Effects of inoculation of corn silage with Lactobacillus spp. or Saccharomyces cerevisiae alone or in combination on silage fermentation characteristics, nutrient digestibility, and growth performance of growing beef cattle

Abstract

This study evaluated the effects of a novel silage inoculant containing Saccharomyces cerevisiae strain 3 as a direct fed microbial (DFM) on the ensiling, aerobic stability, and nutrient digestibility of whole-crop corn silage and growth performance of beef cattle. Treatments included uninoculated corn silage (CON) or corn silage inoculated with a mixture of 1.1 × 10⁵ cfu g⁻¹ fresh forage Lactobacillus plantarum and Lactobacillus buchneri (INOC1) or 1.0 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC2) or a mixture of INOC1 and INOC2 (INOC3). Silage in INOC1 had lower (P = 0.03) proportion of lactate, with acetate (Ac) proportion ranking as INOC1 > INOC3 > INOC2 (P < 0.01). In terminal silage, numbers of lactic acid bacteria were greater (P = 0.05) for INOC1 than CON and INOC2, while yeast counts tended (P = 0.08) to be greater for INOC2 than INOC3 on day 3 of aerobic exposure. Aerobic stability of corn silage was not impacted by inoculation with S. cerevisiae strain 3. Heifers fed INOC2 and INOC3 had lower (P < 0.01) ruminal Ac concentration than those fed CON. Apparent total tract digestibilities of DM, OM, ADF, and NDF were greater (P ≤ 0.03) for heifers fed INOC2 than those fed CON. Growth performance was similar across treatments, excepting DMI as percent of BW tended to be lower (P = 0.08) for INOC2 steers compared to CON steers. These results suggest that S. cerevisiae strain 3 has potential as a component in a fourth generation DFM silage inoculant.

Introduction

Use of bacterial inoculants during ensiling can make a significant contribution to forage conservation. These inoculants can contribute to the preservation of nutrients in silage, improving its feed value and enhancing the production efficiency of cattle (Fellner et al., 2001; Addah et al., 2011; Jin et al., 2015). A number of direct fed microbials (DFM) including Saccharomyces cerevisiae have been shown to improve the health status and production performance of cattle (McAllister et al., 2011; Buntyn et al., 2016). Direct fed microbials are generally administered as an encapsulated bolus or mixed with the feed (McAllister et al., 2011). Introduction of DFM into the diet through silage could be a unique approach to administrating these additives to cattle.

Our previous studies showed that inoculation of corn forage with strains of S. cerevisiae and/or Saccharomyces paradoxus at rates of 10³ to 10⁵ cfu g⁻¹ fresh weight did not affect nutrient composition, aerobic stability, or the in vitro ruminal fermentation of corn silage. Moreover, the inoculated strains did survive ensiling and, in fact, proliferated during aerobic exposure (AE;Duniere et al., 2015; Xu et al., 2019). Therefore, these yeast strains have potential as a component of the next generation of silage inoculants (fourth generation) that express a probiotic effect in addition to those effects achieved by the preceding 3 generations of silage inoculants (Duniere et al., 2015). However, to date, evaluation of fourth generation inoculants has been limited to laboratory scale silos and in vitro laboratory procedures. The objectives of the study were to assess the effects of inoculation of S. cerevisiae alone or in combination with Lactobacillus spp. on the ensiling and aerobic stability of corn silage, and to determine its effect on ruminal fermentation, total tract nutrient digestibility, and the growth performance of feedlot cattle.

Materials and Methods

Animal Care and Management

Experimental protocols involving animals were reviewed and approved by the Lethbridge Research and Development Centre (LeRDC) animal care committee (ACC protocol # 1710) and animals were cared for as per the guidelines of Canadian Council on Animal Care (CCAC, 2009).

Forage

Corn (Zea mays; P7958 AM; Pioneer Inc., Johnston, IA) was planted on May 12, 2017 under irrigation near Lethbridge, AB, and harvested on September 20, 2017 at two-thirds milk line maturity (32.7% DM). Corn was planted at a depth of 4 to 5 cm, with 38 cm row spacing at a density of 75,000 seeds ha⁻¹. The forage was harvested and chopped to 9.5 mm with a forage harvester (Claas Jaguar 950 Forage Harvester, Harsewinkel, Germany) equipped with a kernel processor with the rollers adjusted to 1.0 mm clearance. Chopped forages were ensiled on the same day in both minisilos and Ag-Bag silos (Ag-Bag Int. Ltd., Warrenton, OR).

Determination of the Effects of Inoculants on Ensiling Fermentation of Whole-crop Corn

Minisilos were prepared as described by Duniere et al. (2015). Briefly, chopped and kernel processed corn forage was divided into four 25-kg lots and placed on separate clean plastic sheets. Lots were either sprayed with distilled water (uninoculated control corn; CON) or with strains of lactic acid producing bacteria (LAB)Lactobacillus [11C33 Rapid React (DuPont Pioneer, Johnston, IA) containing 1.0 × 10⁵ cfu g⁻¹ fresh forage Lactobacillus buchneri and 1.0 × 10⁴ cfu g⁻¹ fresh forage Lactobacillus plantarum; INOC1], inoculant containing S. cerevisiae strain 3 (DuPont Pioneer, Johnston, IA; INOC2) or a mixture of equal portions of INOC1 and INOC2 (INOC3). The Saccharomyces strain used in the present study was screened from about 1,700 yeast isolates of Pioneer Hi-Bred’s microbial culture collection and was isolated from whole-plant corn silage from Quebec, Canada in 1998 (Duniere et al., 2015). All inoculants were applied at a rate of 25 mL per lot, providing 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB, 1.0 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3, and 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB + 1.0 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 for INOC1, INOC2, and INOC3, respectively. Each lot was thoroughly hand-mixed to ensure uniform inoculation. The mixing was repeated for a total of 3 mixes per treatment. Approximately 2.5 to 3.0 kg of forage from each lot was packed into 4 minisilos (10.4 cm in diameter × 35.6 cm in height) using a hydraulic press to achieve a density of ~ 240 kg m⁻³. Each of the 4 minisilos served as sample for each sampling day. There were a total of 48 minisilos prepared generating 3 replicate minisilos per treatment for each of the 4 sampling days. Each labeled minisilo was weighed before and immediately after filling and sealing. Silos were stored at ambient temperature (20 °C) and opened after 7, 30, 60, and 90 d of ensiling. Before ensiling (day 0), triplicate samples of fresh forage were collected per treatment for chemical and microbial analysis. Silos were weighed before opening on each sampling day and recorded weights used to calculate DM loss. At sampling, the contents of each minisilo were thoroughly mixed and subsampled for chemical and microbial analyses.

Determination of the Effects of Inoculants on Aerobic Stability of Whole-crop Corn Silage

Silage samples from minisilos opened on day 90 of ensiling were used for the assessment of aerobic stability. Subsamples of silage from triplicate minisilos per treatment were combined (~1.2 kg) and placed into triplicate 4-L insulated containers (13.5 cm in diameter × 30.9 cm in height), covered with 2 layers of cheese cloth and stored at ambient temperature (20 °C) for 21 d. Two Dallas Thermochron iButton sensors (Embedded Data Systems, Lawrenceburg, KY) were embedded in the silage at ~ 9.0 cm and ~ 18.0 cm within each container to measure the temperature every 4 h for 21 d. Two sensors were also placed in the room to measure ambient temperature. The contents of each container were subsampled after 3, 7, 14, and 21 d of AE for determinations of pH, chemical, and microbial compositions. Aerobic stability was defined as the number of hours before the temperature of aerobically exposed silage exceeded ambient temperature by 2 °C (Teller et al., 2012).

