The use of live yeast to increase intake and performance of cattle receiving low-quality tropical forages

Abstract

The objective was to evaluate the effects of a specific strain of live yeast (LY) on growth performance, fermentation parameters, feed efficiency, and bacterial communities in the rumen of growing cattle fed low-quality hay. In experiment (exp.) 1, 12 Droughtmaster bull calves (270 ± 7.6 kg initial body weight [BW]) were blocked by BW into two groups, allocated individually in pens, and fed ad libitum Rhodes grass hay (8.4% of crude protein [CP]) and 300 g/bull of supplement (52% CP) without (Control) or with LY (8 × 10⁹ colony-forming unit [CFU]/d Saccharomyces cerevisiae CNCM I-1077; Lallemand Inc., Montreal, Canada) for 28 d, followed by 7 d in metabolism crates. Blood and rumen fluid were collected before feeding and 4 h after feeding. In exp. 2, for assessment of growth performance, 48 Charbray steers (329 ± 20.2 kg initial BW) were separated into two blocks by initial BW and randomly allocated into 12 pens. The steers were fed Rhodes grass hay (7.3% CP) and 220 g/steer of supplement (60% CP) without or with LY (8 × 10⁹ CFU/d) for 42 d, after a 2-wk adaptation period. In exp. 1, fiber digestibility was calculated from total fecal collection, and, in exp 2, indigestible neutral detergent fiber (NDF) was used as a marker. Inclusion of LY increased (P = 0.03) NDF intake by 8.3% in exp. 1, without affecting total tract digestibility. No changes were observed in microbial yield or in the efficiency of microbial production. There was a Treatment × Time interaction (P < 0.01) for the molar proportion of short-chain fatty acids, with LY increasing propionate before feeding. Inclusion of LY decreased rumen ammonia 4 h after feeding (P = 0.03). The addition of LY reduced rumen bacterial diversity and the intraday variation in bacterial populations. Relative populations of Firmicutes and Verrucomicrobia varied over time (P < 0.05) only within the Control group. At the genus level, the relative abundance of an unclassified bacterial genus within the order Clostridiales, a group of cellulolytic bacteria, was reduced from 0 to 4 h after feeding in the Control group (P = 0.02) but not in the LY group (P = 1.00). During exp. 2, LY tended to increase average daily gain (ADG) (P = 0.08) and feed efficiency (P = 0.10), with no effect on NDF intake or digestibility. In conclusion, S. cerevisiae CNCM I-1077 reduced the intraday variation of rumen bacteria and increased the amount of NDF digested per day. These observations could be associated with the tendency of increased ADG and feed efficiency in growing cattle fed a low-quality forage.

Introduction

Low digestibility of tropical pastures during the dry season is an important factor restricting the growth performance of cattle in the dry tropics (McLennan et al., 2017). A common strategy is to feed grazing cattle with small amounts of supplements high in nitrogen (N), such as urea molasses mineral blocks, to promote rumen microbial growth and fiber degradation (McLennan et al., 2017). Further benefits could be achieved with the use of feed additives to improve fiber degradation, such as specific strains of live yeast (LY). The hypothesis of the current study was that because of potential positive effects on the rumen cellulolytic activity, LY added as a supplement would increase fiber digestibility, increase feed intake, and improve cattle performance.

LY has shown beneficial effects on dairy cattle performance and dry matter (DM) intake by promoting a more stable rumen environment with lower lactate, oxygen concentration, and higher pH (Desnoyers et al., 2009). In addition, LY enhanced the abundance of fibrolytic and lactate-utilizing bacteria, promoting greater fiber degradation activity (Pinloche et al., 2013). The specific LY strain used in the current study, Saccharomyces cerevisiae CNCM I-1077, has been shown to promote rumen bacterial growth and activity with a subsequent effect on total fiber digestibility. This strain enhances fiber colonization by fibrolytic fungi species (Chaucheyras-Durand et al., 2016), which have been reported to increase the breakdown of polysaccharides linked with lignin by the secretion of esterases.

In beef cattle, little is known about the effects of LY on tropical high-forage diets, as most of the studies were conducted with cattle fed high-concentrate diets (Crossland et al., 2019). In a recent study, Smith et al. (2020) fed the I-1077 LY strain to Angus steers during the receiving and backgrounding periods on diets of approximately 60% roughage and reported improved performance during the receiving period (day 1 to 47) with no LY effects on cumulative performance from day 1 to 77.

In high-grain diets, the mechanism on intake regulation and the potential impact of small increases in the rate of fiber digestibility are very different than in high-forage diets (Allen, 2014). Also, the prevalent bacterial communities and the relative abundance of cellulolytic bacteria are strongly regulated by the amount of grain in the diet (Thoetkiattikul et al., 2013). Therefore, the objective of the current work was to evaluate LY effects on growth performance, rumen fermentation parameters, feed efficiency, and bacterial communities in the rumen of growing cattle fed a low-quality tropical grass hay.

Materials and Methods

All procedures involving animals were performed in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes with approval from The University of Queensland Animal Ethics Committee: QAAFI/293/18/LALLEMAND.

