Animals and experimental feeds

Nine Nellore steers (body weight [BW], 300 ± 30 kg; age, 18 ± 2 months) fitted with 10-cm (internal diameter) silicone-type ruminal cannulas and duodenal T-type cannulas (Kehl^®, São Carlos, Brazil) were used in a triple Latin square design 3 × 3 to evaluate dietary intake, apparent total tract digestibility, ruminal pH, ammonia-N (NH₃-N) concentration, ruminal BH, duodenal flow of FAs and rumen microbiota. The steers were fitted with the cannulas at about 10 months of age and underwent at least 8 weeks of recovery before the start of the study. The steers were treated for internal and external parasites at the beginning of the experiment and maintained in individual pens of ~21 m² for adaptation with protected feeders and water. Initially, the animals were weighed, identified and housed in individual pens with feeders and drinkers. The animals spent 21 days adapting to the facilities and management. The experiment consisted of three consecutive 20-day periods. Each study period consisted of 15 days for adaptation, 4 days for recording dry matter intake (DMI) and faeces collection, and 1 day for ruminal fluid sampling. Ruminal fluid was collected for measuring ruminal pH and NH₃-N and quantifying bacteria and protozoa on 1 day during each 20-day period (see later).

The diets were formulated to provide DMI of 22 g/kg of BW (6·6 kg per animal/day), calculated using the method described by Alderman (Reference Alderman1993) and fNDF was fixed at 0·15 of DMI to ensure ruminal fibre requirements and avoid ruminal upsets (Goulart & Nussio Reference Goulart and Nussio2011).

The dietary treatments were as follows: maize silage (MS), sugar cane (SC) and sugar cane bagasse (SB, Table 1). Maize silage used in the present study was obtained from the UNESP farm. A whole plant of the maize hybrid 2B688Hx (Dow AgroSciences, Indianapolis, IN, USA) was harvested and ensiled at about 0·31 DM and chopped into pieces 5 cm long. The chopped maize material was placed in a trench-type silo, covered with black plastic and ensiled for at least 2 months. Sugar cane was obtained from a local farm and chopped into pieces of about 2–3 cm in length, while SB was obtained from a private biofuel plant and chopped into pieces 3–4 cm long. The differences in the length of the forage pieces was due to current practice. Crude glycerine was acquired from the soybean oil-based biodiesel production company ADM, Rondonópolis, Brazil (803·4 g glycerol/kg; 15·9 g ether extract/kg; 50·3 g ash/kg and 120·2 g water/kg).

Table 1.

Chemical composition of maize silage (MS), sugar cane (SC) and sugar cane bagasse (SB) (g/kg on a dry matter basis)

DM, dry matter; OM, organic matter; CP, crude protein; NDF, neutral detergent fibre; iNDF, indigestible neutral detergent fibre; EE, ether extract.

The diets were isonitrogenous and the concentrate was composed of ground maize, soybean meal, crude glycerine and mineral supplements (Table 2). Ingredients for the concentrate were ground using a hammer mill fitted with strainers and 2-mm sieves. Continuous homogenization of the diets was performed in a horizontal mixer for 15 min.

Table 2.

Composition of the experimental diets (g/kg on a dry matter basis)

MS, maize silage; SC, sugar cane; SB, sugar cane bagasse; DM, dry matter; OM, organic matter; CP, crude protein; NDF, neutral detergent fibre; fNDF, NDF from forage; EE, ether extract; NFC, non-fibre carbohydrates; ME, metabolizable energy; FAME, fatty acid methyl esters; SFA, saturated fatty acid; USFA, unsaturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid.

* Composition = 210 g calcium; 20 g phosphorus; 37 g sulphur; 80 g sodium; 490 mg copper; 1·424 mg manganese; 1·830 mg zinc; 36 mg iodine; 29 mg cobalt; 9 mg selenium;333 mg fluorine (max).

† Non-fibrous carbohydrate (NFC) estimated as: NFC = 100 − (CP + EE + ash + NDF).

