.2
NDICP7, %DM	1.4
ADF7, %DM	19.7
NDF7, %DM	31.9
Lignin, %DM	5.0
Soluble Sugar, %DM	6.6
Soluble Fiber, %DM	34.2
Crude Fat, %DM	3.1
Ash, %DM	6.53
TDN3, %DM	70
NEm4, Mcal/kg DM	1.61
NEg4, Mcal/kg DM	0.99
GE5, Mcal/kg	4.06
peNDF6, %DM	21.4

¹Custom Mineral-Vitamin Mix with a Ca:P ratio of 2:1.

²Nutrient densities were determined by Cumberland Valley Analytical Services (Waynesboro, PA).

³Calculated by Weiss et al. (1992).

⁴Calculated by NASEM (2016).

⁵GE measured by bomb calorimetry.

⁶peNDF measured by the Penn State Particle Separator method with three sieves (4, 8, and 19 mm; Lammers, 1996).

⁷DM, dry matter; CP, crude protein; SCP, soluble crude protein; ADICP, acid detergent insoluble crude protein; NDICP, neutral detergent insoluble crude protein; ADF, acid detergent fiber; NDF, neutral detergent fiber.

Experimental design

The effects and interactions of live yeast supplementation and environmental conditions were investigated using a duplicated 2 × 2 crossover design consisting of five 14-d periods. In order to create two dietary treatments (DIET), the base ration was fed with no live yeast supplementation (NOY) or top-dressed with 10 g live yeast supplement/d (LY; S. cerevisiae, Actisaf HR+, Phileo Lesaffre Animal Care, Milwaukee, WI). This dose of LY provided each animal with 1.2 × 10¹² CFU/d to optimize ruminal function (Cagle et al., 2019). Two indirect respiration calorimetry chambers were temperature controlled to create two environmental conditions (ENV). One chamber was designated as a thermoneutral environment (TN; 18.4 ± 1.1 °C, 57.6 ± 2.8% RH, THI = 63.5 ± 1.5), while the other was selected as a heat stress environment (HS; 33.8 ± 0.6 °C, 55.7 ± 2.7% relative humidity [RH], THI = 84.5 ± 1.5). Four treatment arrangements (TRT) between DIET and ENV (HS-LY, HS-NOY, TN-LY, and TN-NOY) were arranged in a sequence such that each carryover interaction (CARRY) from the previous treatment arrangement could be evaluated eight times (Figure 1). Each animal was randomly assigned to a treatment sequence, fed its assigned DIET for the entire period, and underwent its assigned ENV on days 13 and 14 of the period.

Figure 1.

Treatment sequence.

Calorimetry equipment and data collection

On day 13 of each period, a pair of animals were weighed to obtain SBW at 0600, prior to the first feeding and entered into their assigned calorimetry chamber for 48 h of data collection. Two side-by-side, single-stall open-circuit, indirect respiration calorimetry chambers, outfitted with a mass flow system (Flowkit model FK-500; Sable Systems Int., Henderson, NV), were utilized. Ambient air and air from each chamber were sampled by a multiplexer (Respirometry Multiplexer V 2.0, Sable Systems Int.), rotating every 2 min, measuring oxygen (O₂) consumption by difference and CO₂ and CH₄ production (FC-1B O₂ analyzer, CA-2A CO₂ analyzer, and MA-10 CH₄ analyzer, Sable Systems Int.). The SBW, dietary energy density, and dimensions of the calorimetry chamber were utilized to calculate the bank time and flow rate needed for data recording. The assumed gas concentrations of ambient air (O₂ = 20.95%, CO₂ = 0.04%, and CH₄ = 0.00%) were used to calibrate the O₂, CO₂, and CH₄ analyzers using known gasses, span gas (19.4% O₂, 1.1% CO₂, and 0.1% CH₄), and nitrogen (99.999% N₂; zeroing gas) before each animal entered a chamber for data collection. Water vapor was removed from the sampled gas before measurement using fresh desiccant (Hammond Drierite Co Ltd., Xenia, OH). The rate of O₂ consumption (VO₂) and CO₂ and CH₄ production (VCO₂ and VCH₄, respectively) were determined for 48 h of data collection (Lighton, 2008). Each chamber was equipped with a line voltage thermostat (Model# ETC-111000-000, Ranco Enterprises, Inc., Medford, MA), temperature and relative humidity logger (Model# UX100-003, Onset, Bourne, MA), dehumidifier (Model# DH7019KP1WG, Hisense USA, Suwanee, GA), analog water meter (Model# T10-DR-075-G-F, Neptune Technology Group, Inc, Tallassee, AL), and security cameras (Model# LBV1511W, Lorex, Inc., Markham, Canada). Additionally, each chamber was fitted with a metabolism stand to allow the total collection of fecal and urinary output over the 48 h of data collection. After the conclusion of data collection, animals were restrained in a squeeze chute and approximately 1 liter of rumen fluid was extracted via oral tubing and transported in a pre-warmed, air-tight thermos to the laboratory for further analysis. Animals were then returned to their outdoor pens to transition to their subsequent DIET in their assigned treatment sequence.

Water, feed, fecal, and urine analysis

Water intake for each data collection period was calculated by the analog water meter reading difference before and after an animal entered and exited a chamber. As rations for each data collection period were pre-allocated on day 13 before animals were entered into their assigned calorimetry chambers, a feed subsample (approximately 500 g) was collected at the beginning of each data collection period. Subsamples were composited, ground to pass through a 2-mm screen, and sent to Cumberland Valley Analytical Services (Waynesboro, PA) for chemical analysis of dry matter (DM; Goering and Van Soest, 1970), crude protein (Method #990.03; AOAC, 2019; Leco FP-528 Nitrogen Combustion Analyzer, Leco Corporation, St. Joseph, MO), soluble crude protein (Krishnamoorthy et al., 1982), acid detergent insoluble crude protein (Licitra et al., 1996), neutral detergent insoluble crude protein (Licitra et al.), acid detergent fiber (Method #973.18; AOAC, 2019), neutral detergent fiber (NDF; Van Soest et al., 1991), lignin (Goering and Van Soest, 1970), soluble sugars (Dubois et al., 1956), starch (Hall, 2009), fat (Method #2003.05; AOAC, 2019), ash (Method #942.05; AOAC, 2019), and mineral content (Method #985.01; AOAC, 2019; Perkin Elmer 5300 DV ICP, Perkin Elmer, Shelton, CT). Feed was also analyzed for DM, physically effective NDF (peNDF; Lammers et al., 1996), gross energy (GE), and nitrogen content (%N; Method #990.03; AOAC, 2019). As the animals were limit-fed, there were no orts to be collected.