Experiment 1: metabolism trial

Animals, experimental design, and treatment diets

Twelve Droughtmaster bull calves (270 ± 7.6 kg of initial body weight [BW]) were blocked by BW into two groups and randomly distributed to the two treatments, Control or LY. The bull calves were housed in individual pens and fed ad libitum Rhodes grass hay (Chloris gayana Kunth.) and 300 g/bull of supplement without (Control) or with LY for 28 d, followed by 7 d in the metabolism crates. A specific strain of S. cerevisiae (strain CNCM I-1077, Levucell SC10 ME Titan, Lallemand Animal Nutrition, QLD, Australia) was added to supply 8 × 10⁹ colony-forming unit [CFU]/300 g of supplement. The LY supplement was prepared once every 2 wk using a premix with 1 × 10¹⁰ CFU of S. cerevisiae I-1077/kg. The chemical composition of hay and supplement is presented in Table 1.

Table 1.

Ingredients and chemical composition of hay and supplement offered during the metabolism trial for 28 d (12 young bulls into two treatments)

Items	Rhodes grass hay	Supplement1
Ingredients (% of the diet as fed)
Cottonseed meal	—	43.5
Copra meal	—	29.0
MDCP2	—	10.9
Limestone	—	7.6
Urea	—	7.5
Vegetable oil	—	1.0
Trace minerals premix3	—	0.5
Chemical composition
DM, %	87.4	91.9
CP, % of DM	8.4	52.1
NDF, % of DM	72.0	31.3
ADF2, % of DM	41.4	20.9
Lignin, % of DM	5.0	6.4
Ether extract, % of DM	0.9	6.5
Ash, % of DM	12.3	26.2

¹For LY treatment, Levucell SC10 ME Titan (Lallemand Animal Nutrition, Qld, Australia) was added at 0.27% of the supplement to supply 8 × 10⁹ CFU of S. cerevisiae I-1077/300 g of supplement.

²MDCP, mono-dicalcium phosphate; ADF, acid detergent fiber.

³Guaranteed levels per kilogram: calcium 200 g; cobalt 1,200 mg; copper 1,800 mg; iodine 1,000 mg; iron 10,000; manganese 30,000 mg; selenium 400 mg; sulfur 80 g; and zinc 50,000 mg.

Intake

Hay was offered ad libitum throughout the experiment, allowing for 5% orts on a DM basis. Animals were offered the allocated supplements at 0700 hours, and, approximately 20 min later, hay was offered. The time allowed between supplement and hay offer was an attempt to reach complete consumption of supplement prior to hay allocation. During the collection period in metabolism crates, hay was restricted to 90% of ad libitum intake to account for the natural decrease in intake inside the metabolism crates and to minimize diet selection. Samples of hay, supplement, and orts were collected daily, bulked within each week, and dried in forced-air oven at 60 °C for 72 h and at 105 °C for 15 h to calculate DM intake (DMI). Samples were analyzed for neutral detergent fiber (NDF), organic matter (OM) and nitrogen (N) to calculate NDF intake (NDFI), OM intake (OMI), and crude protein intake (CPI).

Rumen fluid and blood collection

Prior to going into metabolism crates, blood and rumen fluid were collected at two time points (0 and 4 h after feeding) for two consecutive days (days 26 and 27). Blood was withdrawn from the jugular vein into a 10-mL vacutainer (Becton Dickinson; Franklin Lakes, NJ, USA) coated with lithium heparin. The vacutainers were inverted six to eight times and placed on ice for approximately 20 min before centrifugation using a Heraeus Labofuge 400R (Thermo Fisher Scientific, Seventeen Mile Rocks, QLD, Australia). Plasma was collected after centrifugation at 4,000 × g for 10 min, in duplicate, and stored at −20 °C for later urea analysis.

Rumen fluid was collected using a stomach tube attached to a dual-action hand pump (Coleman, Sydney, NSW, Australia). The fluid was strained through knitted nylon material and pH was determined immediately (pHep 5, Hanna Instruments; Melbourne, VIC, Australia). Subsamples of rumen fluid were then placed into tubes for the analysis of rumen ammonia nitrogen (NH₃-N) (6 mL rumen fluid + 2 mL 0.5 M H₂SO₄) and rumen short-chain fatty acids (SCFA) (4 mL rumen fluid + 1 mL 20% metaphosphoric acid + internal standard, i.e., 4 methyl n-valeric acid – approximately 0.48 g/L) and were stored at −20 °C until analysis. Furthermore, 1 mL subsamples of rumen fluid were immediately frozen in liquid nitrogen, and all samples were stored at −80 °C until DNA extraction and analysis.

Fecal and urine collections

On day 28, the animals were allocated into individual metabolism crates. The first 2 d were considered as adaptation to crates, followed by five consecutive days of total daily feces and urine output collection of each animal. Feces were weighed, mixed, and a 10% subsample was stored at 4 °C. At the end of the collection period, the fecal samples for each steer were mixed to provide a representative sample for the collection period, weighed, dried, and then stored for further analyses of N, NDF, and OM content.

Laboratory analyses

Samples of hay, supplements, refusals, and feces collected throughout the trial were dried at 60 °C for 72 h in a forced-air oven and ground through a 2-mm screen (Retsch ZM 200; Haan, NW, Germany). DM was analyzed (AOAC, 2005; method 934.01) and OM content was calculated after ash determination (AOAC, 2005; method 942.05) at 550 °C for 8 h (Modutemp; Perth, WA, Australia). Ash-free NDF was determined according to Van Soest et al. (1991) using alpha-amylase, adapted to an Ankom 200 fiber analyzer (Ankom Technology Corporation; Fairport, NY, USA). Fiber insoluble in acid detergent and lignin were determined according to the procedures of the AOAC (2005; method 973.18). Ether extract content was determined after extraction with petroleum spirit with the SER148 Solvent Extraction Unit (VELP Scientifica S.R.L.; Usmate, MI, Italy; AOAC, 2005; method 2003.05).