Determination of the Effects of Inoculants on Ruminal Fermentation and Total Tract Nutrient Digestibility of Whole-crop Corn Silage

Preparation of whole-crop corn silage in Ag-Bag silos

Whole-crop corn described above was harvested and inoculants were applied with ATV sprayers (AG Spray Equipment, Hopkinsville, KY) at a rate of 40 mL t⁻¹, providing the same number of cfu g⁻¹ fresh forage of LAB and Saccharomyces for INOC1, INOC2, and INOC3 as for the minisilo study. Corn sprayed with distilled water at a rate of 40 mL t⁻¹ served as CON. To reduce the cross contamination of the baggers and trucks when the inoculant treatments were switched, the spraying system was flushed with tap water each time the treatments were changed. The treated corn forage was ensiled in 2 Ag-Bag silos (3.0 m × 45.7 m) using 2 Ag-Bag baggers (Ag-Bag, St. Nazianz, WI) with each bag for 2 treatments. Untreated corn forage was introduced into the bag (~3 m) between treatments to avoid cross contamination of treatments and this region was marked on the outside of the bags. To minimize differences in forage composition due to harvest location and time of harvest, forage was chopped at random within the field and loads were randomly delivered to each of the Ag-Bag baggers. Upon delivery to the baggers, forage from each truck load was sampled. Approximately 200 t of each treatment were compressed into the Ag-Bag silos. The forage was ensiled for a minimum of 75 d before being fed.

Animals, treatments, experimental design, and feeding management

Eight ruminally cannulated Angus × Hereford cross-bred beef heifers (442.4 ± 21.9 kg; Mean ± SD) were individually housed at the metabolism barn. The treatments were 4 total mixed rations (TMR) composed of 65.0% of the 4 above prepared corn silages, plus 17.0% barley grain, 13.0% canola meal, and 5.0% vitamin-mineral supplement (DM basis). All diets were formulated to meet or exceed the NASEM (2016) nutrient requirements for CP, energy, minerals, and fat-soluble vitamins for growing beef heifers. Calcium to phosphorus ratio was formulated to range from 1.5:1 to 2:1 across treatments. Monensin sodium was provided to achieve 33 mg kg⁻¹ (DM basis) in all the diets and was incorporated in the vitamin–mineral supplement. The experiment was designed as a replicate 4 × 4 Latin square with 2 heifers per treatment and with heifers in each square receiving one of the 4 corn silage-based diets (i.e., CON, INOC1, INOC2, or INOC3).

Heifers were housed in tie stalls and were provided with daily exercise in an open dry pen, except during the sample collection period. Heifers were moved to the sampling pens during total collection in each period where they were tethered to the pen, but were provided with adequate space to stand, eat, drink, and lie down.

The experiment was conducted for 84 d with 4 periods of 21 d each. Each period consisted of 14 d of dietary adaptation, 5 d of total collection of feces and urine, and 2 d of ruminal sampling. Voluntary intake was measured from day 9 to 14. Diets were mixed using a Calan Data Ranger (American Calan, Northwood, NH) and delivered once daily at 0900 h for ad libitum intake with a target of 5% feed refusal during dietary adaptation. Bunks were cleaned before the morning feeding, orts weighed and recorded and subsampled for determination of DM content during the 14-d adaptation period. From day 15 of each period, heifers were fed at 90% of voluntary intake to ensure consumption of all feed. All heifers were weighed at the beginning and end of each adaptation period to calculate DMI as percent of BW. Silage samples were collected weekly and DM was determined and adjustments for DM content were undertaken as necessary. Dry rolled barley grain, canola meal, and supplement samples were collected every 2 wk and subsamples of the TMR were collected during the sample collection period. All samples of feed, TMR, and orts were composited on a period basis and subsamples were stored for later analysis.

Determination of total tract digestibility and N retention

Total collection of urine and feces were carried out from day 15 to day 19 of each period. Heifers were fitted with bladder catheters (Bardex 75 cc Lubricath 2-way Foley catheter, C. R. Bard Inc., Covington, GA) 24 h before the start of total collection. Urinary catheters were attached to 20-L plastic cans containing 500 mL of 4.0 N sulfuric acid so as to prevent volatilization of urinary ammonia. Urine output was recorded daily, mixed thoroughly and 1% was subsampled throughout the total collection period and stored at −20 °C. At the end of each period, the composite urine sample was thawed, mixed thoroughly, and subsampled for analysis of urinary N. Total fecal collection was carried out by scraping feces off the floor every 2 h from 0600 to 2200 h and every 4 h thereafter. Daily fecal output was recorded, mixed thoroughly and 2.5% was subsampled throughout the total collection period into plastic bags. At the end of each period, the composite fecal samples were thawed, mixed thoroughly, and subsampled per animal per period for later analysis.

Measurement of rumen metabolites

Ruminal fluid samples were collected at 1, 3, 8, and 24 h after feeding on day 20 and 21 of each period. Ruminal fluid (~250 mL) was collected from 4 different regions of the rumen (ventral, anterior, posterior sacs, and the ruminal mat) and strained through 4 layers of cheese cloth. The pH of the ruminal fluid was measured and recorded immediately using a portable pH meter (Model 265A, Orion Research Inc., Beverly, MA). Two 5-mL samples of ruminal fluid were collected with one sample being mixed with 1 mL of 25% (wt/vol) metaphosphoric acid solution for VFA analysis and the second with 1% (vol/vol) aqueous solution of 18.4 M sulfuric acid for measurement of ammonia.

Determination of the Effects of Inoculants on Growth Performance of Feedlot Cattle

Animal, treatments, and experimental design

A total of 60 Angus × Hereford cross-bred steers (260.9 ± 14.8 kg; Mean ± SD) were purchased from commercial sources and housed in individual feeding pens at the LeRDC Individual Feeding Barn. Upon arrival, steers were ear-tagged and processed as described by Wang et al. (2017). Steers were stratified by weight and assigned to 2 blocks of 30 steers each. Steers in each block were randomly assigned to one of the same 4 diets used in the digestibility study, resulting in 15 steers per treatment. The treatments were arranged as completely randomized block design.

Feeding management and measurements

Feed was delivered using a Calan Data Ranger (American Calan, Northwood, NH) to each pen once daily starting at 0900 h. The steers were fed for ad libitum intake with a target of 5% feed refusal. Feeders were examined each morning and the daily feed allotted was based on the residual feed in the feeder and the amount fed the previous day. Orts were collected weekly, weighed, and subsampled for DM to adjust for DMI. Steers were weighed on 2 consecutive days at the beginning and end of an 84-d growth study and every 28 d throughout the study period, before the morning feeding, to calculate the growth rate. Samples of diets and corn silage were collected weekly, with the diet samples being stored for chemical analysis and the silage being used to measure DM to ensure that the DM content of the diet remained constant throughout the experiment. Dry rolled barley grain, canola meal, and supplement samples were collected every 2 wk. All samples of feed and TMR were composited on a monthly basis and representative samples saved for chemical analysis.

