The use of Lentilactobacillus buchneri PJB1 and Lactiplantibacillus plantarum MTD1 on the ensiling of whole-plant corn silage, snaplage, and high-moisture corn

Experiments were conducted over a 3-yr period to evaluate the effects of bacterial inoculants on the fermentation profile and aerobic stability of whole-plant corn silage (WPC), snaplage (SNA), and high-moisture corn (HMC). Whole-plant corn was inoculated with Lentilactobacillus buchneri PJB1 in combination with Lactiplantibacillus plantarum MTD1 or with Lpb. plantarum alone (experiments 1 and 2). Snaplage (experiment 3) and HMC (experiments 4 and 5) were inoculated with Len. buchneri in combination with Lpb. plantarum or with Len. buchneri alone. After inoculation, the feedstuffs were ensiled in 7.57-L silos and stored at 21 ± 2°C for 30 or 90 d. In experiment 5, silage was subjected to air stress for 2 h every 2 wk through 42 d and then for 2 h/wk until 90 d and had samples analyzed for their bacterial community composition by metagenomics. Overall, in all experiments, silages inoculated with Len. buchneri alone or in combination with Lpb. plantarum had more acetic acid and 1,2-propanediol and fewer yeasts than uninoculated silages. After 30 d of ensiling, inoculation with Len. bu-chneri alone or in combination with Lpb. plantarum did not affect the aerobic stability of SNA, but it slightly increased the stability of WPC and markedly improved the stability of HMC. After 90 d of ensiling, inoculation with Len. buchneri alone or in combination with Lpb. plantarum markedly improved the aerobic stability of WPC, SNA, and HMC. In experiment 5, inoculation increased the relative abundance (RA) of Lactobacil-laceae and reduced the RA of Enterobacteriaceae and Leuconostocaceae in HMC at 30 and 90 d and the RA of Clostridiaceae in non-air-stressed HMC at 90 d. Air-stressed HMC inoculated with Len. buchneri had less lactic acid, more acetic acid and 1,2-propanediol, and markedly greater aerobic stability than uninoculated air-stressed HMC at 90 d. In conclusion, inoculation with Len. buchneri PJB1 alone or in combination with Lpb. plantarum MTD1 increased the production of acetic acid and 1,2-propanediol, inhibited yeasts development, and improved the aerobic stability of WPC, SNA, and HMC. In HMC, inoculation markedly improved aerobic stability as soon as after 30 d of ensiling, and after 90 d, inoculation improved stability even under air stress conditions.


INTRODUCTION
Whole-plant corn silage (WPC) is the most widely used type of silage derived from the corn plant.However, the usage of snaplage (SNA) and high-moisture corn (HMC) silages has increased in popularity among dairy farmers due to their high energy value.Snaplage comprises the corn kernels, cob, husk, and shank (Ferraretto et al., 2018), whereas HMC is produced with only the corn kernels, containing between 25% to 35% moisture (Hoffman and Muck, 2011).Snaplage and HMC have larger proportions of grain than forage when compared with WPC.Because of that, SNA and HMC have a higher DM content and more starch, and undergo a restricted fermentation compared with WPC, causing those silages to be more prone to aerobic deterioration than WPC (Saylor et al., 2021).Aerobic deterioration of silages causes DM and nutrient losses, and feeding spoiled silage to animals can negatively affect their health and performance (Borreani et al., 2018).
Yeasts are usually the main initiators of aerobic deterioration in corn-based silages (Pahlow et al., 2003).Inoculants, such as Lentilactobacillus buchneri, that produce acetic acid, an antifungal compound, can be used in corn silage production to reduce or delay aerobic deterioration (Muck et al., 2018).Lentilactobacillus buchneri is an obligate heterofermentative lactic acid The use of Lentilactobacillus buchneri PJB1 and Lactiplantibacillus plantarum MTD1 on the ensiling of whole-plant corn silage, snaplage, and high-moisture corn bacteria (LAB) capable of converting lactic acid into acetic acid, 1,2 propanediol, and minor amounts of ethanol (Oude Elferink et al., 2001).Inoculation with Len.buchneri can slightly increase DM losses in corn silage (Kleinschmit and Kung, 2006).Because of that, inoculants containing Len. buchneri are often formulated in combination with facultative heterofermentative LAB (e.g., Lactiplantibacillus plantarum), which ferment hexoses forming mostly lactic acid, with the main goal of reducing DM and energy losses during fermentation (Muck et al., 2018).Lentilactobacillus buchneri produces acetic acid slowly during ensiling (Schmidt et al., 2009).Because of that, the original recommendation was that forage inoculated with Len.buchneri needed to be ensiled for a minimum of 60 to 90 d before being fed.However, due to feed shortages, producers oftentimes need to feed the silage before that recommended curing period is completed.Therefore, testing the effectiveness of Len.buchneri species that could be beneficial after shorter periods of ensiling is important.
Metagenomics is the direct genetic analysis of the genomes contained in a sample.This technology can be applied to evaluate the effect of silage inoculants on the microbial communities during ensiling and aerobic exposure, with the ultimate goal of improving silage quality and safety (McAllister et al., 2018).However, published studies applying metagenomics to assess the effects of inoculants on the bacterial community composition of HMC during ensiling are few (Wu et al., 2020;da Silva et al., 2021), and studies performed during the aerobic exposure of HMC are lacking.
The objective of the present study was to evaluate the mechanism and effectiveness of treatment with Len.buchneri PJB1, alone or in combination with Lpb.plantarum MTD1, on the fermentation and aerobic stability of various types of corn-based silages.To further the understanding of such mechanism, we also evaluated the effect of Len.buchneri, alone or in combination with Lpb.plantarum, on the microbiota of HMC during fermentation and aerobic exposure using metagenomics.