Urine was collected in plastic trays with 5% sulfuric acid added to maintain pH at around 3 to inhibit microbial growth. A 10% subsample of acidified urine was collected each day and kept at 4 °C, bulked over the 5 d collection period, and stored at −20 °C for subsequent measurement of N and purine derivatives (PD). The N content of feed offered, urine, and feces was determined by the Dumas combustion method (Sweeney, 1989), using a LECO CN928 Carbon/Nitrogen combustion analyzer (LECO Corporation; St Joseph, MI, USA). CP was calculated by multiplying the N content in samples for 6.25.

PD in the urine were analyzed according to the methods of George et al. (2006) and Czauderna and Kowalczyk (2000) using a Prodigy 250 × 46 mm, 5 µm, ODS C18 reverse-phase column (Phenomenex, Torrence, CA, USA). In brief, acidified urine samples were thawed, a buffer and internal standard added, and then this was then filtered through a 0.20-µm cellulose nitrate filter followed by a 300-mg C18 filter and analyzed for PD concentration using high-performance liquid chromatography with quantification at 215 nm (Shimadzu Prominence HPLC with Photo Diode Array Detector; Kyoto, Honshu, Japan). Microbial protein production was calculated using the equation of Chen and Gomes (1992), with the value for excretion of endogenous PD for Bos indicus cattle from Bowen et al. (2006). Efficiency of microbial crude protein (MCP) synthesis (EMPS) was calculated as g of MCP/kg digestible OMI (DOMI).

Plasma urea nitrogen (PUN) was determined using a Beckman Coulter AU480 auto-analyzer (Beckman Coulter Diagnostic Systems Division; Melville, NYC, USA), as described in Marsh et al. (1965) and with intra-assay variation of 2.61%. The NH₃-N concentration was estimated using direct distillation (UDK 139 semi-automatic distillation unit; Rowe Scientific Pty Ltd; Wacol, QLD, Australia) followed by titration (Titralab 840 automatic titration unit; Hach; Dandenong, VIC, Australia), as described in Preston (1995) and with intra-assay variation of 2.02%.

The SCFAs present in rumen fluid were determined by gas liquid chromatography (GC17 Shimadzu; Kyoto, Honshu, Japan) using a polar capillary column (ZB-FFAP; Phenomenex, Lane Cove, NSW, Australia) following the procedures from Cottyn and Boucque (1968). A prepared multi-acid standard was mixed with a protein precipitant/internal standard and used to calibrate the gas chromatography. Samples were then analyzed using the internal standardization method for calibration.

DNA extraction, sequencing, and bioinformatics

The genomic DNA in rumen fluid samples was extracted after a bead-beating procedure and on-column purification (Popova et al., 2010). The quality and quantity of DNA were measured on a Nanodrop 1000 Spectrophotometer (Thermo Fisher Scientific, France). Approximately, 15 µg of DNA extracted was sent to Genome Quebec Innovation Centre (QC, Canada) for fluidigm amplification and MiSeq Illumina sequencing. Bacterial 16S rRNA gene was amplified via 515f (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806r (5′- GACTACHVGGGTWTCTAAT-3′) primers. These sequences were processed using the DADA2 package v.1.4.0 (Callahan et al., 2016) and R software v. 3.6.0. The 16S rRNA bacterial gene sequences (forward and reverse) were trimmed to 240 and 160 bp and merged. After chimera removal (2.5% of total sequences), sequences were aligned to the Green Genes database v. 13.8 (DeSantis et al., 2006) to assign the taxonomy to the operational taxonomic unit (OTU) with 97% of similarity.

Alpha diversity metrics were analyzed using an analysis of variance (ANOVA). Principal coordinates analysis plots were constructed using OTU-based Bray–Curtis dissimilarities distance and function on R software (ggplots2 and vegan packages, respectively; Wickham, 2011). Furthermore, the differences on beta-diversity between control and LY groups were calculated using permutational multivariate analysis of variance (PERMANOVA).

Experiment 2: growth performance study

Animals, experimental design, and treatment diets

Forty-eight growing Charbray steers of 329 ± 20.4 kg BW were blocked by BW into two blocks and randomly allocated to 12 pens (162 m²) with 4 steers per pen. The feedlot pens were designed to maximize steer welfare with large shade cloths provided and varying substrate materials underfoot in a dry lot pen, without any possibility of grazing. The experimental period was 56 d in total, which included a 14 d adaptation to facilities and the control diet. After the 14 d adaptation period, the steers received ad libitum Rhodes grass hay and 880 g/pen per day (220 g/steer) of a commercial protein supplement (Top Country Livestock Nutrition, Queensland, Australia) without (Control) or with LY. The chemical composition of hay and supplements is presented in Table 2. The LY used was S. cerevisiae strain CNCM I-1077 (Levucell SC10 ME Titan, Lallemand Animal Nutrition) added to supply 8 × 10⁹ CFU of LY per day (3.64 × 10¹⁰ CFU/kg supplement), as described on the label.

Table 2.