Laboratory Analyses

Microbial analyses of corn silage

Microbial analyses were conducted as described by Addah et al. (2011). Briefly, samples (10 g) of fresh corn forage collected on the day of ensiling, silage from minisilos, and aerobically exposed silage samples collected on each sampling day were added to 90 mL of sterile 70 mM potassium phosphate buffer (pH = 7.0) and agitated for 30 s at 260 × rpm in a Stomacher 400 Laboratory Blender (Seward Medical Limited, London, UK). The suspension was serially diluted (10⁻² to 10⁻⁷) and 100 µL aliquots of each dilution were spread in triplicate onto de Man, Rogosa, and Sharpe agar (MRS; Dalynn Biologicals, Calgary, AB) for the enumeration of LAB, onto nutrient agar (NA; Dalynn Biologicals, Calgary, AB) for the enumeration of total bacteria (TB) and on to Sabouraud’s dextrose agar (SDA; Dalynn Biologicals, Calgary, AB) for the enumeration of yeast and mold. Both MRS and NA plates contained 200 µg mL⁻¹ of cycloheximide (Dalynn Biologicals, Calgary, AB), while SDA plates contained 100 µg mL⁻¹ each of tetracycline and chloramphenicol. Both MRS and NA plates were incubated at 37 °C for 24 to 48 h while SDA plates were incubated at ambient temperature for 72 h. Colonies were counted from plates containing a minimum of 30 and maximum of 300 colonies.

Chemical analysis.

Corn silage (15 g) from the minisilos and AE containers was mixed with 135 mL of deionized water and blended at full speed for 30 s in a Waring blender (Waring Commercial, Torrington, CT). The suspension was filtered through 2 layers of cheese cloth and the pH of the collected fluid was measured twice using a Symphony pH meter (SB70P,VWR, Mississauga, ON). One portion (7.5 mL) of the filtrate was immediately boiled for 10 min to stop fermentation and stored at −20 °C for subsequent analysis of water-soluble carbohydrates (WSC) by the Nelson–Somogyi method (Nelson, 1944) using a Dynatech MRX microplate reader (Dynatech Laboratories Inc., Chantilli, VA). The second portion was stored on ice until centrifuged at 10,000 × g for 15 min at 4 °C in a Sorvall Legend Mach 1.6R centrifuge (Thermo electron Corporation, Gormley, ON). The collected supernatant was used for analysis of VFA, lactate (LA) and ammonia as described by Addah et al. (2016).

Samples of TMR, barley grain, vitamin-mineral supplements, and orts from the performance and digestibility studies were ground to pass through a 1.0-mm screen using a Wiley Mill (Model 4, Arthur H. Thomas Co., Philadelphia, PA) and analyzed for DM, OM, total nitrogen (N), ADF, NDF, and starch as described by Addah et al. (2016) and Wang et al. (2017). Urinary N was analyzed using NA1500 Nitrogen/Carbon analyzer (Carlo Erba Instruments, Milan, Italy).

Ruminal fluid for VFA was processed and analyzed as described by Addah et al. (2016). The concentration of each VFA (mM) was measured by comparing their peak areas with that of malonic acid as an internal standard. Lactic acid was methylated and quantified as described by Kudo et al. (1987), using the same column and chromatograph used for VFA analysis. Ruminal ammonia concentrations were measured by the phenol-hypochlorite procedure of Broderick and Kang (1980).

Data Calculations and Statistical Analysis

Microbial populations were estimated as cfu g⁻¹ silage DM and were log-transformed before statistical analysis. Dietary NE_m content was calculated based on animal performance using the retained energy formula for medium frame yearling steers [RE = (0.0493 × BW^0.75) × ADG^1.097; NRC 1996] as per Zinn et al. (2002). Net energy of gain was calculated from NE_m assuming NE_g = NE_m × 0.877 − 0.41 as per Zinn and Shen (1998). Apparent total tract nutrient digestibility (ATTD) was calculated using the following equation: ATTD = [(nutrient intake – nutrient output in the feces)/nutrient intake] × 100. To account for gut fill, BW of steers are reported on a shrunk BW basis (BW × 0.96).

All data were statistically analyzed by analysis of variance using the MIXED procedure of SAS (version 9.3.1; SAS Inst. Inc., 2012). For the minisilo experiment, the effect of treatment on nutrient composition, fermentation parameters, and microbial data during ensiling or AE was assessed using a repeated measures analysis with the effect of treatment (T), days of ensiling or AE (D), and T × D included in the model. Minisilos (n = 3) and 4-L insulated containers (n = 3) containing aerobically exposed silage were used as the experimental units for fermentation and aerobic stability parameters, respectively. For the digestibility study, a Latin square design was used for the analysis of nutrient composition, DMI, total tract nutrient digestibility, and N balance data. A repeated measures analysis was carried out for ruminal VFA profile, ruminal pH, and ammonia with the fixed effect of time and treatment × time included in the model. Covariance structure with lowest Akaike’s and Bayesian information criteria value was selected (Littell et al., 1996) for the repeated measure analysis. For the performance study, nutrient composition and performance data were analyzed as a randomized complete block design with treatment as a fixed effect and individual pen as the experimental unit. Denominator degrees of freedom were determined using the Kenward–Roger option. Means were separated using the Tukey’s test. Significant differences and trends were declared at P < 0.05 and 0.10 > P > 0.05, respectively.

Results

Ensiling Fermentation Characteristics and Aerobic Stability of Whole-crop Corn Silage

Chemical composition, fermentation products and microbial populations in terminal (90 d of ensiling), and aerobically exposed (21 d of AE) silages are presented in Table 1. There was a T × D interaction for silage pH (Fig. 1A) with pH of CON (3.75) silage being lower (P = 0.03) than that of INOC1 (3.79) while the pH of INOC2 (3.74) was lower (P ≤ 0.03) than that of INOC 1 and INOC3 (3.78) after 60 d of ensiling. Similarly, pH of CON (3.78) and INOC2 (3.79) was lower (P ≤ 0.01) than that of INOC1 (3.98) after 90 d with the pH of INOC3 silage being intermediate (3.89).

Table 1.

Chemical composition, fermentation products, and microbial populations of fresh corn forage and corn silage inoculated with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination during ensiling in minisilos, and upon aerobic exposure (AE)

		Silage after 90 d of ensiling2						Silage after 21 d of AE2
Item4	Corn forage1	CON	INOC1	INOC2	INOC3	SEM3	P-value	CON	INOC1	INOC2	INOC3	SEM3	P-value
pH	5.72	3.78b	3.98a	3.79b	3.89ab	0.024	<0.01	8.00	8.33	7.82	8.15	0.254	0.57
DM	32.7	31.8	31.9	32.7	32.1	0.91	0.90	30.6	30.7	28.7	32.4	2.23	0.72
DM loss, %	NA	2.99	4.73	1.24	2.59	1.663	0.60	NA	NA	NA	NA	NA	NA
WSC, mg g−1 DM	9.42	0.85	0.59	1.40	0.69	0.202	0.09	1.09	0.98	1.38	1.10	0.154	0.35
NH3-N, mg g−1 DM	0.14	1.09	1.15	1.08	1.22	0.056	0.35	0.45	0.50	0.50	0.43	0.060	0.65
Fermentation products, mg g−1 DM
Acetate	NA	16.0c	27.3a	14.0c	22.3b	0.67	< 0.01	2.02	2.45	2.69	2.36	0.342	0.59
Propionate	NA	0.06	0.11	0.06	0.04	0.025	0.26	0.01	0.12	0.09	0.07	0.028	0.11
Butyrate	NA	0.00	0.12	0.07	0.02	0.043	0.27	0.82	0.91	1.03	0.76	0.126	0.48
Total VFA	NA	16.1c	27.6a	14.2c	22.4b	0.67	< 0.01	2.85	3.49	3.82	3.20	0.447	0.50
Lactate	NA	62.6a	41.9b	64.6a	50.7ab	3.59	0.03	1.08	1.40	1.35	1.32	0.344	0.91
LA:Ac ratio	NA	3.91a	1.56b	4.62a	2.28ab	0.186	< 0.01	0.49	0.58	0.50	0.54	0.109	0.93
Ethanol	NA	10.7	11.0	10.0	11.4	1.49	0.93	0.37b	1.01b	5.67a	4.71a	0.524	< 0.01
Microbiology, log10 cfu g−1 DM
Total bacteria	7.26	8.58	9.31	8.65	9.09	0.131	0.08	10.2	10.1	10.2	10.1	0.10	0.80
LAB	6.36	8.60bc	9.33a	8.57c	9.19ab	0.136	0.05	8.82	8.88	9.45	9.45	0.250	0.21
Yeasts	6.86	4.62	3.84	5.34	3.54	0.350	0.13	9.21	9.04	9.41	9.32	0.130	0.30
Mold	6.33	0.00	1.59	0.00	0.82	0.574	0.23	5.41	5.28	7.42	7.65	1.906	0.73
Aerobic stability5, h	NA	NA	NA	NA	NA	NA	NA	54.7b	68.0ab	50.7b	84.0a	6.04	0.01

