Location

The experiment was developed in the Laboratorio de Microbiología Ruminal y Genética Microbiana, Programa de Ganadería, Campus Montecillo, Colegio de Postgraduados.

Experimental diets

Three diets were evaluated for this study. The diets were formulated in agreement with the NRC nutritional requirements ⁽²⁰⁰⁷⁾ for growing lambs, with 0, 20, and 40 % corn DDGS and with a 15:85 fodder:concentrate ratio. The DDGS partially replaced the corn grain, soybean meal, and corn gluten in the concentrate, maintaining the isonitrogenous diets (Table 1).

Chemical analysis

The chemical composition of the diets (Table 1) was determined using the methods of the ^{AOAC (2005)}: dry matter (MS; method 930.15), ash (method 942.05), ethereal extract (EE; method 954.02), and crude protein (CP, method 984.13). The content of neutral detergent fiber (NDF) and acid detergent fiber (ADF) was determined using the method proposed by ^{Van Soest et al. (1991)}.

Animals and rumen fluid

Three Rambouillet lambs (37 ± 2.5 kg; average six months old) with cannula in the rumen were used to obtain fresh rumen fluid (RF). The lambs were adapted to the experimental diets during 21 d before the in vitro study started. On the 22^nd day, 500 mL of RF were collected from each lamb, 3 h after the morning feeding. The RF was carried in a thermos flask (at 39 °C) to the laboratory, where the biofermenters were inoculated with the experimental diets.

Biogas measurement

Ingredients and diets were ground using a 1 mm mesh in a Willey mill (Arthur H. Thomas, Philadelphia, PA). Samples of 0.5 g MS were weighed and poured into 120-mL serological vials, with 45 mL of culture medium for total bacteria (^{Cobos and Yokoyama, 1995}), in order to replace the energy sources (glucose, cellulose, and starch) with the experimental ingredients or diets according to the treatment. All vials were kept in anaerobic conditions with CO₂, and each vial was considered as a biofermenter and an experimental unit. The proportions of CH₄ and CO₂ were determined in a PerkinElmer^® gas chromatograph equipped with a thermal conductivity detector and a Porapak packed column. The detection conditions were: oven temperatures at 80 °C, packed column at 170 °C, and thermal conductivity detector at 130 °C; retention times were 0.71 and 1.05 min for CH₄ and CO₂, respectively. Helium was used as the carrier gas with a 23 mL m^-1 flow. The CH₄ and CO₂ molar concentration was calculated according to the Ideal Gas Equation (^{Posada and Noguera, 2005}). Volatile fatty acids (AGV) were determined in a PerkinElmer^® gas chromatograph with a flame ionization detector. The working conditions were: oven temperature at 130 °C, and injector and capillary column (15 x 0.32 m) at 250 °C. The retention times were: 1.26 min for acetate, 1.6 min for propionate, and 2.09 min for butyrate.

In vitro biogas production

In order to quantify the biogas (CH₄ and CO₂) amount, the biogas capture technique described by ^{Krabill et al. (1969)} -which uses an acidified saturated solution- was adapted, replacing the 20 % Na₂SO₄ (m/v) with NaCl, in order to prevent CO₂ from mixing with water. This methodology enables the study of different foods, and the practical, simple, and economic quantification of total biogas (CH₄ and CO₂).

The biogas was captured in saturated saline solution traps (370 g NaCl L^-1 of water, and 5 mL of methyl orange at 0.1 % as pH indicator) with pH 2 adjusted with HCl to 1 N. The 120-mL serological vials were completely filled with the saline solution and hermetically sealed with a 2-cm wide neoprene cap and with an aluminum ring with a Wheaton^® crimper. The traps were changed every 24 h during the 72 h fermentation period.

The biofermenters were incubated in a 39 °C baine-marie; afterwards, they were inoculated with fresh rumen fluid centrifuged at 1157 g during 3 min; finally, they were connected to traps by means of a Tygon^® hose (3/32” inside diameter). The hose was adapted with two 20 G x 1” Terumo^® yellow needles at each end. One hose was placed in the biofermenter and the other in the capture trap; the latter had a needle with the above-mentioned characteristics that served as release valve for the saturated saline solution. The trap was placed inverted in a plastic test tube with a V-shaped cut.