MATERIALS AND METHODS
No human or animal subjects were used, so this analysis did not require approval by an Institutional Animal Care and Use Committee or Institutional Review Board.

Inoculation and Ensiling
We conducted 5 experiments over a 3-yr period, to test the efficacy of bacterial inoculants (Volac Inter-national Ltd., Orwell, Hertfordshire, UK) on various types of silages derived from corn plants.The type of corn material, inoculant treatments, pile size, amount of water and water-inoculant solution, ensiling period, and air stress status for each experiment are shown in Table 1.On the day of ensiling, 5 replicated piles of corn material (WPC, SNA, or HMC) were individually prepared and inoculated for each treatment (not blinded).Specific volumes of water for control (CTRL), or inoculant diluted in water were applied to the replicated piles.A hand sprayer was used to apply the water or water-inoculant mix as the piles were thoroughly mixed to ensure a good distribution of the inoculant treatment.After treatment, the material from the piles was packed into 7.57-L food-grade pails (#20256, Encore Plastics Corp.) using a gas-powered hydraulic press (a custom-modified log splitter).The final packing density of the material in the silos was approximately 200 kg of DM/m 3 in experiments 1 and 2, 620 kg of DM/m 3 in experiment 3, and 650 kg of DM/m 3 in experiments 4 and 5. Silos were sealed with plastic lids equipped with O-ring seals (#200002, Encore Plastics Corp.) and stored at 21 ± 2°C until being opened after 30 or 90 d.
For experiment 5 only, we prepared additional silos for air-stressing before being opened at 90 d.The additional treatments in experiment 5 were control (untreated; CTRL-AS) and inoculation with Len.buchneri PJB1 (NCIMB 30139) at 400,000 cfu/g of fresh material (LB400-AS).Air-stressed silos had 3 1.60-cm holes, 2 located on the bottom of the bucket and one on the lid.The holes were plugged with stoppers, which were opened for 2 h every 2 wk through 42 d and then 2 h/wk through 90 d of storage.Subsamples were taken at 50 and 140 h after aerobic exposure following the opening at 90 d.
Fresh material samples after treatments but before ensiling were collected from each of the 5 replicated piles of CTRL and inoculated material for experiments 1, 2, and 4, and only from the CTRL piles for experiments 3 and 5.

Chemical Analysis
The chemical analysis of fresh forage and silages was performed as described by Kung et al. (2021).The DM content of the corn materials before and after ensiling was measured by drying approximately 100 g of material in a forced-air oven at 60°C for 48 h.The DM content was calculated using the weight of the material before and after the drying process.
The concentrations of CP, soluble protein (Sol-N), ADF, NDF, and starch were analyzed on the ovendried samples by Cumberland Valley Analytical Ser- vices (Waynesboro, PA).Dried samples were processed by grinding in a Wiley mill (Thomas Scientific) to pass through a 3-mm screen before being used for the quantification of Sol-N (% of CP) by the methodology of Krishnamoorthy et al. (1982).Another portion of the dried samples was ground to pass through a 1-mm screen using a Udy Cyclone Mill (Udy Corp.) and were then used for the quantification of CP, ADF (not corrected for ash), and NDF (not corrected for ash).The total N values were obtained by the combustion of the samples (Leco CNS 2000 Analyzer;Leco Corp.).The total N values were multiplied by a factor of 6.25 to calculate the CP content.The concentration of ADF was quantified using the method of Goering and Van Soest (1970), with the adjustment of using a 1.5-μm particle retention 7-cm Whatman glass fiber in the place of a fritted glass crucible.Neutral detergent fiber content was determined by the method of Van Soest et al. (1991), using a heat-stable amylase and sodium sulfate.The starch concentration was evaluated and corrected for free glucose using the procedure described by Hall (2009).
Samples containing 25 g of fresh corn material before ensiling or silage were mixed with 225 mL of autoclaved 1/4 strength Ringer's solution (Oxoid BR0052G, Oxoid, Unipath Ltd.) in a Proctor-Silex 57171 blender (Hamilton Beach⁄Proctor, Silex Inc., Washington, DC) for 1 min.The pH of the blended mixture was immediately measured (pH/mV/Ion/°C/°F Ion700 Meter, Oakton.).A portion of this mixture was filtered through 4 layers of autoclaved cheesecloth and used for microbial enumeration (as described in the following section).Another portion was filtered through 4 layers of autoclaved cheesecloth and a Whatman 54 filter paper (Whatman Ltd.).Approximately 10 mL of this filtrate was acidified with 25 μL of 50% vol/vol H 2 SO 4 water solution.Next, the solution was stored at −18°C until being used for quantifying NH 3 -N (Okuda et al., 1965) and reducing sugars (RS; Nelson, 1944) by colorimetric procedures, and lactic acid, acetic acid, butyric acid, 1,2-propanediol, and ethanol by HPLC (Muck and Dickerson,1988).The HPLC (RID-10A, Shimadzu Corp.) was fitted with an Aminex HPX-87H column with a micro-guard Cation H cartridge (Bio-Rad Laboratories Inc.) to measure the concentrations of lactic acid, acetic acid, and butyric acid, 1,2-propanediol, and ethanol.

