Effect of chemical and biological preservatives and ensiling stage on the dry matter loss, nutritional value, microbial counts, and ruminal in vitro gas production kinetics of wet brewer’s grain silage

Abstract

This study evaluated the effects of chemical and biological preservatives and ensiling stage on spoilage, ruminal in vitro fermentation, and methane production of wet brewer’s grain (WBG) silage. Treatments (TRT) were sodium lignosulfonate at 10 g/kg fresh WBG (NaL1) and 20 g/kg (NaL2), propionic acid at 5 g/kg fresh WBG (PRP, 99%), a combination inoculant (INO; Lactococcus lactis and Lactobacillus buchneri each at 4.9 log cfu per fresh WBG g), and untreated WBG (CON). Fresh WBG was treated and then ensiled for 60 d, after which mini silos were opened and aerobically exposed (AES) for 10 d. Data were analyzed as an RCBD (five blocks) with a 5 TRT × 3 stages (STG; fresh, ensiled, and AES) factorial arrangement. Results showed that ensiled PRP-treated WBG markedly preserved more water-soluble carbohydrates and starch than all other ensiled TRT (P < 0.001). Dry matter losses of ensiled PRP-treated WBG were 48% lower than all other ensiled TRT (P = 0.009) but were not different than CON in AES (P = 0.350). Due to its greater concentration of digestible nutrients, PRP-treated AES was less aerobically stable than CON (P = 0.03). Preservation was not improved by INO, NaL1, or NaL2 but the latter prevented the increase of neutral detergent fiber across STG (P = 0.392). Apparent in vitro DM digestibility (IVDMD) decreased only in ensiled CON, INO, and NaL1 relative to fresh WBG and AES NaL2 had greater IVDMD than all other AES TRT (P ≤ 0.032). In vitro ruminal fermentation of fresh WBG resulted in a greater methane concentration and yield than the other STG (P < 0.033). In conclusion, PRP was the most effective at preserving WBG during ensiling but failed to improve aerobic stability under the conditions tested.

Lay Summary

Wet brewer’s grain (WBG) is the most abundant byproduct in the manufacture of beer and its rich nutritional composition makes it a valuable feed for cattle. However, WBG is highly susceptible to spoilage so the application of cost-effective preservatives may be a viable approach to prevent nutrient losses during ensiling and feed out. The present study evaluated the effects of chemical and biological preservatives on the nutritional composition and in vitro fermentation and gas production of WBG across three silage production stages: fresh, ensiled, and aerobically exposed silage (AES). Preservatives tested were propionic acid, a bacterial inoculant, and sodium lignosulfonate (NaL) applied at 1% and 2%. Propionic acid successfully reduced the loss of nutrients and preserved more sugars and starch than all other treatments during ensiling, which resulted in higher digestibility in vitro. However, due to its greater concentration of digestible nutrients, ensiled WBG treated with propionic acid also suffered extensive spoilage in the AES. All other treatments failed to improve the preservation of ensiled or AES WBG, but NaL at 2% prevented the decrease of digestibility for AES.

Wet brewer’s grain for cattle feeding is highly susceptible to loss of quality and nutritive value during storage and feeding. Thus, this study seeks to evaluate the effectiveness of an array of preservatives to extend shelf life and improve digestibility at different stages of storage.

Introduction

Wet brewer’s grain (WBG) is the main by-product of the brewing industry, representing ~85% (fresh basis) of the total by-products generated in both large and small breweries (Buffington, 2014). Each year, in the United States alone, it is estimated that 4.2 million Mg of WBG (fresh basis) are produced if we take into consideration an annual production of 21 billion liters of beer (TTB, 2021) and a ratio of approximately 20 kg (fresh basis) of WBG being generated per every 100 L of beer produced (Mussatto et al., 2006). Due to its high concentration of protein (~28.5% of DM) and energy (~72% total digestible nutrients; NRC, 2001), WBG is regarded as a high-quality feedstuff when original nutrients are preserved adequately. Thus, its abundancy and high nutritional profile make this by-product an attractive option to decrease feeding costs, especially for ruminants.

Studies have shown that incorporating WBG as a replacement for other rumen undegradable protein (RUP) sources in the feed improves the final live weight, average daily gain, and the gain-to-feed ratio of finishing steers and heifers (Belon et al., 2019), and increases milk yield, milk fat content, milk fat yield, and milk total solids content of dairy cows (Belibasakis and Tsirgogianni, 1996; Getu et al., 2020). Moreover, research has shown that it can reduce enteric methane emissions when added as a replacement for canola meal in lactating dairy cattle diets, due to its greater fat concentration (Moate et al., 2011). Its high percentage of RUP and essential amino acids (30% of total protein) make it a valuable source of quality protein that can be combined with inexpensive sources of rumen degradable protein (RDP) to meet the requirements of ruminants (Huige, 2006; Mussatto, 2014). Additionally, utilization of WBG by farmers reduces the environmental impact related to its disposal in landfills (Chanie and Fievez, 2017), as well as saving breweries the expenses related to transporting waste to landfills, estimated at $16 per Mg of WBG for every 8 km traveled (Mussatto, 2014).

The main issue of feeding WBG has been its rapid spoilage when stored unprotected (within 7 to 10 d; Chanie and Fievez, 2017), particularly in summer months when higher ambient temperatures lead to increased microbial activity (Westendorf and Wohlt, 2002). For example, Wang et al. (2014) observed more than twice DM loss (27.8% vs. 11.3%) in WBG stored aerobically for 3 d at 35 °C relative to 15 °C, respectively, and on the first day of storage at 35 °C, the WBG had already suffered 10.6% loss of DM. Its high nutritional value and moisture concentration (~74%; NRC, 2001) make it especially susceptible to undesirable microorganisms, which can result in nutrient losses, offensive odors, and the potential production of mycotoxins (Orosz and Davies, 2015). Nonstructural carbohydrates in WBG are particularly vulnerable to spoilage processes, which often is correlated with a concomitant increase in the concentration of recalcitrant fiber (Wang et al., 2014). Thus, it is crucial to manage and store WBG appropriately in order to reduce spoilage. Drying can extend WBG shelf life considerably but it is costly, and most craft breweries do not have drying facilities (Chanie and Fievez, 2017). Alternatively, ensiling has been a common method used to extend the shelf life of WBG, but it is important to reach anaerobic conditions as close as possible to the time of production (Westendorf and Wohlt, 2002) while avoiding the high temperatures observed after lautering (at least 70 °C; Huige, 2006) that can interfere with preservatives. Due to its high initial moisture concentration, WBG silage suffers from high effluent losses and packing issues (Nishino et al., 2003). Thus, producers often ensile WBG mixed with dry feedstuffs, but previous studies have shown greater spoilage and nutrient losses in mixtures compared with WBG ensiled alone (Moriel et al., 2015; Ferraretto et al., 2018; Parmenter et al., 2018). Furthermore, Wang et al. (2014) observed that increased aerobic spoilage of fresh WBG resulted in decreased digestible nutrients (nonstructural carbohydrates), increased NDF concentration, and increased yeast and mold counts. Consequently, the aerobically spoiled WBG in that study had lower in vitro ruminal digestibility, volatile fatty acid (VFA) production, and higher ruminal butyric acid (Wang et al., 2014). These characteristics of spoiled feed resemble the characteristics of feeds that increase ruminal methane production as evaluated by Benchaar et al. (2001).