Total fecal output from each data collection period was weighed, homogenized, sampled, dried at 55 °C for 48 h, and analyzed for DM (Goering and Van Soest, 1970), GE, and %N (Method #990.03; AOAC, 2019). Urine output was collected using two non-splatter filters to eliminate fecal contamination in a large transmission funnel leading to a catch tub connected to a vacuum system. The catch tub was linked to an external holding tub, and the vacuum system remained primed to eliminate gaseous escape from the chamber. Six hundred milliliters of 3 M HCl was added to each catch tub before animal entry into a chamber to prevent the volatilization of nitrogen. Urine was vacuumed as necessary to avoid the overflow of the catch tub. At the end of each data collection period, total urine output was weighed, homogenized, sampled, and analyzed for GE and %N (Method #990.03; AOAC, 2019); however, pH was not recorded.

Shrunk weight gain (SWG) of the animals was calculated as the difference of SBW between two periods divided by the length of the period (14 d). All estimations of GE were determined by bomb calorimeter (Parr adiabatic calorimeter, Parr Instruments Co., Moline, IL). Total N analysis was performed by Servi-Tech Laboratories (Amarillo, TX) using the Dumas method.

Energy partitioning and nitrogen balance

Gross energy intake (GEI, Mcal/d) was calculated by multiplying GE of the base ration by DMI of each animal. Fecal energy (FE) was calculated by multiplying GE of the feces by daily fecal output. Urinary energy (UE) was calculated by multiplying GE of the urine by the daily urinary production. Gaseous energy (GASE) was determined by multiplying the VCH₄ produced in the chamber by the density of CH₄ (0.6556 g/L at 25 °C) and its energy density (13.3 Mcal/g). Heat production energy (HE) was determined using the Brouwer (1965) equation:

$${\displaystyle \begin{array}{l}{\displaystyle HE=(3.866\times V{O}_2)+(1.2\times VC{O}_2)}\\ -(0.518\times VC{H}_4)-(1.431\times N_{urine})\end{array}}$$ (1)

Digestible energy intake (DEI) was calculated by the difference of GEI and FE. Metabolizable energy intake (MEI) was calculated by the difference between DEI and the sum of GASE and UE. Retained energy (RE) was calculated by the difference between MEI and HE. The respiratory quotient (RQ) was calculated by dividing VCO₂ by VO₂. Retained fat was calculated using the following equation as described by Tedeschi (2019) and Tedeschi and Fox (2020), assuming the caloric density of protein is 5.7 Mcal/kg and fat is 9.4 Mcal/kg (Blaxter and Rook, 1953; Garrett et al., 1959), and that all N_retained in the body is protein and all RE in the body is utilized to deposit protein or fat:

$${\displaystyle \begin{array}{l}{\displaystyle Retainedfat=\frac{RE-(5.7\times N_{retained})}{9.4}}\end{array}}$$ (2)

Total intake nitrogen (N_intake; g/d) was calculated as DMI, in g/d, multiplied by the %N of the base ration. Fecal nitrogen output (N_feces; g/d) was calculated by daily fecal output, in g/d, multiplied by the %N of the feces. Urinary nitrogen output (N_urine; g/d) was calculated by daily urinary output, in g/d, multiplied by the %N of the urine. Dietary retained nitrogen (N_retained; g/d) was calculated by the difference of N_intake, N_feces, and N_urine.

Rumen parameters

One liter of extracted rumen fluid was transported to the laboratory in a pre-warmed, air-tight thermos to maintain an anaerobic environment. Upon arrival, the temperature and redox of the rumen fluid were recorded with a platinum redox/temp electrode (Model# 9180BNMD, ThermoFisher Scientific, Waltham, MA). Rumen fluid was then filtered through 8 μm porosity fiberglass wool with a steady CO₂ flush and transferred to a 150-mL Wheaton bottle which was flushed with CO₂, stoppered with a butyl rubber stopper, and sealed with aluminum crimp. The bottle was then transferred to an in vitro gas production (IVGP) incubation chamber to warm to 39 °C before undergoing the IVGP technique described below.

The remainder of the filtered rumen fluid was tested for pH and underwent the following chemical analyses. Eight milliliters of filtered rumen fluid were combined with 2 mL of 25% meta-phosphoric acid and preserved for volatile fatty acid (VFA) analysis at −20 °C. Acetate, propionate, and butyrate concentrations were determined via gas chromatography following procedures described by Erwin et al. (1961) using an Agilent 6890N (Agilent Technologies, Santa Clara, CA) equipped with a capillary column (HP-FFAP, 10 m × 530 μm × 1.00 μm film thickness, Agilent Technologies). Helium was used as the carrier gas at a flow rate of 27 cm/s. Two milliliters of filtered rumen fluid was acidified with 8 mL of 0.1 N HCl and preserved at −20 °C for ammonia concentration (NH₃-N) determination following colorimetric methods described by Marbach and Chaney (1962). One milliliter of filtered rumen fluid was preserved with 10 mL of absolute ethanol for ­protozoa enumeration following Dehority (1984) methods. Using a Sedgewick Rafter counting chamber, a 1 mL aliquot of the preserved sample was counted at 100× magnification with a 0.5 mm square counting grid. Twenty-five evenly spaced grids from the entire chamber surface were measured using a Nikon Eclipse E200 microscope (Nikon Corporation, Tokyo, Japan).

Enumeration of Fusobacterium necrophorum

Enumeration techniques of Fusobacterium necrophorum in rumen fluid followed procedures outlined by Tan et al. (1994). Each sampled rumen fluid was anaerobically diluted to 10^-11 with sterile water in an anaerobic chamber (Type B Vinyl Anaerobic Chamber, Coy Laboratory Products, Inc., Grass Lake, MI) at 39 °C. Twenty microliters of each dilution was added to 200 μL of previously prepared modified lactate medium such that 12 wells were inoculated per dilution. Plates were anaerobically incubated for 48 h at 39 °C. At the end of incubation, 20 μL of Kovac’s reagent was added to each well. A positive result for indole production indicated positive growth of F. necrophorum. Enumeration of F. necrophorum was completed in quadruplicate using the 3-well Most Probable Number method (Blodgett, 1998).

In vitro gas production

In order to simulate the effects of TRT on subsequent ruminal performance when an animal is returned to TN-NOY conditions, the IVGP technique was utilized. The experimental design was a two-factor randomized complete block design with rumen fluid from animals previously exposed to one of the four TRT described above. One factor was ENV (HS or TN) while the other was DIET (NOY or LY). As the overarching study spanned five periods, there were five periods for this IVGP experiment where each TRT was represented ten times. The IVGP technique has been previously described by Pell and Schofield (1993) and Tedeschi et al. (2009). Approximately 200 mg of the base ration (ground to 2 mm) was weighed into ten 150-mL Wheaton bottles containing equal-sized stir bars. Two additional bottles were prepared per animal: one with no fermentable substrate as a blank for baseline measurements and one with 200 mg of dried alfalfa (ground to 2-mm) as a positive control. To simulate NOY conditions, no live yeast was added to the bottles and to simulate TN conditions, and the incubation occurred at normal rumen temperature (39 °C). Two milliliters of deionized water were added to each bottle to wet the particles in order to prevent scattering during the CO₂ flush. Fourteen milliliters of IVGP buffering media (Goering and Van Soest, 1970) were added to each bottle under constant CO₂ flush. Each bottle was then stoppered with a butyl rubber stopper and sealed with aluminum crimp. Bottles were then placed in an incubation chamber to warm the bottle contents to 39 °C. Four milliliters of filtered and warmed rumen fluid were injected into each bottle, excess gaseous pressure was released, and a pressure sensor was connected to each bottle. Bottles were incubated for 48 h at rumen temperature with gentle stirring to simulate rumen motility. After incubation, bottles were placed on ice for 30 min to stop fermentation. Twelve milliliters of homogenized headspace gas were removed from each bottle with a gas-tight syringe (Trajan Scientific and Medical, Melbourne, Australia) and placed in an evacuated exetainer (Labco, Ceredigion, Wales, UK) for gas chromatography determination of CO₂ and CH₄ concentration (Smith et al., 2020). Bottles were then un-crimped and de-stoppered and redox and pH of the fluid were recorded. Forty milliliters of NDF solution (ANKOM Technology, Macedon, NY) were placed in each bottle. Each bottle was then re-stoppered and re-crimped, autoclaved for 15 min at 121 °C, and filtered through Whatman 54 filter paper to determine in vitro NDF digestibility (ivNDFD).

