Low concentrations of a polyphenolic extract from pine bark in high–concentrate diets decrease in vitro rumen ammonia nitrogen but not methane production

ABSTRACT This study was conducted to assess the effects of small supplemental doses of a polyphenolic extract from pine bark (PBE) on CH4 output and ruminal fermentation parameters when incubated in batch culture with a high–concentrate diet for 24–h. The data from the dietary substrates supplemented with 0.0, 0.3, 0.6, 0.9, 1.2, 1.5 and 1.8% of PBE were evaluated in a randomized complete block design, and compared using ANOVA followed by Tukey's test and polynomial contrasts. Increasing doses of the PBE caused a linear decrease of the NH3–N concentration (p < 0.001), the potentially degradable dry matter (DM) fraction (p = 0.002), the partitioning factor (p = 0.001), CH4 production and proportion (p = 0.001 and p = 0.029, respectively), although only at 6–h, achieving the lowest productions (p = 0.016) with 1.5 and 1.8% PBE. In contrast, the PBE linearly increased asymptotic gas production (p = 0.007), gas yield (p = 0.004), pH (p = 0.002) and the short–chain fatty acid concentration (p < 0.001) at 24–h. Addition of least 1.5% PBE to high–concentrate diets reduces CH4 production by 31% at 6–h, whereas NH3–N concentration is reduced by 31% at 24–h incubations.


Introduction
Animal farming is challenged by a growing demand for quality meat and milk products and, at the same time, by keeping a steady environmental impact (Malik et al. 2017). The livestock industry contributes about 14.5% to the global greenhouse gases (GHG) emissions, with methane (CH 4 ) accounting for 44% of these emissions (Gerber et al. 2013) and representing an energy loss of 2-12%. Additionally, up to 50% of nitrogen (N) intake can be lost by degradation to ammonia (NH 3 ) and excretion via urine (Salem et al. 2015;Brutti et al. 2019), further contributing to GHG emissions because of its potential conversion into nitrous oxide (N 2 O;IPCC 2013). It is possible to reduce CH 4 emissions and N excretion (Bhatta et al. 2015) by manipulating the rumen ecosystem through growth promoters, antimicrobials and hormones. However, the negative perception of consumers on chemical additives has intensified research for natural additives such as polyphenols, flavonoids and tannins with the potential of modulating rumen fermentation and rumen-derived products (Balcells et al. 2012;Brutti et al. 2019;Vasta et al. 2019;Purba et al. 2020a).
Flavonoids, such as tannins, are plant polyphenolic secondary metabolites considered as safe for the environment and for the consumer (Jiménez-Peralta et al. 2011;Bhatta et al. 2015), and have the ability to bind proteins and carbohydrates (Deaville et al. 2010). They have also shown antimicrobial properties (Purba et al. 2020a) and have a profound effect on the outcome of ruminal fermentation of ruminant diets (Aderao et al. 2018). Tannins have been considered antinutritional factors as they can reduce DM intake and protein and carbohydrate digestion (Oliveira et al. 2007). However, depending on the source, concentration, and type of tannin applied to ruminant diets, either naturally or supplemented, they can have an antimethanogenic effect (Malik et al. 2017;Vasta et al. 2019), and improve N use in ruminants by reducing crude protein (CP) degradation (Castro-Montoya et al. 2018), thus improving live weight gain, milk yields and animal fertility and health status (Hatami et al. 2018;Vasta et al. 2019).
Natural polyphenol-rich extracts can be obtained from trees such as acacia (Acacia mearnsii), quebracho (Schinopsis balansae and S. lorentzii) or pine (Pinus radiata; García et al. 2016). Chilés forest industry is based on P. radiata production, with a cultivated area of 1.6 million hectares, equivalent to onethird of the total P. radiata planted globally (Guerrero and Bustamante 2007). Bark represents at least 10% of total pine weight and generates a biomass residue of 1.4-1.5 million tons per year which is generally used to generate electricity by combustion, resulting in negative environmental impacts. However, a polyphenol-rich extract (mixture of flavonoids, stilbenoids and condensed tannins) from P. radiata bark (Berg et al. 2009), reduced ammonia nitrogen (NH 3 -N) concentration in in vitro fermentation by 50%, without affecting diet digestibility or CH 4 production when used at concentrations of 2-4% DM PBE in ruminant forage diets (Vera et al. 2018). Yang et al. (2016) also reported that the supplementation with a moderate concentration (3% DM) of an extract from P. taeda bark decreased NH 3 -N concentration without affecting CH 4 production, suggesting that PBE could affect dietary N use efficiency in vivo. However, the effects of low concentrations of PBE in cattle concentrate diets containing high proportions of quickly degradable protein has not yet been assessed. The purpose of this batch test was to assess the smallest (< 2% DM) effective dose of a PBE as an additive in high-concentrate ruminant diets to decrease NH 3 -N concentration and CH 4 production.

