Experimental site and animal care

The experiment was conducted at Animal Metabolism Laboratory of Faculdade de Medicina Veterinária e Zootecnia, Universidade Federal de Mato Grosso do Sul (Campo Grande, Mato Grosso do Sul, Brazil). The experiment was conducted according to the institutional Ethics Committee on Animal Use under case no.: 577/2013.

Animals, management and treatments

Five male sheep (½ Suffolk + ½ Santa Ines) with a permanent cannula inserted in the rumen and, initial mean body weight (BW) of 46.50 ± 5.45 kg were used in the study. The sheep were housed in special cages for metabolism studies that are suitable for in vivo digestibility assays. These cages had a slatted wood floor and contained a feeder, drinking fountain, and a galvanized steel supplement for urine collection. The cages were housed in a covered shed with good ventilation.

The animals were fed with chopped Coast-cross hay (Cynodon dactylon (L.) Pers.) and concentrate (Table 1). The concentrate formulation contained fine ground corn (700 g kg^-1), soybean meal (260 g kg^-1) and mineral premix (40 g kg^-1). The experimental treatments included growth promoters additives in the concentrate as follows: HAY (hay only); CONT (hay + concentrate); MON (hay + concentrate + monensin), SALI (hay + concentrate + salinomycin), and FLAV (hay + concentrate + flavomycin). Hay was offered ad libitum, concentrate in the amount of 20 g kg^-1 of BW, and additives 0.75 mg kg^-1 of BW.

The additives were weighed on an analytical balance and stored in microtubes until use. Hay and concentrate were provided in separate troughs. The hay and concentrate were provided in two meals a day, at 0007 and 0017 h. Hay intake was adjusted to provide orts of 150 g kg^-1. Additives were provided only with the morning meal and mixed with the concentrate. The amount of hay and concentrate met the nutritional requirements of growing sheep with an average daily gain of 250 g d^-1(1.

The experimental design was a 5 x 5 latin square. Five experimental periods of 21 d each were performed as 10 d of adaptation to treatments and 11 d of data collection. At each new experimental period, the animals were weighed after 16-h solids fasting to adjust the amounts of concentrate and additive.

Intake control and collection of feces and urine

The control of daily intake of feed and water was performed between d 11 and 15 of each experimental period by weighing the amounts of feed offered and orts. Feces and urine were collected in the same period. Water intake was controlled by measuring the quantity supplied in the morning and afternoon. Orts water was measured in the drinking fountains only in the mornings. A control drinking fountain (without animal access) was also used to measure evaporation during the day in order to assess actual water intake in the experiments. A feces collecting bag was adapted to animals to allow total collection of feces. These bags were emptied daily in the morning and afternoon at the same times. Feces were weighed, homogenized, and samples equivalent to 100 g kg^-1 were stored (-20 ºC) for further analysis. Based on this information, the following parameters were evaluated: intake (offered - orts), apparent digestibility coefficients (nutrient intake - excreted nutrient/nutrient intake) of DM, OM, CP, neutral detergent fiber corrected for ashes with the use of amylase (aNDFom), ethereal extract (EE), and NSC. The TDN were estimated by the formula proposed by Sniffen et al¹⁰: TDN= digestible CP + digestible aNDFom + 2.25 * digestible EE + digestible NSC.

In addition to the feces, urine was collected for the evaluation of nitrogen balance. Urine was collected in buckets containing 100 mL of sulfuric acid (100 mL L^-1) placed in the lower part of urine collectors in the metabolic cages. These buckets were emptied daily in the morning and afternoon at the same time. Samples of 100 ml L^-1were collected and stored (-20 °C). For nitrogen (N) balance analysis, absorbed N was calculated by the difference between N intake and excreted in the feces, while the N retained was obtained by the difference between N intake and excreted in feces and urine.

Measurement of in situ degradability

The ruminal degradability of DM and NDF was determined from d 16 to 19 of each period using 5 x 5 cm nylon bags with a porosity of 50 μm, sealed at the edges and properly identified. These bags were weighed empty, filled with 2.5 g of hay (ground and passed through a 2 mm pore sieve), and tied with a rubber band to a metal ring to keep them closed. These bags were first soaked in water for one hour and subsequently attached to a metal chain and anchor weighing approximately 100 g. These bags were subsequently infused into the rumen via cannula at 0007 h (before feeding) and removed after the incubation times (3, 8, 16, 24, 48, 72 and 96 h). These bags were immersed in ice water immediately after removal from the rumen and washed in a washing machine for five minutes for three cycles, changing the water in each cycle. They were subsequently placed in a forced air ventilation oven at 55 ºC for 72 h and weighed after this period.