¹Values for fresh forage were not included in statistical analysis.

²Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh weight L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh weight L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh weight lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC2) or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh weight LAB + 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC3).

³SEM, pooled standard error of mean (n = 3).

⁴LAB, lactic acid bacteria; LA;Ac ratio, ratio of lactic acid to acetic acid; NA, not applicable; NH₃-N, ammonia nitrogen; WSC, water-soluble carbohydrates.

⁵Hours after a 2 °C rise in temperature was recorded in the silage mass after exposure to air.

^a,bWithin a row, means without a common letter differ (P < 0.05) among treatments within ensiling or AE.

Figure 1.

Impact of inoculation of corn silage with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination on (A) pH, (B) lactic, and (C) acetic acid concentrations during ensiling and aerobic exposure (AE; n = 3). Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh weight L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh weight L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh weight lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC2); or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh weight LAB + 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC3). All treatments were ensiled at two-thirds milk line maturity and ensiled for 90 d in minisilos; Area to the left of the dotted vertical line indicates ensiling and to the right indicates the period of AE. (A) *Denote treatment differences in pH during ensiling. pH of CON was lower (P = 0.03) than that of INOC1 while that of INOC2 was lower (P ≤ 0.03) than that of INOC 1 and INOC3 on day 60 of ensiling. pH of CON and INOC2 was lower (P ≤ 0.01) than that of INOC1 on day 90 of ensiling. (B) *Denote treatment differences in lactate concentrations during ensiling. Lactate concentration (mg g⁻¹ DM) of INOC1 was lower (P = 0.03) than that of INOC2 on day 90 of ensiling.

(C) *Denote treatment differences in acetate (Ac) concentrations (mg g⁻¹ DM) during ensiling. Acetate concentration of CON and INOC2 was lower (P < 0.01) than that of INOC1 and INOC3 on day 90 of ensiling. Acetate concentration of INOC1 was also greater (P < 0.01) than that of INOC3 on day 90 of ensiling.

Lactate concentration (mg g⁻¹ DM) of INOC1 was lower (P = 0.03) than that of INOC2 silage after 90 d of ensiling (Fig. 1B). There was a T × D interaction for Ac concentration during ensiling with a tendency (P ≤ 0.09) for CON (12.6 mg g⁻¹ DM) and INOC2 (12.9 mg g⁻¹ DM) to have a lower Ac concentration than INOC1 (18.0 mg g⁻¹ DM) and INOC3 (17.9 mg g⁻¹ DM) after 60 d of ensiling (Fig. 1C). Acetate concentration of CON and INOC2 was lower (P < 0.01) than that of INOC1 and INOC3 after 90 d of ensiling. Acetate concentration of INOC1 was also greater (P < 0.01) than that of INOC3 after 90 d of ensiling. Lactate:acetate ratio (LA:Ac ratio) during ensiling was lower (P < 0.01) for INOC1 than CON and INOC2, with INOC3 being intermediate. Propionate and butyrate concentrations remained low in all silages during ensiling. Total VFA concentration (mg g⁻¹ DM) was greatest (P < 0.01) for INOC1 followed by INCO3 which were all greater (P < 0.01) than CON and INOC2 after 90 d of ensiling. Ethanol concentration (mg g⁻¹ DM) averaged 10.8 ± 2.26% across treatments during ensiling and was greater (P < 0.01) in INOC2 and INOC3 than CON and INOC1 after 21 d of AE. Concentration of LA, individual and total VFA decreased across treatments after 21 d of AE.

There was a T × D interaction for TB counts during ensiling and AE with INOC1 (9.07 log₁₀ cfu g⁻¹ DM) having a greater (P < 0.01) TB count than CON (8.82 log₁₀ cfu g⁻¹ DM) and INOC2 (8.85 log₁₀ cfu g⁻¹ DM) after 30 d of ensiling (data not shown). Similarly, there was a tendency for TB counts to be greater (P = 0.08) in INOC1 (9.31 log₁₀ cfu g⁻¹ DM) silage than CON (8.58 log₁₀ cfu g⁻¹ DM) after 90 d of ensiling. Moreover, TB count of INOC3 (9.44 log₁₀ cfu g⁻¹ DM) was greater (P ≤ 0.03) than CON (8.55 log₁₀ cfu g⁻¹ DM) and INOC2 (8.50 log₁₀ cfu g⁻¹ DM) silage after 3 d of AE. There was a T × D interaction for LAB counts (log₁₀ cfu g⁻¹ DM) during ensiling and AE with INOC1 (9.40 log₁₀ cfu g⁻¹ DM) having greater (P < 0.01) LAB counts than CON (8.93 log₁₀ cfu g⁻¹ DM) and INOC2 (8.98 log₁₀ cfu g⁻¹ DM) after 30 d of ensiling (Fig. 2A). The LAB counts were greater (P < 0.01) for INOC1 (9.44 log₁₀ cfu g⁻¹ DM) and INOC3 (9.53 log₁₀ cfu g⁻¹ DM) than CON (8.49 log₁₀ cfu g⁻¹ DM) and INOC2 (8.50 log₁₀ cfu g⁻¹ DM) after 60 d of ensiling. Similarly, LAB counts were greater (P = 0.05) for INOC1 than INOC2 after 90 d of ensiling. Moreover, LAB counts were greater (P ≤ 0.01) for INOC1 (9.26 log₁₀ cfu g⁻¹ DM) than CON (8.61 log₁₀ cfu g⁻¹ DM) and INOC2 (8.55 log₁₀ cfu g⁻¹ DM) while that of INOC3 (9.15 log₁₀ cfu g⁻¹ DM) was greater (P = 0.02) than INOC2 and tended (P = 0.06) to be greater than CON after 3 d of AE. There was a tendency (P = 0.08) for greater yeast count (log₁₀ cfu g⁻¹ DM) for INOC2 (8.56 log₁₀ cfu g⁻¹ DM) than INOC3 (6.46 log₁₀ cfu g⁻¹ DM) after 3 d of AE (Fig. 2B). Mold counts did not vary (P > 0.10) among treatments and remained low during ensiling, while the numbers increased across treatments by day 21 of AE (Fig. 2C). Aerobic stability of INOC3 was greater (P = 0.01) than that of CON and INOC2 (Fig. 3). There was a tendency (P = 0.07) of higher temperature for CON (29.9 °C) than for INOC3 (20.5 °C) on d 3 of AE (Fig. 3).

