Location and collection of plant material

The collection of plant material was carried out in February 2018 at Xamantún Ranch belonging to the Technological Institute of Chiná of the National Technological Institute of Mexico, located at 19°43' North latitude and 90°24' West longitude. The site has a warm sub-humid A(W1) climate (^{García, 2004}), with mean annual temperature and precipitation of 26°C and 1,138 mm, respectively, and an elevation of 36 m.

One kg of fresh foliage, consisting of leaves and tender stems, was collected, simulating the browsing of an adult bovine at a maximum height of two metres, from the following 10 tree and shrub plant species: B. divaricata, M. calabura, Lysiloma latisiliquum, Tithonia diversifolia, Sida acuta, Guazuma ulmifolia, Moringa oleifera, Acacia farnesiana, Samanea saman and Coccoloba cozumelensis, as well as 5 kg of Pennisetum sp. var. maralfalfa. Samples were dried in a forced-air oven at a temperature of 55°C for 72 h. The samples were then ground in a milling machine. Subsequently, they were ground in a Willey mill with a 1 mm sieve and stored at room temperature (24°C) until use.

Presence of secondary metabolites

The qualitative presence of various secondary metabolites was determined from the foliage of the 10 forage tree and shrub species mentioned above. For this purpose, 25 g of the plant material was taken and placed in glass jars. Subsequently, 3 v/v of 96 % ethanol was added as a solvent to extract MSec. Once the ethanolic extracts were obtained, the presence of MSec was determined following the protocol of Valencia del Toro and Garín (Valencia-Del Toro and Garín-Aguilar, 2010). The results are shown in Table 1.

Treatments

Taking into account the MSec profiles obtained, as well as the scarce information on their effects on ruminal fermentation, two plant species were selected for their high content of metabolites inhibiting microbial activity: M. calabura, which had a high presence of steroidal saponins and flavonoids of the xanthone and flavone type, and B. divaricata, which presented a moderate content of tannins derived from catechol. Based on the above, five treatments were evaluated: P=Pennisetum sp. (100%); Mc=M. calabura (100%); Bd=B. divaricata (100%); McP=M. calabura (30%)/Pennisetum sp. (70%), and BdP=B. divaricata (30%)/Pennisetum sp. (70%).

Chemical analysis

Dry and ground samples from the five treatments were analysed in triplicate for dry matter (DM), ash, crude protein (CP) and ethereal extract (EE) content according to AOAC (2006); while neutral detergent fiber (NDF) and acid detergent fiber (ADF) were determined according to the technique of ^{Van Soest et al. (1991)}. The results are presented in Table 2.

In vitro fermentation

The procedure for the collection of rumen fluid (RF) was carried out in strict accordance with the Mexican Official Standard NOM-062-ZOO-1999 "Technical Specifications for the Production, Care and Use of Laboratory Animals" (NOM-062-ZOO-1999., 1999). The RF used were obtained from four uncastrated male steers, from different crosses of Black Sardinian, Gyr, Brahman and Brown Swiss breeds, one year old and with an average weight of 250 kg. Their feed consisted of an integrated diet (70:30 forage:concentrate) composed of Echynochloa polystachya and Brachiaria brizantha grass straw, ground maize, dried distillers grains, molasses and mineral premix, which was offered twice a day (8:00 and 16:00 hours). Daily the steers consumed the equivalent of 3% DM of their live weight.

RF extraction was performed in the morning, one hour before the first ration of the day using a manual RF collector (Rumen-Mate with RFE, Drench-Mate®). The collected RF was placed in a thermos flask previously heated to 39°C and immediately transported to the laboratory. Once in the laboratory, the RF was filtered through four layers of sky blanket and it was used to prepare the inoculum, which was prepared according to the methodology of ^{Menke et al. (1979)}. Subsequently, 50 mL of inoculum was added to 120 mL amber glass vials containing 0.5 g of each treatment, capped with a rubber stopper, sealed with an aluminium ring and placed in a water bath at 39°C, shaking manually every 2 hours. A total of three fermentations were carried out, with three replicates each.