The DM soluble fraction from hay was determined with nylon bags with samples without incubation in the rumen. These bags were kept in water (38 °C) for one hour, washed in a washing machine, oven dried, and weighed. The difference between the initial and final weight was considered as the soluble fraction for each experimental period, which corresponds to the value at 0 h in the DM degradation curve. The soluble fraction “a”, insoluble fraction “b”, rate of degradation “c”, and effective degradability (ED) were calculated according to Ørskov and McDonald¹¹ with the equation ED = a + (bxc) / (c + k), where “k” is the estimated rumen solids passage rate calculated as 0.02, 0.05, and 0.08 h^-1 in the present study.

Collection of ruminal liquid samples and pH

Ruminal liquid samples were collected for the determination of VFA, pH, and N-NH₃ from d 20 to 21 in each experimental period. The samples were collected at the end of the experimental period after removal of the nylon bags. At the collection time, samples were taken at zero hours (before supplementation) and at 2, 4, 8, 10, 12, 14, 16, 18, 20, 22 and 24 h after feeding in the morning and always following the same sequence among animals. The ruminal fluid collection was performed with the help of a metal shell inserted in a cloth diaper. An aliquot of approximately 100 mL of ruminal fluid was collected.

The pH was measured immediately after the collection of ruminal liquid using a digital potentiometer (B474; Micronal, São Paulo, SP, Brazil). The VFA analysis used four mL of ruminal fluid acidified with 1 mL of metaphosphoric acid (25 %) and stored at -20 °C. The N-NH₃ analysis used 50 mL of ruminal liquid acidified with 1 mL of H₂SO₄ (50 %) and stored at -20 °C.

Chemical analysis

The analysis of the chemical composition of feeds, orts, and feces was conducted according to AOAC¹² as follows: DM - method 967,03; CP- method 981,10; Ashes - method 942,05; and EE - method 920,29. The NDF content was analyzed in a Tecnal TE-149® fiber analyzer (Tecnal, Piracicaba, SP, Brazil) using 5 X 5 cm non-woven fabric (NWF) bags with 100 μm porosity. To these were added 0.5 g of sample (feed or faeces) and followed for neutral detergent analysis according to the methodology of Van Soest et al¹³ without sodium sulfite and using thermostable amylase (Termamyl 120 L Novozymes A/S, Bagsvaerd, Denmark). Subsequently, the NDF was corrected for ashes and the aNDFom content was calculated. The same procedure used for the NDF was used to analyze the material resulting from the in situ ruminal degradation but without the use of amylase and correction for ashes. The NSC content was calculated as proposed by Sniffen et al¹⁰ with the equation: NSC= 100 - (CP + ashes + aNDFom + EE).

The N-NH₃ content analysis used the supernatant of ruminal liquid samples thawed at 4 ºC and distillation with 2N KOH according to Ribeiro et al¹⁴. The concentration of VFAs was determined by gas chromatography (Shimadzu GC-2010, Kyoto, Japan) according to the methodology described by Erwin et al¹⁵.

Statistical analyses

Statistical analyses were performed using the SAS statistical software (SAS Inst., Inc., Cary, NC). The data of intake, apparent digestibility and nitrogen balance were analyzed using ANOVA with a 5 x 5 latin square design. The statistical model used was:

Where:

Yijk = observation of the effect of treatment i in period j, of animal k, where μ is the overall mean, Ti = effect of treatment i, where i = 1 (HAY), 2 (CONT), 3 (MON), 4 (SALI), and 5 (FLAV);

Pj = effect of period j (j = 5 periods);

Ak = effect of animal k (k = 5 animals), and eijk = random error associated with each observation.

The model for the in situ degradation rate included the experimental treatment, incubation time, animal, period, and treatment × time. The experimental design was the latin square with subdivided plots for the ruminal variables data (VFA, pH, and N-NH₃), where the plots were the treatments and subplots were ruminal liquid samples. The statistical model included treatment effects, sampling times, animal, period, and treatment × time. The statistical model used was:

Where:

Yijkl = observation of the effect of treatment i per hours of incubation (rate of degradation) or collection time (ruminal parameters) j in animal k; μ = overall mean; Ti = effect of treatment (i = 1 (Hay), 2 (CONT), 3 (MON), 4 (SALI), and 5 (FLAV);

Hj = effect of incubation hours for degradability (j = 1,....., 7) or collection times for ruminal parameters (j = 1, ....., 13);

Ak = animal effect (k= 1, ..., 5), Pj = the period effect (j= 1, ....., 5);

THij = interaction between treatment i and time j; and eijkl= random error associated with each observation.