Enumeration of Culturable LAB, Yeasts, and Molds
The same extracts used for the pH measurements were used for microbial enumeration after being filtered through 4 layers of autoclaved cheesecloth, as previously mentioned, and diluted in 10-fold serial dilutions of autoclaved 1/4 strength Ringer's solution (Oxoid BR0052G, Oxoid, Unipath Ltd.).We enumerated the culturable LAB, yeasts, and molds in fresh materials and silages using the pour-plating technique.In addition, we enumerated the LAB in the inoculantwater solution applied on the ensiling day to verify if the targeted application rate of the inoculant had been achieved (Supplemental Table S1, https: / / data .mendeley.com/datasets/ 3ypcff25yx/ 1; Benjamim da Silva et al., 2023).Man, Rogosa, and Sharpe agar (CM3651, Oxoid, Unipath) was used for cultivating LAB, and malt extract agar (CM0059, Oxoid, Unipath) acidified with 0.5% vol/vol of 85% lactic acid was used for enumerating yeasts and molds.Man, Rogosa, and Sharpe agar plates were incubated anaerobically at 35°C and malt extract agar plates were incubated aerobically at 30°C, for 48 to 72 h, before colony enumeration.To maintain anaerobic conditions during incubation, 1 WPC = whole-plant corn silage, SNA = snaplage, HMC = high-moisture corn.
the Man, Rogosa, and Sharpe agar plates were kept in airtight containers with an AnaeroPack (Mitsubishi Gas Chemical Co.).

DM Recovery and Aerobic Stability
The DM recovery was calculated using the values of DM content and weight of the material in the silo at the ensiling and silo-opening days.To measure the aerobic stability of the silages, approximately 2 kg of WPC or 4 kg of SNA or HMC were placed into clean 7.5-L buckets, which were then covered with 4 layers of cheesecloth.The temperature of the silages in the buckets was taken by thermocouple wires positioned in the geometric center of the silage masses.The wires were attached to a data logger (DT85 Series 3, Thermo Fisher Scientific Australia Pty. Ltd.), which recorded the temperatures every 15 or 30 min.Aerobic stability was defined as the number of hours taken for the silage to raise 2°C above the silage's baseline temperature after exposure to air.

Analysis of the Bacterial Community Composition
The bacterial community composition analysis by high-throughput sequencing was performed only for experiment 5 in silage samples (5 replicates per treatment) at 30 d, 90 d, and after 50 and 140 h of air exposure after 90 d of ensiling.Twenty-five grams of silage and 225 mL of autoclaved 1/4 strength Ringer's solution (Oxoid BR0052G, Oxoid, Unipath Ltd.) were deposited in a stomacher bag (Stomacher 400 strainer bags, Seward Laboratory) and homogenized for 2 min using a Colworth 400 stomacher (Seward Laboratory).The homogenate was filtered through 4 layers of sterile cheesecloth.Immediately after filtering, 2 mL of the homogenate was centrifuged on a Centrifuge 5424 R (Eppendorf AG) at 21,130.2 × g at 21°C for 3 min.The pellet obtained after centrifugation was resuspended in 100 μL of autoclaved Ringer's solution.Next, the material was frozen at −80°C until subjected to DNA extraction.
The sequencing data were analyzed on Qiime 2 version 2020.6 (Bolyen et al., 2019).Quality check of the demultiplexed sequences was performed using the interactive quality plots.Sequences were quality filtered, denoized, merged, and subjected to chimera removal using the DADA2 pipeline (Callahan et al., 2016), in which the first 20 bp of the sequences were trimmed, and the forward and reverse sequences were truncated at 240 and 220 bp, respectively, Resulting denoized sequences were tabulated as features.A naïve Bayes classifier trained using the 99% identity sequence and classification files from Greengenes version 13.8 (DeSantis et al., 2006) was used for the taxonomy assignment.Features assigned to chloroplast or mitochondria were removed before proceeding.Samples were rarefied to even sequence depth before the analysis of α-diversity (observed features and Shannon, Pielou's evenness, and Faith's phylogenetic diversity indices) and β-diversity (weighted Unifrac dissimilarity).The weighted Unifrac dissimilarity matrices were used to build the principal coordinates analysis plots.

Statistical Analysis
Data were analyzed on JMP Pro 15 (SAS Institute Inc.) using the Fit Model procedure.There were 5 replicates per treatment in all experiments.Data on fresh corn materials before ensiling were analyzed as a completely randomized design with treatment effects in experiments 1, 2, and 4.However, in experiments 3 and 5, data on fresh corn materials were reported as the mean and standard deviation of CTRL replicates.For all experiments, silage data from each length of storage (30 and 90 d) were analyzed separately as a completely randomized design with treatment effects.In experiment 5, data on 90-d ensiled silages exposed to air for 0, 50, and 140 h were analyzed as repeated measurements using univariate ANOVA with a restricted maximum likelihood model with replicates as a random effect and time of aerobic exposure, treatment, and their interaction as fixed effects.Tukey's test (Snedecor and Cochran, 1980) was performed when an effect or interaction was declared significant (P ≤ 0.05).Tendencies were declared at P ≤ 0.10.
In experiment 5, the weighted Unifrac distance matrices were statistically analyzed by permutational multivariate ANOVA and analysis of multivariate homogeneity of group dispersions using Qiime 2 version 2020.6 (Bolyen et al., 2019).
Most of the variables analyzed differed among treatments after 90 d of ensiling, except for the NH 3 -N content and DM recovery (Table 3).The DM content differed only between CTRL and LB200LP, with the latter having a lower DM content than CTRL (29.82 vs. 30.38%,P = 0.031).Uninoculated (3.35) and LP (3.34) silages had the lowest pH, LB200LP intermediate (3.46), and LB400LP (3.58) had the highest pH.The concentrations of RS, lactic acid, acetic acid, and 1,2-propanediol, and the LAB numbers were similar between CTRL and LP and also between LB200LP and LB400LP.However, compared with CTRL and LP, both LB200LP and LB400LP had lower concentrations of RS (average of 0.17 vs. 1.26% of DM, P < 0.001) and lactic acid (average of 3.44 vs. 5.09% of DM, P < 0.001) and greater concentrations of acetic acid (average of 2.63 vs. 1.80% of DM, P < 0.001) and 1,2-propanediol (average of 1.56 vs. < 0.01% of DM, P < 0.001) and more LAB (average of 9.01 vs. 6.89 log 10 cfu/g of fresh weight, P < 0.001).The ethanol content was similar in CTRL, LP, and LB200LP, but it was greater in LB400LP (1.93% of DM) than in CTRL (1.68% of DM, P = 0.002) and LP (1.75% of DM, P = 0.028).
The numbers of yeasts were lower in LP than in CTRL (3.44 vs. 4.27 log 10 cfu/g of fresh weight, P = 0.029), and below the detectable level of 2.00 log 10 cfu/g of fresh weight in LB200LP and LB400LP.The numbers of molds were low in all treatments (<2.71 log 10 cfu/g of fresh weight).Both LB200LP (93 h, P < 0.001) and LB400LP (105 h, P < 0.001) had greater aerobic stability than CTRL (36 h), but LP did not (41 h).A larger improvement in aerobic stability compared with CTRL was observed in LB400LP than in LB200LP (improvement of 69 vs. 57 h, P = 0.030).