The kinetic analysis of 48 h of IVGP fermentation was plotted using nonlinear functions with the lowest sum of squares error (Schofield et al., 1994). Fitting and selecting of nonlinear functions were completed using GasFit (http://www.nutritionmodels.com/gas-fit.html), which executes R scripts to perform convergence of gas production data using the nls function (Bates and Chambers, 1992) and the port algorithm (Fox et al., 1978; Gay, 1990). The GasFit estimated total gas production (TGP, mL), fractional degradation rate (k_d, %/h), and lag time in h using an exponential model. A logarithmic 2-pool nonlinear function (Tedeschi and Fox, 2020) was used to estimate the TGP (P1-TGP) and k_d (P1-k_d) of the non-fiber carbohydrate fermentation pool, the lag time between non-fiber and fiber carbohydrate fermentation (P1>P2), the TGP (P2-TGP) and k_d (P2-k_d) of the fiber carbohydrate fermentation pool. GasFit was also used to estimate the apparent ME of the diet assuming passage rates (k_p) of 4, 6, and 8%/h using equations described by Tedeschi and Fox (2020). Using the TGP from the exponential model with the blank-adjusted molar concentration of CH₄, CH₄ g/L gas, and CH₄ g/kg DM were calculated.

Statistical analysis

Calorimetry, water intake, feed, fecal, and urine metabolism, energy partitioning and nitrogen balance, and VFA and NH₃-N concentrations were assessed for normality using PROC UNIVARIATE and homogeneity of variances using the Brown-Forsythe test of PROC GLM and analyzed using PROC MIXED of SAS (SAS Institute, Inc., Cary, NC) where DIET, ENV, their interaction, and CARRY were the fixed effects and the random error of period and animal were accounted for in the model. The enumerations of protozoa and F. necrophorum were log-transformed and analyzed using the same model above. In vitro data was analyzed using a similar mixed model where DIET, ENV, and their interaction were the fixed effects. The random errors of bottle, animal, and period were accounted for in the model. For all statistical models, pairwise comparison was used to identify statistical differences among ENV, DIET, and TRT. Statistical significance was determined at P ≤ 0.05, while tendencies were determined at P ≤ 0.10. Carryover interactions that were not significant (P > 0.05) were removed from the model.

Results

Water, feed, and nitrogen metabolism

Water, feed, and nitrogen metabolism data are presented in Table 2. There were no carryover interactions (P > 0.05). As all animals underwent each TRT, there were no differences in SBW or SWG for ENV, DIET, or TRT (P ≥ 0.341). There were no effects of DIET or TRT on daily water intake, water intake as a percentage of SBW, and water intake to DMI ratio (P ≥ 0.278). However, during HS, daily water intake (32.66 L/d vs. 17.98 L/d), water intake as a percentage of SBW (8.65%SBW vs. 4.70%SBW), and water intake to DMI ratio (7.21 vs. 3.91) were increased over TN (P = 0.003). Urine output followed a similar pattern with an increase during HS (7.95 kg/d, 2.01%SBW) as compared to TN (5.45 kg/d, 1.42%SBW; P = 0.003) and no effects of DIET or TRT (P ≥ 0.509). As DMI was limited to 1.2%SBW and no orts were observed, there were no effects of ENV, DIET, or TRT on DMI (P ≥ 0.341). Congruently, there were no ENV, DIET, or TRT effects on fecal DM output (P ≥ 0.180). There were no effects of ENV, DIET, or TRT on dry matter digestibility (DMD) (P ≥ 0.181). The lack of effects of ENV, DIET, TRT, or CARRY on DMI, fecal DM output, and urine output were mirrored in N_intake, N_feces, and N_urine (P ≥ 0.252). Of N_intake, approximately 35% was excreted in feces, 21% was excreted in urine, and 44% was retained (approximately 38 g N/d). There were no effects of ENV, DIET, or TRT on N_retained (P ≥ 0.180).

Table 2.

Effect of environmental condition and LY supplementation on water consumption and feed and nitrogen metabolism of steers fed a grower-type diet

	HS1		TN1			P-value
Items	NOY1	LY1	NOY	LY	SEM	ENV2	DIET2	TRT2,3
SBW1, kg	377	380	378	378	12.3	0.769	0.411	0.341
SWG1, kg/d	0.35	0.46	0.40	0.27	0.15	0.660	0.945	0.429
Water intake, kg/d	29.39	35.93	16.27	19.68	7.26	0.003	0.278	0.731
Water intake, % SBW	7.79	9.51	4.29	5.10	1.94	0.003	0.298	0.707
Water intake:DMI1	6.49	7.92	3.57	4.25	1.61	0.003	0.298	0.707
DMI, kg/d	4.53	4.57	4.54	4.54	0.15	0.769	0.411	0.341
Fecal DM1, kg/d	1.86	1.57	1.79	1.82	0.16	0.362	0.221	0.133
Fecal DM, %SBW	0.49	0.41	0.47	0.48	0.38	0.432	0.180	0.138
DMD1, %	58.78	65.53	60.51	60.13	31.64	0.435	0.181	0.136
Urine, kg/d	7.54	7.63	5.83	5.07	0.95	0.003	0.600	0.509
Urine, %SBW	2.01	2.01	1.53	1.32	0.25	0.003	0.556	0.534
Nintake, g/d	84.70	85.39	84.95	84.80	3.46	0.745	0.612	0.424
Nfeces, g/d	30.14	27.27	30.51	30.61	2.42	0.252	0.390	0.356
Nfeces, %Nintake	36.23	32.15	35.96	36.12	31.24	0.373	0.346	0.308
Nurine, g/d	17.76	17.65	15.24	19.56	2.31	0.894	0.356	0.332
Nurine, %Nintake	21.01	20.63	18.17	22.88	25.05	0.902	0.365	0.289
Nretained, g/d	37.10	40.02	38.82	35.78	4.65	0.653	0.982	0.287
Nretained, %Nintake	42.78	47.55	45.52	41.66	4.24	0.620	0.885	0.180

¹HS, heat stress; TN, thermoneutral; NOY, no live yeast supplementation; LY, live yeast supplementation; SBW, shrunk body weight; SWG, shrunk weight gain; DMI, dry matter intake; DM, dry matter; DMD, dry matter digestibility.