In vitro dry matter degradability

In vitro dry matter degradability (DMD, mg/g DM) was determined at 24 and 72 hours. At the end of each time, the vials were removed from the water bath and placed in ice water for 2 hours to stop microbial activity. Subsequently, the contents of each vial were filtered with a vacuum pump; the filtered material was placed in an oven at 70°C for 24 hours until constant weight.

In vitro gas production

Gas production (GP) was recorded at 2, 4, 6, 8, 12, 12, 16, 16, 20, 24, 30, 36, 42, 48, 60 and 72 hours using a manometer (^{Theodorou et al., 1994}). After each recording, the pressure in each bottle was equalized to zero. To obtain the maximum volume (Vmax), fermentation rate (S) and lag phase (L), the NLIN procedure of the statistical software ^{SAS (2004)} version 9.0 was used, using the model mentioned by ^{Kholif et al. (2017)}.

CH₄ and CO₂ determination

The determination of CH₄ and CO₂ was performed according to ^{Kholif et al. (2017}), with the following modifications: every 6 h for 24 h, the gas produced in vials was collected with a 60 mL glass syringe. Subsequently this gas was transferred to another vial containing 50 mL of a NaOH (1N) solution, shaken to ensure the incorporation of the gas into this solution, collected again and the gas was recorded with the same glass syringe. The mixing of the gas with the NaOH solution allowed the absorption of CO2, and the volume of gas collected in the syringe was considered as CH₄.

Determination of VFA

The determination of VFA was carried out by gas chromatography (^{Erwin et al., 1961}). After 72 hours of fermentation, 1.6 mL of liquid was collected from each vial, placed in a microcentrifuge tube with 0.4 mL of 25% metaphosphoric acid and stored at -20°C.

Metabolizable energy and partition factor

The calculation of metabolizable energy was performed following the model proposed by ^{Menke et al. (1979}). The partition factor (ratio of DMD (mg/g) to GP (ml/g DM), was calculated after 24 hours of incubation using the model mentioned by ^{Kholif et al. (2017}).

Experimental design and statistical analysis

A completely randomized block design was used using the following statistical model:

Yij= µ+ Ti+ Fj +Eij

Where: Yij = is each observation of i-th treatment (Ti) of j-th fermentation (Fj); µ =arithmetic mean; Eij = experimental error.

The results were analysed with the GLM procedure of the statistical programme ^{SAS (2004)} version 9.0 and the comparison of means was performed by Tukey's test (P≤0.05).

Presence of secondary metabolites and chemical analysis

The forage species evaluated had the presence of several MSec, but none showed the presence of alkaloids, quinones or laucoanthocyanidins (Table 1). It was observed that M. oleifera, A. farnesiana and L. latisiliquum species had a high content of gallic acid- derived tannins and G. ulmifolia in catechol-derived tannins; while M. calabura and S. saman species had a moderate content in gallic acid-derived tannins, and B. divaricata and T. diversifolia species in catechol-derived tannins. Regarding saponins, the species with the highest presence of steroids were M. calabura, L. latisiliquum, T. diversifolia, S. acuta, A. farnesiana and S. saman, while only M. oleifera had a low presence of terpenoids. From the flavonoids group, the forage species with a high presence of xanthones and flavones were M. calabura, M. oleifera and A. farnesiana, while S. saman had a moderate presence; the rest of the species had a low presence, with the exception of T. diversifolia, which was the only species with a low presence of aurones and chalcones.

Secondary metabolites are compounds produced in different pathways of plant secondary metabolism, which are not essential for growth and reproduction. These biomolecules perform various functions, notably environmental stress response, immunity and protection against pathogenic microorganisms, pests (^{Pang et al., 2021}) and herbivorous animals (^{Ugbogu et al., 2019}). It has been shown that some of these secondary metabolites may have antimicrobial activity on ruminal microorganisms, with decreased ruminal methanogenesis being one of their main effects (^{Patra et al., 2017}), which has been widely reported for tannins and saponins (^{Anantasook et al., 2013}; ^{Ugbogu et al., 2019}; ^{Patra et al., 2017}) and to a lesser extent for flavonoids (^{Patra et al., 2017}). Contrary to ^{Delgado et al. (2012)}, who reported moderate or high presence of alkaloids in mixtures of tree foliages with grasses, this metabolite was not detected in the present study. Alkaloids possess antimicrobial activity but have also been shown to be toxic to both animals and humans. In general, animals do not consume high amounts of plants with high alkaloid content due to their bitter taste (^{Guil-Guerrero et al., 2016}).