Silage preservatives have been evaluated to reduce spoilage of WBG during storage and feeding phases. For instance, the application of propionic acid has reduced DM losses and formation of ammonia-N (NH₃-N) and preserved more soluble carbohydrates than untreated WBG silage stored for up to 90 d (Allen and Stevenson, 1975; Schneider et al., 1995; Moriel et al., 2016). However, its high price and hazardousness (Rotz and Shinners, 2007) have limited its adoption, especially by small producers (Moriel et al., 2016). In the case of lactic acid bacteria (LAB) inoculants, studies have shown varying effects on WBG silage, but most of these studies have focused on homofermentative species (e.g., Lactobacillus plantarum) that have no effect on shelf life after opening (Schneider et al., 1995; Marston, 2007; Anderson et al., 2015). Remarkably, heterofermentative LAB (e.g., Lactobacillus buchneri) has not been directly evaluated with WBG silage despite its capacity to extend shelf life by producing acetic acid, a strong antifungal compound (Arriola et al., 2021). Moreover, Wang and Nishino (2009) stated that the presence of L. buchneri in total mixed ration silage containing WBG played a major role in improving its aerobic stability.

Given the limitations of the aforementioned preservatives, research on new potential preservatives for WBG is warranted. Recent studies on sodium lignosulfonate, a low-cost papermill byproduct, have shown antimicrobial (Núñez-Flores et al., 2012; Jha and Kumar, 2018) and antiproteolytic (Petit et al., 1999) properties. When tested in vitro, sodium lignosulfonate showed promising results at reducing spoilage of high-moisture alfalfa hay due to its fungistatic properties (Reyes et al., 2020). Similarly, Leon-Tinoco et al. (2020) observed that an optimized sodium lignosulfonate matched propionic acid at reducing DM loss of high-moisture alfalfa hay. Consequently, sodium lignosulfonate could have the potential to extend the aerobic stability of WBG. The objective of this study was to evaluate the effects of a wide array of preservatives on the nutritional composition and in vitro ruminal fermentation and methane production of WBG during the fresh, ensiled, and feed-out stages. Our hypothesis is that sodium lignosulfonate will match the preservation effects of propionic acid, especially after aerobic exposure and that a combination inoculant that includes L. buchneri will increase aerobic stability of WBG compared with untreated silage. Also, we hypothesize that ruminal methane production will increase as feed spoilage increases during storage due to a decreased concentration of nonstructural carbohydrates.

Materials and Methods

Protocol A2020-07-03 involving animal handling in this study was approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Maine.

WBG substrate and treatments

Fresh unroasted WBG consisting of mainly barley and wheat were provided by a local brewery from the manufacture of an Indian pale ale beer type in October 2019. The material was collected in five separate batches (blocks) of 80 kg each (fresh basis), directly from the lauter tun after the lautering process was completed. Since the temperature of the WBG coming from the tun was ~70 °C, it was spread on separate clean tarps, covered with another tarp, and allowed to cool overnight to 15 °C (ambient temperature) in an open barn. This was done to avoid the heat negatively affecting the inoculant bacteria or the volatilization of propionic acid. Each batch of WBG was divided into 5 piles which were allocated randomly to the 5 treatments, resulting in a total of 25 mini silos. The treatments (TRT) were untreated (CON) with no additives applied; sodium lignosulfonate generated from a neutral sulfite process applied at a rate of 10 g/kg (NaL1) and 20 g/kg (NaL2) (fresh basis); propionic acid (Alfa Aesar, Ward Hill, MA; 99%) applied at a rate of 5 g/kg (PRP; fresh basis); and a combination inoculant (INO) containing L. buchneri and Lactococcus lactis, applied as a solution at the rate of 1mL/kg of fresh WBG, which resulted in the actual delivery of 4.9 log cfu per fresh WBG g of each bacteria. The sodium lignosulfonate used in this experiment had 957 ± 0.49 g/kg of DM (fresh basis), and, on a dry matter basis, contained 324 ± 4.41 g/kg of ash (AOAC, 2000), 271 ± 5.82 g/kg of water-soluble carbohydrates (Dubois et al., 1956), 109 g/kg of acetic acid (Siegfried et al., 1984), 1.71 ± 0.05 g/kg of nitrogen (Baethgen and Alley, 1989), 19 g/kg of NDF (Van Soest et al., 1991), 0.6 g/kg of ether extract (AOCS, 2017), and 97.3 ± 2.83 µg/mg of total soluble phenolics (Singleton et al., 1999).

Silage preparation

Treated WBG (8.44 kg, fresh basis) was manually packed alone to a density of 233 kg DM/m³ in 8.8 L Leaktite HDPE containers (Leaktite, Leominster, MA) that were previously fitted with a gas trap on the lid, and a rubber gasket in the inner rim. Mini silos were stored at 21 °C ± 0.6 for 60 d in a dark room, after which they were opened, weighed, and aerobically exposed for 10 additional days under the same conditions.

Sampling procedure

Samples were collected from each mini silo during packing (fresh, day 0), after 60 d of ensiling (ensiled), and after exposing the silage to air for 10 d (AES, aerobically exposed silage), which represent critical stages (STG) in the production and feed-out of WBG silage. The empty buckets and full mini silos were weighed at each STG for the estimation of DM losses by subtracting the final from the initial DM weight (Queiroz et al., 2013). From each STG, a 1 kg sample (fresh basis) was taken for nutritional value analysis. Additional samples were taken for bacterial and fungal counts through plating techniques (100 g, fresh basis).

Laboratory analyses

Nutritional analysis

From samples taken at each STG, subsamples were processed for the determination of DM by drying at 60 °C until constant weight in a forced-air oven. Dried samples were ground to pass through a 1 mm screen in a Wiley mill (A. H. Thomas Company, Philadelphia, PA) and ash concentration of ground samples was determined according to AOAC (2000). Concentrations of NDF (Van Soest et al., 1991) and acid detergent fiber (ADF; AOAC, 2000) were measured sequentially using an ANKOM 200 Fiber Analyzer (ANKOM, Macedon, NY). Prior to the NDF procedure, sample bags were soaked in acetone for 10 min twice and air-dried to remove excess fats (Fahey et al., 2018). Heat-stable bacterial α-amylase (FAA, ANKOM Technology) was used for the NDF assay, but sodium sulfite was not used. Nitrogen concentration was determined by using the total Kjeldahl nitrogen digestion procedure and the digested samples were analyzed colorimetrically with the sodium salicylate-nitroprusside method (Baethgen and Alley, 1989). Crude protein was calculated by multiplying N concentration by 6.25 (Church, 1993). Starch concentration was determined by the enzymatic-colorimetric method described by Hall (2015). Ether extract (crude fat) concentration was determined by using an ANKOM XT15 Fat Extractor (ANKOM, Macedon, NY) following the method described by AOCS (2017).

Liquid extracts were prepared by mixing 25 g of WBG from subsamples with 225 mL of 0.1% sterile peptone water in a 400C Stomacher blender for 3 min (Seward Ltd., Worthing, UK). The solution was filtered through two layers of sterilized cheesecloth and the pH of the extract was measured with a calibrated Φ34 Beckman pH meter (Beckman, Brea, CA) fitted with an Accumet Universal pH electrode with an integrated temperature sensor (ThermoFisher Scientific, Waltham, MA). Afterward, a portion of the extracts was centrifuged at 8,000 × g for 20 min at 4 °C and the supernatant was frozen (−20 °C) until further analysis. For d 0 samples, the centrifuged extracts were acidified with 50% H₂SO₄ (1% v/v) before freezing. Extracts were analyzed for lactic, acetic, butyric, and propionic acids, and 1,2-propanediol and ethanol concentrations using an Agilent High-Performance Liquid Chromatograph 1200 series system fitted with an Agilent Hi-Plex H column (Agilent Technologies, Santa Clara, CA) coupled to an Agilent refractive index detector (Siegfried et al., 1984). Ammonia-N concentration was measured using an adaptation of the procedure outlined by Weatherburn (1967). Water-soluble carbohydrates were measured using the protocol outlined by Dubois et al. (1956) using sucrose as the standard as described by Hall (2003).