²ENV, environmental condition (HS or TN); DIET, diet (NOY or LY); TRT, treatment arrangement (HS-LY, HS-NOY, TN-LY, and TN-NOY).

³CARRY was removed from the model if P > 0.05.

Energy partitioning and gas flux

Energy partitioning and gas flux data are presented in Table 3. There were no carryover interactions (P > 0.05) except for VO₂ (P = 0.031). Gross energy intake was the same across ENV, DIET, and TRT (18.5 Mcal/d; P ≥ 0.298). There were no effects of ENV, DIET, or TRT on FE (6.95 Mcal/d; P ≥ 0.205) which resulted in no differences for DEI (11.5 Mcal/d; P ≥ 0.152) and DE of the diet (2.53 Mcal/kg DM; P ≥ 0.202). Approximately 62% of GEI was partitioned for DEI with no differences among ENV, DIET, or TRT (P ≥ 0.202). There were no effects of ENV, DIET, or TRT on UE (0.25 Mcal/d; P ≥ 0.438), with approximately 1.4% of GEI excreted as UE, regardless of ENV, DIET, or TRT (P ≥ 0.378). Although there were no effects of DIET or TRT on GASE (1.80 Mcal/d P ≥ 0.154), approximately 9.8% GEI was lost as GASE. GASE from TN (10.0% GEI) tended to be higher than GASE from HS (9.5% GEI). There were no effects of ENV, DIET, or TRT on MEI (9.46 Mcal/d; P ≥ 0.210) with approximately 51% GEI partitioned for MEI. The ME of the diet was estimated to be 2.08 Mcal/kg DM with no effects from ENV, DIET, or TRT (P ≥ 0.224). There were no effects of ENV, DIET, or TRT on HE (9.37 Mcal/d; P ≥ 0.372), RE (0.09 Mcal/d; P ≥ 0.240), RQ (1.07; P ≥ 0.267), or retained fat (−13.4 g/d; 0.247). Approximately 51% GEI was dissipated as HE and 0.50% of GEI converted to RE.

Table 3.

Effect of environmental condition and live yeast supplementation on energy partitioning and gas flux of steers fed a grower-type diet using indirect calorimetry

	HS1		TN1			P-value
Items	NOY1	LY1	NOY	LY	SEM	ENV2	DIET2	TRT2	CARRY2,3
Energy Partitioning
GEI1, Mcal/d	18.37	18.55	18.45	18.42	0.61	0.804	0.439	0.298	-
FE1, Mcal/d	7.19	6.30	7.09	7.22	0.65	0.306	0.342	0.205	-
DEI1, Mcal/d	11.20	12.23	11.37	11.19	0.70	0.304	0.316	0.152	-
DEI, %GEI	60.57	65.95	61.51	61.07	3.27	0.383	0.276	0.202	-
DE1, Mcal/kg DM	2.46	2.68	2.50	2.48	0.14	0.385	0.277	0.202	-
UE1, Mcal/d	0.27	0.24	0.25	0.25	0.03	0.795	0.438	0.645	-
UE, %GEI	1.47	1.30	1.37	1.33	0.17	0.752	0.378	0.580	-
GASE, Mcal/d	1.75	1.77	1.81	1.86	0.10	0.154	0.543	0.753	-
GASE, %GEI	9.56	9.55	9.86	10.13	0.50	0.097	0.624	0.599	-
MEI1, Mcal/d	9.17	10.12	9.33	9.21	0.71	0.377	0.335	0.210	-
MEI, %GEI	49.56	54.88	50.40	50.04	3.51	0.399	0.295	0.229	-
ME1, Mcal/kg DM	2.01	2.23	2.05	2.03	0.15	0.401	0.294	0.224	-
ME:DE ratio	0.81	0.83	0.81	0.81	0.01	0.228	0.296	0.167	-
HE1, Mcal/d	9.26	9.36	9.38	9.49	0.36	0.372	0.429	0.955	-
HE, %GEI	50.35	50.42	50.86	51.28	1.00	0.357	0.737	0.813	-
RE1, Mcal/d	−0.09	0.80	−0.07	−0.29	0.70	0.259	0.473	0.240	-
RE, %GEI	−0.83	4.33	−0.38	−1.27	3.77	0.327	0.414	0.246	-
RQ4	1.07	1.07	1.05	1.10	0.05	0.731	0.267	0.383	-
Retained fat5, g	−31.75	61.32	−31.24	−52.04	72.97	0.256	0.462	0.247	-
Gas Flux
VO2, L/d	1,835.57a,c	1,906.79a,b	1,971.49a,b	1,703.87c	82.36	0.520	0.068	0.003	0.031
VCO2, L/d	1,947.28	1,975.42	1,953.09	2,014.16	95.77	0.584	0.277	0.685	-
VCH4, L/d	201.04	202.73	207.56	212.86	10.96	0.154	0.543	0.753	-
CH4 production, g/kg DM	29.20	29.17	30.16	30.95	1.54	0.096	0.637	0.613	-

¹HS, heat stress; TN, thermoneutral; NOY, no live yeast supplementation; LY, live yeast supplementation; GEI, gross energy intake; FE, fecal energy; DEI, digestible energy intake; DE, digestible energy; UE, urinary energy; MEI, metabolizable energy intake; ME, metabolizable energy; HE, energy from heat production; RE, retained energy.

²ENV, environmental condition (HS or TN); DIET, diet (NOY or LY); TRT, treatment arrangement (HS-LY, HS-NOY, TN-LY, and TN-NOY); CARRY, carryover interaction

³CARRY was removed from the model if P > 0.05.

⁴RQ = respiratory quotient (CO₂/O₂)

⁵ $Retainedfat=\frac{RE-(5.7\times N_{retained})}{9.4}$ (Blaxter and Rook, 1953; Garrett et al., 1959; Tedeschi, 2019; Tedeschi and Fox, 2020).

^a–cLeast square means within a row with different superscripts differ at P ≤ 0.05.

There was a carryover interaction associated with VO₂ (P = 0.031), which was mainly due to a carryover effect of TN-NOY to HS-LY and of HS-LY to HS-NOY. There was an effect of TRT (P = 0.003), such that HS-LY had a greater VO₂ than TN-LY and TN-LY had a lower VO₂ than TN-NOY. There was a tendency for NOY (1,903 L/d) to have a greater VO₂ than LY (1,805 L/d; P = 0.068). There were no effects of ENV, DIET, or TRT on VCO₂ (1,972 L/d; P ≥ 0.277) and VCH₄ (206 L/d; P ≥ 0.154). There were no effects of DIET or TEMP on g CH₄ produced per kg DM (P ≥ 0.613). However, there was a tendency for animals in TN to produce more g CH₄/kg DM than animals in HS (30.5 vs. 29.2 g/kg DM, respectively; P = 0.096).