Materials and methods
This experiment was conducted at the Livestock Systems and Nutrition Laboratory of the Universidad de Concepción (UdeC), Chillán, Chile. The care and management of the cows were certified by the animal ethics and welfare committee of the UdeC.

Extract from Pinus radiata bark
The polyphenolic PBE was produced by methanolic extraction at the Technological Development Unit, UdeC, according to Berg et al. (2009). This extract is an aqueous solution (38.0% DM) with a concentration of 133.2 g of total polyphenols/kg DM, or 43.5 g of total tannins (TT)/kg DM. It is mainly composed of flavonoids (luteolin, pinocembrin, catechin, procyanidin, gallocatechin, quercetin and taxifolin) and small amounts of stilbenoids (astringin and piceatannol) and phenolic acid.

Incubation substrates and treatments
The substrates simulated a high-concentrate diet with forages containing high concentrations of quickly degradable protein for cattle. Therefore, the incubation substrates were corn grain, mixed hay (Lolium perenne with Trifolium repens) and soybean meal in a ratio of 60:20:20, respectively. Treatments were 0 (Control) or 0.3, 0.6, 0.9, 1.2, 1.5 and 1.8% DM basis PBE replacing equivalent amounts of substrate. All ingredients were ground (2 mm; Grain Mill, Breuer, Temuco, Chile) before adding PBE and thereafter mixed. The ingredients and PBE inclusion are listed in Table 1.

Donor animals and batch incubation
Rumen fluid was obtained from two non-lactating rumen-cannulated adult Aberdeen Angus cows (500 kg body weight), fed a diet containing mixed hay (Lolium perenne with Trifolium repens), ground corn and a vitamin-mineral supplement in a ratio of 70:25:5, respectively, formulated to meet nutritional requirements for maintenance of adult 500-kg cows (NASEM 2016). Animals were fed daily at 7 am and 5 pm Access to fresh water was available at all times. Two hours after the morning feeding rumen fluid was collected and filtered through four cheesecloth layers and immediately transported to the laboratory in a pre-heated thermal flask (39°C). The inoculum was a blend of rumen fluid and mineral buffer (Menke et al. 1979) in a 1:3 ratio (v/v).
Substrates were weighed (0.5 g) into ANKOM F57 filter bags (Ankom Technology Corp., Macedon NY), and each substrate bag was placed individually in a 50 mL amber glass bottle (Avila et al. 2011). For each dose of PBE (n = 7) and sampling time (6, 12 and 24-h) three replicates were incubated, plus two blanks (no substrate), to calculate in vitro net gas production (GP) and the in vitro DM disappearance (IVDMD).
Each bottle was filled with 25 mL of the inoculum (39°C), gassed with CO 2 and sealed with a rubber stopper. Once the bottles were inoculated, they were incubated at 39°C for 24h (Forma Series II 3110 Water-Jacketed CO 2 Incubator, Thermo Fisher Scientific, Waltham, USA) on an orbital shaker set at 90 oscillations/min (Heidolph Unimax, Germany). The incubations (runs) were repeated thrice during separate weeks, resulting in a total bottle number of 207 [(seven doses of PBE × three replicates + two blank bottles) × three sampling times × three runs].