Experiment 2
Overall, inoculation treatment did not affect the chemical and microbiological composition of chopped whole corn plants before ensiling (Supplemental Table S3, https: / / data .mendeley.com/datasets/ 3ypcff25yx/ 1; Benjamim da Silva et al., 2023).Only the RS content (P = 0.014) and LAB numbers (P = 0.023) were affected by inoculation.The RS content in LP was similar to CTRL, but in LB200LP it was lower than in CTRL (3.28 vs. 4.05% of DM, P = 0.011).Both LP (6.40 log 10 cfu/g of fresh weight, P = 0.044) and LB200LP (6.49 log 10 cfu/g of fresh weight, P = 0.024) had more LAB than CTRL (6.02 log 10 cfu/g of fresh weight).
The numbers of LAB in corn silage ensiled for 90 d were not affected by inoculation, but the numbers of yeasts (P < 0.001) and molds (P = 0.004) were affected by inoculation (Table 4).The numbers of yeasts were lower in LB200LP (<2.00 log 10 cfu/g of fresh weight) than in CTRL (5.06 log 10 cfu/g of fresh weight, P < Aerobic stability, defined as number of hours before a 2°C increase above silage baseline temperature after exposure to air.0.001) and LP (4.93 log 10 cfu/g of fresh weight, P < 0.001), and the numbers of molds were lower in LP (2.60 log 10 cfu/g of fresh weight, P = 0.032) and LB200LP (<2.00 log 10 cfu/g of fresh weight, P = 0.003) compared with CTRL (3.87 log 10 cfu/g of fresh weight).Inoculation did not affect the recovery of DM, but it did affect the aerobic stability (P < 0.001) of corn silage ensiled for 90 d (Table 4).Whereas the aerobic stability of LP (48 h) was like that of CTRL (44 h), LB200LP (260 h) had greater aerobic stability than CTRL (P = 0.013) and LP (P = 0.010).
There were only a few differences between LB400 and LB600 HMC after 30 d of ensiling (Table 7).The DM content (72.72 vs. 71.97%,P < 0.001) and DM recovery (97.65 vs. 96.62%,P < 0.001) were greater in LB600 than LB400, and the aerobic stability was lower in LB600 than in LB400 (217 vs. 248 h, P = 0.033).However, after 90 d of ensiling, there were no differences between LB400 and LB600 for any of the items analyzed (Table 8).
3 Dry matter recovery. 4Aerobic stability, defined as number of hours before a 2°C increase above silage baseline temperature after exposure to air.Means in rows with unlike superscripts differ (P ≤ 0.05).
We also analyzed the changes in pH and concentrations of lactic acid, acetic acid, butyric acid, 1,2-pro- panediol, and ethanol during aerobic exposure of HMC after silo opening and found an interaction (P < 0.009) between treatment and time of aerobic exposure for all the items analyzed (Table 11).Uninoculated airstressed HMC was the first treatment to show an increase in pH during aerobic exposure, having a greater pH (5.47 vs. average of 4.18, P < 0.001) than other treatments after 50 h of aerobic exposure.After 140 h, both CTRL-AS and CTRL had a greater pH than inoculated HMC (average of 6.94 vs. average of 4.19, P < 0.001).The pH of inoculated HMC remained low from 0 to 140 h of aerobic exposure (average of 4.20).At 140 h, CTRL and CTRL-AS had very low levels of lactic acid, which were inferior to the levels of inoculated HMC (average of <0.01 vs. average of 0.72% of DM, P < 0.001).In addition, inoculated HMC had more Means in rows with unlike superscripts differ (P ≤ 0.05). 1 CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB400LP = LB400 plus Lactiplantibacillus plantarum MTD1 at 100,000 cfu/g of fresh material. 2Lactic acid bacteria. 3Dry matter recovery. 4Aerobic stability, defined as number of hours before a 2°C increase above silage baseline temperature after exposure to air up to 250 h.Means in rows with unlike superscripts differ (P ≤ 0.05).
1 CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB400LP = LB400 plus Lactiplantibacillus plantarum MTD1 at 100,000 cfu/g of fresh material; CTRL-AS = CTRL subjected to air stress of 2 h every 2 weeks up to 42 d of ensiling and then 2 h every week up to 90 d of ensiling; LB400-AS = LB400 subjected to air stress of 2 h every 2 weeks up to 42 d of ensiling and then 2 h every week up to 90 d of ensiling.
acetic acid (average of 1.13 vs. average of 0.17% of DM, P < 0.001) and 1,2-propanediol (average of 0.73 vs. average of 0.06% of DM, P < 0.001) than in CTRL and CTRL-AS all time points analyzed.