Microbial counts and aerobic stability measures

The aforementioned liquid extracts were also used for the enumeration of bacterial and fungal populations. An aliquot of these extracts was taken immediately after filtering with sterilized cheesecloth and used to perform serial (10-fold) dilutions in 0.1% sterile peptone water, which was then spread-plated (100 µL) on Man, Rogosa, and Sharpe agar (MRS; BD Difco, Franklin Lakes, NJ) with 0.01% cycloheximide (Ha et al., 1995) for LAB counts and in malt extract agar (BD Difco, Franklin Lakes, NJ) with 0.4% lactic acid for yeast and mold counts (Koburger, 1971). Plates were incubated for 24 h at 37 °C for LAB and for 72 to 120 h at 25 °C for yeasts and molds. WBG samples from AES were only analyzed for yeast and mold counts.

Aerobic stability was determined by transferring 2.5 kg of the freshly opened silage (day 60) into 8.8 L containers and placing a temperature sensor (Gemini Data Logger, UK) in the middle of the biomass, previously set to record data every 30 min for 10 d. Two additional sensors were placed in the temperature-controlled room (21 ± 0.67 °C). The buckets were left open, and the silage was covered with two layers of cheesecloth to avoid excessive drying. Aerobic stability was expressed as the amount of time (hours) before the silage was heated 2 °C above ambient temperature (Kung, 2010). The maximum temperature and heat degree-days (HDD) above room temperature of the biomass in the containers were determined according to Coblentz et al. (2013) during the recording period (10 d).

In vitro ruminal digestibility, fermentation, and gas production

In vitro incubations were performed in 250 mL glass bottles using the Ankom Gas Monitoring System (Ankom Technology, Macedon, NY) to determine DM digestibility and gas production kinetics of all 75 samples (25 mini silos sampled across 3 stages). The substrate for incubation consisted of 1.4 g of the aforementioned dried and ground samples, with one bottle per sample and bottles grouped by block (total of five incubation runs). The ruminal fluid was representatively collected by aspiration 3 h after feeding (1200 hours) from two ruminally cannulated Holstein cows in lactation consuming a ration consisting of corn silage (Zea mays L., 12.8 kg), timothy grass haylage (Phleum pratense L., 3.2 kg), and concentrate (10.6 kg, DM basis). The collected ruminal fibrous mat was blended with ruminal fluid under a constant CO₂ flush, filtered through two layers of cheesecloth (de Assis Lage et al., 2020), and mixed with McDougall’s artificial saliva at a buffer:ruminal fluid ratio of 2:1 (McDougall, 1948; Henry et al., 2015). A blank bottle with buffered ruminal fluid but no sample was included in each incubation run and the results of these blanks were subtracted from the sample values. A zero-module recording of the atmospheric pressure was placed in the incubator with the sample bottles. Samples were incubated for 24 h at a temperature of 39 °C and constant agitation at 60 rpm. The cumulative pressure increase was automatically recorded every 30 min by the Ankom modules and the global release of pressure was set to 2 psi. The total gas produced was collected in gas sampling bags (Supel-Inert Multi-Layer Foil, Supelco Inc, Bellefonte, PA) that were connected directly to the Ankom modules, following the protocol outlined by Henry et al. (2015). After incubation, fermentation was stopped by placing the bottles on ice. The contents of the bottles were centrifuged at 8,000 × g for 15 min and filtered through previously dried and weighed Whatman No. 541 ashless filter papers (Fisher Scientific, Pittsburgh, PA). Residues were dried at 105 °C to constant weight to determine apparent in vitro DM digestibility (IVDMD), and subsequently burned at 600 °C in a muffle furnace for apparent in vitro organic matter (OM) digestibility (IVOMD) determination. The filtrate liquid was measured for pH before being acidified with 50% H₂SO₄ (1% v/v) and stored for further analysis of NH₃-N and total VFA (acetate, propionate, butyrate, isobutyrate, isovalerate, and valerate) using the same HPLC as described earlier, but fitted with a diode-array detector (Castillo Vargas et al., 2021). The concentration of methane gas (CH₄) in the sampling bags was measured with gas chromatography (SRI 8610C Multigas #5 GC; SRI, Torrace, CA) using a flame ionization detector and an MXT-1 column (Restek, Bellefonte, PA). Samples were injected splitless through a 1 mL sample loop connected to a VICI auto-injector valve and processed isothermally at 36 °C (Lopez and Newbold, 2007). The resulting CH₄ produced was expressed as absolute mass (mg), concentration (mM), and yield (mmol/g of fermented OM). The volume of gas produced was calculated from the pressure measurements of the Ankom modules multiplied by the atmospheric pressure (zero-module). Gas production kinetics parameters included asymptotic maximal gas production (M_f expressed as per gram of DM), fractional rate of gas production (K_f, % per h), and lag phase (h), which were quantified using the following modified Gompertz model (Schofield et al., 1994; Henry et al., 2015):

$${\displaystyle \begin{array}{l}{\displaystyle V=M_{\mathrm{f}}\mathrm{e}\mathrm{x}\mathrm{p}\{-\mathrm{e}\mathrm{x}\mathrm{p}(1+\frac{K_{\mathrm{f}}e}{M_{\mathrm{f}}}(L-t))\},}\end{array}}$$

where V is the total gas volume produced during 24 h incubation; M_f is the asymptotic maximal volume of gas (corresponding to complete substrate digestion); K_f is the rate of gas produced; L is the lag time; and t is the incubation time.

Statistical analysis

Data were analyzed as a randomized complete block design (5 blocks) with a 5 TRT × 3 STG factorial arrangement of treatments, consisting of a total of 75 experimental units. The model used to analyze the data was:

$${\displaystyle \begin{array}{l}{\displaystyle Y_{ijk}=\mu +T_i+S_j+T{S}_{ij}+{\beta }_k+{\epsilon }_{ijk},}\end{array}}$$

where Y_ijk is the dependent variable, µ is the overall mean, T_i is the treatment effect (TRT), S_j is the stage effect (STG), TS_ij is the effect of the interaction between TRT × STG, β_k is the block effect, and ε_ijk is the residual error. The PROC GLM procedure of SAS v. 9.4 (SAS Institute Inc., Cary, NC) was used to analyze the data. When an interaction was present, the SLICE option was used to analyze simple effects. Mean separation was performed using the PDIFF procedure of LSMEANS. Significance was declared at P ≤ 0.05. Since LAB counts were only measured for fresh and ensiled, a 5 TRT × 2 STG factorial arrangement of treatments was used instead for the same model. Only main effects of TRT were analyzed for aerobic stability measures and DM loss within the applicable STG, using the following model:

$${\displaystyle \begin{array}{l}{\displaystyle Y_i=\mu +T_i+{\beta }_k+{\epsilon }_{ijk},}\end{array}}$$

where the symbols have the same meaning as above.

