Fermentation substrates and chemical analysis

Four forage species commonly used for grazing ruminants in Mexico were utilized as fermentation substrates (Table 1): alfalfa (Medicago sativa L) cv San Miguel, orchardgrass (Dactylis glomerata L) cv Potomac, cocuite (Gliricidia sepium (Jacq.) Kunth ex Walp.) and guinea grass (Panicum maximum Jacq.) cv Tanzania. Alfalfa, guinea grass and orchardgrass were cut at 7 cm above ground level; only leaves of cocuite were collected by hand from branches of several trees. Enough sample material was collected to obtain at least 1 kg of sample (DM basis) for each forage species. Collected forage samples were dried in a forced air oven at 60 ºC for 96 h; they were ground to pass through a 1 mm screen (Wiley Mill, Arthur H. Thomas Co., Philadelphia, PA) and analyzed for crude protein, ash, ether extract²⁴ (methods # 976.06; # 942.05; # 920.39, respectively). Acid detergent fiber (ADF)²⁵, neutral detergent fiber (NDF) were assayed without heat stable amylase and expressed inclusive of residual ash²⁵, and soluble sugars²⁶.

Ruminal fluid collection, preservation and reactivation

In vitro procedures reported for each experiment included three replicates^27,28. Similar to previous studies²⁹, fresh ruminal fluid was collected for each in vitro run from three ruminally fistulated adult Creole rams with an average body weight of 53.0 kg. Donor rams were fed a diet containing 80 % forage and 20 % concentrate. Feed was offered at 0900 and 1500 h every day; ad libitum intake was allowed. In addition, clean fresh water was available ad libitum. Rams were fitted with a ruminal cannula to collect ruminal fluid by suction³⁰. The collected ruminal fluid was filtered through four layers of cheesecloth and equal volumes of ruminal fluid from each donor were combined in order to obtain a representative sample and to prevent animal-to-animal variations^31,32,33.

Five percent (v/v) glycerol (Sigma-Aldrich, St. Louis, MO) was incorporated to serve as a cryoprotectant of ruminal inoculum^12,13. Aliquots were placed in 10-mL sterile glass containers. Containers were hermetically sealed and frozen at -70 ºC for 3 d. Lyophilization was conducted as previously described⁸ (Labconco Lyph Lock, model 195) under vacuum (-0.133 mBar), and inoculum was stored until later use. The reactivation of the preserved inoculum was conducted by reconstituting lyophilized samples in a cysteine solution to a volume equal to that of the original strained ruminal fluid. This solution contained 2.5 g of L-cysteine, 2.5 g of sodium sulfite and 0.1 mL of resazurin (1%) dissolved in 15 mL of sodium hydroxide (2N), distilled water was added to make a total volume of 100 mL (Table 2), which served as a buffer and created a reduced environment in the media, simulating the reduced conditions of the rumen. Reconstituted inoculum was incubated at room temperature for 10 min to allow rehydration, the reconstituted inoculum was transferred to 100 mL of culture medium, then it was pre-incubated at 39 °C for 24 or 12 h. The culture medium used and the pre-incubation time varied depending on the experiment, as described below.

Ruminal inocula evaluated

Experiment 1. Four types of ruminal inoculum were evaluated for measurements of fermentation kinetics and IVDMD. A fresh ruminal fluid (Control) was compared to lyophilized inocula reactivated by pre-incubation for 24 h in 1 of 3 culture media (Table 2). Specifically, treatments were 1) CONTROL, fresh ruminal fluid; 2) LOW24, inoculum reactivated in a medium containing 100 mL of a basal culture solution (composed of 50 % distilled water, 29 % clarified ruminal fluid, 14 % mineral solutions I and II, 5 % sodium carbonate, 2 % cysteine solution) and 0.1 % resazurin; 3) MODE24, inoculum reactivated in a medium containing 100 mL of the basal culture solution, 0.1 % resazurin, 0.5 g of yeast extract (Sigma-Aldrich, St. Louis, MO) and 0.5 g of peptone from casein (Bioxon Becton Dickinson, Mexico); and 4) HIGH24, inoculum reactivated in a medium containing 100 mL of a basal culture solution, 0.1 % resazurin, 0.5 g of yeast extract (Sigma-Aldrich, St. Louis, MO), 0.5 g of peptone from casein (Bioxon Becton Dickinson, Mexico), 0.3 g of glucose, 0.3 g of cellobiose and 0.25 g of starch. Media also included 0.25 g of ground forage on a DM basis (Table 2).