Compared with inoculate HMC, CTRL-AS had more butyric acid before aerobic exposure (0 h; 0.03 vs. average of <0.01% of DM, P < 0.002), and CTRL had more butyric acid at 50 h (0.04 vs. average of <0.01% of DM, P < 0.001).At 140 h, there were no differences in butyric acid concentration among treatments.The ethanol content was similar among all treatments at 0 h.However, after 50 h, CTRL-AS had less ethanol than other treatments (0.12 vs. average of 0.76% of DM, P < 0.001) and, at 140 h, CTRL and CTRL-AS had less ethanol than inoculated HMC (average of <0.01 vs. average of 0.50% of DM, P < 0.006).At all time points, CTRL and CTRL-AS had more yeasts than inoculated HMC (average of 7.24 vs. average of 2.70 log 10 cfu/g of fresh weight, P < 0.001).The yeasts numbers were low in LB400 and LB400-AS and only slightly increased from 0 to 140 h (average of <2.00 to 3.59 log 10 cfu/g of fresh weight, P < 0.002).The yeasts numbers were lower in LB400LP than in LB400 and LB400-AS at 50 h (<2.00 vs. average of 3.51 log 10 cfu/g of fresh weight, P < 0.013) and 140 h (2.10 vs. average of 3.59 log 10 cfu/g of fresh weight, P < 0.009).
We also evaluated how the inoculants affected the bacterial community composition of HMC during ensiling and after aerobic exposure.The α diversity analysis is shown in Supplemental Tables S7, S8, and S9 (https: / / data .mendeley.com/datasets/ 3ypcff25yx/ 1; Benjamim da Silva et al., 2023).At 30 d of ensiling, LB400 and LB400LP had less observed features and lower Shannon, Pielou's evenness, and Faith's phylogenetic diversity indices than CTRL.Similar results were found at 90 d with inoculated silages having less observed features and lower Shannon and Pielou's evenness indices than CTRL and CTRL-AS, except that LB400LP had similar numbers of observed features to CTRL and CTRL-AS.The Faith's phylogenetic diversity index was lower in LB400-AS than in CTRL and CTRL-AS but similar among CTRL, CTRL-AS, LB400, and LB400LP.After 50 and 140 h of aerobic exposure, CTRL had more observed features than CTRL-AS, LB400, LB400LP, and LB400-AS.The β diversity analysis (Supplemental Figure S1, https: / / data .mendeley.com/datasets/ 3ypcff25yx/ 1; Benjamim da Silva et al., 2023) shows that inoculated silages clustered separated form uninoculated silages at 30 and 90 d, indicating that inocu- Means in columns with unlike subscripts differ (P ≤ 0.05).
1 Sampling times after aerobic exposure.
2 CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB400LP = LB400 plus Lactiplantibacillus plantarum MTD1 at 100,000 cfu/g of fresh material; CTRL-AS = CTRL subjected to air stress of 2 h every 2 weeks up to 42 d of ensiling and then 2 h every week up to 90 d of ensiling; LB400-AS = LB400 subjected to air stress of 2 h every 2 weeks up to 42 d of ensiling and then 2 h every week up to 90 d of ensiling.
When analyzing the RA of bacterial families in the silos ensiled for 90 d after 0, 50, and 140 h of aerobic exposure, we found an interaction between treatment and time after aerobic exposure for the RA of Lactobacillaceae (P < 0.001), Enterobacteriaceae (P < 0.001), Leuconostocaceae (P < 0.001), Bacillaceae (P < 0.001), and Streptococcaceae (P < 0.001; Figure 3).Inoculated HMC had a high RA of Lactobacillaceae at 0 h, and this RA remained high throughout aerobic exposure.The RA of Lactobacillaceae in CTRL-AS was lower than inoculated silage at 0 h (59.17 vs. average of 91.73%, P < 0.001), but at 50 h it increased and was similar to that of inoculated HMC.The RA of Lactobacillaceae was lower in CTRL than in inoculated and CTRL-AS at 0 h (43.33 vs. average of 83.59%, P < 0.026) and remained lower at 50 h (31.99 vs. 85.21%,P < 0.001).However, at 140 h, the RA of Lactobacillaceae in CTRL was similar to that of other treatments.At 0 h, inoculated HMC (average of 5.24%) had a lower RA of Enterobacteriaceae than CTRL (42.62%,P < 0.001) and CTRL-AS (28.16%,P < 0.001), and this RA remained low throughout aerobic exposure.In CTRL-AS, the RA of Enterobacteriaceae decreased (28.16% to 8.79%, P < 0.001) from 0 to 50 h reaching the levels found in Relative abundances of the bacterial families with relative abundance greater than 0.10% in high-moisture corn ensiled with or without air stress for 90 d in experiment 5. CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB400LP = LB400 plus Lactiplantibacillus plantarum MTD1 at 100,000 cfu/g of fresh material; CTRL-AS = CTRL subjected to air stress; LB400-AS = LB400 subjected to air stress.Air stress occurred for 2 h every 2 wk up to 42 d of ensiling and then 2 h every week up to 90 d of ensiling.