Results

Nutritional composition

All nutritional components were affected by a TRT × STG effect (P < 0.01; Table 1). In fresh WBG, DM concentration was higher for both NaL1, NaL2, and PRP compared with CON (P ≤ 0.036), and in ensiled, all TRT except INO had higher DM than CON (P ≤ 0.046). No differences in DM were observed across STG for either of the NaL, unlike the other TRT which increased DM from fresh to AES ($\overline{\mathrm{x}}=+12.6\mathrm{g}/\mathrm{k}\mathrm{g};$ P ≤ 0.024). It should be noted that the sodium lignosulfonate used in this experiment contained 957 g/kg of DM. Ensiled PRP successfully preserved 46% more WSC (Figure 1) and 37% more starch (Figure 2) than all other ensiled TRT (P < 0.001). In AES, PRP continued to preserve 34% more starch than all the other TRT (P ≤ 0.049), but no differences were observed between TRT for WSC (P ≥ 0.141). In fresh WBG, ether extract was reduced by NaL2 relative to other TRT (P < 0.001) but was increased after the ensiling period for all TRT relative to fresh (P ≤ 0.03). Within ensiled, PRP had lower ether extract levels compared with CON and INO (P ≤ 0.01) and only NaL1 and NaL2 increased levels further at AES, relative to the other TRT (P ≤ 0.028). Notably, after calculating the proportion of ether extract recovered in ensiled compared with fresh, it was observed that 111 ± 3.19% of fats were recovered, indicating some fat synthesis during ensiling, but no effects of TRT were observed (P = 0.105; data not shown). Furthermore, it is also important to mention that the concentration of NDF in NaL2-treated WBG did not increase across STG ($\overline{\mathrm{x}}=384;$ P = 0.392) and for NaL1, it only increased from fresh to ensiled (P < 0.001). In comparison, NDF of both CON and INO increased from fresh to ensiled and again at AES (P < 0.001). In the case of PRP, no difference in NDF were observed between fresh and ensiled ($\overline{\mathrm{x}}=407\mathrm{g}/\mathrm{k}\mathrm{g};$ P = 0.898), but the increase of NDF from ensiled to AES (P < 0.001) was large enough to match the AES levels observed for CON and INO ($\overline{\mathrm{x}}=525\mathrm{g}/\mathrm{k}\mathrm{g}$). The extent of proteolysis, measured as NH₃-N, during ensiling was the same across all TRT relative to fresh (P = 0.09). However, at AES more proteolysis was observed for NaL1 and NaL2 compared with PRP (P ≤ 0.003), although none of them were different from CON (P ≥ 0.056).

Table 1.

Effect of treatment and stage of storage on the nutritional composition of wet brewer’s grain silage^1,2

Item3	Treatment					SEM	P-value
	CON	INO	NaL1	NaL2	PRP		TRT	STG	TRT × STG
DM, g/kg WBG						2.81	<0.001	<0.001	0.003
Fresh	212Bc	215Abc	229a	235a	220Bb
Ensiled	221Ab	214Ab	229a	234a	236Aa
AES	227A	228B	230	236	230A
OM, g/kg DM						0.92	<0.001	<0.001	<0.001
Fresh	970Aab	969Ab	963Ac	953Ad	972Aa
Ensiled	963Ba	960Bb	960Bb	946Bc	964Ba
AES	957Cb	960Bab	950Cc	932Cd	961Ca
Ether extract, g/kg DM						2.17	0.008	<0.001	<0.001
Fresh	73.7Ba	72.7Ba	71.3Ca	62.4Cb	76.9Ba
Ensiled	91.8Aab	94.0Aa	86.4Bbc	80.6Bc	83.7Ac
AES	87.6Ab	92.2Ab	102Aa	99.1Aa	87.3Ab
NDF, g/kg DM						9.00	<0.001	<0.001	<0.001
Fresh	389C	385C	391B	376	406B
Ensiled	480Ba	473Ba	435Ab	394c	408Bc
AES	536Aa	518Aa	437Ab	382c	522Aa
ADF, g/kg DM						1.02	<0.001	<0.001	0.008
Fresh	12.5C	12.5C	12.2C	11.8C	12.6C
Ensiled	14.9Ba	14.5Ba	14.2Bab	13.2Bc	13.7Bbc
AES	18.2Aa	16.7Abc	16.5Ac	15.0Ad	17.4Aab
Hemicellulose, g/kg DM						0.83	<0.001	<0.001	<0.001
Fresh	26.4B	26.1B	26.8	25.9A	28.0B
Ensiled	33.1Aa	32.8Aa	29.3b	26.2Ac	27.0Bbc
AES	35.3Aa	35.1Aa	27.2b	23.2Bc	34.8Aa
CP, g/kg DM						4.21	<0.001	<0.001	0.01
Fresh	247Ca	247Ca	238Cab	230Cb	249Ba
Ensiled	274Ba	276Ba	259Bb	245Bc	246Bc
AES	315A	316A	312A	302A	306A
NH3-N, g/kg N						0.47	0.783	<0.001	0.002
Fresh	0.31B	0.35B	0.30C	0.20C	0.29C
Ensiled	5.20A	8.08A	4.65B	4.77B	8.58A
AES	6.37Aabc	5.72Abc	9.17Aab	11.2Aa	3.65Bc

CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis.

Fresh, fresh WBG prior to ensiling; ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening.

OM, organic matter; NDF, neutral detergent fiber; ADF, acid detergent fiber; CP, crude protein; NH₃-N, ammonia-N.

Means with different uppercase superscripts within a column are significantly different (P ≤ 0.05).

Means with different lowercase superscripts within a row are significantly different (P ≤ 0.05).

Figure 1.

Water-soluble carbohydrates (WSC; g/kg DM) of WBG at the different stages of storage (STG) as affected by TRT. CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis. Fresh, fresh WBG prior to ensiling; ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening. Bars represent means ± SEM. Uppercase letters depict differences across STD within TRT (CON: A, B, C; INO: F, G, H; NaL1: K, L, M; NaL2: R, S, T; PRP: X, Y, Z) and lowercase letters depict differences across TRT within STG (P ≤ 0.05).

Figure 2.

Starch (g/kg DM) of WBG at the different stages of storage (STG) as affected by TRT. CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis. Fresh, fresh WBG prior to ensiling; ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening. Bars represent means ± SEM. Uppercase letters depict differences across STD within TRT (CON: A, B, C; INO: F, G, H; NaL1: K, L, M; NaL2: R, S, T; PRP: X, Y, Z) and lowercase letters depict differences across TRT within STG (P ≤ 0.05).

Microbial populations, DM losses, and aerobic stability

The counts of LAB were higher for ensiled relative to fresh (P < 0.001) and, across STG, were lower for PRP vs. all other TRT (P < 0.001), but higher for INO vs. CON (P = 0.036; Table 2). Yeast counts increased from fresh to ensiled and again at AES (P < 0.001), but no effects of TRT were observed (Table 2). No molds were observed in any of the mini silos at any STG (<2 log cfu per g fresh WBG). During ensiling, only PRP decreased DM loss relative to CON (P = 0.009) but after aerobic exposure, INO, NaL1, and NaL2 had higher DM loss ($\overline{\mathrm{x}}=18.2\mathrm{\%}$) than both CON and PRP ($\overline{\mathrm{x}}=12.5\mathrm{\%};$ P < 0.002). Overall, the accumulated DM loss that occurred between fresh and AES WBG was higher for INO and NaL2 ($\overline{\mathrm{x}}=27.4$ ) relative to CON and PRP ($\overline{\mathrm{x}}=18.5\mathrm{\%};$ P = 0.003; Table 2). None of the TRT managed to increase the aerobic stability of WBG silage, and PRP was even less stable than CON by 28.2 h (P = 0.03). NaL1 had a higher HDD (P = 0.019) and NaL2 had a higher maximum temperature (P = 0.032) compared with CON.

Table 2.

