Effects of glucose spraying and lactic acid bacteria inoculation applied to high-moisture alfalfa during pre- and post-harvest periods on silage fermentation and feed quality

Silage preparation

For the pre-harvest applications, alfalfa at the 10% flowering period (DM about %20) was theoretically inoculated with a density of 10⁸ cfu/g fresh weight with two bacterial inoculants and three glucose doses (%0, 3 and 6) were applied prior to 24 h from cutting on the field. For pre-harvest, field inoculations were conducted in three replications, with a 3-meter interval maintained between each application to minimize the contamination risk. Fresh forage yield was assessed by cutting 2 m² of alfalfa, which was weighed immediately prior to inoculation. Application dosages were subsequently calculated, and inoculation was performed using an atomizer for spraying.

For post-harvest, plant material was chopped into 2–4 cm fragments and separated into groups. Each strain and glucose was added to 4,000 g of fresh alfalfa plant material at a theoretical concentration of 10⁸ cfu/g ensuring thorough mixing by hand in sterile gloves.

Alfalfa materials were ensiled into plastic vacuum packages weighing 500 ± 40 g, with three replications conducted for all sampling times. A vacuum sealer was employed to remove 99.9% of O₂ from the silage bags.

Microbiological and chemical analysis

To monitor the microbial growth and pH changes for the first T₀: before ensiling, and final silage T₆₀: 60 days after ensilage. For T₆₀ openings, two harvest times (pre-pro harvest), three inoculants (control, L. citerum, L. plantarum), three glucose doses (%0, 3 and 6) and three vacuum silages for each application (2 × 3 × 3 × 3 = 54 silage packets) were made separately. After the ensiling period, the silage was opened, and extracts were obtained through filtration using Whatman 54 filter paper (Whatman, Florham, NJ, USA).

The pH of the silage extracts was measured immediately following filtration to assess the acidity resulting from the fermentation process. Microbial counts were conducted using ten-fold serial dilutions. The enumeration of lactic acid bacteria was performed by pour plating on MRS agar with a double overlay for anaerobic conditions, followed by incubation at 36 °C for 48 to 72 h. Enterobacteria were quantified using pour plating on violet red bile glucose agar (VRBD) with a single overlay, and the plates were incubated at 36 °C for 18 h. Yeast and mold counts were assessed by pour plating on malt extract agar (MEA) acidified to pH 4 with lactic acid, using a single overlay and incubating at 32 °C for 48 h. The dry matter (DM) content of both fresh forage (designated as T0) and the resulting silage (designated as T60) was determined by drying samples at 70 °C in a forced-air oven for 48 h. After 60 days of ensiling, the silages were analyzed for several parameters, including pH, neutral detergent fiber (NDF), acid detergent fiber (ADF) ash crude protein (CP). Nitrogen (N) content was measured using the Kjeldahl method, with crude protein content calculated as N × 6.25 (AOAC, 1990) method. The cell wall fiber components, including NDF, ADF and ADL, were analyzed according to the methodology described by Van Soest, Robertson & Lewis (1991).

Dry matter digestibility (DMD), dry matter intake (DMI), and relative feed value (RFV) of silages were calculated by the following formulas developed by Van Dyke & Anderson (2000). DMD values were used to calculate the RFV.

DMD (%) = 88.9−(0.779 × ADF %)

DMI (%) = 120/NDF %

RFV = DMD %× DMI %× 0.775

Statistical analysis

A factorial ANOVA model was used to analyse the effects of harvest time, glucose dose and inoculation type as fixed effects and their interactions. Repeats were considered as random effects. JMP statistical software was used for all analyses and mean comparisons were performed using the LSD test at a significance level of p < 0.05.

Effects of harvest, inoculant, and glucose dose on ph, dry matter content, and microbial population of fresh alfalfa material before ensiling (T0)

Table 1 shows the effect of glucose doses and inoculant treatments on microbial population, pH levels, and dry matter (DM) content of fresh material before and after harvest. pH (T₀) values were statistically significantly affected by the two-way interactions of Harvest x Inoculant, Harvest x Glucose, Glucose x Inoculant, and the three-way interaction of Harvest x Glucose x Inoculant (p < 0.05). Harvest x Glucose x Inoculant interaction significantly affected pH (T₀) (p < 0.001), and the highest pH(T₀) value of 6.60 was determined at post-harvest L. plantarum and 6% glucose dose, while the lowest pH(T₀) value of 6.30 was determined at pre-harvest L. plantarum and 0% glucose dose.

According to the research findings, both Harvest × Inoculant (p < 0.001) and Harvest x Glucose x Inoculant interactions were found to be statistically significant (p < 0.05) on dry matter (DMT₀) content. When considering the interaction between harvest and inoculant, the highest dry matter (%DM) contents were determined in L. plantarum and L. citerum inoculants applied pre-harvest (25.17% and 25.26%, respectively). All other applications were statistically similar and did not show a significant difference when compared to these two applications (p < 0.001) (Table 2). Harvest x Glucose x Inoculant interaction significantly affected DM(T₀) (p < 0.001), and the highest DM(T₀) content of 25.78% was determined at pre-harvest L. citerum and 3% glucose dose. The lowest DM (T₀) contents were found to be 20.86% for post-harvest L.plantarum inoculant and 0% glucose dose. This indicates that both pre-harvest applications and the combined effects of the inoculant and glucose doses applied play a decisive role in the dry matter accumulation of silage.

