Changes in in vitro ruminal and post-ruminal degradation of tropical tannin-rich legumes due to varying levels of polyethylene glycol

Introduction

In the tropics, legumes are used in ruminant nutrition in combination with grasses to increase feed protein concentration (Lascano and Carulla, 1992). Among these, Calliandra calothyrsus (var. San Ramón CIAT 22310 and var. Patulul CIAT 22316) and Flemingia macrophylla (CIAT 17403) (CIAT, 1993) have a high protein level (Lascano et al., 2003) and an agronomic potential both in acidic, low-fertility soils and under dry conditions (Tiemann et al., 2008). However, the dry matter digestibility of these legumes is low (30–60%) due to their high content of condensed tannins (Maasdorp et al., 1999; Hess et al., 2003; Lascano et al., 2003).

High levels of condensed tannins (>50 g/kg plant DM) have antinutritional effects such as reduction in dry matter intake (Barahona et al., 1997), but lower concentrations (<50 g/kg plant dry matter) can be useful due to their capacity to bind protein in the rumen (Saminathan et al., 2014), increasing nitrogen flow to the small intestine (Waghorn et al., 1987; Wang et al., 1994; Min et al., 2003). Moderate concentrations of CT (20–45 g CT/Kg DM) could improve milk production, ovulation rate and wool growth and reduce internal parasite burdens (Min et al., 2003)

The potential of condensed tannins to bind proteins depends not only on their diet concentration but also on their molecular weight and monomeric composition (Hagerman et al., 1992; Siebert et al., 1996; Saminathan et al., 2014). These properties may be influenced by species and by environmental conditions (Hess et al., 2006) and even differ among varieties within the same species of plants (Jackson et al., 1996). Several in vitro and in vivo studies have shown that tanniferous legumes grown in Colombia added to grass decreased ruminal protein and dry matter degradation and ammonia production (Lascano et al., 2003; Hess et al., 2006; Stürm et al., 2007). Polyethylene glycol (PEG) binds condensed tannins reducing their biological activity. It has been used in these experiments to evaluate the effects of tannins. It contains a large number of oxygen atoms capable of forming hydrogen bonds with phenolic groups in tannins (Silanikove et al., 2001). Supplementation with PEG increased nutrient intake and total tract crude protein (CP), neutral detergent fibre (NDF) and acid detergent fibre (ADF) digestibility (Yisehak et al., 2014).

Tiemann et al. (2008) suggested a PEG:CT ratio of 1:1 is not always sufficient to bind and inactivate all condensed tannins present in tanniferous plants, and Cortés et al. (2009) showed that tannins have a different potential to inhibit protein degradation. Therefore, the ratio of tannin to PEG is relevant to evaluate the effect of tannins on protein degradation. Different tannins may require different amounts of PEG to inactivate them. Whether this protein is fully available post-ruminally requires further investigation (Getachew et al., 2000)

The objective of this work was to determine the minimal PEG:tannin ratio required to completely bind tannins from F. macrophylla and C. calothyrsus and to avoid negative effects on total (ruminal and post-ruminal) in vitro degradation of DM and CP (rdDM; rdCP and tdDM; tdCP) as well as ammonia concentration.

Materials and methods

Samples of legume foliage were obtained from the germplasm collection held at the International Center of Tropical Agriculture (CIAT) in Colombia. Legumes used were C. calothyrsus CIAT accession No. 22310 (var. San Ramón), C. calothyrsus CIAT accession No. 22316 (var. Patulul) and F. macrophylla (CIAT accession No. 17403) grown at CIAT's experimental station in Santander de Quilichao (Cauca, Colombia, 990 m elevation, 1700 mm average annual precipitation and 24 °C average temperature) on an Ultisol. Foliage was harvested in June from mature plants 8 weeks after a standardization cut, immediately stored at −20 °C (Luchini et al., 1996; Chaudhry and Mohamed, 2012) and later lyophilized and ground in a Wiley mill using a 1-mm screen.

Amounts of 0.5 g of DM for each legume were weighed into one of 117 plastic tubes (100 ml). Each tube was considered as an experimental unit. For each tannin source (legume), a solution of PEG (8000 Da average molecular weight, Sigma-Aldrich, USA) in McDougall buffer (McDougall, 1948) was added to achieve polyethylene glycol: condensed tannin (PEG:CT, g/g) ratios of 0:3, 1:3, 2:3 and 3:3, similar to the ones reported by Barry et al. (1986). Each tube containing one of each of the three legumes mixed with the PEG solution in the four ratios was prepared in triplicate (n = 3) and stored for 12 h at 39 °C to allow tannins and PEG to interact. A total of three sets of tubes (39 each set) were used for the subsequent determination of (i) apparent ruminal in vitro degradation of DM and CP (rdDM and rdCP, respectively), (ii) ruminal and post-ruminal in vitro degradation of DM and CP (tdDM and tdCP, respectively) and (iii) ruminal ammonia concentration.