Figure 2.

Impact of inoculation of corn silage with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination on (A) lactic acid bacteria (LAB), (B) yeast, and (C) mold counts during ensiling and aerobic exposure (AE; n = 3). Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh weight L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh weight L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh weight LAB (INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC2); or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh weight LAB + 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC3). All treatments were ensiled at two-thirds milk line maturity and ensiled for 90 d in minisilos; Area to the left of the dotted vertical line indicates ensiling and to the right indicates the period of AE. (A) *Denote treatment differences in LAB counts (log₁₀ cfu g¹ DM) during ensiling and AE. LAB counts for INOC1 were greater (P < 0.01) than that of CON and INOC2 on day 30 of ensiling. LAB counts were greater (P < 0.01) for INOC1 and INOC3 than CON and INOC2 on day 60 of ensiling. LAB counts were greater (P = 0.05) for INOC1 than INOC2 on day 90 of ensiling. LAB counts were greater (P ≤ 0.01) for INOC1 than CON and INOC2 while that of INOC3 was greater (P = 0.02) than that of INOC2 on day 3 of AE. (B) *Denote treatment differences in yeast counts (log₁₀ cfu g⁻¹ DM) during aerobic exposure (AE). There was a tendency (P = 0.08) for yeast count of INOC2 to be greater than that of INOC3 on day 3 of AE.

Figure 3.

Impact of inoculation of corn silage with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination on aerobic stability (n = 3). Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh weight L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh weight L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh weight lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC2) or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh weight LAB + 1 × 10⁴ cfu g⁻¹ fresh weight S. cerevisiae strain 3 (INOC3). All treatments were harvested at two-thirds milk line maturity and ensiled for 90 d in minisilos and exposed to air for 21 d for evaluation of aerobic stability. *Denote treatment differences in temperature (°C) during aerobic exposure (AE). There was a tendency (P = 0.07) for temperature of CON to be greater than that of INOC3 on day 3 of AE.

Rumen Metabolites, Nutrient Digestibility, and N Retention

Dietary OM content of CON, INOC2, and INOC3 used in the digestibility study was greater (P < 0.01) than that of INOC1 (Table 2). The CP content of INOC2 diet was lower (P < 0.01) than that of CON, and INOC1, with INOC3 intermediate. The ADF (P = 0.23) and NDF (P = 0.13) content did not differ among diets and averaged 22.4 ± 1.03% and 41.1 ± 1.55%, respectively.

Table 2.

Chemical and nutrient composition of corn silage, inoculated with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination and ensiled in Ag-Bag silos, and diets, used for the digestibility and performance studies

Item	Treatment1
	CON	INOC1	INOC2	INOC3	SEM2	P-value
Nutrient composition of corn silage, % DM basis
DM	30.7 ± 1.88	30.3 ± 2.09	33.0 ± 2.12	32.1 ± 1.47	-	-
OM	94.3 ± 0.34	94.3 ± 0.25	95.0 ± 0.17	94.8 ± 0.29	-	-
CP	8.65 ± 0.74	8.61 ± 0.18	8.21 ± 0.42	8.28 ± 0.55	-	-
ADF	23.1 ± 2.83	23.8 ± 2.85	22.5 ± 2.76	23.2 ± 1.30	-	-
NDF	44.2 ± 2.90	45.5 ± 3.26	42.4 ± 0.85	43.4 ± 1.53	-	-
Starch	33.6 ± 2.16	31.2 ± 4.52	33.9 ± 2.53	32.3 ± 2.95	-	-
Diet formulation for the digestibility and performance study, % DM basis
Corn silage	65.0	65.0	65.0	65.0
Barley grain	17.0	17.0	17.0	17.0	-	-
Canola meal	13.0	13.0	13.0	13.0	-	-
Supplement3	5.0	5.0	5.0	5.0	-	-
Nutrient composition of the diets for the digestibility study, % DM basis
OM	93.0a	92.5b	93.2a	93.1a	0.13	< 0.01
CP	15.5a	15.5a	14.7b	15.1ab	0.19	< 0.01
ADF	22.0	22.7	21.8	22.7	0.35	0.23
NDF	41.1	41.9	39.8	41.0	0.49	0.13
Starch	23.9	23.6	24.9	23.8	0.68	0.50
Nutrient composition of the diets for the performance study, % DM basis
OM	92.0	92.2	92.6	92.5	0.17	0.08
CP	16.7	16.9	16.2	16.2	0.33	0.33
ADF	22.6	24.5	21.6	22.9	0.87	0.18
NDF	40.7	41.5	39.5	39.9	0.61	0.12
Starch	22.2	21.3	22.8	21.4	1.11	0.39

¹Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh weight L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh forage L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh forage lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC2) or combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB + 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC3).

²SEM, pooled standard error of mean (n = 4).

³Supplement (as fed basis) contained 30.0% ground barley, 29.1% canola meal,26.0% limestone, 6.5% urea, 5.0% salt, 1.5% canola oil, 1.5% vitamin-mineral premix, 0.3% Rumensin 200 and 0.1% Vitamin E 500.

^a,bWithin a row, means without a common letter differ (P < 0.05).

There was no treatment × sampling time interaction, but time postfeeding did differ (P < 0.01) for a number of ruminal fermentation parameters (Table 3). Ruminal pH did not differ (P = 0.99) among heifers and averaged 6.37 ± 0.39 across treatments and time. Proportion (mol/100 mol) of ruminal Ac was greater (P < 0.01) for CON than INOC2 and INOC3 with INOC1 being intermediate. Proportions of propionate (P = 0.35), butyrate (P = 0.17), isobutyrate (P = 0.79), valerate (P = 0.18), and isovalerate (P = 0.90) did not differ among treatments. Total VFA concentrations did not differ (P = 0.98) among treatments and time and averaged 97.5 ± 23.4 mM. Similarly, ruminal ammonia concentration did not differ (P = 0.98) among treatments and across sampling times with an average of 9.83 ± 5.53 mg dL⁻¹.

Table 3.

Ruminal pH, proportions of VFA, total VFA concentrations, acetate to propionate ratio (A:P), and ruminal ammonia concentration of heifers fed corn silage inoculated with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination

Item	Treatments1					P-value3
	CON	INOC1	INOC2	INOC3	SEM2	TRT	TIME	TRT × TIME
Ruminal pH	6.40	6.38	6.36	6.35	0.049	0.90	< 0.01	0.99
VFA, mol/100 mol
Acetate (A)	67.1a	65.8ab	63.6b	64.1b	0.74	< 0.01	< 0.01	0.15
Propionate (P)	18.4	19.2	20.3	19.9	0.78	0.35	< 0.01	0.27
Butyrate	10.5	10.9	11.9	11.7	0.50	0.17	< 0.01	0.65
Isobutyrate	1.02	0.96	0.99	1.00	0.046	0.79	< 0.01	0.90
Valerate	1.08	1.13	1.20	1.17	0.042	0.18	< 0.01	0.69
Isovalerate	1.71	1.68	1.70	1.77	0.095	0.90	< 0.01	0.97
Total VFA, mM	96.5	97.7	98.8	97.0	3.18	0.96	< 0.01	0.98
Lactate, mM	0.04	0.04	0.03	0.04	0.005	0.78	< 0.01	0.58
A:P Ratio	3.70	3.53	3.23	3.31	0.165	0.19	< 0.01	0.30
Ruminal NH3-N, mg dL−1	10.9	10.3	8.90	9.20	1.056	0.54	< 0.01	0.98

¹Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh forage L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh forage L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh forage lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC2) or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB + 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC3).