Estimation of ruminal gas production, CH 4 and dry matter disappearance
Starting at 6-h of incubation, and then at 12 and 24-h, a sample of gas (15 mL) was collected with a syringe from each bottle and immediately transferred to a vacuumed exetainer (5.9 mL; Labco Ltd., Wycombe, Bucks, UK) and then analysed for CH 4 concentration by gas chromatography (Avila et al. 2011). The gas chromatograph (GC; Agilent 7890B, Agilent Technologies, Inc., Santa Clara, CA, USA) was equipped with a thermal conductivity detector (TCD) and a 30-m column (GS-CarbonPLOT, Agilent Technologies, Italy) using helium as carrier gas with a flow rate of 1.33 mL/ min, and an isothermal oven temperature of 35°C. The injector and detector temperature were set to 185°C and 150°C, respectively. A subsample of gas (2 mL) was removed from each exetainer and injected manually into the GC. Methane gas of analytical quality (99.5%) was purchased from Linde (Santiago, Chile) to prepare the standards. Standards of CH 4 (15, 10, 7.5, 5.0, 2.5 and 1.0%) were prepared by diluting stock CH 4 gas with N gas at room temperature (≈22-24°C).
The total volume of gas produced in each bottle was measured using a water displacement apparatus according to Fedorak and Hrudey (1983). After gas sampling, the ANKOM F57 filter bags were removed from the bottles and washed with distilled water, followed by drying at 60°C for 24-h to estimate IVDMD (Avila et al. 2011).

Determination of culture pH and NH 3 -N
The inoculum pH of every bottle was measured on a portable pH metre (Orion Star A121, Thermo Scientific, USA), and incubation fluid was sampled in a screw cap vial (2 mL; Biologix Research Company, USA) with trichloroacetic acid (150 μL; 0.65 w/v) to determine the NH 3 -N concentration in a UV-VIS spectrophotometer (Merck, Spectroquant Pharo 300, Germany) at 625 nm. The cryotubes were stored at -20°C until analysis. At the beginning of each incubation, inoculum samples were collected and used to correct NH 3 -N concentration (Avila et al. 2011).
To estimate IVDMD kinetics, the calculated values were fitted using the non-linear Gompertz model (Lavrenčič et al. 1998; Equation 1): where y = IVDMD (%) at t time; B = potentially degradable DM fraction (%); C = relative rate of degradation (/h); and A = describes a constant factor of microbial efficiency. The parameters B, C and A were used to calculate the first and second derivatives of the Gompertz model to obtain the degradation rate at the inflexion point (maximum degradation rate, MDR), and the time when 95% of the substrate is fermented (time of maximum degradation rate, TMDR), allowing a more comprehensive evaluation of small doses of PBE as feed supplement (Lavrenčič et al. 1998).
Organic matter (OM) in the substrates was estimated by the difference between DM and total ash, hemicellulose (HC) was estimated by the difference between the NDF and ADF, while DM digestibility (DMD) was estimated by Rohweder et al. (1978;Equation 2). Based on the value of DMD, digestible energy (DE) value was obtained using Fonnesbeck et al. (1984;Equation 3), and from the DE, metabolizable energy (ME) content was determined according to Equation 4 (NASEM 2016): To estimate GP and CH 4 kinetics, recorded volumes were fitted using the non-linear Gompertz model (Schofield et al. 1994;Equation 5): where y = gas (mL/0.5 g DM incubated) or CH 4 (mg/0.5 g DM incubated) production; b = asymptotic gas (mL/0.5 g DM incubated) or CH 4 (mg/0.5 g DM incubated) production; c = production rate (/h); Lag = initial delay (h) before gas or CH 4 production begins; and t = time of measurement. The half-life (t 1/2 ) is the time (h) taken for gas or CH 4 production to reach 50% of its b value. The average production rate (APR) was defined as the average gas (mL/0.5 g DM incubated) or CH 4 (mg/0.5 g DM incubated) production rate between the start of the incubation and the t 1/2 (García-Martínez et al. 2005).
To estimate fermentation efficiency, the partitioning factor (PF) at 24-h of incubation was determined as the ratio between degraded DM (mg) and the total GP (mL; Blümmel et al. 1997). Gas and CH 4 yields were estimated as the net gas (mL) or CH 4 (mg) volume at each sampling times (6, 12 or 24-h of incubation), divided by the corresponding g of degraded DM.
Estimation of microbial CP production (MCP) was based on Blümmel et al. (1997;Equation 6): MCP (mg/g DM) = mg DM degraded-(GP 24 × 2.2mg/mL) (6) where GP 24 = net gas production (mL/0.5 g DM) at 24-h; and 2.2 mg/mL is a stoichiometric coefficient of the amounts (mg) of C, H 2 and O 2 required for the production of volatile fatty acids associated with 1 mL of GP (Blümmel et al. 1997).