Ingredients and chemical composition of hay and supplement offered during the growth performance trial for 42 d (12 pens divided into two treatments)

Items	Rhodes grass hay	Supplement1
Ingredients (% of the diet AF)
Cottonseed meal	—	62.5
MDCP2	—	10.9
Limestone	—	7.6
Urea	—	7.5
Salt	—	5.0
Ammonium sulfate	—	5.0
Vegetable oil	—	1.0
Trace minerals premix3	—	0.5
Chemical composition
DM, %	91.0	93.7
CP, % of DM	7.3	60.0
NDF, % of DM	69.2	11.6
ADF2, % of DM	41.4	8.3
iNDF, % of DM	27.7	12.9
Lignin, % of DM	5.8	3.2
Ether extract, % of DM	0.9	3.5
Ash	10.4	25.4

¹For LY treatment, Levucell SC10 ME Titan (Lallemand Animal Nutrition, Qld, Australia) was added at 0.36% of the supplement to supply 8 × 10⁹ CFU of S. cerevisiae I-1077/220 g of supplement.

²MDCP, mono-dicalcium phosphate; ADF, acid detergent fiber.

³Guaranteed levels per kilogram: calcium 200 g; cobalt 1,200 mg; copper 1,800 mg; iodine 1,000 mg; iron 10,000; manganese 30,000 mg; selenium 400 mg; sulfur 80 g; and zinc 50,000 mg.

Intake and BW gain

The amount of offered hay was adjusted daily by visual score to assure ad libitum intake. After two consecutive weeks, refusals of hay accumulated in feed bunks were swept, collected, and weighed. A subsample of approximately 500 g from each bunk was dried in a forced-air oven at 60 °C for 72 h for initial DM determination, with residual moisture removed in a forced-air oven at 105 °C for 24 h. If present, supplement refusals were collected daily and bulked in a container to represent 14 d of feeding. Each bulk sample of supplement refusals for the individual pen was mixed, and a subsample was collected and dried as per the hay samples, to estimate DM content.

The steers were weighed at the beginning of the adaptation period and every 14 d until the end of the trial. At each of these time points, steers were weighed for two consecutive days, before feeding, to minimize variation in rumen fill. DMI, average daily gain (ADG), and gain:feed ratio (G:F) were calculated to represent the 42 d of the performance trial.

Feed digestibility

Feed digestibility was estimated using indigestible NDF (iNDF) as an internal marker. On the final 5 d of the feeding period, fecal grab samples were collected from the rectum of all animals, twice a day (0600 and 1500 hours). Fecal samples were stored at −20 °C until the final day of collections when samples were thawed and combined as one bulk sample per animal. All samples were ground using a 2-mm sieve, and hay, supplement, and fecal samples were analyzed for NDF as described for the metabolism trial. The iNDF was analyzed using 240 h in vitro incubations using 0.3 to 0.4 g of sample. Samples were placed into labeled acetone-washed Ankom fiber filter bags (F57, Ankom Technology, Macedon), which were then heat sealed. Each sample was incubated in triplicate. Fermentations of the samples utilized Daisy Incubators (D200, Ankom Technology, Macedon) set at 39 °C using media according to Goering and Van Soest (1970). The rumen fluid was taken from two fistulated steers fed Lucerne and tropical grass. Renovation of media was performed after 120 h. After 240 h, bags were removed, rinsed until clear, and analyzed for NDF content. The remaining NDF is iNDF, as all potentially digestible NDF would have been removed during the fermentation process. The NDF digestibility (NDFD) was then calculated using the formula below:

$${\displaystyle \begin{array}{l}{\displaystyle NDFD=100-\left[100\left(\frac{DietiNDF(\mathrm{\%})}{FecesiNDF(\mathrm{\%})}\right)\left(\frac{FecesiNDF(\mathrm{\%})}{DietiNDF(\mathrm{\%})}\right)\right]}\end{array}}$$

Statistical analysis

Data were analyzed as a completely randomized block design for both trials, using the MIXED Procedure of SAS 9.4 (SAS Institute Inc., Cary, NC). Normality of residuals was evaluated by the Shapiro–Wilk test, and homogeneity of variances was assessed by the Levene test.

For the metabolism trial, intake data were analyzed as a randomized block design with repeated measures in time. Model included the fixed effects of treatment and the random effects of block and pen within treatment as the experimental unit. Day was included as repeated measures with AR(1) as the covariance structure, selected based on the Bayesian information criterion. The degrees of freedom and tests were adjusted by the Kenward–Roger option. Total tract digestibility was analyzed as a randomized block design including the fixed effects of treatment and the random effects of block. For the rumen fermentation data, with two time points (0 and 4 h after feeding), time was included as repeated measures and Time × Treatment was included as fixed term.

For the growth performance trial, pen was used as experimental unit (n = 6 pens/treatment), and data were analyzed as a completely randomized block design. BW gain was determined by regressing BW with days of experiment for each animal and analyzed in a model including the fixed effect of treatment and random effects of block and pen within treatment. For intake data, period was included in the model as a repeated measure assuming no specific variance-covariance structure (unstructured). The degrees of freedom and tests were adjusted by the Kenward–Roger option. For both trials, treatment effects were compared using Fisher’s least significant difference (option PDIFF from LSMEANS) and significance declared at P ≤ 0.05 and trends at P ≤ 0.10.

Microbial community composition was compared across treatments using the R package DESeq2 (Love et al., 2014). The effect of Time (0 and 4 h after feeding) was also compared within each treatment using the same package. False discovery rate corrected P-values (FDR) were calculated using R Statistical Software v.3.6.0. (R Core Team, Vienna, Austria).