The procedure of Tilley and Terry (1963), as modified by Carulla and Pabón (2004), was used to determine DM and CP degradation. After 48 h of incubation, a set of tubes were filtered, in order to estimate apparent ruminal DM and apparent CP degradation (rdDM and rdCP), a second set of tubes were incubated for 24 h to determine ammonia concentration and a third set of tubes were incubated for another 24 h with HCl/pepsin to estimate apparent total in vitro DM and CP degradation (tdDM and tdCP).

Ruminal fluid was obtained from a cannulated steer grazing kikuyu grass (Pennisetum clandestinum). The ruminal fluid was filtered through four layers of gauze and the filtrate was placed in a separation funnel, later gassified with CO_2. After standing for 15 min at 39 °C, the upper and lower parts of the ruminal fluid in the funnel were discarded for content of fibre residues. The remainder ruminal fluid was mixed with McDougall buffer in a ratio of 1:4 to obtain the inoculum. This inoculum was added to tubes containing either only the legume foliage (control; three per set of tubes) or PEG-treated legume foliage. After gassifying with CO₂, the tubes were sealed with rubber stoppers provided with Bunsen valves and incubated at 39 °C for 48 h.

After 48 h, the fermentation was stopped in the first set of tubes by adding 1 ml of mercury chloride solution (50 mg/ml). To the second set of tubes, 6 ml of HCl solution 6.21 N and 2 ml pepsin solution (50 g/l) were added and the incubation continued for another 24 h, simulating abomasal digestion (Tiemann et al., 2010). After incubation, contents of the tubes were filtered using nitrogen-free paper filters (no. 10300 012, S&S Whatman, Dassel, Germany). The residue was dried and subjected to Kjeldahl analysis (Tiemann et al., 2010). For the first set of tubes, rdDM and rdCP were calculated from the previous amounts of DM and CP and the amount recovered on the filter paper after incubation. Additionally, the residue of the third set of tubes was dried and tdDM and tdCP were determined.

To determine ammonia concentration, samples of 20 ml were taken from each incubation tube and transferred to tubes containing 0.64 ml of sulphuric acid solution (80%).

Forages were analysed for moisture (930.15, AOAC, 2005), protein (2001.11, AOAC, 2005) and neutral detergent fibre (NDF) (Van Soest et al., 1991). For the quantification of tannins, the HCl/butanol method was used (Terrill et al., 1992) modified by Barahona (1999). Ammonia concentration was determined using the distillation part of the Kjeldahl method (2001.11A.O.A.C. 2005) (Table 1).

The rdDM and tdDM were calculated as the difference of DM supplied and recovered after degradation with ruminal fluid or degradation with ruminal fluid and HCl/pepsin, respectively, and were corrected for the DM recovered from the blanks. In the treatments with PEG, the dry matter of the residue was corrected as for the amount of PEG added. The rdCP and tdCP were calculated as the difference of CP supplied and recovered after degradation with ruminal fluid or degradation with ruminal fluid and pepsin respectively. The N determined in the tube residue from degradation of protein was assumed to be only from undegraded plant CP, although microbia attached to solid forage particles might have led to an overestimation of CP in the residue. The difference between tdCP and rdCP was assumed to be equivalent to digestible bypass CP, that is CP that which was undegraded under ruminal conditions but degraded under post-ruminal conditions, pdBCP (protein undegraded in ruminal fluid but degraded with HCl/pepsin). From this value, the pepsin/HCl digestibility of the CP which was not degraded in ruminal fluid was calculated (pdCP).

Data were subjected to analysis of variance based on a completely randomized design within a 4 × 3 factorial arrangement of treatments, Y_ijk=μ+β_i+τ_j+βτ_ij+Є_ijk. Tree legume species, four PEG:CT ratios and their interactions were included as sources of variation using GLM procedure of sas (version 9.2; SAS Institute Inc., Cary, NC, USA). Multiple comparisons were made using the Tukey's honest (PDIFF) significant differences test (Tukey's HSD) and Levene's test for variance homogeneity. Differences were considered significant at p < 0.05.

Results

Higher crude protein levels were found in F. macrophylla than in either accession of C. calothyrsus. Content of extractable condensed tannins (ECT) was twice as much in C. calothyrsus accessions than in F. macrophylla (Table 1).