Statistical analyses
Data were analyzed with Stata 14 statistical software (College Station, StataCorp LP, TX, USA). Shapiro-Wilk's and Levene's tests were used to verify the assumptions of normality and homogeneity of variances, respectively. All the data were analyzed in a randomized complete block design using the model: where Y ij is every observation of the PBE dose i on run j, μ is the general mean of observations, α i is the fixed effect of the PBE dose (i = 0.0, 0.3, 0.6, 0.9, 1.2, 1.5, 1.8% DM basis), β j is the random effect of the incubation run ( j = 1, 2, 3) and ε ij is the residual error. The results are presented as average values with the standard error of mean. The averages were compared by Tukey's test, being statistically significant when p < 0.05 and considered a trend when 0.05 < p < 0.09. In addition, polynomial contrasts were used to recognize linear and quadratic effects of increasing PBE concentrations.

Dry matter disappearance and kinetics
The inclusion of increasing concentrations of PBE did not affect the IVDMD at 6-h (p = 0.865), but resulted in a linear reduction at 12 (p = 0.001) and 24-h (p = 0.003), being from 3 to 5% lower than control at 12-h (p = 0.004) with the highest doses of PBE (1.5 and 1.8%), and a 6% lower at 24-h (p = 0.032) with a 1.8% PBE ( Table 2). The potentially degradable DM fraction (B) also decreased linearly (p = 0.002) with increasing PBE concentrations, being in average a 4% lower with a 1.5 and 1.8% PBE (p = 0.023), whereas relative rate of degradation (C ), microbial efficiency (A), TMDR and MDR were unaffected (p ≥ 0.507) by the extract.

In vitro ruminal gas production and kinetics
Net gas production at 6-h was unaffected by PBE (p = 0.730; Table 3), but at 12 and 24-h it was decreased linearly (p < 0.001), being lower with a 1.5 and 1.8% PBE at 12-h (p = 0.011), and with 1.2, 1.5 and 1.8% PBE at 24-h (p < 0.001). Total GP at 6 and 12-h was unaffected by treatments (p = 0.971 and p = 0.432, respectively). However, increasing concentrations of PBE linearly increased in vitro GP at 12 (p = 0.035) and 24-h (p = 0.006), and trended (p = 0.073) to a higher GP with 1.5 and 1.8% PBE at 24-h than control. These results agree with those of gas yield (GY), which increased linearly at 12 (p = 0.017) and 24-h (p = 0.004), with a trend (p = 0.067) to a higher GY with 1.5 and 1.8% PBE at 24-h whereas asymptotic GP (b) also increased linearly (p = 0.007) with PBE addition, trending to be higher with 1.5 and 1.8% (p = 0.072). Gas production rate (c), lag time, t 1/2 and the APR were not affected by PBE (p ≥ 0.327).

In vitro fermentation
Increasing PBE concentrations linearly increased pH (p = 0.002) and SCFA (p < 0.001), being higher (p = 0.007) in the dietary substrates supplemented with 1.5 and 1.8% PBE (Table 5). Microbial Crude Protein (p < 0.001) and the PF 24 (p = 0.001) were linearly decreased, being both parameters lower (p < 0.005) with the highest PBE supplementation (1.5 and 1.8%). The in vitro NH 3 -N concentrations linearly decreased at 24-h (p < 0.001), and were reduced by 31% as compared to control (p = 0.001) with the highest doses of PBE (1.5 and 1.8%).   Different letters in same column indicate significant differences (p < 0.05). a b = asymptotic production (mL gas/0.5 g DM incubated); c = rate of gas production (/h); Lag = initial delay before gas production begins (h); t 1/2 = half-life (h); APR = average production rate (mL/g DM incubated per h). b SEM = Standard error of mean. c Probability of differences between treatments (T), or of a linear (L) or quadratic (Q) effect by PBE concentration.