Results

Experiment 1: metabolism trial

During the 28 d of trial, young bulls fed LY had greater DMI (P = 0.03) than their counterparts (Table 3). Additionally, NDFI, OMI, and CPI were also increased (P ≤ 0.05). There were no treatment effects on total OM (P = 0.94), CP (P = 0.38), or NDF (P = 0.73) digestibilities. Despite this, the total OM, CP, and NDF digested per day were greater (P ≤ 0.05) in young bulls with LY supplement because of improvements in hay intake. There was no effect of LY on MCP yield (P = 0.62) or on EMPS (P = 0.85).

Table 3.

Effect of LY supplementation on intake, diet digestibility, and microbial protein production during the metabolism trial for 28 d (12 young bulls into two treatments)

Items	Control	LY1	SEM	P-value
BW2, kg	276	268	19.4	0.50
DMI3, kg/d	4.54	4.81	0.085	0.03
OMI4, kg/d	3.53	3.85	0.084	0.03
NDFI5, kg/d	2.81	3.08	0.069	0.03
CPI6, kg/d	0.47	0.51	0.010	0.02
OM digestibility, %	59.9	60.0	0.86	0.94
NDF digestibility, %	60.1	59.6	1.04	0.73
CP digestibility, %	61.4	60.5	0.75	0.38
Total OM digested, kg/d	2.11	2.31	0.047	0.02
Total NDF digested, kg/d	1.69	1.83	0.040	0.03
Total CP digested, kg/d	0.29	0.31	0.006	0.03
MCP yield, g/d	114	122	11.2	0.62
EMPS, g MCP/kg DOMI	54.4	52.9	5.43	0.85

¹LY strain used is S. cerevisiae I-1077.

²Average BW during the 28 d evaluation trial.

³DMI during the 28 d evaluation trial.

⁴OMI during the 7 d in the metabolism crate.

⁵NDFI during the 7 d in the metabolism crate.

⁶CPI during the 7 d in the metabolism crate.

The LY supplementation did not change the rumen pH (P = 0.27) or total rumen SCFA concentration (P = 0.32; Table 4). However, there was a Time × Treatment interaction for most of the individual SCFA proportions. For instance, LY supplementation tended to decrease acetate (P = 0.08, Figure 1A) and tended to increase propionate (P = 0.06, Figure 1B) proportions before feeding, yet no effect was observed 4 h after feeding. Hence, adding LY in the supplement tended to decrease (P = 0.06) acetate:propionate ratio in the rumen before feeding but not after feeding (Figure 1C).

Table 4.

Effects of LY on rumen fermentation parameters and PUN in metabolism trial for 28 d (12 young bulls into two treatments)

Items	Control	LY1	SEM	Trt	Time	Trt * Time
pH	6.97	7.14	0.101	0.27	<0.01	0.52
Total SCFA, mM	71.6	64.8	4.74	0.32	0.54	0.99
Acetate, molar %	77.5	76.9	0.53	0.42	0.10	<0.01
Propionate, molar %	14.0	14.7	0.37	0.23	<0.01	<0.01
Butyrate, molar %	6.3	6.2	0.22	0.75	<0.01	0.02
Acetate:propionate	5.5	5.3	0.17	0.26	<0.01	<0.01
Valerate, molar %	0.56	0.57	0.023	0.96	<0.01	<0.01
BCFA2, molar %	1.55	1.61	0.078	0.64	<0.01	0.34
Rumen NH3-N, mg/L	58.8	48.8	3.41	0.06	<0.01	0.05
PUN, mmol/L	4.50	4.17	0.275	0.33	<0.01	0.60

¹LY strain used is S. cerevisiae I-1077.

²BCFA, branched-chain fatty acids.

Figure 1.

Effect of LY supplementation on rumen fermentation characteristics for samples taken at 0 and 4 h after feeding. Values are presented as mean ± SEM. ^a–cMeans without common letters differ at P < 0.10.

There was also a Time × Treatment interaction on rumen butyrate (P = 0.02) and valerate (P ≤ 0.01) proportions (Table 4). Butyrate proportion increased from 0 to 4 h after feeding in the Control group but not in the LY group (Figure 1D). Valerate proportion tended (P = 0.09) to be greater with LY supplementation before feeding but not 4 h after feeding (Figure 1E). The proportion of branched-chain fatty acids was not affected (P = 0.64) by treatment.

The effect of LY on rumen NH₃-N was influenced by time, as there was a Trt × Time interaction (P = 0.05, Table 4). Feeding LY reduced (P = 0.03) rumen NH₃-N concentration 4 h after feeding, with no effect (P = 0.97) before feeding (Figure 1F). However, LY supplementation did not affect (P = 0.33) PUN concentration.

The effect of LY on the rumen microbiota showed differences in ruminal OTU abundance and diversity between treatments. The beta-diversity analysis revealed a tendency (P = 0.08) for a distinction between rumen bacteria from the two treatments (Figure 2). In addition to that observation, alpha diversity, as measured using Simpson and Inverse Simpson indices, representing ruminal bacterial population uniformity, demonstrated that young bulls consuming LY had lower diversity (P ≤ 0.05) than the Control group (Figure 3). Furthermore, the Shannon index, which accounts for both abundance and evenness of the community, tended (P = 0.08) to be lower with LY addition.

Figure 2.