Experiment 2. In the second experiment, two types of ruminal fluid were compared: 1) CONTROL, fresh ruminal fluid, and 2) HIGH12, inoculum reactivated using a medium previously described for HIGH24. However, in this experiment, preserved ruminal inoculum was reactivated by pre-incubation for only 12 h as an attempt to find a more practical and faster approach for the reactivation process.

Fermentation kinetics and IVDMD

In each experiment, fresh and reactivated ruminal inocula were combined with a diluting agent containing the reduced mineral solutions I and II and the cysteine solution³⁴ at a ratio of 1:9 (v/v, ruminal fluid: diluting agent, Table 2). CO₂ was flushed while adding the diluting agent to the ruminal fluid, which was maintained at 39 ºC.

Afterwards, the fermentation kinetics and forage IVDMD were determined by combining 90 mL of diluted ruminal inoculum with 0.5 g of fermentation substrate using 125-mL glass bottles. The determination of the parameters of fermentation kinetics was based on the procedure utilized for the measurement of gas^35,36. More specifically, gas pressure (kg/cm²) was recorded at 1, 2, 4, 6, 10, 14, 18, 24, 30, 38, 48 and 72 h of incubation. After recording this value at each time point, the gas pressure was reset to cero. The values of pressure were then converted to volume of gas (mL/g DM of substrate); to do so, a standard curve was first generated by injecting known volumes of CO_2. The equation of this standard curve was generated by adding a linear regression line, and this equation was: Gas volume (mL/g of substrate) = pressure (kg/cm²)* 39.46 + 0, with an R² of 0.94. This standard curve was generated at room temperature. The use of this technique has also been recently reported by other investigators²⁹.

The accumulated volume of gas at each time point was used to estimate the parameters of the fermentation kinetics: maximum volume of gas (Vm; mL/g), lag phase (L; h), and the rate of gas production (S; h^-1). This was conducted using a logistic model described by Schofield et al³⁷:

Where:

Vm is maximum volume;

S is the rate of gas production;

t is the time point of measurement;

L is the lag phase.

In addition to parameters of fermentation kinetics, the IVDMD of substrates was determined at 24 and 72 h fermentation (IVDMD₂₄, and IVDMD₇₂, respectively). At each time point, the content of corresponding fermentation bottles was filtered through Whatman filter paper No. 4. The residue was dried at 100 ºC for 12 h in a forced air oven and the dry weight was recorded. Then, IVDMD was calculated relative to the amount of original sample used.

Statistical analysis

The GLM procedure of SAS³⁸ was used. In each experiment (n =3), the mean values within each fermentation substrate were considered as the experimental unit. Given the controlled experimental conditions, significant effect was declared at P<0.01; this level of significance may also contribute to reduce the type I error risk. Only when first order interaction was not significant, mean separation for main effects was conducted by Tukey; otherwise, paired mean separation by t-test was performed. Dispersion parameter reported is the largest standard error of the mean (SEM).

Data from Exp 1 were analyzed as a completely randomized experimental design with a 4 × 4 factorial arrangement of treatments (4 inoculum types and 4 fermentation substrates). Data from Exp 2 were analyzed according to a completely randomized experimental design with a 2 × 4 factorial arrangement of treatments (2 inoculum types and 4 fermentation substrates). In both experiments, the main effects of inoculum type and fermentation substrate were analyzed. The interaction of inoculum type × fermentation substrate was also evaluated. The statistical model for the analyses was:

Where:

Y_ijk represents the observation of the ijk treatment;

µ represents the overall mean;

τ_i represents the inoculum type i;

β_j represents the fermentation substrate j;

τβ_ij represents the interaction effect of the inoculum type i and the fermentation substrate j.

The residual term ε _ijk was assumed to be normally, independently, and identically distributed, with variance σ² _e .

Chemical composition of forages used as fermentation substrates

Analyzed chemical composition of the four forages used is listed in Table 1. Guinea grass was low in crude protein and high in fiber contents (65.0 and 779.0 g/kg DM for crude protein and NDF, respectively); whereas alfalfa was high in protein and low in fiber content (206.0 and 442.0 g/kg DM for crude protein and NDF, respectively), with these nutrients having intermediate values for orchardgrass and cocuite.