Effect of treatment and stage of storage on dry matter loss, microbial counts, and aerobic stability measures^1,2

Item3,4	Treatment					Mean	SEM	P-value
	CON	INO	NaL1	NaL2	PRP			TRT	STG	TRT × STG
DMloss,%
Ensiled	9.07a	12.7a	9.82a	10.2a	5.06b		1.26	0.009	NA	NA
AES	13.3b	17.3a	18.3a	18.9a	11.7b		1.19	0.002	NA	NA
Accumulated	20.9bc	27.7a	26.2ab	27.1a	16.1c		1.97	0.003	NA	NA
LAB, log cfu/g							0.23	<0.001	<0.001	0.240
Fresh	-	-	-	-	-	5.42B
Ensiled	-	-	-	-	-	7.57A
AES	-	-	-	-	-	ND
Mean	6.46b	6.95a	6.74ab	6.70ab	5.62c
Yeast, log cfu/g							0.38	<0.001	0.112	0.178
Fresh	-	-	-	-	-	3.21C
Ensiled	-	-	-	-	-	4.19B
AES	-	-	-	-	-	7.41A
Aerobic stability measures
Aerobic stability, h	90.1ab	94.1ab	75bc	103.2a	61.9c		8.86	0.03	NA	NA
Max. Temp, °C	36.8bc	35.6c	38.4ab	39.1a	36.0c		0.71	0.01	NA	NA
HDD, °C-d	62.2bc	55.8c	86.0a	77.8ab	72.0abc		6.42	0.03	NA	NA

CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis. Fresh, fresh WBG prior to ensiling; Ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening.

Whenever the interaction TRT × STG was not significant (P > 0.05), the means of significant main effects (P ≤ 0.05) are presented instead.

Only TRT effect was evaluated for DM loss and aerobic stability measures. The accumulated DM loss was calculated at AES relative to the initial DM of fresh.

Aerobic stability, time (h) taken for the silage internal temperature to increase 2 °C above room temperature. HDD, heat-degree days; calculated as the sum of the daily temperature increments above room temperature. Max Temp, maximum temperature during aerobic exposure. NA, not applicable; ND, not determined.

Means with different uppercase superscripts within a column are significantly different (P ≤ 0.05).

Means with different lowercase superscripts within a row are significantly different (P ≤ 0.05).

Silage fermentation profile

All fermentation components were affected by a TRT × STG effect (Table 3). As expected, PRP decreased the pH of fresh WBG upon application compared with all other TRT (P < 0.001). After ensiling, pH was decreased to the same extent by all TRT relative to fresh (P < 0.001). Following aerobic exposure, pH was the highest for NaL2, followed by PRP and NaL1, and lowest for CON and INO(P ≤ 0.02). Taking into consideration that the WBG in this study was collected directly after the lautering process, we did not observe any lactic acid, ethanol (EOH), or 1,2-propanediol in fresh WBG above their detection limits ($\overline{\mathrm{x}}<0.6$, $\overline{\mathrm{x}}<2.5$, and $\overline{\mathrm{x}}<0.2\mathrm{g}/\mathrm{k}\mathrm{g}\mathrm{D}\mathrm{M}$, respectively). After ensiling, lactic acid was increased across all TRT to the same extent ($\overline{\mathrm{x}}=67.4\mathrm{g}/\mathrm{k}\mathrm{g}\mathrm{D}\mathrm{M};$ P < 0.001) relative to fresh. However, at AES, lactic acid levels decreased across all TRT relative to fresh (P < 0.001), albeit to a lesser extent for NaL1 and NaL2 relative to CON (P ≤ 0.002). Remarkably, acetic acid was higher for NaL2 vs. CON, INO, and PRP in fresh WBG (P ≤ 0.006). After ensiling, acetic acid levels increased relative to fresh only for CON and INO (P < 0.001) but decreased at AES for the same treatments (P ≤ 0.013). Consequently, the lactic-to-acetic acid ratio was highest for PRP, followed by NaL1, relative to all other TRT in ensiled WBG (P ≤ 0.029), and highest only for PRP relative to all other TRT at AES (P ≤ 0.002). Propionic acid was only detected in PRP-treated fresh WBG, which maintained the same levels at ensiled (P = 0.953) but decreased at AES (P < 0.001). Propionic acid was also increased by INO after ensiling (P = 0.008) and further at AES (P = 0.044). Relative to fresh WBG, all ensiled TRT showed increased ethanol levels (P < 0.001) except PRP (P = 0.093). After aerobic exposure, all TRT experienced large ethanol losses that resulted in trace levels (Table 3).

Table 3.

Effect of treatment and stage of storage on the fermentation profile of wet brewer’s grain^1,2

Item		Treatment				SEM	P-value
	CON	INO	NaL1	NaL2	PRP		TRT	STG	TRT × STG
pH						0.08	<0.001	<0.001	<0.001
Fresh	6.07Aa	6.04Aa	6.17Aa	6.22Aa	4.63Ab
Ensiled	3.39C	3.37C	3.50C	3.70C	3.49B
AES	4.27Bc	4.28Bc	4.58Bb	4.96Ba	4.56Ab
Lactic acid, g/kg DM						2.50	<0.001	0.004	0.021
Fresh	<0.56C	<0.56C	<0.56C	<0.56C	<0.56C
Ensiled	57.0A	65.2A	78.3A	69.6A	66.7A
AES	10.9Bc	15.7Bbc	22.3Bab	28.3Ba	15.9Bbc
Acetic acid, g/kg DM						2.13	<0.001	<0.001	<0.001
Fresh	<1.50Bb	<1.50Cb	5.28ab	8.68a	<1.50b
Ensiled	16.1Ab	24.7Aa	10.8b	14.8b	4.02c
AES	6.16Bc	16.9Ba	10.5bc	13.4ab	3.57c
Propionic acid, g/kg DM						0.95	0.103	<0.001	<0.001
Fresh	<1.50b	<1.50Cb	<1.50b	<1.50b	15.7Aa
Ensiled	2.02bc	3.70Bb	<1.50c	<1.50c	15.8Aa
AES	2.00b	6.46Aa	1.84b	2.77b	2.89Bb
Ethanol, g/kg DM						3.24	<0.001	0.035	0.005
Fresh	<2.50B	<2.50B	<2.50B	<2.50B	<2.50
Ensiled	31.3Aa	33.2Aa	31.0Aa	24.8Aa	9.35b
AES	<2.50B	<2.50B	<2.50B	<2.50B	<2.50
1,2-propanediol, g/kg DM						0.20	<0.001	0.001	<0.001
Fresh	<0.20B	<0.20B	<0.20B	<0.20B	<0.20
Ensiled	1.51Ab	4.00Aa	0.95Ab	3.28Aa	<0.20c
AES	<0.20B	<0.20B	<0.20B	<0.20B	<0.20

CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis.

Fresh, fresh WBG prior to ensiling; ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening.

Means with different uppercase superscripts within a column are significantly different (P ≤ 0.05).

Means with different lowercase superscripts within a row are significantly different (P ≤ 0.05).

In vitro ruminal digestibility, gas production, and fermentation profile

An effect of TRT × STG was observed only for IVDMD, IVOMD, pH (Table 4), lag (Table 5), and acetic and propionic acids (Table 6). All TRT had similar IVDMD in fresh WBG ($\overline{\mathrm{x}}=519\mathrm{g}/\mathrm{k}\mathrm{g}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{M}$), which was then decreased (P ≤ 0.034) in ensiled WBG only by CON, INO, and NaL1 ($\overline{\mathrm{x}}=451\mathrm{g}/\mathrm{k}\mathrm{g}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{M}$). Both NaL2 and PRP had higher IVDMD than CON in ensiled WBG but only NaL2 surpassed CON in AES WBG (P ≤ 0.017). Similarly, relative to fresh, ensiling decreased IVOMD only for CON, INO, and NaL1 (P ≤ 0.016). Aerobic exposure decreased IVOMD for all TRT but to a lesser extent for NaL2 relative to CON and NaL1 (P ≤ 0.009). Notably, ruminal NH₃-N levels were lowest at fresh, increased after ensiling, and were highest for AES (P < 0.001). The PRP-treated WBG had lower ruminal NH₃-N than CON and INO (P ≤ 0.009). In regards to gas production kinetics, M_f and K_f were highest for fresh, decreased after ensiling, and were lowest for AES (P < 0.001). Across STG, NaL1 had the lowest M_f relative to PRP, CON, and NaL2 (P ≤ 0.05), and the lowest K_f relative to CON, INO, and PRP (P ≤ 0.020). Ruminal methane production, concentration, and yield were greater in fresh vs. other STG (P ≤ 0.033).