Beta diversity in rumen samples from young bulls fed Control (C, ▲) or LY supplementation (LY, ● ).

Figure 3.

Between-sample diversity indices for rumen samples from young bulls fed Control (C) or LY supplementation.

At the phylum level, Bacteroidetes dominated the bacterial community in both treatments, followed by Fibrobacteres, Firmicutes, and Verrucomicrobia (Figure 4). Adding LY to the supplement did not cause major changes in the rumen bacterial abundance. However, the relative abundance of both Firmicutes and Verrucomicrobia varied over time (FDR ≤ 0.05), only within the Control animals, with no effect of time (FDR ≥ 0.50) on the relative abundance of LY-fed bulls.

Figure 4.

Relative abundance of Bacterial phylum in Control or LY treatments in rumen fluid samples taken at 0 and 4 h after feeding. ^a,bMeans without common letters differ at P < 0.05.

When analyzed at the genus level, Prevotella dominated the bacterial community, followed by Fibrobacter (Figure 5). The next dominant genus was comprised of unclassified rumen bacterium belonging to the Bacteroidales order, followed by an unclassified member of the Verrucomicrobia RFP12 family, Bacteroidales RF16, and an unclassified bacteria member of the Clostridiales order. As for the phylum analysis, the relative abundance of bacteria changed over time but only in the Control animals. The abundance of an unclassified bacterial member of the Verrucomicrobia RFP12 family was increased from 0 to 4 h after feeding, but only in the Control bulls (FDR ≤ 0.05), with no difference (FDR = 0.43) in the LY-fed group. Also, the relative abundance of an unclassified bacterial member of the order Clostridiales was reduced from 0 to 4 h after feeding in the Control group (FDR = 0.02) and not in the LY group (FDR = 1.00).

Figure 5.

Relative abundance of Bacterial genus in Control (C) or LY treatments in rumen fluid samples taken at 0 and 4 h after feeding. ^a–cMeans without common letters differ at P < 0.05.

Experiment 2: growth performance trial

There were no differences in final BW between treatment groups (P = 0.75); however, steers fed LY supplement tended (P = 0.08) to have greater ADG than their counterparts (Table 5). Despite the tendency for a 27.6% increase in ADG, there was no effect of LY on hay (P = 0.40) or total DM (P = 0.52) intakes, leading to a tendency (P = 0.10) for increased feed efficiency in steers consuming LY. LY supplementation did not increase total OM (P = 0.73) nor total NDF (P = 0.23) digestibilities, when measured by iNDF in the feces. Furthermore, there were no effects of LY supplementation on the total amount of OM (P = 0.87) or NDF (P = 0.64) digested per day.

Table 5.

Effect of LY supplementation on BW gain, intake, and feed efficiency in growing crossbred steers in the growth performance trial for 42 d (n = 6 pens/treatment)

Items	Control	LY1	SEM	P-value
Performance parameters
iBW2, kg	333	326	20.0	0.24
fBW3, kg	351	349	22.9	0.75
ADG, g/d	424	541	75.2	0.08
Hay DM intake, kg/d	6.15	6.36	0.171	0.40
Supplement DM intake, g/d	172	167	12.1	0.79
Total DM intake, kg/d	6.35	6.51	0.173	0.52
G:F	0.066	0.086	0.008	0.10
Digestibility parameters
OM digestibility4, %	53.1	52.9	0.32	0.73
NDF digestibility4, %	48.4	49.1	0.38	0.23
Total OM digested, kg/d	3.30	3.33	0.314	0.87
Total NDF digested, kg/d	2.09	2.14	0.229	0.64

¹LY strain used is S. cerevisiae I-1077.

²iBW, initial BW after 14 d of adaptation to control diet.

³fBW, final BW.

⁴Estimate using iNDF as an internal marker.

Discussion

LY-containing additives have been used to modulate rumen pH and increase fiber degradation in high-grain diets (Ovinge et al., 2018). The specific strain of S. cerevisiae, CNCM I-1077, has been shown to promote enzymes that degrade cellulose and hemicellulose, such as xylanase and CMCase, which could lead to greater fiber degradation (Chaucheyras-Durand and Fonty, 2001). The latter authors also observed a decrease in rumen NH₃-N concentration due to the enhanced ammonia utilization by cellulolytic bacteria. This strain of yeast also exhibited greater oxygen-scavenging capacity in comparison to other S. cerevisiae strains, which could create a more suitable rumen environment for cellulolytic bacteria. There is limited information on the potential benefits of including LY to protein supplements for cattle eating poor-quality tropical forages. The central hypothesis of this study was that LY supplementation would modify the rumen environment increasing the population of fiber-degrading bacteria, fiber digestibility, daily intake, and, ultimately, BW gain of beef cattle receiving low-quality forages. The results from both trials did not corroborate the initial hypothesis that LY supplementation would increase total tract fiber digestibility. However, because fiber intake was increased and total tract digestibility was unchanged, LY supplementation increased the total amount of fiber being digested and, therefore, the energy supply to the growing steers.

The intake of poor-quality forages is usually limited by physical rumen fill, and faster rate of fiber degradation usually increases total NDF intake (Allen, 1996). Despite no significant change in total tract NDF digestibility with LY supplementation in exp. 1, LY resulted in a 9.6% increase in NDF intake and an 8.2% increase in the total amount of NDF being digested over a 24-h period, in comparison to the control treatment. However, these positives effects were not corroborated by exp. 2.