²SEM, pooled standard error of mean, n = 4 heifers/treatment.

³ P-value for the effect of treatment (TRT), TIME and the TRT × TIME interaction.

^a,bWithin a row, means without a common letter differ (P < 0.05).

Dry matter intake and intake as a % of BW averaged 11.4 ± 1.44 kg and 2.30 ± 0.24% respectively across heifers (Table 4). The digestibilities of DM (P = 0.03), OM (P = 0.02), ADF (P = 0.03), and NDF (P = 0.03) of heifers fed INOC2 were all greater than those fed CON. In contrast, starch digestibility of heifers fed INOC1 was greater (P = 0.02) than those fed INOC2. The CP digestibility averaged 64.4 ± 3.17% and did not differ (P = 0.62) among treatments.

Table 4.

Dry matter intake and apparent total tract nutrient digestibility of beef heifers fed silage inoculated with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination

Item	Treatments1
	CON	INOC1	INOC2	INOC3	SEM2	P-value
DMI
kg d1	11.1	11.4	11.5	11.2	0.50	0.69
% of BW	2.25	2.32	2.40	2.22	0.087	0.31
Apparent nutrient digestibility coefficient, %
DM	63.1b	64.3ab	66.5a	65.1ab	0.93	0.03
OM	65.5b	66.7ab	69.1a	67.4ab	0.92	0.02
CP	63.1	64.6	64.7	64.5	1.15	0.62
ADF	33.3b	38.0ab	41.0a	35.9ab	1.91	0.03
NDF	47.4b	48.3ab	52.2a	49.0ab	1.54	0.03
Starch	96.3ab	98.0a	96.0b	97.5ab	0.71	0.02

¹Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh forage L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh forage L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh forage lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC2) or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB + 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC3).

²SEM, pooled standard error of mean, n = 4 heifers/treatment.

^a,bWithin a row, means without a common letter differ (P < 0.05).

There was a tendency (P = 0.06) for greater total N intake for heifers fed INOC1 than those fed INOC2 (Table 5). Fecal N excretion of heifers fed INOC2 and INOC3 was lower (P = 0.03) than that of CON and INOC1. However, urinary N as a percent of either total N ingested (P = 0.48) or total N digested (P = 0.60) did not differ among treatments. Total N excretion did not differ (P = 0.22) among treatments and averaged 169.9 ± 31.2 g d⁻¹. Similarly, apparent total N retained (P = 0.31) and N efficiency (P = 0.26) did not differ among treatments and averaged 54.9 ± 15.6 g d⁻¹ and 24.9 ± 7.3%, respectively.

Table 5.

Nitrogen (N) excretion and apparent N retention of beef heifers fed corn silage inoculated with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination

Item	Treatments1
	CON	INOC1	INOC2	INOC3	SEM2	P-value
Fecal output, kg DM d−1	3.37	3.33	3.18	3.12	0.192	0.11
Urine output, kg d−1	7.98	8.79	8.07	7.69	0.785	0.34
Nitrogen, g d−1
Total intake N	222.2	232.0	213.5	214.8	9.54	0.06
Total excreted N	171.6	182.2	165.4	160.3	11.20	0.22
Fecal N	83.4a	82.6a	76.4b	76.2b	4.50	0.03
% of excreted N	48.6	44.6	48.4	47.8	1.79	0.38
Urinary N	89.5	102.0	86.7	86.3	7.91	0.38
% of intake N	39.1	44.0	38.8	39.1	2.75	0.48
% of digested N	61.4	68.0	60.2	61.1	4.49	0.60
% of excreted N	51.4	55.4	51.8	52.2	1.79	0.38
Apparent total retained N	64.3	47.9	56.3	53.0	5.75	0.31
Retained as a % of intake N	28.1	20.8	25.8	25.2	2.53	0.26
Retained as a % of digested N	38.6	32.0	39.8	38.9	4.49	0.60

¹Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh forage L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh forage L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh forage lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC2) or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB + 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC3).

²SEM, pooled standard error of mean, n = 4 heifers/treatment.

^a,bWithin a row, means without a common letter differ (P < 0.05).

Growth Performance of Feedlot Cattle

There was a tendency (P = 0.08) for greater OM content for diet containing INOC2 than CON (Table 2). Contents of other nutrients did not differ among diets with CP (P = 0.33), ADF (P = 0.18), NDF (P = 0.12), and starch (P = 0.39) content of the diets averaging 16.5 ± 0.68%, 22.9 ± 1.89%, 40.4 ± 1.31%, and 21.7 ± 1.77%, respectively.

Initial (P = 0.98) and final shrunk BW (P = 1.00), ADG (P = 0.99), DMI (P = 0.34), and ADG:DMI (P = 0.44) did not differ among diets and averaged 250.6 ± 15.57 kg, 371.5 ± 26.12 kg, 1.44 ± 0.20 kg d⁻¹, 6.91 ± 0.59 kg, and 0.208 ± 0.019, respectively (Table 6). The DMI as a % BW tended (P = 0.08) to be lower for steers fed INOC2 than those fed CON. The NE_m (P = 0.20) and NE_g (P = 0.18) content of diets calculated based on animal performance did not differ among treatments and averaged 2.00 ± 0.13 Mcal kg⁻¹ DM and 1.35 ± 0.12 Mcal kg⁻¹ DM, respectively.

Table 6.

Performance parameters of growing beef steers fed corn silage inoculated with Lactobacillus spp. or Saccharomyces cerevisiae strain 3 alone or in combination

Item	Treatment1
	CON	INOC1	INOC2	INOC3	SEM2	P-value
Number of steers	14	14	15	14	-	-
Initial shrunk BW3, kg	249.4	250.0	251.3	251.6	4.21	0.98
Final shrunk BW3, kg	371.3	370.9	372.6	371.2	7.11	1.00
ADG, kg d−1	1.45	1.44	1.44	1.43	0.054	0.99
DMI, kg	7.11	6.79	6.76	6.98	0.155	0.34
DMI as % BW	2.30	2.19	2.16	2.24	0.038	0.08
ADG:DMI	0.203	0.210	0.213	0.204	0.0049	0.44
NEm, Mcal kg−1 DM4	1.96	2.03	2.05	1.97	0.035	0.20
NEg, Mcal kg−1 DM4	1.31	1.37	1.39	1.32	0.030	0.18

¹Treatments included corn silage with no inoculant (CON); corn silage inoculated with 1 × 10⁵ cfu g⁻¹ fresh forage L. buchneri and 1 × 10⁴ cfu g⁻¹ fresh forage L. plantarum for a total of 1.1 × 10⁵ cfu g⁻¹ fresh forage lactic acid bacteria (LAB; INOC1); corn silage inoculated with 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC2) or a combination of INOC1 and INOC2 at equal proportions containing 1.1 × 10⁵ cfu g⁻¹ fresh forage LAB + 1 × 10⁴ cfu g⁻¹ fresh forage S. cerevisiae strain 3 (INOC3).

²SEM, pooled standard error of mean, n = 15 steers per treatment.