Table 4.

Effect of treatment and stage of storage on the in vitro fermentation profile of wet brewer’s grain silage^1,2

Item3	Treatment					Mean	SEM	P-value
	CON	INO	NaL1	NaL2	PRP			TRT	STG	TRT × STG
IVDMD, g/kg of DM							23.9	0.035	<0.001	0.032
Fresh	501A	538A	555A	524A	479AB
Ensiled	421Bc	451Bbc	481Babc	505ABab	523Aa
AES	367Bbc	401Babc	350Cc	449Ba	425Bab
IVOMD, g/kg of OM							18.2	0.034	<0.001	0.026
Fresh	543A	567A	573A	551A	525A
Ensiled	463Bc	490Bbc	510Babc	528Aab	545Aa
AES	399Cbc	438Cab	386Cc	468Ba	440Bab
pH							0.02	<0.001	<0.001	<0.001
Fresh	6.47C	6.48C	6.50B	6.52C	6.48B
Ensiled	6.59Bab	6.57Bb	6.51Bc	6.62Ba	6.49Bc
AES	6.70Acd	6.74Ac	6.81Ab	6.86Aa	6.69Ad
NH3-N, mg/dL							1.07	0.026	<0.001	0.114
Fresh	-	-	-	-	-	27.2C
Ensiled	-	-	-	-	-	34.4B
AES	-	-	-	-	-	40.7A
Mean	36.2a	35.4a	33.9ab	32.7ab	30.3b

CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis. Fresh, fresh WBG prior to ensiling; Ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening.

Whenever the interaction TRT × STG was not significant (P > 0.05), the means of significant main effects (P ≤ 0.05) are presented instead.

IVDMD, in vitro dry matter digestibility; IVOMD, in vitro organic matter digestibility; NH₃-N, ammonia-N.

Means with different uppercase superscripts within a column are significantly different (P ≤ 0.05).

Means with different lowercase superscripts within a row are significantly different (P ≤ 0.05).

Table 5.

Effect of treatment and stage of storage on methane production and gas kinetics of wet brewer’s grain silage fermented in vitro for 24 h^1,2

Item3	Treatment					Mean	SEM	P-value
	CON	INO	NaL1	NaL2	PRP			TRT	STG	TRT × STG
Mf, mL/g incubated DM							7.81	<0.001	<0.001	0.058
Fresh	-	-	-	-	-	215A
Ensiled	-	-	-	-	-	181B
AES	-	-	-	-	-	155C
Mean	184b	182bc	169c	183b	200a
Kf, %/h							0.847	0.006	<0.001	0.09
Fresh	-	-	-	-	-	14.6A
Ensiled	-	-	-	-	-	12.6B
AES	-	-	-	-	-	9.81C
Mean	13.4a	13.0ab	11.0c	11.8bc	12.6ab
Lag, h							0.228	0.400	0.007	0.007
Fresh	0.15B	0.274	0.522	0.498	0.448B
Ensiled	0.148B	0.250	0.910	0.230	0.174B
AES	1.39Aa	0.644bc	0.430c	0.230c	1.15Aab
Methane production, mg							0.270	0.058	<0.001	0.775
Fresh	-	-	-	-	-	3.16A
Ensiled	-	-	-	-	-	2.43B
AES	-	-	-	-	-	2.00C
Methane concentration, mM							0.069	0.087	0.033	0.696
Fresh	-	-	-	-	-	0.944A
Ensiled	-	-	-	-	-	0.850B
AES	-	-	-	-	-	0.833B
Methane yield, mmol/g fermented OM							0.026	0.072	0.025	0.855
Fresh	-	-	-	-	-	0.273A
Ensiled	-	-	-	-	-	0.234B
AES	-	-	-	-	-	0.233B

CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis. Fresh, fresh WBG prior to ensiling; ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening.

Whenever the interaction TRT × STG was not significant (P > 0.05), the means of significant main effects (P ≤ 0.05) are presented instead.

M_f, asymptotic maximal gas production in a 24 h in vitro fermentation. K_f, rate of gas production.

Means with different uppercase superscripts within a column are significantly different (P ≤ 0.05).

Means with different lowercase superscripts within a row are significantly different (P ≤ 0.05).

Table 6.

Effect of treatment and stage of storage on volatile fatty acids of wet brewer’s grain silage fermented in vitro for 24 h^1,2

Item	Treatment					Mean	SEM	P-value
	CON	INO	NaL1	NaL2	PRP			TRT	STG	TRT × STG
Total VFA							0.06	0.618	<0.001	0.710
Fresh	-	-	-	-	-	109.4A
Ensiled	-	-	-	-	-	107.9B
AES	-	-	-	-	-	96.1C
Acetic acid, g/kg DM							1.02	0.255	<0.001	0.003
Fresh	60.9A	60.4A	58.6A	60.2A	59.6A
Ensiled	56.0Bb	57.3Bb	60.9Aa	56.4Bb	58.5Aab
AES	54.5Ba	53.6Cab	51.4Bb	51.0Cb	55.3Ba
Propionic acid, g/kg DM							0.50	<0.001	<0.001	0.001
Fresh	30.7Aa	31.1Aa	28.9Ab	28.9Ab	30.6Aa
Ensiled	28.8Bb	29.0Bb	29.9Ab	29.2Ab	31.6Aa
AES	24.4Ca	24.2Cab	21.3Bc	22.8Bb	24.8Ba
Isobutyric acid, g/kg DM							0.12	<0.001	0.164	0.702
Fresh	-	-	-	-	-
Ensiled	-	-	-	-	-
AES	-	-	-	-	-
Mean	1.46bc	1.38c	1.60b	2.08a	1.46bc
Butyric acid, g/kg DM							0.51	0.010	<0.001	0.078
Fresh	-	-	-	-	-	11.9B
Ensiled	-	-	-	-	-	12.7A
AES	-	-	-	-	-	9.67C
Mean	11.1bc	10.8c	11.2bc	12.3a	11.6ab
Isovaleric acid, g/kg DM							0.25	0.019	0.783	0.976
Fresh	-	-	-	-	-
Ensiled	-	-	-	-	-
AES	-	-	-	-	-
Mean	3.26b	3.13b	3.24b	3.82a	3.32b
Valeric acid, g/kg DM							0.27	<0.001	0.484	0.754
Fresh	-	-	-	-	-
Ensiled	-	-	-	-	-
AES	-	-	-	-	-
Mean	3.01b	2.97b	2.89b	3.86a	3.09b
A:P Ratio							0.04	<0.001	<0.001	0.201
Fresh	-	-	-	-	-	2.00B
Ensiled	-	-	-	-	-	1.95C
AES	-	-	-	-	-	2.28A
Mean	2.05b	2.04b	2.16a	2.13a	2.00b

CON, control (no additive); INO, inoculant delivering L. lactis and L. buchneri 150,000 cfu per g of fresh WBG each; NaL1, sodium lignosulfonate applied at a rate of 10 g/kg fresh basis; NaL2, sodium lignosulfonate applied at a rate of 20 g/kg, fresh basis; PRP, propionic acid (Alfa Aesar; 99%) applied at a rate of 5 g/kg, fresh basis. Fresh, fresh WBG prior to ensiling; ensiled, ensiled WBG after 60 d of storage; AES, aerobically exposed WBG silage 10 d after opening.