Ruminal parameters

Ruminal parameters are presented in Table 4. There were no carryover interactions (P > 0.05). There were no effects of ENV, DIET, or TRT on pH and redox (6.71 and −232.8 mV, respectively; P ≥ 0.157). There was a tendency of TRT to have an effect on acetate production (P = 0.076), where TN-LY tended to have increased acetate production over HS-NOY, HS-LY, and TN-NOY. This same tendency is ­present in TRT effects on propionate production (P = 0.074), with TN-LY ­having greater production than HS-NOY, HS-LY, and TN-NOY. Yeast supplementation increased propionate production over NOY (12.88 vs. 11.32 mM; P = 0.034) and animals in TN conditions had more propionate production than animals in HS (12.94 vs. 11.25 mM; P = 0.023). There were no effects of ENV or TRT on butyrate production (P ≥ 0.370). However, animals receiving LY had a tendency to have greater butyrate production than NOY (13.66 vs. 12.08 mM; P = 0.090). The tendency of TRT on total VFA (P = 0.065) can mainly be attributed to the tendency of TRT on acetate and propionate production: TN-LY had greater total VFA than HS-NOY, HS-LY, and TN-NOY. Although there was no effect of DIET on total VFA production (P = 0.111), animals in TN conditions tended to produce more total VFA than animals in HS conditions (99.34 vs. 90.94 mM; P = 0.087). There were no effects of ENV, DIET, or TRT on acetate:propionate ratio (P ≥ 0.190) or NH₃-N (P ≥ 0.490). There were no effects of ENV or TRT on CFU of F. necrophorum (P ≥ 0.508). However, animals receiving LY had a tendency to have about 20 times the number of CFU when compared with animals receiving NOY (1.75 × 10⁷ vs. 8.95 × 10⁵ CFU/mL; P = 0.053). There were no effects of ENV, DIET, or TRT on the enumeration of protozoa (P ≥ 0.712).

Table 4.

Effect of environmental condition and live yeast supplementation on ruminal parameters of steers fed a grower-type diet

	HS1		TN1			P-value
Items	NOY1	LY1	NOY	LY	SEM	ENV2	DIET2	TRT2,3
pH	6.75	6.65	6.70	6.75	0.06	0.700	0.659	0.157
Redox, mV	-262.25	-221.10	-224.25	-223.64	32.97	0.817	0.358	0.758
Acetate, mM	62.78	60.26	61.84	72.80	5.48	0.123	0.257	0.076
Propionate, mM	11.12	11.38	11.51	14.37	1.05	0.023	0.034	0.074
Butyrate, mM	11.81	13.10	12.34	14.22	1.20	0.370	0.090	0.744
Total VFA1, mM	91.58	90.29	90.90	107.77	7.33	0.087	0.111	0.065
Acetate:propionate ratio	5.65	5.43	5.45	5.12	0.27	0.218	0.190	0.786
NH3-N, mg/dL	13.28	12.62	12.21	13.30	1.66	0.875	0.867	0.490
F. necrophorum, CFU × 106/mL	1.10	8.29	0.65	36.98	0.004	0.746	0.053	0.508
Protozoa, organisms × 105/mL	1.33	1.33	1.30	1.38	.378	0.958	0.712	0.718

¹HS, heat stress; TN, thermoneutral; NOY, no live yeast supplementation; LY, live yeast supplementation; VFA, volatile fatty acid.

²ENV, environmental condition (HS or TN); DIET, diet (NOY or LY); TRT, treatment arrangement (HS-LY, HS-NOY, TN-LY, and TN-NOY).

³CARRY was removed from the model if P > 0.05.

In vitro gas production

The results of the in vitro fermentation study are detailed in Table 5. The initial pH and redox (6.71 and −232.8 mV, respectively) of the rumen inoculum were not affected by ENV, DIET, or TRT (P ≥ 0.157). There was an effect of TRT (P = 0.046) on final pH, with HS-LY finishing with a greater final pH than TN-LY (P = 0.032). There were no effects of ENV or DIET on final pH (P ≥ 0.296), but there was an effect of TRT (P < 0.001) on final redox, where TN-LY had a greater redox than HS-NOY, HS-LY, and TN-NOY. Animals receiving LY had a greater redox than animals receiving NOY (−301.33 vs. −315.10 mV; P = 0.007) and animals in HS conditions had a lower redox than animals in TN conditions (−313.47 vs. −302.96 mV; P = 0.004). There were no differences in TGP in the exponential model due to ENV, DIET, or TRT (P ≥ 0.101). TRT had an effect (P = 0.013) on the exponential model’s k_d where TN-LY had a greater k_d than HS-LY and TN-NOY. There were no effects of ENV or DIET on k_d (P ≥ 0.451). There was a TRT effect (P < 0.001) on exponential lag time, where TN-NOY had the shortest lag time when compared with HS-NOY and TN-LY. Additionally, HS-NOY had a longer lag time than HS-LY. Part of these differing lag times can be attributed to the tendency (P = 0.094) of HS to have a shorter lag time than TN (−0.81 vs. −1.28 h).

Table 5.

In vitro performance of rumen fluid, with previous TRT exposure, when returned to TN-NOY

	HS1		TN1			P-values
Item	NOY1	Y1	NOY	Y	SEM	ENV2	DIET2	TRT2
Initial pH	6.75	6.65	6.70	6.75	0.06	0.700	0.659	0.157
Final pH3	6.40a,b	6.48a	6.44a,b	6.37b	0.08	0.296	0.881	0.046
Initial redox, mV	−262.25	−221.10	−224.25	−223.64	32.97	0.817	0.358	0.758
Final redox3, mV	−310.15a	−316.80a	−320.06a	−285.86b	17.70	0.004	0.007	<0.001
Exponential model
TGP1, mL	11.26	11.72	10.20	10.73	2.14	0.101	0.430	0.963
kd, %/h	28.21a,b	20.40a	21.27a	33.32b	6.14	0.451	0.593	0.013
lag time, h	−0.13a	−1.49b,c	−1.83b	−0.73a,c	0.44	0.094	0.634	<0.001
Log, two-pool model
P1-TGP, mL	4.81	6.12	5.21	5.03	1.12	0.474	0.247	0.127
P1- kd, %/h	56.90	44.82	61.71	63.45	11.14	0.161	0.534	0.408
P1>P2, h	0.82a	0.72a	0.42b	0.83a	0.25	0.149	0.135	0.011
P2-TGP, mL	6.81a	5.72b	5.37b	5.93a,b	1.13	0.081	0.450	0.022
P2-kd, %/h	7.25	5.95	5.73	6.63	1.08	0.664	0.837	0.253
ivNDFD, %	35.86a	50.26b	40.95a,c	45.35b,c	4.55	0.967	<0.001	0.022
ME4, Mcal/kg DM
kp 4%/h	2.51a	2.46b	2.47b,c	2.50a,c	0.02	0.732	0.236	<0.001
kp 6%/h	2.47a	2.41b	2.42b,c	2.45a,c	0.02	0.751	0.242	<0.001
kp 8%/h	2.44a	2.38b	2.39b,c	2.42a,c	0.02	0.779	0.260	<0.001
CH4, g/L gas	0.21	0.21	0.18	0.19	0.02	<0.001	0.398	0.757
CH4, g/kg DM1	11.80a	10.36a,b	9.36b	11.01a,b	2.64	0.209	0.888	0.028

¹HS, heat stress; TN, thermoneutral; NOY, no live yeast supplementation; LY, live yeast supplementation; TGP, total gas production; DM, dry matter.