When significant (P≤0.05) F-statistics were noted, means were separated using a multiple comparison test (Tukey’s method) and treatment differences were considered significant at P≤0.05.

Effect of concetrate

The inclusion of concentrate in the diet, increased the intake of MS and nutrients (Table 2). The lowest intake observed in the HAY treatment, could be the result of a ruminal physical limitation in sheep. The hay intake may have caused greater filling and ruminal distension that, according to Grovum¹⁶, provoke inhibitory neural stimulus from the hunger center reducing feed intake. The addition of concentrate, is known to increase the passage rate, providing limiting nutrients, such as nitrogen and sulfur, to ruminal microorganisms and leading to a high rate of multiplication of microorganism¹⁷ and consequently reflecting increased intake.

The inclusion of concentrate in the diet, increased the digestibility of DM and other nutrients (with the exception of aNDFom). According to Hagos and Melaku¹⁸, the lowest concentration of CP and highest of aNDFom in the HAY treatment may have reflected low microbial efficiency, leading to a low level of nutrients’ fermentation and consequently lower digestibility.

The absence of improvement in the aNDFom digestibility as well as the reduction in the degradability of DM and NDF by the inclusion of concentrate may have occurred because the inclusion of rapidly fermentable carbohydrates, leads to a reduction in ruminal pH, and increased development of amylolytic, and decreased development of cellulolytic bacteria^19,20. These microbial alterations reduce the production of fibrolytic enzymes and consequently negatively affect fiber degradability and digestibility.

The inclusion of concentrate in the diet increased the total production of VFA and proportions of propionate and butyrate, and a reduction in the acetate proportion and acetate:propionate ratio. The alteration in the VFA production with the inclusion of concentrate in the diet was probably the result of changes in the microbial population of the rumen, which is altered according to the type of substrate available. According to Wanapat and Khampa²⁰, the inclusion of concentrate in the diet, increases the number of protozoa and amylolytic and proteolytic bacteria, and reduces the number of fungi and cellulolytic bacteria. These alterations cause changes in the type and quantity of VFA produced.

The lowest ruminal pH was observed in treatments containing concentrate. Rumen microorganisms ferment carbohydrates producing VFA and lactate, which have an acidogenic effect. Therefore, the inclusion of rapidly fermentable carbohydrates in the diet increases the fermentation rate and production of these compounds, reducing the pH^3,17. In addition, the inclusion of the concentrate is usually associated with less rumination and chewing with the consequent low production of saliva and diet buffer capacity^19,21. The treatment HAY, showed a low daily pH variation because the animals showed gradual intake throughout the day. Conversely, the treatments with the concentrate, presented a high pH variation during the day because the supplied concentrate was quickly consumed, reflecting subsequent drops in pH values, with the smallest values observed between 2 and 4 h after the feeding. The lowest pH values of 5.68 and 6.27 were observed in the treatments with the concentrate and hay, respectively. According to Hoover and Stokes²², these values are within the range suitable for maximum microbial growth and maximal ruminal digestion of fibers (5.5 to 7.0) and the great pH range for fiber digestion is from 6.7 to 7.1.

The highest concentration of ruminal N-NH₃ was observed with the inclusion of concentrate. According to Van Soest¹⁹, protein degradation in the rumen occurs through the action of enzymes secreted by ruminal microorganisms. These bacteria digest part of the protein, and others microorganisms on site use amino acids, peptides, and ammonia for cell multiplication. When the rate of ammonia production is greater than the rate of use, it is absorbed by the animal through the rumen wall passing into the blood stream and converting into urea in the liver. Urea can be recycled or lost through the urine. Excessive values are reached by a high inclusion of concentrate in the diets and are not desired.

The treatment HAY presented a low daily variation of N-NH₃ (Figure 2B) because the animals gradually consumed the feed throughout the day. On the other hand, as observed in the pH analysis, the treatments with concentrate showed a high variation due to the fast intake of concentrate causing peaks of ammonia production between 2 and 4 h after feeding. These increased levels occurred as the result of high amounts of CP from the concentrate (188 g kg^-1), increasing its rumen degradation rate and producing N-NH₃ peaks.