Fermentation kinetics and IVDMD for inocula reactivated by 24 h pre-incubation

Figure 1 illustrates in vitro gas production for the CONTROL and inocula reactivated by pre-incubation for 24 h. CONTROL displayed the fastest and greatest maximum gas production compared to the other treatments. Even with the HIGH24 treatment, fermentation was reduced by around 50 % during the first hours of incubation. Within the preserved inocula, HIGH24 displayed the greatest maximum gas production, followed by MODE24 and by LOW24. In addition, Figure 2 displays gas production for each fermentation substrate. Alfalfa had the greatest and fastest maximum gas production compared to the rest of forages. Orchard grass displayed intermediate values for gas production, with cocuite and guinea grass having the lowest values.

CONTROL= fresh ruminal fluid; LOW24= inoculum reactivated by 24 h pre-incubation in a basal culture solution; MODE24= similar to LOW24, but included yeast extract and peptone from casein; HIGH24= similar to MODE24, but included carbohydrates. Control: Vm=387.15 mL/g, L=4.06 h, S=0.041 h^-1; LOW24: Vm=266.11 mL/g, L=13.70 h, S=0.018 h^-1; MODE24: Vm=288.22 mL/g, L=7.62 h, S=0.023 h^-1; HIGH24: Vm=332.83 mL/g, L=8.05 h, S=0.32 h^-1.

Figure 1 In vitro gas production for the control and the preserved inocula reactivated by pre-incubation for 24 h in different culture media 

Alfafa: Vm=357.90 mL/g, L=5.22 h, S=0.030 h^-1; orchardgrass: Vm=333.50 mL/g, L=11.77 h, S=0.029 h^-1; cocuite: Vm=291.50 mL/g, L=3.83 h, S=0.030 h^-1; guinea grass: Vm=291.20 mL/g, L=12.20 h, S=0.025 h^-1.

Figure 2 In vitro gas production for four forages when values for fresh ruminal fluid and inocula reactivated by pre-incubation for 24 h were averaged 

More specifically, fermentation kinetics, IVDMD₂₄ and IVDMD₇₂ were affected by treatment and fermentation substrate (Table 3). Regardless of the nutrient composition of the medium used for reactivating the inoculum, Vm decrease (P<0.01) when reactivated inoculum was used compared to fresh ruminal fluid, with this difference being greater during the first 24 h (Figure 1). However, treatment HIGH24 displayed greater Vm (P<0.01) compared to treatments MODE24 or LOW24 regardless of fermentation substrate. In addition, both alfalfa and orchardgrass had the highest (P<0.01) Vm across treatments with an average of 345.7 ± 14.40 mL/g. The interaction of treatment × fermentation substrate was significant (P<0.01) for L and S. Specifically, L was highest (P<0.01) for LOW24 when orchardgrass or guinea grass was used as fermentation substrates with an estimate of 22.29 h. However, there was no difference in L among treatments when cocuite was used as fermentation substrate. Furthermore, S was lowest (P<0.01) for treatment LOW24 regardless of fermentation substrates with an average of 0.018 h^-1. However, S reached highest (P<0.01) values in most fermentation substrates when CONTROL was used as inoculum followed by HIGH24.

The treatment × fermentation substrate interaction was significant (P<0.01) for IVDMD₂₄ and IVDMD₇₂. Specifically, when alfalfa was used as fermentation substrate, IVDMD₂₄ for treatments MODE24 and HIGH24 was similar (P≥0.1) to CONTROL with an average of 521.3 ± 10.8 g/kg. However, IVDMD₂₄ for alfalfa was lower (P<0.01) for treatment LOW24 compared to CONTROL with estimates of 432.0 and 558.0 ± 10.8 g/kg for LOW and CONTROL, respectively. Likewise, IVDMD₇₂ for treatments MODE24 and HIGH24 were similar (P≥0.1) to CONTROL with an average of 642.0 ± 10.5 g/kg. However, IVDMD₇₂ for alfalfa was lower (P<0.01) for LOW24 compared to CONTROL24 with estimates of 590 and 622.0 ± 10.5 g/kg for LOW24 and CONTROL, respectively. The lowest (P<0.01) IVDMD₇₂ was observed for treatment LOW24 when cocuite was used as fermentation substrate with an average of 439.0 ± 10.5 g/kg. Overall, regardless of fermentation substrate, there was a depression (P<0.01) in IVDMD for LOW24 compared to any of the other treatments.