To prevent the obstruction of the biogas flow, the hose was blocked with a plastic clamp that released the gas at 24, 48, and 72 h of fermentation. The total biogas production was quantified measuring the movement of trap fluid in the test tube. To measure the amount of CH₄ and CO₂, a 500 μL sample was taken from the inverted space in the traps.

Evaluation of the rumen fermentation variables

Samples were taken from each biofermenter’s liquid phase at 72 h, using 2-mL Eppendorf^® vials, with a 25 % metaphosphoric acid (4:1 ratio), and they were frozen at -4 °C; subsequently, the AGV molar concentration was measured by gas chromatography (^{Erwin et al., 1961}). The concentration of rumen bacteria was evaluated by direct counting in a Petroff Houser^® chamber and using a 1000 X Olympuss^® contrast microscope. The pH was measured directly from the biofermenter at 72 h, using an ORION^® potentiometer, calibrated at pH 4 and 7. Biofermenter content was filtered through Whatman paper in order to recover non-degraded MS, and in vitro degradation of MS (DIVMS) was calculated by weight difference.

Design and statistical analysis

The experimental design was fully randomized, with five repetitions per variable. The variables (pH, DIVMS, AGV total molar concentration, and acetate, propionate and butyrate production) were analyzed at 72 h of fermentation; the production of total biogas, cumulative CH₄ and CO₂, and CH₄ and CO₂ production per 100 g of MS, were also analyzed at 24, 48, and 72 h. The results were analyzed with GLM from SAS (^{SAS Inc., 2011}) and the means were compared using the Tukey test (p≤0.05). To fulfill the normality and homogeneity of the variances, the bacteria concentration data were transformed to Log 10, prior to statistical analysis.

Rumen fermentation variables

The degradations were lower than expected for a completely mixed diet (85 % concentrate and 15 % fodder). In 20 and 40 % DDGS diets, DIVMS was 17.3 and 25.1 % lower (p≤0.05), respectively; this result is similar to that reported by ^{Félix et al. (2012)} (Table 2).

The difference between the DIVMS of the experimental diets is attributed to the partial substitution of corn kernel and soybean meal that have higher DIVMS than DDGS (Table 3). However, ^{Buckner et al. (2008)} and ^{Depenbusch et al. (2009)} show an optimal use of 20 to 30 % MS of DDGS in cattle fattening, and up to 40 %, without affecting neither feed conversion nor growth (^{Amat et al., 2012}; ^{Gibb et al., 2008}; ^{Klopfenstein et al., 2008}). According to ^{Schauer et al. (2008)}, the productive behavior of lambs fed with 20 % MS with DDGS diets does not change. ^{Avila-Stagno et al. (2013)} reported a linear reduction of the DIVMS when DDGS of wheat was used to replace soybean meal in the diet of lambs. When canola meal and barley grain were replaced with 20 % MS of wheat DDGS, DIVMS decreased 6.17 %; however, there was no difference, when they were substituted with 20 % MS of corn DDGS (^{McKeown et al., 2010}).

The AGV concentration did not change as a result of the inclusion of DDGS in diets, which is similar to what was reported by ^{Miśta et al. (2014)}, but ^{McKeown et al. (2010)} mention that with 20 % MS of DDGS in lamb growth diets, the propionate concentration increased without affecting the total AGV concentration. The increase of wheat DDGS in diet decreased the total AGV concentration, and the propionate had a quadratic behavior (^{Avila-Stagno et al., 2013}). According to ^{Pecka Kiełb et al. (2015)}, the butyrate molar concentration decreases after 24 h of incubation, since it includes 30 % MS of DDGS in the diet of lambs. Furthermore, including 40 % MS of DDGS in diets for finishing cattle decreases total AGV concentration (^{Hünerberg et al., 2013}). The differences of our results with those of other studies may be caused by the different ingredients used and their individual rumen fermentation (Table 3), in addition to the levels at which each ingredient was included in the diets.

The corn kernel has higher DIVMS than the corn DDGS (p≤0.05). DDGS, wheat bran, and corn gluten meal produce a higher total AGV concentration (p≤0.05), while the corn stover produces a smaller amount (Table 3).

The pH (6.7 average) was similar from one treatment to another (p>0.05). The other studies reached contrasting results: ^{Hünerberg et al. (2013)} and ^{Pecka-Kiełb et al. (2015)} reported that pH decreased, while ^{Avila-Stagno et al. (2013)} mentioned that pH increased.