³Shrunken BW calculated as 96% of live weight (NRC 2000).

⁴Calculated based on performance of beef steers (Zinn and Shen 1998; Zinn et al., 2002).

Discussion

First-generation silage inoculants consisted of homofermentative LAB such as L. plantarum, which predominantly produced LA and accelerated the decline in silage pH during ensiling (Pahlow et al., 2003; Addah et al., 2014). Second-generation silage inoculants consisted, in addition to homofermentative LAB, heterofermentative LAB such as L. buchneri which improved aerobic stability of silages during feedout through the production of Ac during ensiling (Reich and Kung 2010; Addah et al., 2014). Silage inoculants, where L. buchneri possessed ferulic acid esterase activity for improved digestion of forage fiber in ruminants, constituted the third generation of silage inoculants (Addah et al., 2012). The fourth generation of silage inoculants, in addition to possessing the characteristics of previous 3 generations, could act as DFM to benefit the health and production efficiency of ruminants (Duniere et al., 2015). Saccharomyces cerevisiae and its metabolites are often included in the diets of growing and finishing beef cattle (Hinman et al., 1998; Keyser, 2006; Ovinge et al., 2018). In a meta-analysis of the effect of live yeast supplementation on beef cattle performance, Sartori et al. (2017) reported that although live yeast supplementation did not increase the ADG, feed efficiency was improved due to a reduction in DMI.

The S. cerevisiae strain 3 used in the present study was the same as that used by Duniere et al. (2015). The strain was selected based on its ability to survive the ensiling process, fermentation end product profiles, and effects on ruminal fermentation in vitro. Our previous studies indicated an increase in counts of Saccharomyces in aerobically exposed corn silages (Duniere et al., 2015; Xu et al., 2019). Moreover, these strains did not affect the LAB population or nutrient composition during ensiling, or the aerobic stability of silage. The L. buchneri used in the present study did not possess ferulic acid esterase activity.

Effects of Inoculant on Ensiling Characteristics of Whole-crop Corn Silage

The decrease in silage pH during ensiling was similar across treatments with all treatments having a pH below 3.9 after 7 d of ensiling. The drop in silage pH corresponded to a concurrent decrease in WSC from 9.42 mg g⁻¹ DM in fresh corn forage to 1.04 ± 0.16 mg g⁻¹ DM after 7 d of ensiling (data not shown). Water-soluble carbohydrates are the principal substrates for epiphytic microorganisms during ensiling. The pH of INOC1 silage showed a slight increase from 3.79 on day 60 to 3.98 by day 90 of ensiling, while pH remained relatively constant for other treatments between day 7 and day 90 of ensiling. This is likely due to the fact that heterolactic L. buchneri can convert lactic acid to acetic acid (Ranjit and Kung, 2000; Kung, 2010) resulting in an increase in silage pH in the later stage of ensiling. A decrease in LA and an increase in Ac concentration in INOC1 silage from day 60 to day 90 is a reflection of the conversion of LA to Ac by L. buchneri. Lower LA and greater Ac concentrations in INOC1 silages resulted in lower LA:Ac ratio than CON and INOC2 silages. In a meta-analysis of the effects of L. buchneri on the fermentation and aerobic stability of silages, Kleinschmit and Kung (2006) reported that the LA:Ac ratio of corn silage ranged from 2.3:1.0 to 1.3:1.0 for low (≤ 1.0 × 10⁵ cfu g⁻¹ fresh forage) and high (≥ 1.0 × 10⁵ cfu g⁻¹ fresh forage) doses of L. buchneri respectively, with control silage having LA:Ac ratio of 3.0:1.0. High yeast numbers are generally associated with high concentrations of ethanol as yeasts ferment WSC to ethanol during ensiling. However, terminal silage in this study contained similar ethanol concentrations across treatments, likely indicating that yeasts in INOC2 and INOC3 remained dormant and metabolically less active during ensiling. Similar concentrations of ethanol for uninoculated control and S. cerevisiae inoculated corn silage during ensiling have also been reported (Duniere et al., 2015; Xu et al., 2019). Mold were undetected in all silages on day 7 and day 30 of ensiling. A similar decline in mold counts during ensiling was also reported by Duniere et al. (2015), where molds were undetected by plate counts after 7 and 28 d of ensiling in inoculated and uninoculated corn silages.

Silage fermentation parameters, TB and LAB counts for corn silage inoculated with INOC2 were similar to that of CON. These results indicate that S. cerevisiae did not adversely impact the fermentation of corn silage. Similar results were also reported by Duniere et al. (2015) and Xu et al. (2019) where the TB and LAB counts did not change in corn silage inoculated with strains of S. cerevisiae relative to control silage. The greater TB and LAB counts for INOC1 and INOC3 than for INOC2 can be explained by the fact that both INOC1 and INOC3 contained strains of LAB whereas INOC2 contained only yeast.

Effects of Inoculants on the Aerobic Stability of Whole-crop Corn Silage

Aerobic stability of corn silage inoculated with S. cerevisiae alone (INOC2) was similar to that of CON (with no inoculant), but lower than that of INOC3, with INOC1 being intermediate. The greater aerobic stability of INOC1 and INOC3 is likely due to the fact that these 2 treatments were inoculated with L. buchneri and contained a greater Ac concentration relative to CON and INOC2 silages. Acid-utilizing yeast can degrade LA in the presence of oxygen (Kung, 2010). Degradation of LA increases the silage pH and facilitates proliferation of opportunistic bacteria and mold (McDonald et al., 1991). As yeast are the principal microorganisms that cause spoilage in silages and as Ac acts as an antifungal agent, silages with high concentration of Ac (INOC1 and INOC3) are expected to be aerobically more stable than treatments with low Ac (CON) or high yeast counts (INOC2). However, yeast associated with aerobic deterioration of silages which utilize lactic acid as a substrate are generally Pichia or Candida genera (Jonsson and Pahlow, 1984). Apparently, LA assimilating Pichia and Candida spp. have a competitive advantage over non-LA assimilating S. cerevisiae upon AE (Pahlow et al., 2003) and tend to outgrow this species under aerobic conditions when WSC are limited. Similar observations were also reported by Xu et al. (2019) in a recent evaluation of impact of S. cerevisiae and L. buchneri on microbial communities during ensiling and AE, where Kazachstania, not Saccharomyces, were primarily associated with the aerobic deterioration of corn silage. Lack of any adverse effect on the aerobic stability of corn silage inoculated with Saccharomyces increases the potential for these strains being used as component of DFM silage inoculants.

Yeast counts tended to be greater for INOC2 than INOC3 in aerobically exposed (day 3 of AE) silages. This is likely attributable to the acetic acid produced by L. buchneri in INOC3 inhibiting the proliferation of yeast. An increase in counts of S. cerevisiae upon AE of corn silage was also reported by Xu et al. (2019), who reported 10³ to 10⁵ fold increase in the numbers of yeast in corn silage inoculated with S. cerevisiae as compared to uninoculated corn silage. It is likely that the S. cerevisiae strain 3 survived the acidic conditions during ensiling and proliferated upon AE. These observations increase the possibility of delivering greater numbers of this Saccharomyces strain as a DFM through silage. Although this Saccharomyces strain may possess DFM properties, it does not appear to contribute significantly to silage fermentation. Combining Saccharomyces with third generation LAB, such as L. buchneri with ferulic acid esterase activity, could also improve the aerobic stability and fiber digestibility in silage.