Whenever the interaction TRT × STG was not significant (P > 0.05), the means of significant main effects (P ≤ 0.05) are presented instead.

Means with different uppercase superscripts within a column are significantly different (P ≤ 0.05).

Means with different lowercase superscripts within a row are significantly different (P ≤ 0.05).

The concentration of total VFA produced in vitro sequentially decreased at each STG (P < 0.001) but no effects of TRT were observed (Table 6). Relative to fresh ($\overline{\mathrm{x}}=59.9$), ruminal acetic acid levels were only maintained after ensiling by NaL1 and PRP ($\overline{\mathrm{x}}=59.7;$ P ≥ 0.117) but decreased for all other TRT ($\overline{\mathrm{x}}=56.6\mathrm{m}\mathrm{M};$ P ≤ 0.033). However, both NaL1 and PRP showed lower levels at AES relative to their previous values (P ≤ 0.034). Moreover, AES INO and NaL2 underwent further decrease in ruminal acetic acid relative to their ensiled levels (P ≤ 0.014). In the case of ruminal propionic acid, fresh NaL1 and NaL2 ($\overline{\mathrm{x}}=28.9$) had lower levels relative to all other fresh TRT ($\overline{\mathrm{x}}=30.8;$ P ≤ 0.024) and after ensiling, PRP increased propionic acid compared with CON (P < 0.001). Both NaL2 and NaL1 had lower propionic acid than CON at AES (P ≤ 0.032). Consequently, the acetic-to-propionic acid ratio was the highest in AES (2.28), followed by fresh (2.00), and the lowest in ensiled (1.95; P < 0.001), while both NaL increased it relative to all the other TRT ($\overline{\mathrm{x}}=2.18$ vs. $\overline{\mathrm{x}}=2.03$, respectively; P < 0.001). Butyric acid was the highest in ensiled, followed by fresh, and lowest in AES (12.7, 11.9, 9.67 mM, respectively; P < 0.001). Across STG, NaL2 had higher levels of butyric acid relative to INO (P < 0.001).

Discussion

Characteristics of WBG

During the brewery mashing process, malted cereal grains are mixed with warm water to stimulate the enzymatic breakdown of starch into fermentable sugars such as maltose and maltotriose. The resulting sugary liquid (wort) is then separated from the grain residues (lautering process) and proceeds to the brewing process while the spent grains (WBG) are discarded (Mussatto et al., 2006). However, the final concentration of sugars in the WBG will depend on the type of grain, harvest time and conditions, sparging procedures, and types of machines (Huige, 2006; Gupta et al., 2010). Thus, high variability in nutritional composition of WBG between and even within breweries is expected and must be addressed as it can affect the quality of silage fermentation (Westendorf and Wohlt, 2002; Ferraretto et al., 2018). Furthermore, in smaller craft breweries, the extraction of sugars tends to be less intensive than in commercial breweries to prevent the extraction of tannins that affect the flavor of the beer (Mosher and Trantham, 2017). The WBG evaluated in this study was freshly sourced from a craft brewery and thus contained a higher concentration of WSC ($\overline{\mathrm{x}}=289\mathrm{g}/\mathrm{k}\mathrm{g}\mathrm{D}\mathrm{M}$) than the average reported in feed databases (0.0 to 82.0 g WSC per kg of DM; Westendorf et al., 2014; DairyOne, 2019), but was similar to other fresh WBG generated by several craft breweries in Northern Maine ($\overline{\mathrm{x}}=210\pm 7.26\mathrm{g}/\mathrm{k}\mathrm{g}\mathrm{D}\mathrm{M}$).

Preservation of nutritional composition

It is evident that the high concentration of WSC, along with optimal DM density (213 kg ± 12.2 DM per m³; Ruppel et al., 1995; Kung and Muck, 2017) and anaerobic conditions in the mini silos resulted in an adequate fermentation of the WBG, which was characterized by a decrease in pH and increases in LAB and lactic acid (Silva et al., 2017). Nevertheless, the loss of nutrients (especially WSC and starch) during ensiling was still deleterious for all TRT except PRP, which managed to preserve twice as many WSC and around 50% more starch than the other TRT. Similar results were reported by Moriel et al. (2016), who observed a successful decrease in DM loss compared with untreated WBG when applying a propionic acid-based product (66% acid) at a rate of 1 g/kg of fresh WBG and ensiled for 28 d. This is a consequence of the strong antimicrobial properties of PRP, which are conferred by its ability to cross the membrane of microbial cells and dissociate internally to release protons, leading to depletion of the microbe’s energy reserves in an attempt to eliminate the excess protons and restore internal pH (Pearlin et al., 2020). Additionally, PRP also produces oxidative conditions through the accumulation of reactive oxygen species and induces apoptotic cell death of microorganisms (Yun and Lee, 2016). Consequently, the inhibition of microorganisms by PRP resulted in decreased DM losses, lower proteolysis, and increased preservation of nutrients, leading to improved digestibility. Furthermore, although we did not observe differences in yeast counts across TRT, the concentration of ethanol in ensiled WBG was considerably decreased by PRP, which indicates a decrease in the metabolism of sugars to ethanol by the yeasts (Kung et al., 2018). Propionic acid is also added to silages to extend their aerobic stability due to the inhibition of aerobic microorganisms, particularly fungi and acetic acid bacteria, that readily degrade the silage (Silva et al., 2017). However, due to its effective preservation and greater concentration of digestible nutrients after ensiling, the aerobic exposure of PRP-treated WBG silage resulted in a considerable loss of nutrients, it did not prevent proteolysis relative to CON and INO, and was less aerobically stable than CON, although it still preserved more starch than all other TRT. This indicates that WBG silage may require higher doses of PRP to extend aerobic stability efficiently due to its great efficacy at preserving nutrients during ensiling. It is noteworthy that the concentration of propionic acid measured in fresh WBG (15.7 g/kg, DM basis) is lower than the theoretical PRP application (22.7 g/kg on a DM basis). This is because the high volatility of non-neutralized propionic acid results in considerable losses during application (Lord et al., 1981a; Baah et al., 2005; Gad, 2014). Furthermore, the concentration of propionic acid in PRP-treated WBG decreased considerably when exposed to air relative to ensiled and fresh, due to volatilization and possibly some degradation by the spoilage microbes (Lord et al., 1981b; Garcia et al., 2021). It is also noteworthy that INO had a greater concentration of propionic acid in AES than all other TRT including PRP, possibly due to the conversion of 1,2-propanediol into propionic acid by bacteria such as Lactobacillus diolivorans (Kung et al., 2018), which may have contributed to some extent to the inhibition of microorganisms by INO.