Feed disappearance from the rumen is affected by two major factors, digestion and passage (Poppi et al., 1981; Allen, 1996; Mertens and Grant, 2020). Previous studies have shown that LY can increase the rumen rate of fiber degradation both in vivo and in vitro (Plata et al., 1994; Chaucheyras-Durand et al., 1996; Guedes et al., 2008). It is logical to speculate that LY may have increased both rates of digestion and passage, thereby increasing total NDF flow but not total tract digestibility (Mertens and Grant, 2020). However, rates of fiber digestion and passage through the rumen were not directly measured in this trial.

Greater hay intake and faster flow of diet from the rumen could potentially result in greater MCP yield and improved EMPS, favoring faster-growing species, and decreasing protozoal predation rate (Firkins et al., 2007), which was not the case in the present experiment. The association between faster rumen flow rate and greater EMPS has been found when comparing forages differing in CP content, intake, digestibility, and rumen retention time (Panjaitan et al., 2010; Bowen et al., 2017). Small differences in intake and NDF flow, such as in the present study, were likely not enough to increase MCP and EMPS. The average EMPS in the present study, 53.6 g MCP/kg DOMI, was well below the values proposed in feeding standards, such as the 130 g MCP/kg DOMI used by NRC (2000). However, it is consistent with the range (26 to 90 g MCP/kg DOMI) in EMPS reported for steers grazing low-quality tropical grasses (Bowen et al., 2017).

Lack of rumen NH₃-N can also limit MCP yield, but only in low-protein diets with CP content below 7% to 8% (Bowen et al., 2017), as microbial growth is maximized when rumen NH₃-N concentration is between 20 and 50 mg/L (Satter and Slyter, 1974; Slyter et al., 1979). Therefore, although LY-supplemented bull calves had lower rumen NH₃-N at 4 h after feeding, it is unlikely that rumen nitrogen was limiting MCP yield, as NH₃-N was above 65 mg/L at that time point. There are two possible explanations for the lower rumen NH₃-N in LY-supplemented animals. Firstly, LY has been shown to decrease the proliferation of proteolytic bacteria, such as Prevotella ruminicola (Wallace et al., 1997), as LY competes with proteolytic bacteria for energy substrates and limits the action of this type of bacteria on proteins and peptides (Chaucheyras-Durand et al., 2005). Secondly, with greater hay intake, there was more carbohydrate for microbial fermentation and incorporation of NH₃-N into microbial cells (Hristov et al., 2010). The lower NH₃-N concentration may have resulted from greater cellulolytic bacteria abundance and activity, as this type of bacteria mainly utilizes NH₃-N as the source of N (Ogunade et al., 2019).

Despite the increased hay intake and the total amounts of NDF digested per day in the metabolism trial, LY supplementation had no effects on total rumen SCFA concentration. SCFAs concentration in the rumen fluid has a direct relationship with microbial activity, but, as previously highlighted, EMPS was similar between treatments. Similarly, Cagle et al. (2019) studied growing beef cattle fed either a backgrounding or transition diets using 5, 10, or 15 g/d of S. cerevisiae CNCM I-4407 at 1.1¹⁰ CFU/g and found no differences in total rumen SCFA concentrations. In contrast, S. cerevisiae CNCM I-1077 supplied daily at 8 × 10⁹ CFU/animal throughout four seasons of the year to grazing cattle in the work of Sousa et al. (2018) increased total SCFA during autumn; although, as emphasized by the authors, NDF digestibility was significantly lower during that season and S. cerevisiae CNCM I-1077 was shown to benefit fiber digestibility and fiber-degrading bacteria.

In the current experiment, there was a tendency for LY to increase propionate and decrease acetate proportion but only before feeding. Sousa et al. (2018) reported that LY decreased acetate and increased propionate over the summer in cattle grazing tropical pastures. This effect of LY supplementation enhancing propionate proportions in high-forage diets should not be confounded with the more prevalent LY effect in increasing rumen pH and acetate proportion in high-grain, low-pH diets (AlZahal et al., 2017). Rumen fermentation favoring propionate production potentially increases metabolizable energy availability and the gluconeogenic potential of the diet, as there are fewer energy losses as CO₂ and CH₄ when carbohydrates are fermented to propionate instead of to acetate or butyrate and because propionate is the main gluconeogenic substrate in cattle (Ungerfeld and Kohn, 2006).

Yeast supplementation has been shown to increase the relative abundance of propionate-producing bacteria, such as Fibrobacter succinogenes (Beauchemin et al., 2003; Uyeno et al., 2017), which could explain the increase in propionate proportion. It is well known that F. succinogenes synthetizes succinate as a principal end product of cellulose degradation (Suen et al., 2011), and succinate can be decarboxylated into propionic by propionic-producing bacteria species (Reichardt et al., 2014; Deusch et al., 2017). However, there was no significant effect of LY on Fibrobacteres abundance in the present study.

Both treatments presented distinct rumen bacterial compositions and diversity as observed through the beta diversity. Shannon and Simpson diversity indices, as measures of richness and evenness, respectively, demonstrated that LY supplementation tended to reduce the number of species present while decreasing the evenness. Previously, LY inclusion decreased bacterial diversity in the rumen of dairy cows (Ogunade et al., 2019) or had no influence (AlZahal et al., 2017; Chaucheyras-Durand et al., 2019). Although, Chaucheyras-Durand et al. (2019) mentioned that S. cerevisiae I-1077 enhanced richness and diversity of other rumen eukaryote and fungi species in preweaned lambs. Less rumen microbial diversity has been associated with more feed-efficient dairy cows, with the assumption that substrates would be utilized in a limited number of metabolic pathways generating fewer energy losses (Shabat et al., 2016).