The animals were fed twice daily at 08.00 and 16.00 and feed refusals were recorded daily for each pen. The concentrates and forage for each animal were weighed individually and mixed manually in the troughs. Throughout the entire experimental period, refusals of ~100 g/kg were allowed in relation to the total amount consumed on the previous day. Feed refusals were collected and weighed before feeding; sub-samples of ~100 g/kg of feed refusals were collected for each animal and frozen at −20 °C.

Data collection and sampling procedures

Total faeces were collected for 4 days between days 15 and 18 (Barbosa et al. Reference Barbosa, Valadares, Valadares Filho, Pina, Detmann and Leão2011). Faeces were collected immediately after each spontaneous defecation and stored in 20-litre buckets. At the end of each 24-h collection period, the buckets were changed and the faeces were weighed, blended manually and aliquots (~300 g) were collected. Afterwards, these samples were dried in a forced-ventilation oven at 55 °C for 72 h and ground in a mill with a 1-mm mesh sieve, and a composite sample was prepared for each animal per period on the basis of the air-dry weight.

Duodenal samples were collected over 2 days (fourth and fifth data collection days during the study period) at 6-h intervals. The sampling was delayed for 3 h on the second day to represent the 24-h period (de Oliveira et al. Reference de Oliveira, Berchielli, Pedreira, Primavesi, Frighetto and Lima2007). The samples were maintained at −10 °C and at the end of the period, the samples for each animal were pooled for each period. Indigestible NDF (iNDF) was used as an indicator of daily DM flows in the duodenum (Harvatine & Allen Reference Harvatine and Allen2006) and obtained via an in situ method after 240 h (Casali et al. Reference Casali, Detmann, Valadares Filho, Pereira, Henriques, De Freitas and Paulino2008) with incubated samples of feed offered, feed refusals, faeces and ground duodenal digesta (Wiley mill, 2-mm screen). Indigestible NDF was analysed using an Ankom 200 Fibre Analyser (Ankom Technology Fairport, NY, USA).

Rumen pH and NH₃-N were measured on day 18 of each period. To assess rumen fermentation parameters, rumen fluid samples (c. 80 ml) were collected manually before supplying the diet (time zero) and 1, 2, 4, 6, 8, 10 and 12 h after feeding. Immediately after collection, pH of the rumen fluid was determined using a digital potentiometer (ORION 710A, Boston, MA, USA). The ruminal fluid (40 ml) was placed in a plastic bottle and frozen at −20 °C for NH₃-N analysis, according to the methodology adapted by Fenner (Reference Fenner1965). In brief, ruminal fluid NH₃ was analysed by distilling with 2 m potassium hydroxide (KOH) in a micro-Kjeldahl system.

The feed offered, feed refusals and faeces samples were dried at 55 °C for 72 h and ground in a Wiley mill (Thomas Scientific, Swedesboro, NJ, USA) to pass through a 1-mm screen. Samples of Tifton 85 hay, concentrates and feed refusals were analysed for DM (934.01), mineral matter (MM, 942.05) and ether extract (EE, 920.39), according to AOAC (1990). Ether extract contents in faeces were determined by extraction in ether (EE, 920.39). Nitrogen was determined using an LECO FP-528 nitrogen analyser (LECO Corp., St. Joseph, MI, USA).

The NDF was determined using α-amylase and without the addition of sodium sulphite, according to the method reported by Van Soest et al. (Reference Van Soest, Robertson and Lewis1991) and adapted for the Ankom200 Fibre Analyser (Ankom Technology, Fairport, NY). Acid detergent fibre was determined using the method described by Goering & Van Soest (Reference Goering and Van Soest1970) and adapted for the Ankom200 Fibre Analyser (Ankom Technology, Fairport, NY). Acid detergent lignin was determined by solubilisation of cellulose with sulphuric acid, according to Van Soest & Robertson (Reference Van Soest and Robertson1985). Gross energy was obtained by the combustion of samples in an adiabatic bomb calorimeter IKA^® model 2000 Basic.