The fermentation process of the WBG in this study may resemble that of sugarcane silage due to its high concentration of WSC. In a meta-analysis evaluating the effects of L. buchneri inoculants on the chemical composition and aerobic stability of sugarcane silage, Rabelo et al. (2019) described that even though the application of the inoculant significantly improved the preservation of nutrients and increased production of acetic acid, the aerobic stability was not extended compared with untreated silage due to the high availability of residual WSC and lactic acid, which serve as substrates for the growth of spoilage microorganisms. Noticeably, in this study, the concentration of residual WSC in ensiled WBG decreased to the same extent for all TRT when exposed to air, and lactic acid decreased to the same extent in CON, INO, and PRP, indicating similar metabolization of these substrates by spoilage microorganisms. Furthermore, the consistent levels of acetic acid detected in NaL-treated WBG might not be related to fermentation processes but to possible residual acetate compounds present in the lignosulfonate deriving from the use of oxidizing reactants such as peroxyacetic acid during its manufacture (Aro and Fatehi, 2017). As described by Kung et al. (2003), the small differences in preservation observed between CON and INO during ensiling may be due to the adequate fermentation obtained in the control treatment. This is reflected in the extended aerobic stability of these two TRT ($\overline{\mathrm{x}}=92.1\mathrm{h}$) compared with other studies performed with WBG (<60 h; Wang and Nishino, 2009; Ferraretto et al., 2018). Also, the presence of acetic acid in ensiled CON may be explained by the presence of naturally occurring L. buchneri or other similar obligate heterofermentative LAB capable of converting lactic acid into acetic acid in anaerobic conditions (Wang and Nishino, 2009). Nevertheless, it should also be considered that L. buchneri utilizes lactic acid as a substrate in sugar-limited anaerobic conditions but, when readily available, utilizes the sugars instead of the lactic acid for growth and thus the production of acetic acid would be restricted (Oude Elferink et al., 2001). Thus, the high concentration of WSC in the presently studied WBG may have delayed the onset of lactic acid metabolism and acetic acid production (Johanningsmeier and McFeeters, 2015), which may also explain the failure of INO to extend aerobic stability compared with CON. Furthermore, Wilkinson and Davies (2013) suggested that at least 8 g/kg (fresh basis) of acetic acid is required in silage to ensure aerobic stability, especially when levels of WSC and yeast populations are high, but that concentration of acetic acid was not achieved by any TRT in this study (i.e., 5.28 g/kg, fresh basis for INO). Given the limitations of L. buchneri in high-WSC feeds, we speculate that extending the ensiling time to >90 d may have resulted in greater acetic acid concentrations and improved aerobic stability in INO-treated WBG.

The increase in the concentration of NDF, ADF, CP, and ether extract is likely due to the decreased concentration of WSC and starch, rendering a proportional increase in these less degradable nutrients. However, we also observed an increased recovery of ether extract during ensiling for all TRT ($\overline{\mathrm{x}}=111\mathrm{\%}$), which may suggest bacterial de novo synthesis of fatty acids from carbohydrates, which were used for the formation of cell membranes as microbial counts increased (Vlaeminck et al., 2006; White and Fuqua, 2012). This has been described in some corn and alfalfa silage studies, where palmitic (C16:0) and stearic acids (C18:0) were found to increase during ensiling at the expense of WSC (Han and Zhou, 2014; Liu et al., 2018). We believe this process would be expanded in a WSC and starch-rich substrate like WBG, especially under high moisture conditions favorable for microbial growth (Kung et al., 2018). Further research is needed to understand microbial fat synthesis in silages rich in nonstructural carbohydrates.

In vitro digestibility and gas production

The spoilage process, which occurred to different extents depending on TRT, resulted in a decrease in IVDMD and IVOMD, rate (K_f) and extent (M_f) of gas production, and an increase in ruminal NH₃-N. The latter indicates that WBG N fractions contributing to the rumen degradable protein pool increased as STG progressed. This may reduce the efficiency of N utilization, especially in animals fed high-forage rations because of excessive supply of rumen degradable protein, which is eventually excreted as urea in the urine (Savari et al., 2018). The overall preservation of IVDMD by PRP compared with CON is expected due to its more extensive protection of digestible nutrients (i.e., WSC and starch) across storage stages. However, given that NaL treatments did not prevent the loss of nutrients in WBG, their positive effect on IVDMD was unexpected and may be attributed to its surfactant properties, which improves the adsorption of microbial enzymes onto feed particles, thus increasing the rate of digestion of cellulose (Kamande et al., 2000; Zheng et al., 2020). Similarly, Reyes et al. (2020) observed increased DM and NDF digestibility when testing sodium and magnesium lignosulfonates as preservatives of high-moisture alfalfa hay in vitro, but the latter increased digestibility despite not having an effect on the preservation of nutrients compared to the former.

Contrary to our hypothesis, methane production decreased for ensiled and AES WBG compared with fresh. It is understood that high-starch diets favor the activity of amylolytic bacteria in the rumen, resulting in increased production of propionic acid, which acts as a hydrogen sink, and thus limits the availability of hydrogen for methanogenesis (Nagaraja, 2016). Even though starch concentrations were decreased in ensiled and AES WBG, samples from these stages also had increased concentrations of fats compared with fresh WBG, which may have contributed substantially to the decreased production of methane. Dietary fats, especially polyunsaturated fatty acids (PUFA), inhibit methane production by reducing the accumulation of hydrogen in the rumen through fatty acid biohydrogenation, as well as exerting a direct toxic effect on methanogens and reducing fiber digestibility (McAllister et al., 1996; Patra et al., 2017; Alvarez-Hess et al., 2019). Around 50% of the total fatty acids in WBG is linoleic acid (C18:2), a PUFA, followed by palmitic acid (C16:0), which represents ~25% of total fatty acids (Fărcaş et al., 2015; Lordan et al., 2019). This is in agreement with Moate et al. (2011), who described that the addition of WBG to the diet of lactating cows as a fat supplement decreased methane emissions per unit of milk produced. Studies evaluating the effects of silage inoculants on enteric methane production have shown variable results. Cases of successful methane mitigation by silage inoculants have been related to the improved nutritional composition of the silage and the higher concentrations of lactic acid, which acts as a hydrogen and CO₂ sink when converted into propionic acid by lactate-utilizing bacteria (Cao et al., 2011). However, these responses may be strain, dose, and substrate-specific (Ellis et al., 2016; Arriola et al., 2021) and thus further research on these factors is required.

Conclusions

Propionic acid was the most efficient at preventing the loss of nutritional value in ensiled WBG because it reduced DM loss and preserved twice as many WSC and 50% more starch than all other TRT. Consequently, PRP-treated ensiled WBG had higher IVDMD, IVOMD, M_f, and K_f than other TRT. However, due to its greater concentration of digestible nutrients, PRP failed to prevent spoilage when aerobically exposed and was less stable than CON. Even so, it still prevented proteolysis and had more starch relative to all other TRT. Both NaL treatments failed to prevent spoilage of ensiled or AES WBG, but they prevented the increase of NDF and increased IVDMD possibly due to its surfactant properties. INO did not improve the preservation of ensiled WBG and failed to extend aerobic stability of AES compared with CON, even though it produced more acetic acid than CON. The high concentration of WSC may have limited the conversion of lactic acid to acetic acid to the extent necessary to improve aerobic stability. Contrary to our hypothesis, methane production, concentration, and yield were lower for ensiled and AES, compared with fresh, possibly due to the increased concentration of ether extract in ensiled and AES as a result of other nutrient losses and microbial synthesis of fatty acids. The high spoilage susceptibility of WBG, particularly when aerobically exposed, remains a challenge that requires further research.

Glossary

Abbreviations

ADF

acid detergent fiber

AES

aerobically exposed silage

CFU

colony-forming units

CH₄

methane

CON

control

CP

crude protein

DM

dry matter

HDD

heat degree-days above room temperature

INO

microbial inoculant

IVDMD

in vitro dry matter digestibility

K_f

rate of gas production

LAB

lactic acid bacteria

M_f

asymptotic maximal gas production

NaL

sodium lignosulfonate

NDF

neutral detergent fiber

NH₃-N

ammonia nitrogen

PRP

propionic acid

STG

stage of storage

TRT

treatment

VFA

volatile fatty acids

WBG

wet brewer’s grain

WSC

water-soluble carbohydrates

$\overline{\mathrm{x}}$

mean

Conflict of interest statement

The authors declare that they have no conflict of interest.