In the triple interaction, the highest LAB count was determined in the post-harvest L. citerum inoculant and 3% glucose dose treatment, while the lowest LAB count was determined in the pre-harvest control treatment. The count of enterobacteria was not statistically affected by harvest, glucose dose, and inoculant treatments. Glucose doses and inoculant applications did not statistically affect the yeast count of fresh material. It was determined that harvest and harvest x glucose x inoculant interaction statistically affected the count of yeasts. According to the harvest*inoculant*glucose dose application, the lowest yeast count was detected in the post-harvest control application and 3% glucose dose application (p < 0.01).

Effects of harvest timing, inoculant type, and glucose dose on fermentation parameters and microbial population of alfalfa silage after 60 days (T60)

Microbial population count, dry matter content and pH value of ensiling silages with different treatments after 60 days are shown in Table 3. pH values ranged from 4.91 to 5.50, with Glucose x Inoculant interactions having a particularly significant effect on pH values (p < 0.02). The highest pH value was observed in the 0% glucose + control application, while the lowest pH value was detected in the 6% glucose + L. plantarum combination (4.91), indicating that the interaction between the inoculant and glucose rapidly stabilised the fermentation process (Table 2).

DM (T_{6 0}) contents ranged from 19.67% to 24.00%, and the interaction between harvest and inoculant had a significant effect on DM contents, as shown in Tables 3 and 4. According to Table 2, the highest DM content was obtained with the application of L. citerum (24.00%) and L. plantarum (23.20%) pre-harvest. Other treatments were statistically similar.

Table 3 shows that glucose doses statistically affected the count of LAB, and as the glucose dose increased, the count of LAB increased significantly. Post-harvest applications were found to have a lower count of enterobacteria. But the count of yeasts was lower in the pre-harvest treatments. Glucose doses did not affect enterobacteria and yeast counts in a statistically significant way. L. citerum inoculant significantly reduced yeast counts when compared to control and L. plantarum. As shown in Table 3, L. citerum was the most effective inoculant. Mold could not be determined on the 60th day of silage.

Chemical composition and relative feed value of alfalfa silage as effected by harvest timing, inoculant type, and glucose dose

Table 4 shows the chemical composition of alfalfa silage. Although the interaction effects in Table 4 were not statistically significant, noticeable mean differences among treatments were still observed. In the present study, NDF contents of harvest x inoculant x glucose doses interaction from 32.25 to 35.50% and no statistically significant difference occurred. ADF and ADL value was lower in pre-harvest treatments. A significant harvest x inoculant x glucose doses interaction was observed for ADL content (p < 0.03). The lowest ADL value was recorded in the L. citerum and 3% glucose treatment pre-harvest, while higher values were observed in the post-harvest L. citerum and 3% glucose treatment. Combining glucose supplementation and bacterial inoculant before harvest can effectively reduce the lignin content. The average RFV value was found to be 191.91 RFV value was not statistically affected by the treatments.

Post-harvest (20.17%) exhibited significantly lower CP content compared to pre-harvest (21.99%) of silage alfalfa (p < 0.001). Glucose doses (p < 0.0021) significantly affected CP content. The highest protein content was determined at 6% glucose dose with 21.76% and the lowest was determined in the control group. Inoculation treatment and Harvest*Glucose*Inoculant interaction did not affect protein content (Table 4).

Correlogram of microbial and chemical traits in alfalfa silage

Figure 1 shows the correlation matrix between various chemical and microbial parameters affecting the quality of alfalfa silage after 60 days of fermentation. Positive correlations are indicated by red circles and negative correlations by blue circles, with the size of the circles proportional to the strength of the relationship. The lower triangle shows Pearson correlation coefficients with statistical significance levels (^∗p <  0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001). In particular, a strong positive correlation was observed between dry matter (DM) and dry matter conservation (DDM), emphasizing the contribution of moisture control to dry matter conservation. In contrast, neutral detergent fiber (NDF) and acid detergent fiber (ADF) were negatively correlated with relative feed value (RFV), suggesting a trade-off between fiber content and digestibility. Similarly, a positive correlation was observed between crude protein (CP) and relative feed value (RFV). This suggests that feeds with higher protein content generally have higher feed quality. Negative correlations were also found between the enterobacteria parameter and pH, suggesting that especially enterobacteria species are suppressed in the acidic environment of fermentation. At day 60, a positive and statistically significant relationship (r = 0.30, p < 0.05) was determined between the yeast (T60) population and the final pH (T60) value of silage. This result indicates that yeast growth was higher in silage samples with higher pH levels. The fact that yeast species become more active in low acidic environments also supports this finding biologically. Insufficient acid formation during the fermentation process, pH not decreasing sufficiently and therefore yeast growth cannot be controlled may adversely affect the stability of silage. This finding shows that low pH is important not only for bacterial fermentation but also for the control of yeast populations. There was a moderate negative and statistically significant relationship (r = −0.39, p < 0.01) between crude protein content and yeast population at 60 days. This shows that yeast growth was suppressed in silage samples with high protein content. In other words, we can say that there was no dehydration in protein content in silages with low yeast growth. Probably, fermentation progressed more effectively in these samples and the acidic environment was better formed. No statistically significant relationship was observed between crude protein and enterobacteria population (T60) (r = 0.01, p < 0.93).