In the current experiment, LY decreased the variation in ruminal bacterial populations over time (0 and 4 h after feeding). In contrast, there was a reduction in the proportion of Firmicutes and an increase in the proportion of Verrucomicrobia phyla in the rumen fluid of Control bulls between collections. The phylum Firmicutes includes bacterial species linked to fiber degradation (Sandri et al., 2014), such as Ruminococcus albus and other members of the order Clostridiales. In the present study, less variation was observed in the abundance of an unclassified bacteria member of the order Clostridiales, with the use of LY supplementation. Bacterial species within this order are equipped with a multi-enzymatic complex (cellulosome) and cellulose-adhesion mechanisms, allowing them to efficiently adhere and degrade fiber, which accounts for why it is one of the predominant orders in the rumen fluid of grazing cattle (Dassa et al., 2014; Noel et al., 2017).

LY-supplemented animals had lower Verrucomicrobia phylum abundance. This type of bacteria is known to release H₂ from complex polysaccharides degradation (Martinez-Garcia et al., 2012; Detman et al., 2018). The Verrucomicrobia family RFP12, seen in greater abundance in Control animals in our work, was prevalent in the rumen of high methane yielding animals (Kamke et al., 2016). Therefore, decreases in Verrucomicrobia family RFP12 after feeding could also have beneficial effects on rumen fermentation and contribute to methane mitigation.

A possible cause for a more stable bacterial abundance in the rumen fluid of LY-treated animals could be related to the higher DMI. Higher intake observed in supplemented steers could decrease the variation in rumen bacteria communities by providing more substrate available for fermentation throughout the day. Previous studies have reported a positive correlation between DMI increments from LY supplementation and the relative abundance of F. succinogenes, R. albus, and unclassified genus belonging to Bacteroidales and Clostridiales orders (Jiang et al., 2017; Bach et al., 2019). In diets with a higher concentration of grains, LY reduced changes in pH and the redox potential, making it a more stable rumen environment with less variation of rumen bacterial abundance, especially fibrolytic species (Chaucheyras-Durand and Fonty, 2001; Koike and Kobayashi, 2001).

The results from the exp. 2 suggested that the described effects of LY on rumen fermentation and total fiber intake observed during the metabolism trial resulted in tendencies for 27.6% increase in ADG and 30% increase in G:F ratio. In high-forage diets, contrary to when animals are receiving high-grain diets, an increase in the rate of fiber degradation would represent a significant increase in available energy for muscle growth. In diets with circa 69% forage, Smith et al. (2020) observed a 4.3% increase in ADG of Angus steers supplemented with LY for 47 d. In contrast to the present study, previous studies with high-grain diets have reported no additional increase in ADG of feedlot steers receiving LY as an additive in high-grain diets (Magrin et al., 2018; Ovinge et al., 2018).

The positive LY effects observed during the metabolism trial in exp. 1, that is, greater intake and total amount of hay degraded per day, associated with higher propionate concentration in the rumen before feeding, could potentially be translated into more glucose and nutrients available for muscle deposition and hence contribute to higher production performance in growing steers receiving low-quality tropical forages. In exp. 2, as there was no effect of LY on intake or digestibility, the observed tendency for increased ADG and G:F ratio would be suggesting increased efficiency of rumen fermentation (Cagle et al., 2019). However, because of the errors associated with the use of internal markers, such as iNDF, to estimate fiber digestibility, and because intake was measured at the pen level and only every 2 wk, these results need to be considered with caution. In contrast, during exp. 1, intake was measured daily for each individual animal and fiber digestibility was measured with total feces collection, a more precise method.

Conclusions

There is limited information on the effect of LY supplementation on intake and fiber digestibility of low-quality tropical forages. Saccharomyces cerevisiae CNCM I-1077 supplementation reduced the intraday variation of rumen bacteria and increased the total amount of NDF digested per day in exp. 1. Based on the increase in the amount of fiber being digested, it appears LY supplementation was increasing both rates of fiber digestion and passage, therefore allowing greater intake with the same total digestibility. The lesser variation in bacterial abundance throughout the day is proposed as the reason for the increased fiber intake, in exp. 1, and the tendency for increased ADG and feed efficiency observed in exp. 2. Further research is needed to investigate the effect of LY supplementation on the rumen kinetics of poor-quality forages.

Glossary

Abbreviations

ADF

acid detergent fiber

ADG

average daily gain

BCFA

branched-chain fatty acids

BW

body weight

CFU

colony-forming unit

CP

crude protein

DM

dry matter

DMI

dry matter intake

EMPS

efficiency of microbial protein synthesis

fBW

final body weight

FDR

false discovery rate corrected P-values

G:F

gain to feed

iBW

initial body weight

iNDF

indigestible neutral detergent fiber

LY

live yeast

MCP

microbial crude protein

MDCP

mono-dicalcium phosphate

NDF

neutral detergent fiber

NDFD

NDF digestibility

OM

organic matter

OTU

operational taxonomic unit

PD

purine derivatives

PUN

plasma urea nitrogen

SCFA

short-chain fatty acids

Conflict of interest statement

No conflicts of interest, financial, or otherwise are declared by the authors.