The selection of M. calabura and B. divaricata was due to their phytochemical characteristics and the scarce information on their effects on ruminal fermentation. The nutritional value of the two tree species evaluated in this study is similar to that reported in other studies (Table 2). According to ^{Gómez-Fuentes-Galindo et al. (2017)}, B. divaricata has lower NDF content (46%) and higher CP content (12.8%) than grasses such as Panicum maxima or Paspalum langei which have on average 70% NDF and 6.5% CP, characteristics that could enhance a higher voluntary intake and a faster passage rate. Other authors have reported that B. divaricata has between 12 and 18% CP, between 30 and 40% in vitro ruminal digestibility, between 1.5 and 3.8% tannins, and the presence of saponins and alkaloids (^{Albores-Moreno et al., 2018}; ^{Cab-Jiménez et al., 2018}; ^{Gómez-Fuentes-Galindo et al., 2017}; ^{Sosa-Rubio et al., 2004}). Regarding M. calabura, nutritional information is practically non-existent. ^{Kongvongxay et al. (2011}) reported that M. calabura leaves contain 13% CP, but they point out that little is known about its nutritional value as it is little used for animal feed. Regarding the content of secondary metabolites, its fruits (^{Pereira et al., 2018}) and leaves (^{Pujaningsih et al., 2018}) have been reported to contain bioactive compounds with antioxidant and antimicrobial activity, among which several phenolic compounds, including anthocyanins and flavonoids, stand out.

In vitro dry matter degradability

Table 3 shows the results of the DMD at 24 and 72 h of M. calabura and B. divaricata foliage alone and combined with Pennisetum sp. and shows differences between treatments (P≤0.05). The treatments with the highest DMD at 24 h were BdP, followed by P and Bd; however, at 72 h they were P, followed by BdP and McP. M. calabura foliage alone had the lowest DMD at both times, which is in agreement with ^{Puspitaning (2012)}, who mentions that 20% M. calabura in the diet decreases DMD. On the other hand, the DMD of B. divaricata was higher (53%) than that reported by other authors (32 and 39%) ((^{Albores-Moreno et al., 2018}; ^{Sosa-Rubio et al., 2004}).

^{Kamalak et al. (2004}) conclude that cell wall content in forages negatively affects ruminal fermentation parameters. ^{Zhang et al. (2017)} emphasized that the incorporation of foliage of some tropical tree and shrub species in grass-based ruminant diets increases CP content and decreases total structural carbohydrate content in the diet. Such statements were partially observed in this study. Treatments Mc and Bd (corresponding to foliages alone) had ≈33% lower NDF content and ≈57.5% higher CP content, compared to P, which is similar to that reported by ^{Gómez-Fuentes-Galindo et al. (2017)}, who reported NDF and CP contents of 46.5% and 12.8%, respectively, for B. divaricata. However, the lower NDF and higher CP content of M. calabura and B. divaricata did not improve DMD, which contrasts with those reported for other foliages, such as L. leucocephala, which increases DMD by 18% when included at 25% in grass-based cattle diets (^{Molina et al., 2015}). In the present study, the inclusion of M. calabura negatively affected DMD (P≤0.05), as it was 10% lower in the McP treatment than in P.

In vitro gas production

GP is presented in Figure 1. There were no significant differences (P>0.05) during the 72 h of fermentation between treatments Bd, BdP, McP and P, which had an average cumulative GP of 439 mL/g DM. However, there were significant differences (P≤0.05) with Mc, which had a cumulative GP of 283.49 mL/g DM. ^{Albores-Moreno et al. (2018)} reported a GP of 200.59 mL/g DM at 72h with B. divaricata foliage, which is 53% lower than that observed in the present study. On the other hand, the GP of M. calabura foliage reported by ^{Silivong et al. (2013}) was 269 mL/g DM at 24 h, 166% higher than that obtained with Mc. These variations in GP can be attributed to differences in forage composition, generally when they present high concentrations of CP, fiber and polyphenol content (^{Vargas et al., 2012}). In our study, Mc was the treatment with the lowest GP and degradability.