The apparent rdDM increased with the increase in PEG:CT ratio, and the extent of such increase varied with legume type (interaction legume×ratio, p < 0.001; Table 2). For both Calliandra accessions, the highest rdDM value was obtained with a PEG:CT ratio of 3:3. In contrast, no differences were found for Flemingia between PEG:CT ratios of 2:3 and 3:3 (p > 0.05), indicating a ratio of 2:3 might be enough to completely inactivate CT and to avoid negative effects on rdDM. For all legumes, a linear relationship was found between PEG:CT ratio and rdDM (p < 0.001). The tdDM also increased with the increase in PEG:CT ratio and the extent of this increase varied with the legume specie (interaction legume × ratio, p < 0.0001; Table 1). The highest tdDM values were obtained for ratios of PEG:tannin of 3:3 for both Calliandra varieties. For F. macrophylla, no differences were found between 2:3 and 3:3 ratios (p > 0.05 (Table 2).

The effect of increasing PEG:CT ratio on rdCP was similar to the effect on rdDM, with an increase in degradability as PEG:CT ratio increased (p < 0.001). Nevertheless, the increases differed between legumes (interaction species × ratio; p < 0.0001; Table 3). For Calliandra, the highest rdCP values were obtained with a PEG:CT ratio of 3:3, but for F. macrophylla, no differences were found between PEG:CT ratios of 2:3 and 3:3 (p > 0.05). The tdCP also increased with the increase in PEG:CT ratio, and the extent of this increase varied with legume species (interaction species × ratio, p < 0.001; Table 3). The highest tdCP values were obtained for ratios of PEG:tannin of 3:3 for all species, but for F. macrophylla, no differences were found between 2:3 and 3:3 ratios (p > 0.05).

For Calliandra, increasing the PEG:CT ratio from 0:3 to 2:3 increased the proportion of total CP degraded by HCl/pepsin, but at PEG:CT ratio of 3:3, there was a different response for each Calliandra variety (pdBCP; Table 3) For Patulul, there was a decrease when PEG:CT ratio increased from 2:3 to 3.3 (p < 0.05), but for San Ramon, 2:3 and 3:3 ratios did not differ (p > 0.05). Adding PEG to Flemingia increased pdBCP, but the extent of this increase was independent from the PEG:CT ratio and did not vary between ratios of 1:3, 2:3 or 3:3 (p > 0.05). These differences among legumes were also reflected by a highly significant interaction between legume and PEG:CT ratio (p < 0.0001).

The digestibility of the protein that reached the abomasum (not degraded in the rumen) dBCP was influenced by both legume type and PEG:CT ratio (interaction legume × ratio) (p < 0.0001). For Calliandra Patulul, as PEG:CT ratio increased, dBCP increased, but after 2:3 ratio, there was not an additional significant effect, and for Calliandra San Ramón, dBCP degradation increased as PEG:CT ratio increased up to 2:3, but at 3:3 ratio, there was a decrease in dBCP. Also for Flemingia, PEG addition increased (p < 0.05) dBCP, but such increase did not vary (p > 0.05) between the different PEG:CT ratios.

Legume species as such had no effect on ammonia levels (p > 0.05), but an interaction was found between legume and PEG:CT ratio (p < 0.001). For Calliandra Patulul and Flemingia, the highest ammonia level was obtained with a PEG:CT ratio of 3:3, and for Calliandra San Ramón, ammonia concentration was highest with a PEG:CT ratio of 2:3.

Discussion

Tannins in ruminant diets can have negative or positive effect depending on concentration, type of tannin, chemical structure and structural flexibility. The chemical differences include their molecular weight, monomeric units and hydroxylation (Egan and Ulyatt, 1980; Barry and Manley, 1984; Waghorn et al., 1994). Tannin structure may be different depending on plant species, plant parts or environment (Stewart et al., 2000; Tiemann et al., 2010). Additionally, tannins bind salivary proline-rich proteins forming aggregates that are supposed to be the origin of the astringency sensation (Soares et al., 2012). PEG has been used to study the effects of tannins because it binds them by forming inert tannin–PEG complexes (Makkar et al., 1995). It has been used to increase intake and digestibility of tannin-rich legumes and shrubs by grazing ruminants (Getachew et al., 2000). Different ratios of tannin to PEG have been reported to inactivate tannins, and optimal ratio to inactivate them may depend on tannin structure. Using different ratios (PEG:CT) may also allow to study the biological response to variation of active tannins.

In this experiment, each one of three tropical shrub legumes was combined with PEG in ratios PEG:tannin of 0:3, 1:3, 2:3 and 3:3, and incubated with ruminal fluid or with ruminal fluid + pepsin/HCl. Dry matter degradation increased as more PEG was added because it binds the tannins present in the legume, and consequently, there is more protein and other plant constituents available for ruminal micro-organisms. Similar results have been reported by others with tropical legumes where addition of PEG improved fermentability of plant constituents (Getachew et al., 2001; Milad et al., 2014; Olivares-Perez et al., 2014).