The flows of DM to the duodenum were calculated as the ratio of iNDF consumed to the concentration of iNDF in duodenal contents. Ruminal BH was estimated by the differences in USFA between the intakes and flows to the duodenum (Xu et al. Reference Xu, Jin, He, Li, Beauchemin and Yang2014).

Fatty acid quantification

Fatty acid composition was analysed according to the protocol described by Shingfield et al. (Reference Shingfield, Ahvenjarvi, Toivonen, Arola, Nurmela, Huhtanen and Griinari2003). Briefly, lipids were extracted from 100 mg of freeze-dried samples in a mixture (3 : 2, v/v) of hexane and isopropanol after adjustment of pH to 2 by using 2 m hydrochloric acid and methylation with an acid-based catalyst using methanolic sulphuric acid (1%, v/v). The FA profile was determined using a gas chromatograph (Focus CG – Finnigan, Thermo Finnigan, San Jose, CA, USA) with a fused silica capillary column (100 m × 0·250 mm, 0·20-m film thickness; CP-Sil 88 for FA methyl esters – FAME) by using hydrogen as a carrier gas at a constant flow of 1·5 ml/min). The flame ionization detector was set at 300 °C. The time–temperature programme used was as follows: initial temperature of 70 °C for 4 min before increasing to 175 °C (13 °C/min) with waiting time of 27 min; 215 °C (4 °C/min) with waiting time of 9 min; and finally 230 °C (7 °C/min) for 4 min (total 65 min). Identification of FAs was performed by comparing their retention times with those observed in commercial standards (C4–C22) such as Supelco 18919-1AMP, a methyl ester mixture of 37 FAs. The major FAs were identified using the following pure commercial standards: 05632-SIGMA, methyl ester mixture of the conjugated linoleic acid cis-9, trans-11 and trans-10, cis-12; V1381-SIGMA and methyl ester vaccenic acid.

Microbiological analysis

Samples for bacteria and protozoa quantification were collected on day 19. For quantification and identification of rumen ciliate protozoa, samples of ruminal contents were collected via cannula, 3 h after the morning feeding. Cell counts were performed using rumen content aliquots preserved in formalin (a solution of equal parts water and 370 ml/l formaldehyde), according to D'Agosto & Carneiro (Reference D'Agosto and Carneiro1999). Ciliate protozoa species were identified and quantified in a Sedgewick-Rafter chamber, according to Dehority (Reference Dehority1984). Each sample was homogenized and 1 ml of ruminal content was pipetted and transferred to vials with Lugol's solution, according to the methodology of D'Agosto & Carneiro (Reference D'Agosto and Carneiro1999). After 15 min, 9 ml of 30% glycerine was added to the vials. To quantify the protozoa, 1 ml was transferred from each vial to fill the Sedgewick-Rafter chamber.

Bacterial population was quantified using quantitative polymerase chain reaction (qPCR). Samples of ruminal contents were collected via cannula before the morning feeding (Saro et al. Reference Saro, Ranilla, Cifuentes, Rosselló-Mora and Carro2014). Fifty grams of the rumen contents was weighed and immediately added to 50 ml of phosphate-buffered saline (pH 7·4), stirred vigorously for 3 min, and then filtered with a mesh fabric (100 µm). The filtrate was centrifuged at 16 000  g for 10 min at 4 °C. The supernatant was discarded and the precipitate was resuspended in 4 ml of tris-ethylenediaminetetraacetic acid buffer (10×, pH 8). The re-suspended content was centrifuged at 16 000  g for 10 min at 4 °C, the supernatant was discarded, and the precipitate stored immediately in a freezer (–20 °C) for 2 months.

A ‘Fast spin kit for soil’ (MP Bio^®, Biomedicals, Illkirch, France) extraction kit was used to extract DNA from 250 mg of the sample, according to the manufacturer's instructions. Integrity and quantity of the DNA was checked using agarose gel electrophoresis (0·8%) and complementary DNA was assessed using a spectrophotometer (Thermo Scientific NanoDrop 1000, Thermo Fisher Scientific, MA).