Figure 3.
Relative abundances of the bacterial families with relative abundance greater than 0.10% in high-moisture corn, ensiled with or without air stress for 90 d, after 0, 50, and 140 h of aerobic exposure in experiment 5. CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB400LP = LB400 plus Lactiplantibacillus plantarum MTD1 at 100,000 cfu/g of fresh material; CTRL-AS = CTRL subjected to air stress; LB400-AS = LB400 subjected to air stress.Air stress occurred for 2 h every 2 wk up to 42 d of ensiling and then 2 h every week up to 90 d of ensiling.
in CTRL and CTRL-AS (average of 1.67% vs. average of 5.37%, P < 0.001) at 0 h, and lower than in CTRL at 50 (average of 1.56 vs. 5.69%, P < 0.001) and 140 h (average of 2.30 vs. 6.67%,P < 0.001).Bacillaceae were not detected in any treatment at 0 and 50 h of aerobic exposure, but it was found in CTRL (7.23%) and CTRL-AS (0.53%) after 140 h.There was a treatment (P < 0.001) effect on the RA of Clostridiaceae.with inoculated HMC having a lower RA of Clostridiaceae than CTRL (average of <0.01 vs. 4.23%, P = 0.002).

DISCUSSION
There are only a few published studies where an inoculant has been tested over a period of several years (Xu et al., 2020;Diepersloot et al., 2021).In the present study, we investigated the effects of Len.buchneri 30139 alone or in combination with Lpb.plantarum MTD1 on the fermentation and aerobic stability of WPC, SNA, and HMC after 30 and 90 d of ensiling over a 3-yr period.Higher doses of the inoculants were applied to SNA and HMC than to WPC as it is known that SNA and HMC are more prone to aerobic deterioration than WPC (Saylor et al., 2021) and because most manufacturers of LB-based inoculants suggest a higher application dose for HMC and SNA than for WPC.We evaluated the dose effect for WPC and HMC in the first experimental year of the study.Because the results of the first year showed that the lower dose was effective we followed the next year's trials using only the lower dose, to validate its effectiveness.In experiment 5, with HMC, we also evaluated the effect of air stress during storage on the effectiveness of the inoculant and the influence of the inoculant on the bacterial community composition during ensiling and aerobic exposure.

Fermentation Profile and DM Recovery
Inoculants based on facultative heterofermentative LAB, such as Lpb.plantarum, are applied to produce lactic acid and quickly reduce the silage's pH to prevent undesirable bacteria proliferation and reduce DM and nutrients losses (Muck et al., 2018).Whole corn plant is an easy to ensile crop in which most of the drop in pH occurs in the first week of ensiling (Allen et al., 2003).Because of that, the use of Lpb.plantarum in corn silage production does not only aim to improve acidification.As such, we likely missed the window where the Lpb.plantarum effect could have been detected as we analyzed the silage only after 30 d of ensiling.
Obligate heterofermentative LAB, such as Len.buchneri, are mainly used to improve aerobic stability (Muck et al., 2018).In anaerobic conditions Len. buchneri can convert lactic acid into acetic acid and 1,2-propanediol (Oude Elferink et al., 2001).Acetic acid is a weak acid able to inhibit the development of yeasts (Woolford, 1975), which are usually the main initiators of aerobic deterioration in corn silage (Pahlow et al., 2003).Overall, across all experiments, all types of silage inoculated with Len.buchneri alone or in combination with Lpb.plantarum had less lactic acid, more acetic acid and 1,2-propanediol, and fewer yeasts than CTRL silages, indicating that the inoculants tested effectively modified the fermentation.
Inoculation of WPC with Len.buchneri can cause small increases in DM losses due to acetic acid production (Kleinschmit and Kung, 2006).To prevent such increases in DM losses, Len.buchneri can be applied in combination with Lpb.plantarum, which produces almost exclusively lactic acid (Muck et al., 2018).Inoculation did not affect the DM recovery of WPC (experiments 1 and 2).However, inoculation of SNA with the combination of Len.buchneri and Lpb.plantarum, but not with Len.buchneri alone, tended to improve DM recovery compared with CTRL after 90 d of ensiling (experiment 3).In contrast, in HMC, inoculation with Len.buchneri alone or in combination with Lpb.plantarum slightly decreased DM recovery (approximately 1%) after 90 d of ensiling (Experiments 4 and 5).The increase in acetic acid concentration by inoculation after 90 d was 4 times greater in HMC than in WPC.Such difference in acetic acid production might explain why inoculation reduced DM recovery in HMC but not in WPC.Hence, when acetic acid was produced in significant amounts, the association with Lpb.plantarum was not capable of counteracting the effects of Len.buchneri on DM loss.

Proteolysis
We found that after 90 d of ensiling, silages inoculated with Len.buchneri alone or in combination with Lpb.plantarum had a greater concentration of NH 3 -N in Experiments 2, 4, and 5.In addition, there was a tendency for LB400LP to have more NH 3 -N than CTRL at 90 d in experiment 3.In corroboration, da Silva et al. ( 2021) found that inoculation with Len.buchneri 40788 increased NH 3 -N content in HMC ensiled for 92 d.An increased concentration of NH 3 -N might indicate an intensification of proteolysis in the silo (Der Bedrosian et al., 2012).Therefore, inoculation with Len.buchneri alone or in combination with Lpb.plantarum apparently increased proteolysis in most experiments.Lactic acid bacteria generally have complex proteolytic systems but foster low proteolytic activity (Sasaki et al., 1995).The low pH of silage can restrict proteolysis (Rooke and Hatfield, 2003).Silage inoculated with Len.buchneri usually presents a higher pH than uninoculat-ed silage, as these bacteria convert moderate amounts of the stronger acid lactic acid into the weaker acid acetic acid (Kung et al., 2018).Therefore, the higher pH in silages inoculated with Len.buchneri could have been the reason of the more pronounced protein degradation in those silages, as previously suggested by Junges et al. (2017).The enhanced proteolysis in the silo can increase the degradation of the starch-protein matrix of corn kernels, facilitating the access of rumen microorganisms to starch grains (Hoffman et al., 2011;Fernandes et al., 2021).Silage proteolysis increases with the extension of storage time (Fernandes et al., 2021).This fact elucidates why we mostly detected an increase in proteolysis after 90 d of ensiling.An early effect (at 30 d) of inoculation on NH 3 -N content occurred only in experiment 5, in which LB400 and LB400LP had more NH 3 -N than CTRL.