²ENV, environmental condition (HS or TN); DIET, diet (NOY or LY); TRT, treatment arrangement (HS-LY, HS-NOY, TN-LY, and TN-NOY).

³Final pH and redox were measured after 48 h of fermentation.

⁴Estimated apparent ME of the ration assuming k_p of 4, 6, and 8%/h.

^a–cLeast square means within a row with different superscripts differ at P ≤ 0.05.

In the log two-pool model, there were no differences due to ENV, DIET, or TRT on P1-TGP (P ≥ 0.127) or P1-k_d (P ≥ 0.161). There was a TRT effect (P = 0.011) on P1>P2 such that animals assigned TN-NOY experienced the shortest lag time when compared with HS-NOY, HS-LY, and TN-LY. There were no effects of ENV or DIET on P1>P2 (P ≥ 0.138). There was a tendency (P = 0.081) of previously HS rumen inoculum to produce more P2-TGP than TN (6.27 vs. 5.65 mL). This is represented in the effect of TRT on P2-TGP (P = 0.022), where HS-NOY had a greater P2-TGP than HS-LY and TN-NOY. Interestingly, there were no effects of ENV, DIET, or TRT on P2-k_d (P ≥ 0.253).

In vitro NDFD experienced an effect from TRT (P = 0.022), where HS-NOY had the lowest ivNDFD when compared with HS-LY and TN-LY. Rumen fluid from TN-NOY had less ivNDFD than HS-LY. Rumen fluid with previous LY had a greater ivNDFD than rumen fluid with previous NOY (47.80% vs. 38.40%; P < 0.001). Regardless of k_p, the calculated ME density of the diet experienced an effect from TRT (P < 0.001) but not from ENV or DIET (P ≥ 0.242). There was an effect of ENV on CH₄ g/L gas, with HS (0.21 g/L gas) producing more CH₄ than TN (0.19 g/L gas; P < 0.001), but there was no effect of DIET or TRT on CH₄ g/L gas (P ≥ 0.398). There was an effect of TRT (P = 0.028) on CH₄ g/kg DM such that HS-NOY had a greater CH₄ g/kg DM than TN-NOY, but there was no effect of ENV or DIET on CH₄ g/kg DM (P ≥ 0.209).

Discussion

This study utilized a 2 × 2 crossover design to evaluate the effects of ENV, DIET, and TRT, and their corresponding CARRY, on many factors. There was a 12-d DIET-adaptation period between data collection periods. With a lack of discernable CARRY effects, except for VO₂, this adaptation period was appropriate for this trial. This is supported by Machado et al. (2016), who found that an 11-d adaptation period was the longest time required to realize changes in ruminal fermentation characteristics of animals fed tropical forage diets, after which the rumen became adapted to the new diet. As k_d and k_p are much higher for rations with increased amounts of concentrate, like the base ration in this study, the 12 d used in this experiment was enough time to adapt the rumen microbiome given the ruminal turnover rate (Tedeschi and Fox, 2020). One of the setbacks of this study was the maturity of the animals. At about 3 yr old, the animals had an adequate adaptation to HS environments, such as the one in College Station, TX (Tedeschi and Fox, 2020). This may be one of the leading causes of the lack of effects on metabolism as elaborated below.

Water intake increased for animals during HS, which is consistent with ruminant behavior during greater THI (Kadzere et al., 2002; Mader and Davis, 2004). Water intake did not increase for animals receiving LY and there were no interactions of TRT on water intake. This is consistent with findings by McGinn et al. (2004), Zerby et al. (2011), and Crossland et al. (2018). As expected, an increased water intake during HS resulted in increased urine output. The amount of water intake, as %SBW, did not exceed the amount of urine output, which is expected even during HS conditions (McDowell et al., 1969; Collier et al., 2006). Because the base ration was limit-fed to the animals, with no orts observed, there were no effects present for DMI. Interestingly, there was also no effect on fecal output. This contrasts with previous HS observations as high ambient conditions are known to slow the digestive tract, reducing fecal output (Beede and Collier, 1986; Miaron and Christopherson, 1992). Contrasting to previous research, DMD was not affected by LY (Desnoyers et al., 2009). This could be partially attributed to the limited DMI. However, there was a trend toward increasing it between HS-NOY and HS-LY. This could be attributed to the consistency of DMD from the yeast product used in this study (Cagle et al., 2019). With no effects measured on feed intake or fecal and urine output, no effects were found on N in the feed or in fecal and urinary excretions. The animals in this study are estimated to have retained about 44% of the intake N, which is consistent with values presented by Crossland et al. (2018). The values of nitrogen metabolism in this study fall in the lower range of values Dong et al. (2014) utilized to predict nitrogen metabolism in beef cattle but remain within one standard deviation of the mean. Although nitrogen emissions were not measured from the manure or urine, a low excretion percentage of nitrogen in either could lead to lesser nitrogen volatilization and emissions (Klopfenstein and Erickson, 2002).

Due to a limited DMI, there were no differences in GEI offered to each animal. Of the 62% of GEI retained as DEI, approximately 82% DEI was retained as MEI. This is consistent with the ME:DE ratio presented by NASEM (2016). There were no effects of ENV, DIET, or TRT on FE and UE. There was a tendency for the conversion of gross energy to GASE to be greater for animals in TN when compared with HS. This is reflected in the production of CH₄/kg DM suggesting HS slowed fermentation. The estimates for the conversion of GE to GASE are over the recommended conversion rate from the Intergovernmental Panel on Climate Change (3.0 ± 1.0% GEI for feedlot cattle fed a 90% concentrate diet; IPCC, 2006). The diet in this study has 10% less concentrates than the baseline diet set forth by IPCC and the guidelines give a disclaimer that the conversion rate of GE to GASE may be elevated for diets with more fibrous portions. When comparing the GASE to GE conversion rate to the example given by IPCC for more fibrous rations, the GASE to GE conversion rate in this study is in line with appropriate ranges for restricted diets (8% to 11%).