Effects of Inoculants on Rumen Metabolites, Total Tract Nutrient Digestibility, and N Retention

Our laboratory previously assessed the effects of S. cerevisiae on ensiling and microbial ecology using laboratory minisilos (Duniere et al., 2015, Xu et al., 2019). This study evaluated the effects of this strain at a farm production scale and its potential impact on nutrient digestion and the growth performance of feedlot cattle.

The similar DMI (kg d⁻¹ or % of BW) of heifers across treatments indicated that the inoculants used in this study did not affect silage palatability or feed intake. The average value of DMI in this study was 2.26% of BW (2.16 to 2.40), which is within the normal range for growing feedlot cattle fed a corn silage-based diet (Jesse et al., 1976; Tjardes et al., 2002). Ruminal pH remained ≥ 6.35, indicating that it remained well above the levels associated with the inhibition of fiber digestion or the development of subclinical ruminal acidosis (Nocek, 1997; Penner et al., 2007). Acetate:propionate ratio was also > 3.0 across all treatments, reflecting the relatively high level of silage in the diet. Ruminal microbial activity was unlikely impacted by treatment as indicated by an average ruminal ammonia concentration of 9.83 ± 5.53 mg dL⁻¹, which is more than the recommended minimum concentration of 5 mg dL⁻¹ for normal ruminal microbial activity (Satter and Slyter, 1974).

An improvement in apparent total tract DM, OM, ADF, and NDF digestibility for heifers fed INOC2 than CON suggests that the yeast in the silage inoculant likely acted as DFM, modulating ruminal fermentation and enhancing the ruminal environment so as to promote fiber digestibility. An increase in total tract nutrient digestibility in dairy cattle with the addition of live Saccharomyces to the diet has been previously reported (Marden et al., 2008). These authors reported that feeding 57:43 forage:concentrate diet along with 5 g of live S. cerevisiae (10¹⁰ cfu g⁻¹ DM, Biosaf Sc 47, Lesaffre Feed Additives, Marquette-Lez-Lille, France) tended to improve apparent total tract DM and OM digestibility and increased total tract ADF and NDF digestibility in lactating dairy cattle. It was reported that yeast supplementation stabilized the ruminal environment by decreasing ruminal LA and increasing ruminal pH as compared to cows fed the control diet. However, ruminal pH and LA concentration were similar across treatments in the present study. Allen and Ying (2012) reported that lactating cows with high DMI supplemented with S. cerevisiae fermentation product had lower rate of ruminal starch digestion. Similarly, Dias et al. (2018) reported numerically lower total tract starch digestibility for lactating cows fed low (23% diet DM) or high (29% diet DM) starch diets along with yeast culture relative to those fed control diets. It was concluded that yeast supplementation stabilized the ruminal environment by reducing the rate of ruminal starch digestion. Total tract starch digestibility of heifers fed INOC2 was lower than that of INOC1 in the present study. Moreover, an increase in apparent total tract ADF and NDF digestibility in heifers fed INOC2 supports the possibility that yeast enhanced the activity of ruminal cellulolytic bacteria (Dawson et al., 1990; Newbold et al., 1995).

The tendency for the high N intake of heifers fed INOC1 likely reflects the greater CP content of INOC1 (15.5% vs. 14.7%) as compared to INOC2 diets, while the DMI of heifers remained similar across these treatments. Similar amounts of total N excretion for heifers likely reflect the similar apparent total tract N digestibility across treatments. Apparent total N retained and N efficiency in the present study were greater than expected for the lean tissue deposition observed in the performance study. Apparent total N retention of 48 to 86 g d⁻¹ corresponds to a gain in excess of 2 kg d⁻¹ (Walter et al., 2012). However, the ADG of steers fed similar treatments in the concurrent performance study averaged 1.63 ± 0.32 across treatments. Similar N retention values (49.4 to 95.8 g d⁻¹) were also reported by Zenobi et al. (2015) where heifers were fed 50:50 forage:concentrate diet. Over estimates for apparent total N retention were likely attributed to potential sampling errors including underestimation of fecal N caused by incomplete sample collection, volatile losses of NH₃-N from the pen, and loss of N during drying of the fecal samples (Spanghero and Kowalski, 1997). Kohn et al. (2005) reported that volatilization of urinary NH₃-N during total collection may overestimate retained N. However, in the present study, urine was acidified with 4 N sulfuric acid during total collection to keep the urine pH below 2.0. It should also be noted that the average CP content of diets in the present study was 15.1 ± 0.61% (DM basis) which is greater than what is required for growing beef cattle. High dietary N is associated with increased fecal and urinary N excretion (Vasconcelos et al., 2009).

Effects of Inoculants on Growth Performance of Feedlot Cattle

Similar to the digestibility experiment, DMI across the silage diets indicated that the inoculants did not affect the feed intake of feedlot cattle. This was contrary to Sartori et al. (2017) who in a meta-analysis of the effect of live yeast on beef cattle performance reported that they resulted in a decrease in DMI, had no effect on ADG, and improved feed efficiency of beef cattle depending on diet composition, strain of yeast, and dosage. These authors reported that steers fed high-forage diets (51 to 75%, DM basis) along with live yeast at 10⁹ cfu g⁻¹, tended to have lower DMI. However, yeast concentration in the diets of this study would be much lower given that the silage was inoculated with 1 × 10⁴ cfu g⁻¹ and that silage only accounted for 65% of the TMR. An improvement in backgrounding performance was expected for steers fed INOC2 based on results of the digestibility study as apparent total tract DM, OM, ADF, and NDF digestibility was increased for heifers fed silage inoculated with Saccharomyces. An increase in total tract DM digestibility is often associated with an increase in DMI (Zobell et al., 2005). However, the DMI of steers fed INOC2 in the present study did not differ from those fed the other silages. It should be noted that the average DMI among steers in the present study was 6.76 ± 0.16 kg, while that for heifers in the concurrent digestibility study was 11.4 ± 0.50 kg. This difference is attributable to the difference of the BW between the 2 sets of beef cattle.

Results indicate that there is potential for using silage inoculants containing Saccharomyces in a manner that positively impact growth performance, ruminal fermentation, and total tract nutrient digestibility of growing beef cattle. Moreover, an increase in yeast count observed during AE for INOC2 likely indicates that the number of yeast will increase during AE, but it is clear that there is a balance between increasing DFM at the expense of a potential loss in silage quality. Increasing the dosage rate of Saccharomyces during ensiling, along with the observed increase during AE, will further increase the daily dose of this strain. Direct supplementation of yeast in the diet could be one approach to make up for any shortfalls in response as a result of inadequate levels of DFM in silage. A greater yeast count in terminal INOC2 indicated that the yeast strain survived the ensiling process and proliferated upon AE. Moreover, inclusion of Saccharomyces in the silage inoculant did not affect the LAB counts during ensiling or silage stability upon AE. Heifers fed INOC2 displayed an increase in apparent total tract DM, OM, ADF, and NDF digestibility. However, there was no improvement in feedlot performance of steers fed INOC2 relative to those fed other treatments, likely due to comparatively lower DMI of steers and therefore lower yeast intake, relative to that of heifers in the digestibility study. These results indicate that there is potential for using S. cerevisiae strain 3 in fourth generation silage inoculants to provide live yeast to serve as DFM upon feedout. However, further research is needed for the combination of strains of LAB and Saccharomyces, number of cfu, and rate of application for optimizing the beneficial effects for the development of the next generation silage inoculant.

Conflict of Interest

Authors declare that there are no conflicts of interests.