Aerobic Stability
Inoculants containing Len. buchneri are principally intended to improve the aerobic stability of silage (Muck et al., 2018).After 30 d of ensiling, inoculation of SNA resulted in no significant improvement in aerobic stability (experiment 3), and while inoculation of WPC resulted in a small but significant improvement in aerobic stability (experiment 1), this improvement would likely be of minimal practical value.This was not unexpected as the conversion of lactic acid to acetic acid and 1,2-propanediol by Len.buchneri is a relatively slow process that does not occur immediately after inoculation (Schmidt et al., 2009).However, marked improvements in aerobic stability were observed after 30 d of ensiling in HMC inoculated with LB400, LB600, and LB400LP (experiments 4 and 5).Similarly, Saylor et al. (2020) observed that Len.buchneri LB1819 at 75,000 cfu/g plus Lactococcus lactis O224 at 75,000 cfu/g improved the aerobic stability of HMC after 28 d of ensiling.Uninoculated HMC had a lower base concentration of acetic acid than WPC likely due to a more restricted fermentation, but inoculation promoted a greater relative increase in acetic acid in HMC than in WSC.Still, the reason why inoculation promoted more pronounced improvements in aerobic stability after short-term ensiling in HMC than in WPCS and SNA is unclear, but it could be because the inoculant might have faced a less competitive environment in HMC than in the other types of silage.Such fact is corroborated by the finding that the numbers of observed bacterial operational taxonomic units in HMC were low, indicating reduced diversity and competitive pressure.Low bacterial diversity in HMC has been previously reported by da Silva et al. ( 2021), whereas studies on WPC usually show high bacterial diversity (Gharechahi et al., 2017;Hu et al., 2018).After 90 d of ensiling, inoculation with Len.buchneri alone or in combination with Lpb.plantarum markedly improved the aerobic stability of WPC, SNA, and HMC in all experiments, likely due to the consistently higher levels of acetic acid produced relative to untreated CTRL silages.
Regarding WPC ensiled for 90 d, the lower dose of the inoculant containing Len. buchneri plus Lpb.plantarum increased more the aerobic stability in experiment 2 than the higher inoculant dose in experiment 1.This finding can be explained by the fact that the forage used in experiment 2 had a higher DM content than the forage used in experiment 1 (~37% vs. ~32%).In agreement, Benjamim da Silva et al. ( 2022) reported that inoculation with Lentilactobacillus hilgardii, Len.buchneri, and Pediococcus pentosaceus promoted greater improvements in aerobic stability after 60 d of ensiling when the whole corn plant had 39% DM than when it had 32% DM.Benjamim da Silva et al. ( 2022) raised the hypothesis that the wetter the silage, the more prone it is to the development of undesirable bacteria, such as acetic acid bacteria, which are not inhibited by acetic acid but instead can use this acid to develop and initiate the aerobic deterioration of WPC.

Dose Effect
We tested Lpb.plantarum in combination with 2 different application rates of Len.buchneri in WPC (experiment 1) and 2 rates of Len.buchneri in HMC (experiment 4).Ranjit et al., (2002) showed that inoculation of WPC with Len.buchneri at a lower dose of 250,000 cfu/g of fresh material extended the aerobic stability in 30 h compared with uninoculated silage, but inoculation with a higher dose (500,000 cfu/g of fresh material) increased the aerobic stability by more than 536 h.In the present study, while WPC treated with the higher dose of Len.buchneri (400,000 cfu/g of fresh material) had more 1,2-propanediol (after 30 and 90 d of ensiling) and slightly greater aerobic stability (after 90 d of ensiling) than that treated with the lower dose (200,000 cfu/g of fresh material), such increments were small, making any additional cost associated with higher doses economically unjustifiable.Published studies evaluating the ideal application rate of Len.buchneri for HMC are rare.However, a study from Basso et al. (2012) compared the effects of Len.buchneri when applied in HMC at 50,000, 100,000, 500,000, and 1,000,000 cfu/g of ground corn and observed that an application rate of more than 500,000 cfu/g of ground corn promoted the larger improvements in aerobic stability.In the present study, in HMC ensiled for 30 d, the lower dose (400,000 cfu/g of fresh material) of Len.buchneri promoted lower DM recovery but greater Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE aerobic stability than the higher dose (600,000 cfu/g of fresh material).However, after 90 d of ensiling, there were no differences in response between applying the lower or higher dose of Len.buchneri in HMC.Therefore, in this study, the lower dose of Len.buchneri was sufficient to improve the aerobic stability of both WPC (200,000 cfu/g of fresh material) and HMC (400,000 cfu/g of fresh material) after 30 and 90 d of storage without air stress.