Discussion
As the PBE concentration increased in the substrate, DM decreased numerically given the aqueous state (38.0% DM) of the extract. Both NDF and ADF substrates values were altered by PBE supplementation. If both had been increased, it could have indicated that the increase, although numerical, may be associated to the NDF and ADF contents of the extract (48.6 and 64.9% DM, respectively). However, NDF increased and ADF decreased, which could be related to the influence of condensed tannins in the PBE which form complexes with fibre and hinder the use of conventional detergent method of fibre analysis (Guglielmelli et al. 2011;Vera et al. 2018). The substrates were high in CP to simulate the use of spring forages and concentrate diets with high protein values, of which an important fraction is lost as NH 3 -N (Guglielmelli et al. 2011).
The linear decrease in the potentially degradable DM fraction (B) and the IVDMD by increasing polyphenolic PBE concentrations can be attributed to reduced fibre digestibility by the formation of complexes of tannins and lignocellulose, or by the inhibition of cellulolytic microorganisms or their enzymes (Deaville et al. 2010;Vasta et al. 2019), as reported by Ahnert et al. (2015) in heifers receiving increasing concentrations of a quebracho tannin extract or by Brutti et al. (2019), who used a mixture of chestnut and quebracho tannins under in vitro conditions.
The increased asymptotic GP (b) was negatively correlated with ADF in the substrate, concurring with Kafilzadeh and Heidary (2013), who reported that with increasing incubation time, the medium conditions vary by releasing cell wall components that could affect rumen microbial activity. It is possible that as PBE increased in the substrate, the numerical decrease of the ADF improved the microbial activity, increasing b, through favourable environmental conditions as incubation time progressed. After 12 and 24-h of batch incubation, GP and GY increased with PBE inclusion, suggesting that it contains fermentable compounds. Gas production depends on the available nutrients for microorganisms (Elghandour et al. 2016) and therefore, it is possible that PBE provided an additional nutritional contribution to the inoculum microorganisms. Flavonoids, such as quercetin, are metabolized in the rumen by hydrolysis of the glycoside moiety and cleavage of the heterocyclic ring, producing di-and monohydroxyphenolics, phloroglucinol and SCFA such as acetate and butyrate (McSweeney et al. 2002). Alternatively, according to Jiménez-Peralta et al. (2011), some rumen bacteria can metabolize various phenolic compounds cross-linking polysaccharides and lignin, thus increasing fermentation and GP. However, this is unlikely to have occurred in this study, as IVDMD was not increased. Moreover, the most common response to tannin or polyphenol inclusion in ruminant diets is a decrease in in vitro GP (Rira et al. 2015;Brutti et al. 2019).
At 6-h of incubation, a 1.8% PBE in DM reduced the CH 4 production (-26.7%) but not at 12 and 24-h, concurring with previous studies (Oliveira et al. 2007;Szczechowiak et al. 2016) Table 4. Effect of a polyphenolic extract from pine bark (PBE) at different concentrations (% of DM basis) as an additive for feed in in vitro CH 4 output and kinetics parameters. Different letters in same column indicate significant differences (p < 0.05). a b = asymptotic production (mg methane/0.5 g DM incubated); c = rate of methane production (/h); Lag = initial delay before methane production begins (h); t 1/2 = half-life (h); APR = average production rate (mg/g DM incubated per h). b SEM = Standard error of mean. c Probability of differences between treatments (T), or of a linear (L) or quadratic (Q) effect by PBE concentration. where the tannin content in the diet did not reduce CH 4 production. However, mitigation of CH 4 production by flavonoids has been reported previously in several reviews (Patra and Saxena 2010;Vasta et al. 2019). In addition, leaves of different tannin-rich tropical trees suppressed CH 4 production in different magnitudes in in vitro (Bhatta et al. 2015) and in vivo conditions (Malik et al. 2017). Because the CH 4 lag time was increased by PBE supplementation, while the 6-h IVDMD and GP were unaffected, along with the fact that the SCFA profile could not be determined in this study, we suggest that the PBE polyphenols reduce CH 4 production in the first hours of incubation by a delayed microbial colonization and growth rate (Firkins et al. 1998), an initial methanogen inhibition (Elghandour et al. 2016), or by a change in the SCFA profile, since fermentation to propionate increases hydrogen consumption, whereas acetate formation produces hydrogen (Patra and Saxena 2010;Vasta