The RQ of the animals, regardless of ENV, DIET, or TRT, was about 1.07, which suggests complete oxidation of absorbed carbohydrates for energy consumption (Ferrannini, 1988). The measured HE and calculated RE of this study are not consistent with expected outcomes for animals in heat stress; HS animals should have greater HE and lesser RE when compared with TN animals due to energy expenditure for thermoregulation and altered tissue metabolism (Crossland et al., 2018). The values in this study suggest that the animals were likely acclimatized to the simulated hot environment. With the energy density of the diet meeting maintenance energy requirements, a near-zero value for RE, a negative value for retained fat, and positive values for retained nitrogen and shrunk weight gain, animals in this study might have been using digested carbohydrates for maintenance in addition to utilizing body fat to support protein deposition. It is known that ruminants, with minimal or no exogenous energy consumption, fall into this contradictory energy versus protein balance, where endogenous body fat is utilized as energy for protein accretion (Meehan et al., 2021). This conundrum has been presented in comparative slaughter studies and, although rarely studied and discussed, it could explain the values found in this study (Fattet et al., 1984; Chowdhury et al., 1990; Drouillard et al., 1991). Because RE was near zero, the partial efficiency of the utilization of maintenance energy for gain was not calculated in this study, as any estimation would incorrectly describe the energy balance of these animals. It is important to note that the indirect calorimetry method yields greater RE than the comparative slaughter technique, suggesting that the RQ and RE of this study might be inflated and the HE might be underestimated (Johnson et al., 1997).

The oxygen consumption rate was the primary variable to indicate a significant CARRY effect. This is mainly attributed to a CARRY effect from LY such that animals in HS-LY were at a deficit when coming from the TN-NOY treatment and excess when going into the HS-NOY treatment. It is well known that LY can alter rumen digestive patterns by stabilizing rumen pH and providing a favorable environment for cellulolytic bacteria and that HS can inhibit fiber digestion (Baek et al., 2020; Burdick Sanchez et al., 2021). It is plausible that LY supplementation during HS altered ruminal cellulolytic bacteria profiles so considerably that a carryover effect could be quantified in this study. Direct assessment of the microbiome was not completed in concurrence with this study and could be a point of further investigation to understand this CARRY. With the CARRY effect included in the statistical model, there was a main effect of TRT such that HS-LY animals had a greater VO₂ than animals in TN-LY. This could be attributed to the increased panting rate during HS events (Bernabucci, 2019; Lacetera, 2019). There was also an interesting interaction between TN-NOY and TN-LY, such that animals in TN-NOY consumed more oxygen than their LY counterparts. This is mirrored in the direct effect of DIET on VO₂. As live yeast supplementation is known to improve digestion and VFA production in the rumen, perhaps the absence of supplementation resulted in a change of metabolism for NOY animals. With less VFA supporting metabolism, NOY animals turned to white adipose tissue lipolysis, a high oxygen-consumptive process, for energy to continue protein accretion (Cruz et al., 2018). From a gross energy perspective, these animals were in maintenance, but from a molecular perspective, LY animals’ maintenance was supported by ruminal fermentation products, and NOY animals’ maintenance was supported by endogenous lipolysis. Further investigation into empty body characteristics is required to understand this relationship. There were no effects of ENV, DIET, or TRT on VCO₂ or VCH₄. The CH₄ emissions in this study are below the average emissions of beef cattle (250 L/d; Johnson and Johnson, 1995). Methane production per kilogram DMI was increased for animals in TN conditions as compared to HS conditions. This is consistent with the literature; as ambient temperature increases, digestion slows, causing a decrease in CH₄ production (Ngwabie et al., 2011; Barnett et al., 2015; Yadav et al., 2016).

There were no effects of ENV, DIET, or TRT on ruminal pH or redox. This is not consistent with previous literature. Live yeast supplementation is known to increase ruminal pH and decrease redox in order to provide a more favorable fermentation environment for rumen microorganisms (Cagle et al., 2019). Additionally, HS conditions, with a stall of rumen motility, result in a lower pH as less acid escapes the rumen and buffering capability is decreased due to reduced intake and rumination (Das et al., 2016; Zhao et al., 2019). These differences may be explained by the propionate concentrations in the rumen fluid. Animals receiving LY had greater propionate concentrations, suggesting that there may have been an increase of propionate-producing bacteria (Desnoyers et al., 2009; Humer et al., 2018). It is also important to note that rumen fluid was extracted more than 12 h after the final feeding. This is an adequate amount of time for rumen functionality to return to normal conditions. Animals in TN conditions had greater propionate production and tended to have greater total VFA than animals in HS conditions. This can be attributed to more favorable ruminal conditions for fermentation. The increased propionate production during TN and LY resulted in a tendency where TN-LY rumen fluid tended to have the greatest acetate, propionate, and total VFA concentrations compared to all the other TRT. As LY supplementation is known to increase fiber digestion in the rumen, there was a tendency for LY to increase butyrate production, a VFA mainly attributed to fiber digestion. This relationship between LY and butyrate has been described once before by Křížová et al. (2011), but in general, LY has an inverse effect on acetate and butyrate production (Armato et al., 2017). There were no effects of ENV, DIET, or TRT on the acetate-to-propionate ratio, NH₃-N production, and protozoa enumeration. The concentration of F. necrophorum in rumen fluid tended to increase with LY. F. necrophorum is a lactate-fermenting bacteria and the lactate produced from LY’s anaerobic fermentation of starch provides a perfect environment for the proliferation of F. necrophorum. As this enumeration came from rumen fluid and it is known that F. necrophorum in rumen contents is capable of causing liver abscesses in beef cattle, further investigation is required to understand if LY has the capability of increasing these numbers, and furthermore, increasing liver abscessation in cattle (Kanoe et al., 1978; Amachawadi and Nagaraja, 2016).

The IVGP portion of this study sought to model the carryover effects of ENV, DIET, and TRT on the performance of rumen fluid in vitro. Usually, beef animals travel from TN conditions to HS conditions as they grow through the production system. However, animals originating from ranches in sub-tropical and tropical climates to grow- and feedyards in temperate climates may undergo this paradoxical weather pattern. Therefore, it is important to understand the effects of previous heat stress and yeast supplementation on rumen fermentation patterns. The initial pH and redox of the rumen fluid are the same as the measurement values from the overarching study as it is the same rumen inoculum. The final pH, after 48 of fermentation, experienced a TRT effect such that TN-LY rumen fluid had the lowest pH. As LY is capable of altering ruminal bacteria pools and increasing digestion, a lower pH is to be expected from this rumen fluid. Interestingly, it seems as if HS conditions had a deleterious effect on LY’s capability to alter bacteria profiles enough to see a carryover interaction. The final redox for TN-LY was higher than the rest of the TRT. This could be due to the microbiome relying on LY to scavenge O₂ and once LY was removed, there were no microbes to fill that niche. This theory is supported by the DIET effect. There was also an effect of ENV on final redox such that HS rumen fluid had a lower redox than TN rumen fluid. This suggests that oxygen-scavenging bacteria in the rumen play an important role in maintaining anaerobic environments during heat stress events and are capable of performing at a greater capacity once the heat stimulus is removed. In the exponential model of IVGP, there was no effect of ENV, DIET, or TRT on TGP. However, there was an effect of TRT on k_d where TN-LY had greater k_d than HS-LY and TN-NOY. This is related to HS’s inhibitory effect and LY’s stimulatory effect on bacteria pools in the rumen. The k_d of the IVGP experiment may seem elevated as compared to other k_d of other rations, but the base ration consists of several rapidly fermentable feedstuffs, such as ground corn, molasses, and ground cottonseed meal. The exponential lag time of HS-NOY was far reduced compared to the exponential lag time of HS-LY and TN-NOY. There was also a tendency for HS rumen fluid to have a shorter lag time than TN rumen fluid, possibly due to HS rumen fluid acclimating to a lower, more comfortable temperature for fermentation.