Fermentation kinetics and IVDMD for inoculum reactivated by 12 h pre-incubation

Figure 3 illustrates in vitro gas production for the CONTROL and inoculum reactivated by incubation for 12 h. CONTROL displayed a faster and greater maximum gas production compared to HIGH12. In addition, Figure 4 illustrates gas production for each fermentation substrate. Alfalfa displayed the greatest and fastest maximum gas production compared to the rest of forages, with cocuite and guinea grass having the lowest values.

CONTROL= fresh ruminal fluid; HIGH12, inoculum reactivated by 12 h pre-incubation in a basal culture solution, yeast extract, peptone from casein and carbohydrates. Control= Vm=410.80 mL/g, L=5.42 h, S=0.032 h^-1; HIGH12= Vm=264.97 mL/g, L=4.29 h, S=0.017 h^-1.

Figure 3 In vitro gas production for the control and the preserved inoculum reactivated by pre-incubation for 12 h in a culture medium 

Alfafa: Vm=368.30 mL/g, L=2.13 h, S=0.032 h^-1; pasto de ovillo: Vm=352.51 mL/g, L=9.19 h, S=0.033 h^-1; cocuite: Vm=334.16 mL/g, L=3.79 h, S=0.027 h^-1; pasto de Guinea: Vm=296.56 mL/g, L=4.31 h, S=0.025 h^-1.

Figure 4 In vitro gas production for four forages when values for fresh ruminal fluid and inoculum reactivated by pre-incubation for 12 h were averaged 

Specifically, fermentation kinetics, IVDMD₂₄ and IVDMD₇₂ were affected by treatment and fermentation substrate (Table 4). Vm was greater (P<0.01) for CONTROL compared to HIGH12, with estimates of 410.80 and 264.97 ± 13.050 mL/g, respectively. The interaction of inoculum type × fermentation substrate was significant for L (P<0.011); orchardgrass and alfalfa incubated in HIGH12 had the greatest and lowest (P<0.01) L, respectively; with estimates of 12.4 and 0.07 ± 0.700 h for orchardgrass and alfalfa. Likewise, an interaction (P<0.01) was detected for S; alfalfa incubated in CONTROL and cocuite incubated in HIGH12 had the greatest and lowest S, respectively; with estimates of 0.047 and 0.013 ± 0.007 h^-1 for alfalfa and cocuite. Regardless of forage type, the IVDMD₂₄ was greater for CONTROL compared to HIGH12 with estimates of 520.6 and 374.3 ± 12.70 g/kg, respectively. There was an inoculum type × fermentation substrate interaction (P<0.01) for IVDMD₇₂ with alfalfa and orchargrass incubated in CONTROL having the greatest IVDMD₇₂, and cocuite and guinea grass having the lowest IVDMD₇₂ values.

Ruminal fluid sampling and the use of glycerol as a cryoprotectant

Studies^31,32,33 have found differences in in vitro fermentation patterns between ruminal fluid from different donors. The source of ruminal fluid can influence in vitro fermentation and digestibility trials^31,39. Differences in fermentation patterns among animals observed by those researchers can be partially attributed to differences in the composition of the established bacterial community among host animals^40,41. Therefore, in the present study, ruminal fluid samples from three donors were pooled to obtain a representative sample to prevent bias due to ruminal fluid source.

The use of glycerol improves the preservation of the ruminal bacterial community^12,13,14. Benefits of glycerol may be explained by peripheral vitrification providing protection to the bacterial cytoplasmic membranes from potential damage that can be caused by ice crystal formation⁴². More specifically, glycerol penetrates the cells, which can then protect them from damage by maintaining a semi-fluid state^43,44. Consequently, the use of glycerol not only protects the integrity and viability of the ruminal bacterial cells, but may also prevent degradation of the microbial DNA¹³.