There was an increase in the amount of absorbed and retained N (g d¹ or g kg^-1 of CP) with the inclusion of the concentrate in the diet. This happened probably due to the high N intake and high N digestibility contained in the concentrate compared to those in HAY. However, when the amount of N retained in g kg^-1 of the N ingested was analyzed, the absence of increased efficiency in the use of ingested N with the inclusion of concentrate was probably resulting from the low flow rate displayed in the HAY treatment, which improved the efficiency of N use, producing results similar to those in the concentrate treatments. The inclusion of concentrate resulted in high losses of N, mainly through urine. This may have been due to a low efficiency of the urea cycle, which is a reflection of peaks of N-NH₃ production and less saliva production with the inclusion of concentrate¹⁹. Because the inclusion of concentrate reflects less chewing and rumination time²¹, less saliva is produced, and less nitrogen is recycled in this way, which ends up being eliminated in the urine.

Effect of additives

The inclusion of additives did not alter the DM and nutrients intake. However, ionophores are known to have an effect on DM ingestion leading to reduced intake when used in diets with a high proportion of concentrate^23,24. This effect may occur as a consequence of an increase in the concentration of ruminal propionic acid, which reflects an increase in energy efficiency, allowing nutritional requirements to be reached with a smaller amount of feed intake⁷. This intake is also affected by pathophysiological aspects such as ruminal pH. Thus, a diet that induces rumen acidosis, with the addition of ionophores, may lead to a smaller pH drop, increasing intake²⁵. However, in this experiment, besides the high hay intake (approximately 430 g kg^-1 of the diet), the diets were all cationic, which produces an alkalizing effect that may have caused the lack of effects of additives on the intake.

The inclusion of additives did not alter the digestibility of the diet. However, ionophores can increase dietary digestibility by increasing the DM retention time in the rumen as the result of lower voluntary intake, stimulating rumination, improving the ruminal environment, and allowing increased digestibility²⁶. However, this effect was not observed in this study because the additives did not alter the intake.

The additives did not influence the degradability of hay. Ionophores are known to alter the microbial population of the rumen and act on cellulolytic bacteria, which can lead to a reduction in fiber degradability²⁷. Nevertheless, it is probably easier to observe these effects in diets with higher proportions of roughage. Edward et al⁵ report in an in vitro experiment that cellulolytic Gram-negative bacteria of the genus Fibrobacter are among the most sensitive to the action of flavomycin. According to these same authors, flavomycin normally does not decrease fiber degradation in in vitro and in vivo studies. Therefore, it is probable that the cellulolytic activities of bacteria of the genus Ruminococcus, fungi, and protozoa, which are not affected by this antibiotic and are in the microorganism's rumen population, compensate decreased numbers of Fibrobacter bacteria that could affect fiber degradation.

The additives altered the production of VFA. Ionophores cause such alterations in the production of VFAs by modifying bacterial populations of the rumen. Gram-positive bacteria that produce acetate, butyrate, and H₂ are inhibited by ionophores, and Gram-negative bacteria that produce propionate find better conditions to reproduce²⁸. The production of butyrate was increased in FLAV. (Thus, the response from the effect of flavomycin on the proportion of VFA differs from that from ionophores, which promotes an increase in the proportion of propionate. The action of flavomycin, not exactly on the same microorganisms, possibly reflects these alterations.

The additives did not alter the ruminal pH. This happened probably due to the high proportion of hay in the diet (approximately 430 g kg^-1 diet) inducing longer intake time, regurgitation, and saliva production resulting in a small drop in ruminal pH and not allowing the demonstration of these antibiotics’ effects.

The reduction in the concentration of N-NH₃ during the first peak of production (only 4 h after the morning feeding) was only detected in the diet with MON. This response is associated with the reduction in the number of bacteria that use amino acids and peptides as an energy source for their growth, and consequently, release ammonia in the ruminal environment. This reduction in the use of amino acids and peptides favors their passage into the small intestine and absorption, increasing the efficiency in the use of nitrogen²⁹.

No effect was observed from additives on nitrogen balance. Nevertheless, the ionophores may promote an improved utilization of dietary nitrogen, as a result of reduced DM intake, consequent reduction of nitrogen intake, and reduced rumen deamination³⁰. In this study, the use of additives did not alter DM intake, and salinomycin and flavomycin did not alter the concentration of N-NH₃. Therefore, the absence of an effect on the total nitrogen balance was expected. Although MON reduced the ammonia peak production after the morning feeding, this was not enough to alter the nitrogen balance.