In our study, increases in PEG:CT ratio also resulted in an increase in ruminal protein degradation, resulting in more ammonia in the solution for all legumes suggesting that as the concentration of active tannins decreased, more protein was available for degradation, as have been observed by others (Barry et al., 1986; Waghorn et al., 1994; Carulla et al., 2001, 2005; Getachew et al., 2001). Ammonia concentration in rumen fluid resulting from degradation of Calliandra San Ramón was lower than for Calliandra Patulul or Flemingia which could be due to its lower protein content or differences in escape protein between the two provenances, which in turn were associated with differences in tannin structure (Stewart et al., 2000; Lascano et al., 2003).

In our experiment, maximal rumen degradation was for the two Calliandra varieties when PEG/CT ratio was 3:3 and in Flemingia when it was 2:3 suggesting that PEG has a higher affinity for Flemingia tannins or less PEG is required to inactivate them and therefore that tannin between these species differs. Differences in CT content, molecular weight, chemical structure, chemical composition and astringency have been reported (Cano et al., 1994; Lascano et al., 2003)

In this experiment, San Ramón and Flemingia maximal protein digested by pepsin CP were higher when PEG/CT ratio was 3:3 and Patulul with PEG:CT ratio was 2:3. Higher protein digestion by pepsin was found for San Ramón than for Patulul regardless of the amount of PEG added. Higher apparent digestibility of protein in the small intestine for San Ramon than for Patulul may be inferred from the in vivo study of Lascano et al. (2003). Differences in the response due to tannins within a species have been related to differences in tannin structure (Hedqvist et al., 2000; Stewart et al., 2000). The extractable condensed tannins (ECT) in Patulul are comprised mainly of catechin/epicatechin subunits (producing procyanidin on treatment with butanol/HCl), whereas in San Ramón, the ECT are mainly gallocatechin/epigallocatechin (producing prodelphinidin with butanol/HCl) (Stewart et al., 2000). F. macrophylla contains low levels of extractable CT (Andersson et al., 2006) and is rich in polyphenols such as daidzin, daidzein, genistin, genistein, flemingin A and flemingin (Shiao et al., 2005).

Addition of PEG (inactivation of tannins) increased the proportion of protein digested in HCl/pepsin in all legumes. This suggests that tannins decreased not only ruminal protein degradation but also the degradation of CP in HCl/pepsin (abomasum). Work by Jones and Mangan (1977) in vitro and Waghorn et al. (1987) in vivo suggested that protein digestion in the lower tract is not affected by tannins. However, feeding tropical legumes rich in tannins to sheep have shown an increase in N excretion in their faeces (Waghorn et al., 1987; Carulla et al., 2001) and reduced apparent protein digestibility in goats (Fagundes et al., 2014). However, increased faecal N excretion not necessarily means lower true protein digested in the intestines. The argument of Jones and Mangan (1977) is that even though tannins increase faecal N excretion, this is compensated by a higher protein flux to the duodenum due to lower ruminal CP fermentation and losses of CP as ammonia. Our results showed that tannins of these tropical legumes do increase protein not digested in the rumen (by pass) due to reduced CP loss as ammonia has been demonstrated elsewhere (Barry et al., 1986; Waghorn et al., 1994; Caygill and Mueller-Harvey, 1999). However, at higher levels of tannins (no PEG added), most of this protein is not digested by pepsin (less than 5%). This suggests that complexes formed between CT and CP are stable even at low pH (abomasum) in contradiction to the often-stated hypothesis that CT may have beneficial effects on CP utilization. Low levels of tannins (PEG:CT 2:3) had similar amounts of protein digested in pepsin suggesting no beneficial effects of tannins. The only exception was C. calothyrsus Patulul. For this legume, low levels of tannins may be beneficial for the ruminant because data suggest that more protein will be digested in the abomasum as compared with no tannins.

Some evidence of the beneficial effects of tannins comes from research in New Zealand with the legume Lotus pedunculatus. Positive effects of tannins in animal production are explained partially by a reduction of protein degradation in the rumen increasing the flow of amino acids to the intestine (Waghorn et al., 1998; Carulla et al., 2001; Frutos et al., 2004; Hess et al., 2006). Contrary to our findings, Cortés et al. (2009) reported that purified tannins from tropical plants increased the amount of protein that will escape ruminal degradation increasing its availability for digestion and absorption in the intestine.

Data from our in vitro study suggest that tropical legumes containing tannins negatively affect protein digestion by pepsin even though they may increase dietary protein flow to the abomasum. This effect was similar regardless of the legume, and only small differences were observed among them. However, in vitro studies have limitations and in vivo experiments will be needed to confirm our in vitro results.