In the log two-pool model, there were no differences attributed to ENV, DIET, or TRT for P1-TGP and P1-k_d. As Pool 1 in IVGP models the pattern of the non-fiber degrading bacteria, P1-k_d was elevated like the exponential k_d. This, again, is attributed to the rapid fermentability of the base ration. The P1>P2 was shorter TN-NOY rumen fluid compared to TN-LY, most probably due to the differences in redox. Pool 2 in IVGP models the pattern of fiber degrading bacteria which are functionally superior in more anaerobic environments with higher pH; a condition the TN-LY rumen fluid did not end on. Additionally, P1>P2 was shorter for TN-NOY compared to HS-LY and HS-NOY due to the HS effect on rumen microbes. The longer lag time did not hinder P2-TGP as there was a tendency of fibrolytic bacteria in HS-exposed rumen fluid to produce more gas than TN rumen fluid. In fact, HS-NOY rumen fluid produced the most P2-TGP from the fermentation of fiber. There were no effects of ENV, DIET, or TRT on P2-k_d, suggesting that fibrolytic bacteria fermented fiber at the same rate, regardless of previous TRT. As the quantity of pool 2 bacteria was not measured in this IVGP experiment, it could be the reason for greater P2-TGP in HS-NOY. There was an effect of DIET on ivNDFD such that LY rumen fluid had greater ivNDFD than NOY rumen fluid. This relationship has been previously demonstrated by Cagle et al. (2019) where LY, at the same dose, had a greater ivNDFD compared to the control (NOY). The ME predicted by GasFit with k_p of 4, 6, and 8%/h are much lower than the actual ME calculated in the overarching study. This suggests that the k_p of the animals, regardless of TRT, was much faster than the three predicted rates. There was an effect of ENV on CH₄ g/L gas such that previous exposure to HS had a tendency to increase CH₄ production. This, and the similar pattern of CH₄ g produced/kg DM, is directly related to the increase of P2-TGP from HS as fiber fermentation is the direct driver of CH₄ in the rumen. The in vitro results from this experiment prove interesting points: the rumen fluid with previous HS-exposure functions equal or superior to TN-NOY rumen fluid, LY during HS allowed for the rumen fluid to function better than with NOY, and removing LY during TN may cause a detriment to the functionality of the rumen fluid.

In conclusion, live yeast supplementation is a plausible method to reduce multiple stresses in newly received, transitioned, and feedlot cattle. This study showed a lack of effects of LY on energy partitioning, nitrogen metabolism, and some ruminal parameters during HS. This could partly be attributed to the animal’s maturity and its previous adaptation to HS environments but may be more related to the restricted DMI in the study. Previous studies had shown greater benefits of LY when DM was fed at a greater %SBW or ad libitum. The relationship between lipolysis and protein accretion when exogenous carbohydrates are minimal, or zero, is poorly described in the literature. This interaction is prominent during comparative slaughter studies and now there is a supporting indirect calorimetry experiment. This metabolic decision must be defined to better understand the metabolic effects of ENV or DIET on the ruminant animal. This study further solidified the benefits of feeding LY during TN conditions but showed an interesting relationship between LY and butyrate production. As more butyrate production is attributed to greater rumen papillae growth and health, perhaps LY, regardless of ENV, would provide a health benefit to the animal. Although there are no known studies relating LY to an increase in liver abscesses, this study suggests an investigation into the relationship between LY and F. necrophorum before LY is considered a production enhancement feed additive. The IVGP portion of this study provides many platforms for further research. The ruminal microbes seemed to be optimized during HS conditions which allowed them to perform equally or better than the ruminal microbes from TN conditions and LY supplementation during HS only further optimization of this relationship. Live yeast supplementation proved to alter the rumen microbial dynamics so far that removing LY from the ration was almost detrimental to fermentative performance. As this study is only the second of its kind, further research should be completed to investigate the benefits of LY during HS on energy partitioning, nitrogen metabolism, and ruminal fermentation dynamics.

Glossary

Abbreviations

%N

nitrogen content

ADF

acid detergent fiber, %DM

ADICP

acid detergent insoluble crude protein, %DM

CARRY

carryover interactions from the previous treatment arrangement

CH₄

methane

CO₂

carbon dioxide

CP

crude protein, %DM

DE

digestible energy

DEI

digestible energy intake

DIET

dietary treatment; LY or NOY

DM

dry matter

DMD

dry matter digestibility, %

DMI

dry matter intake, kg/d

ENV

environmental condition; HS or TN

FE

fecal energy

GASE

gaseous energy of CH₄

GE

gross energy

GEI

gross energy intake

HE

energy from heat production

HS

heat stress environmental conditions (33.8 ± 0.6 °C, 55.7 ± 2.7% RH)

IVGP

in vitro gas production

ivNDFD

in vitro NDF digestibility

kd

fractional rate of degradation, %/h

kg

partial efficiency of the use of ME for growth

km

partial efficiency of the use of ME for maintenance

kp

assumed fractional rate of passage, %/h

LY

base ration with live yeast supplementation

ME

metabolizable energy

MEI

metabolizable energy intake

Nintake

total nitrogen intake

Nfeces

fecal nitrogen output

Nretained

retained nitrogen

Nurine

urinary nitrogen output

NEg

net energy for growth

NEm

net energy for maintenance

NDF

neutral detergent fiber, %DM

NDICP

neutral detergent insoluble crude protein, %DM

NH3-N

ammonia concentration, %

NOY

base ration with no live yeast supplementation

O₂

oxygen

P1>P2

lag time between nonfiber and fiber carbohydrate fermentation, h

P1-kd

fractional rate of degradation of nonfiber carbohydrates, %/h

P1-TGP

total gas production of the nonfiber carbohydrate fermentation pool, mL

P2-kd

fractional rate of degradation of fiber carbohydrate, %/h

P2-TGP

total gas production of the fiber carbohydrate fermentation pool, mL

peNDF

physically effective neutral detergent fiber, %DM

RE

retained energy

RH

relative humidity, %;

RQ

respiratory quotient

SBW

shrunk body weight, kg

SCP

soluble crude protein, %DM

SWG

shrunk weight gain, kg/d

TGP

total gas production, mL

TDN

total digestible nutrients, %DM

TN

thermoneutral environmental conditions (18.4 ± 1.1 °C, 57.6 ± 2.8% RH)

TRT

four treatment arrangements (HS-LY, HS-NOY, TN-LY, and TN-NOY)

UE

urinary energy, Mcal/d

VCH4

rate of CH₄ production

VCO₂

rate of CO₂ production

VFA

volatile fatty acid

VO₂

rate of O₂ consumption

Conflict of Interest Statement

The authors declare no real or perceived conflicts of interest.