Where: P=Pennisetum sp. (100%); Mc=M. calabura (100%); Bd= B. divaricata (100%); McP= M. calabura (30%)/Pennisetum sp. (70%), and BdP= B. divaricata (30%)/Pennisetum sp. (70%)

Figure 1 Foliage gas production of Muntingia calabura and Bauhinia divaricata alone and combined with Pennisetum sp 

In vitro fermentation characteristics and CH₄ and CO₂ production

The results of Vmax, S, L, CH₄ and CO₂ production, ME and PF are shown in Table 4. The Vmax and S of the Mc treatment was significantly lower (P≤0.05) than the rest of the treatments; however, the L of Bd and Mc (the foliage-only treatments) was about 20% lower than the L of the McP, BdP and P treatments. Regarding CH4 production there were also significant differences (P≤0.05), with Mc being the treatment with the highest CH4 production, while Bd and BdP had the lowest production. In relation to ME content, it was observed that the highest values were presented by Bd and BdP, being close to 2 Mcal/kg DM, which were significantly different (P≤0.05) to Mc and McP treatments, with 1.5 Mcal/kg DM on average. There were no differences (P>0.05) in PF at 24 h, but there were differences at 72 h (P≤0.05) between treatments with foliage alone, being higher with Mc and lower with Bd.

The behavior of ruminal fermentation depends on the availability of carbohydrates present in the forages (^{Zhang et al., 2017}), so in this study, treatments with higher Vmax of GP could be related to a higher degradation of structural carbohydrates present in the foliage. While the low Vmax in Mc may be due to a low concentration of the soluble fraction of the feed, decreasing the amount of gas produced. In addition, it is likely that the presence of MSec such as saponins and flavonoids may have affected DM fermentation and hence GP. These MSec interact directly on the cytoplasmic membrane function of the rumen microorganisms by inhibiting cell wall synthesis, creating a defaunation effect, thereby decreasing the protozoa and consequently the associated methanogenic archaea and thus decreasing rumen methanogenesis. In the rumen, protozoa are associated with methanogenic archaea and their relationship can generate between 9 and 37% of the total enteric CH₄ emissions produced by ruminants (^{Rivera et al., 2018}; ^{Vargas et al., 2012}). Meanwhile, Velez-Terranova et al. (2014) and ^{Lakhani & Lakhani, 2018} note that plants with high flavonoid content decrease CH₄ production and induce extensive stimulation of microbial metabolism in the rumen, which increases both CP degradability and cell wall constituents by enhancing fermentation by up to 50%. In this regard, ^{Silivong et al. 2013} reported that CH₄ production in the in vitro fermentation of M. calabura was only 3.8%, a concentration considerably lower than that obtained for L. leucocephala (7.8%) or G. sepium (15. 6%). It contrasts with what was reported in this study, in which the preliminary chemical analysis showed that the M. calabura foliage alone has a high presence of saponins and flavonoids, and a moderate presence of tannins derived from gallic acid; however, in the in vitro test, the gas produced with Mc had a higher CH₄ concentration, close to 15%.

Volatile fatty acids

No significant differences (P>0.05) were observed in VFA concentration, except for acetic acid between Mc and P with Bd, and isovaleric between Mc and Bd (P≤0.05; Table 5). There are no studies evaluating VFA production from ruminal fermentation using M. calabura or B. divaricata as substrate; however, the modification in ruminal VFA production may be influenced by the presence of MSec in the diet. In this regard, ^{Broudiscou & Lassalas, (2000}) conducted a study where they evaluated the effect of dry extracts of Lavandula officinalis and Equisetum arvense. Ttwo species known for their high flavonoid content, on ruminal fermentation in vitro and found that the use of the extract of both species improved the fermentation rate by 50% by increasing the production of acetate and propionate, thus reducing CH₄ production. In the present study, the foliage of M. calabura showed a high flavone content, while in the foliage of B. divaricata it was low, which may have influenced the higher concentration of acetic acid in the Mc treatment. On the other hand, the higher concentration of isovaleric acid in Bd could have been because of the higher CP content of this foliage (almost 15% in BS), which is produced from the decarboxylation and deanimation of leucine, a branched-chain amino acid (^{Apajalahti et al., 2019}).