Biogas production of experimental diets and ingredients used

Cumulative gas production is an indicator of the MS fermentation degree of diets in inverse ratio to digestibility. In this study, biogas production was higher in the 40 % DDGS diet, which matches its lower DIVMS (Table 4). According to ^{McKeown et al. (2010)}, biogas production decreased 5.96 % using 20 % corn DDGS to replace all the canola meal and part of the barley grain in the diet of lambs. The results of this study are similar: by including 20 % DDGS, the gas production diminished 11.37 %; but with 40 %, production was 44.4 % higher than control diet, and 62.5 % more than the 20 % DDGS diet. This could be the result of the different proportions in which corn kernel and soybean meal were substituted with DDGS in diets, which reduced the DIVMS.

However, the ingredients’ individual biogas production is not reflected in the diets, since corn kernel produces a greater amount of biogas (p≤0.05) than DDGS and the other ingredients(Table 5).

Production of CH₄ and CO₂ of experimental diets and ingredients used

As expected, the biogas production was higher after a longer fermentation time; however, the biogas production of CH₄ per each 100 mL of produced biogas decreased (22.8 %), and the biogas production of CO₂ increased (13.2 %) in the 40 % DDGS diet in relation to control (Table 6).

^{Avila-Stagno et al. (2013)} reported an increase in CH₄ production when 40 % DDGS was included, and they related this with the DIVMS decrease and the total biogas production increase. According to ^{Pecka-Kiełb et al. (2015)}, the biogas production increases in the diet of lambs with 30 % MS of DDGS, after 24 h of incubation. In our study, the CH₄ production at 72 h was 21.1 % greater in the DDGS-free diet compared to the 20 % DDGS diet, but was similar to the 40 % diet; however, the cumulative CO₂ was 22.2 % lower in the 20 % DDGS diet, compared to the 40.0 % DDGS diet (Table 7).

These results show that, as total gas production increases, so does the fermentation time, while the digestibility of diets is reduced; however, the CH₄ and CO₂ proportions of gas composition may vary, depending on the ingredients used to prepare the diets. ^{Johnson and Johnson (1995)} indicated that methane emissions -expressed as a percentage of energy consumption or per kg of MS consumed- are lower in livestock fed with diets high in concentrate than in fodder. However, in this study, with a 40 % DDGS diet, the CH₄ production was reduced by 22.8 % in relation to the control diet, but CO₂ production increased 13.1 %. Substituting corn kernel and soybean meal with DDGS in diets contributed to the decrease of DIVMS, perhaps as a result of a lower amount of starch and the increase in levels of fat, NDF, and ADF. The methane production of each ingredient indicates that ingredients with high protein content -such as DDGS, soybean meal, and corn gluten- produce more CH₄ at 72 h of incubation (Table 8); however, in fully mixed diets, including 20 % DDGS decreased CH₄ production.

According to ^{Grainger and Beauchemin (2011)}, the type and level of fat in diet can be the most important factor affecting methanogenesis, and that a 5-6 % increase of fat in diet decreases methane production by 5.1 % in livestock. In this study, the diets had 3.2, 3.7, and 4.5 % fat expressed as ethereal extract -with 0.0, 20, and 40 % DDGS, respectively-, which could have contributed to reduce methane production by 22.8 %, and to increase CO₂ by 6.1 %, in the 40 % DDGS diet, with respect to the control. According to ^{Hünerberg et al. (2013)}, when 40 % barley grain is replaced with corn DDGS, the fat in the diet reached 5.4 %, and CH₄ production was reduced by 6.3 % for every 1 % of fat; they concluded that CH₄ production -as a result of adding corn DDGS- will depend directly on the fat content in DDGS. The results of this study match those already mentioned: fat increase in diet is related to the amount of DDGS included in the diets and the reduction of CH₄ production. Fat content is associated with the decrease of MS rumen digestion, because high levels of AG are toxic for some bacteria (such as methanogens) and protozoa. According to ^{Martin et al. (2010)}, this decrease is related to a lower methane production in rumen. The protozoa population was not determined in our study; however, a total bacteria count was carried out and their population was 37 % lower in the 40 % DDGS diet, in relation to the control, which may indicate the negative effect of fat in diet on the microbial population, since the pH of the diets was higher than 6.5.