et al. 2019). The concentrations of NH 3 -N decreased when inclusions of PBE reached 1.5% (-31% at 24-h incubation). These concentrations were never lower than 5 mg/dL of NH 3 , which is the minimum level required for adequate DM digestion (Junior et al. 2017). In vitro NH 3 -N concentration represents the balance between degraded dietary protein, absorption through the rumen wall and use for microbial protein synthesis (Hariadi and Santoso 2010). Decreased NH 3 -N ls can be attributed to an NH 3 inhibition by increasing doses of polyphenols in supplemented substrates, which can affect the deamination process (Purba et al. 2020b), decrease dietary protein degradation by inhibiting protease activity (Brutti et al. 2019) and/ or create a pH-dependent (6.0-6.5) tannin-protein complex. In the abomasum, at pH 2.0, the complex dissociates, allowing peptidase action (Junior et al. 2017). The reduction of rumen CP degradation leads to a more efficient dietary protein use through the generation of 'by-pass protein', which may result in improved animal performance and decreased urea N excretion. However, the N losses via faeces (more resilient to environmental loss) should be slightly increased (Deaville et al. 2010). Despite the above, tannin supplementation lowers urinary N excretion ( Figure  1), but has had detrimental effects on animal performance (Aguerre et al. 2016). This suggests that by decreasing the protein degradation of the diet, the MCP could be affected by tannins via enzyme inhibition, affecting the viability of essential metal ions, and/or by changes in the bacterial cell wall, which would compensate for the increase of by-pass protein (Ahnert et al. 2015).
The average pH value varied from 6.3-6.4 in 24-h; an optimal range of 6.7 ± 0.5 is required to maintain normal cellulolytic activities, and pH values above 6.0 are required for microbial protein synthesis (Hariadi and Santoso 2010). Flavonoids are the main PBE components and can prevent pH decrease by having a direct buffer effect or by increasing the activity of lactate-consuming bacteria (Balcells et al. 2012;Goto et al. 2016).
The calculated SCFA increase with PBE addition is related to a GP increase, since there is a good association between calculated SCFA and in vitro GP; the degraded substrate in a closed in vitro gas system is converted into gases, SCFA, water and microbial mass (Makkar 2005). Higher SCFA concentrations reflect a greater amount of fermented substrate by rumen microorganisms (Purba et al. 2020c), however, as the IVDMD decreased by PBE inclusion in the substrate, the increasing concentration of SCFA in supplemented substrates can be attributed to the presence of some flavonoids in the extract, as quercetin, which have been shown to increase SCFA concentration (Purba et al. 2020b). This increase in SCFA can be beneficial, as these are the main end products of fermentation and represent the major supply of energy for ruminants (Salem et al. 2015). By contrast, both the PF 24 and the calculated MCP decreased with increasing PBE doses. A decrease in PF indicates that less substrate was converted into microbial biomass (Elghandour et al. 2016), possibly due to the tannins presents in the PBE that form tannin-protein or -fibre complexes (Jiménez-Peralta et al. 2011); in addition, CT can bind to ruminal microorganisms or their enzymes, inhibiting their growth (Castro-Montoya et al. 2018).
Our results indicate that the supplementation of PBE at a concentration of 1.5% DM basis in high-concentrate diets could decrease N excretion in ruminants, given the NH 3 -N concentration reduction (-31%). In addition, PBE can delay CH 4 production (-31%), but this delay is offset after 12 or 24-h. This suggests that PBE supplementation in ruminant diets has the potential to contribute to improve sustainability of environmentally friendly animal production systems. However, before it can be used in in vivo conditions and to validate its effects, it is necessary to perform long-term incubations (Rumen Simulation Technique, RUSITEC) and to clarify the effect of supplementing with 1.5% PBE on rumen microorganisms.

Conclusion
Supplementing high-concentrate diets with least 1.5% DM of a polyphenolic extract from pine bark (PBE) reduces CH 4 production at 6-h of incubation, but not at 24-h. However, at 24-h incubations NH 3 -N concentration can be reduced by 31% with a slight reduction in digestibility at 1.8% PBE inclusion. Our results warrant future research in long term incubations and under in vivo conditions to confirm PBE potential contribution to livestock systems sustainability.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by the Fondo de Fomento al Desarrollo Científico y Tecnológico under Grant ID14I10370. FONDEF