Fermentation kinetics and IVDMD of preserved inocula

In vitro fermentation kinetics and IVDMD revealed differences between fresh and lyophilized ruminal fluid. When compared to fresh ruminal fluid, the greatest depression in fermentation kinetic parameters was observed when lyophilized inocula were reactivated in media without sugars or growth promoters. These observations are in line with other studies¹⁵, indicating a depression on fermentation parameters with frozen inocula, which can be explained by a decrease in microbial activity due to microbial death or nutrient limitation⁹. In addition, researchers have reported⁸ that protein degradation rates with preserved ruminal microorganisms were 4 to 8 times slower than when using fresh ruminal fluid. Furthermore, the use of inoculum preserved through freezing affects fermentation parameters during the first hours of fermentation³⁰, and deep freezing may represent a better preservation method compared to freezing at -20 °C. Consequently, in agreement with recent reports, the reactivation of preserved bacteria is one of the most critical steps in obtaining active and effective microorganisms for in vitro fermentation trials^21,22,23.

In this study, compared to fresh ruminal fluid, the negative effects of lyophilization on fermentation kinetics and IVDMD was less severe when ruminal inocula were reactivated in a nutrient-rich medium including a basal culture solution, growth promoters and sugars. These observations indicate that growth promoters such as yeast extract and peptone from casein, and carbohydrates such as glucose, cellobiose and starch enhance the reactivation of ruminal microorganisms, thus, improving in vitro fermentation. The need for yeast extract in the medium for adequate bacterial reactivation and growth may be attributed to the absence of the genes for the synthesis of some proteinogenic amino acids such as arginine and asparagine in the genome of some ruminal bacterial species^45,46, which indicates that these amino acids contained in the yeast extract need to be included in the medium. Additionally, peptone from casein and carbohydrates provide readily available nitrogen and energy stimulating microbial reactivation, growth and activity^19,47. It is interesting to note the different patterns (gas production at different time points) among the fermentation curves, which suggests that different microbial populations may be acting on the substrates at each fermentation time point. Further research should aim at investigating shifts in the structure of the microbial community^48,49 using techniques such as high-throughput DNA sequencing^50,51,52, which allows a broad evaluation of the profile microbial community from highest to lowest taxonomic levels.

It has been found that, in comparison to fresh ruminal fluid, digestibility of alfalfa decreased 17.63 % when using frozen inoculum or lyophilized inoculum reactivated by 24 h pre-incubation in McDougall’s solution⁹. In contrast to previous observations⁹, in the present study, IVDMD₇₂ was not affected when using lyophilized inoculum reactivated by 24 h pre-incubation in a nutrient-rich medium; indicating that, compared to the use of McDougall’s solution, using a nutrient-rich medium containing a wider range of nutrients represents a better approach for stimulating the reactivation and activity of ruminal microorganisms.

When averaged across fermentation substrates (i.e. alfalfa, orchard grass, cocuite and guinea grass), the IVDMD for any of the reactivated inocula was negatively affected, compared to the values obtained with the control. However, within lyophilized inocula, the reactivation by 24 h pre-incubation in a nutrient-rich medium displayed the best performance. In addition, when the inoculum was reactivated by 12 h pre-incubation, IVDMD values were lower compared to the control fresh ruminal fluid. It is important to note that, at 72 h fermentation all forages but cocuite reached almost 100 % of the total gas produced. This indicates that 72h-fermentation rates are not suitable for measuring the effectiveness of the treatments, which also suggests that a better approach would be to measure fermentation rates and IVDMD at 24 or 48 h of fermentation.

Effect of fermentation substrate on fermentation kinetics and IVDMD

The use of fermentation substrates with a wide range of nutrient composition facilitated the evaluation of our hypothesis under different scenarios. Overall, our results revealed that forages from temperate zones, namely alfalfa and orchardgrass, had higher Vm and IVDMD compared to their counterparts from the tropical regions. These observations were likely due to differences in the structural components of the plant cell-wall existing between forages from temperate and those from tropical zones⁵³.

Furthermore, alternative inoculum sources have been suggested for in vitro fermentations. One of these sources is ruminant feces; however, results have been inconsistent. For example, fecal inoculum has been demonstrated to be effective for in vitro gas production studies⁵⁴; nonetheless, fecal inoculum from sheep was not comparable to fresh ruminal fluid when evaluating in vitro dry matter digestibility⁵⁵. In addition, other studies, have revealed that fecal inoculum does not perform as good as ruminal fluid in in vitro fermentation techniques^55,56, which may be due to differences in the bacterial populations between the rumen and the lower gastrointestinal tract⁵⁷.