Effects of Air Stress During the Storage of HMC in Experiment 5
One of the most important factors determining the success of silage fermentation is the amount of air in the silo (Woolford, 1990).In experiment 5, applying air stress to HMC during ensiling clearly affected the process, resulting in significantly higher yeast levels, lower DM recovery, and lower aerobic stability than when no air stress was applied.Previous studies have also indicated that the degree of aerobiosis in the silo can influence inoculants' effectiveness (Kung et al., 2021).Even under air stress conditions, inoculation with Len.buchneri was able to modify the fermentation profile, by decreasing the concentration of lactic and butyric acids and increasing the concentration of acetic acid and 1,2-propanediol, and markedly improve aerobic stability.An important finding was that the aerobic stability of LB400 and LB400-AS were similar.This indicates that whereas air stress reduced the aerobic stability in CTRL, it did not reduce the aerobic stability in inoculated HMC (up to 250 h of aerobic exposure).

Bacterial Community Composition of HMC During Ensiling and Aerobic Exposure in Experiment 5
In experiment 5, we analyzed the effects of inoculation on the bacterial community composition of HMC at 30 d of ensiling and at 90 d after 0, 50, and 140 h of aerobic exposure.Inoculation reduced the bacterial diversity and affected the bacterial community composition of HMC after 30 and 90 d of ensiling.The high RA of the Lactobacillaceae in treated silages implied good establishment of the inoculated lactobacillus strains in these silages, having outcompeted the Enterobacteriaceae, Leuconostocaceae, and Streptococcaceae, which were far more abundant in uninoculated HMC.Considering only non-air-stressed HMC, Clostridiaceae was detected in CTRL but not in inoculated HMC after 30 and 90 d of ensiling.Regarding air-stressed HMC at 90 d, the RA of Clostridiaceae was 4.82% in CTRL but only 0.01% in inoculated HMC.Such findings indicated that inoculation with Len.buchneri alone or in combi-nation with Lpb.plantarum inhibited the development of clostridia in HMC.Similarly, da Silva et al., (2021) detected Clostridiaceae in uninoculated HMC but did not detect this family in HMC inoculated with Len.buchneri, L. hilgardii, or with a combination of Len.buchneri and L. hilgardii after 30 and 92 d of ensiling.Even though we observed differences in Clostridiaceae RA among treatments, only small amounts of butyric were detected.
Although the family Bacillaceae was present in nonair-stressed and air-stressed CTRL HMC after 140 h of aerobic exposure, it was not detected in inoculated HMC.The Bacillaceae have been associated with the second wave of spoilage that occurs after yeasts have consumed lactic acid causing the pH to increase, which allows the development of less acid-tolerant microorganisms, such as bacilli (Pahlow et al., 2003).This finding is supported by the fact that non-air-stressed and air-stressed CTRL HMC had extremely high yeasts numbers after 50 h of aerobic exposure and very high pH after 140 h.In contrast, inoculated HMC maintained low numbers of yeasts and low pH throughout the entire aerobic exposure period, even when subjected to air stress.

CONCLUSIONS
Few published studies have evaluated an inoculant over a multiple-year period.We found repeatable evidence that Lentilactobacillus buchneri PJB1 combined with Lpb.plantarum MTD1 applied to WPC and Len.buchneri alone or combined with Lpb.plantarum applied to SNA and HMC modified the fermentation profile by increasing the production of acetic acid and 1,2-propanediol in studies conducted over 3 years.The increased concentration of acetic acid inhibited the development of yeasts, markedly improving the aerobic stability of all ensiled materials after 90 d, and of HMC after only 30 d of ensiling.Additionally, in HMC ensiled for 90 d, inoculation with Len.buchneri diminished the detrimental effects of air stress during storage, as inoculated air-stressed HMC had considerably greater aerobic stability than uninoculated air-stressed HMC.shire, UK).The authors have not stated any conflicts of interest.
Figure 1.Relative abundances of the bacterial families with relative abundance greater than 0.10% in high-moisture corn ensiled for 30 d in experiment 5. CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB400LP = LB400 plus Lactiplantibacillus plantarum MTD1 at 100,000 cfu/g of fresh material.

Table 1 .
Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE Ensiled forage crops, inoculant treatments, replicate forage pile size, inoculation volume, ensiling period, and inclusion of air stress for the 5 experiments in this study Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE

Table 5 .
Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE Dry matter content, pH, chemical composition, microbiological composition, DM recovery, and aerobic stability of snaplage after 30 d of ensiling in experiment 3

Table 7 .
Dry matter content, pH, chemical composition, microbiological composition, DM recovery, and aerobic stability of high-moisture corn after 30 d of ensiling in experiment 4 a-c Means in rows with unlike superscripts differ (P ≤ 0.05). 1 CTRL = control; LB400 = Lentilactobacillus buchneri PJB1 at 400,000 cfu/g of fresh material; LB600 = Len.buchneri PJB1 at 600,000 cfu/g of fresh material. 2Reducing sugars. 3Lactic acid bacteria. 4Dry matter recovery. 5Aerobic stability, defined as number of hours before a 2°C increase above silage baseline temperature after exposure to air.

Table 8 .
Dry matter content, pH, chemical composition, microbiological composition, DM recovery, and aerobic stability of high-moisture corn after 90 d of ensiling in experiment 4 Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE

Table 9 .
Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE Dry matter content, pH, chemical composition, microbiological composition, DM recovery, and aerobic stability of high-moisture corn after 30 d of ensiling in experiment 5

Table 10 .
Dry matter content, pH, chemical composition, microbiological composition, DM recovery, and aerobic stability of high-moisture corn ensiled for 90 d with or without air stress in experiment 5

Table 11 .
Benjamim da Silva et al.: LENTILACTOBACILLUS BUCHNERI ON CORN SILAGE The pH, chemical composition, and numbers of yeasts of high-moisture corn ensiled with or without air stress for 90 d after 0, 50, and 140 h of aerobic exposure in experiment 5