Absolute quantification of total bacteria and relative quantification of cellulolytic bacteria (Ruminococcus albus, Fibrobacter succinogenes and Ruminococus flavefaciens) was achieved by qPCR. The primers used in the present study are listed in Table 3. Four concentrations (200, 400, 600 and 800 nM) of forward and reverse primers were tested to determine minimum primer concentration giving the lowest threshold cycle (Ct) and to reduce non-specific amplification before the start of the reaction. Validation of the chosen concentration of primers was performed with different concentrations of DNA (150, 125, 100, 50 and 25 ng). The value of the ‘slope’ was determined and efficiency was calculated using the following equation:

$${\rm Efficiency} = 10\left( { - 1/{\rm slope}} \right){\rm} -{\rm} 1$$

Table 3.

Primers used in the present study

F, forward; R, reverse.

* Temperature of annealing.

† Denman & McSweeney (Reference Denman and McSweeney2006).

‡ Khafipour et al. (Reference Khafipour, Li, Plaizier and Krause2009).

§ Denman et al. (Reference Denman, Tomkins and Mcsweeney2007).

The amplifications were performed in triplicate. Real-time PCR was performed using the Applied Biosystems 7500 Real-time PCR System (Applied Biosystems, Foster City, CA). Rox was used as a passive reference dye. The qPCR was performed using 100 ng of total DNA in a reaction containing 6·25 µl of SYBR^® Green PCR Master Mix (Bio-Rad, Hercules, CA), 400 or 600 nM of the primer pair and water to a final volume of 12·5 µl. Cycling conditions were 50 °C for 2 min and 95 °C for 10 min, followed by 35 cycles of denaturation at 95 °C for 15 s, pairing at 60 °C for 60 s and extension at 78 °C for 1 min. After one cycle of amplification, a step was added to increase the temperature from 60 to 95 °C to obtain the dissociation curve of the reaction products for analysing amplification specificity.

Relative quantification was used to determine species proportion. The results were expressed as a 16S rDNA ratio of general bacteria:

$${\rm Relative}\;{\rm quantification} = {2^{-({\rm Ct}\;{\rm target}-{\rm Ct}\;{\rm total}\;{\rm bacteria})}}$$

where Ct is defined as the number of cycles required for the fluorescent signal to cross the threshold.

Statistical analyses

The data of intake, digestibility, ruminal BH, duodenal FA flows and protozoa populations were analysed using a triple 3 × 3 Latin square design with the PROC MIXED procedure of SAS (version 9.2). The general mathematical model was as follows:

$${Y_{ijk}} = \mu + {\alpha _i} + {\beta _j} + {s_k} + \alpha {\beta _{ij}} + {e_{ijk}}$$

where Y _ijk represents the observation of steers k given diet i at period j; μ represents overall mean; α _i represents the fixed effect of the ith diet, i = 1, 2,…, nt; β _j represents the fixed effect of the jth period, j = 1, 2,…, np; and s _k represents the random effect of the kth steers, k = 1, 2,…, ns, μ; αβ _ij represents the interaction between diet i and period j; and e _ijk represents the residual experimental error with variance component σ2c (Tempelman Reference Tempelman2004).

Ruminal pH and NH₃-N data were analysed using a triple 3 × 3 Latin square design with repeated measures over time and the PROC MIXED procedure of SAS (version 9.2). The model included fixed effects of diet, time, diet × time interaction and random effects of steers and periods. The structure of errors that best fitted the data according to the Bayesian information criterion was used. Differences between treatment means were determined using Tukey's test. Differences among means with P < 0·05 were accepted as representing statistically significant differences.

Relative quantities of 16S rRNA, as determined using real-time PCR, were analysed using R software (version 3.1.1), with three treatments and three periods. The fixed effects were treatments and Latin Square, and random effects were time, animal and error. When significant, the means between treatments were compared using Tukey's test and the significance was set at P < 0·05.