Content of phenolic , extractable and bound condensed tannins and their effect on in vitro gas production from browse leaves

Two studies were conducted to screen leaves of four browse species (Dichrostachys cinerea, Flagea villosa, Harrisonia abyssinica and Piliostigma thorningii): 1. to quantify levels of total extractable phenolics (TEP), extractable tannins (TET), total condensed tannins (CT), extractable and bound CT fractions and proanthocyanidins (PAs) using chemical assays; 2. to investigate effect of tannin anti-nutritive activity on in vitro gas production assessed by polyethylene glycol (PEG) tannin bioassay. Crude protein (CP) varied (P<0.05) between fodder species from 109 (P. thorningii) to 160 g/kg DM (D. cinerea). The fodders had detectable TEP that varied from 112 (P. thorningii) to 234 mg/g DM (F. villosa). TET varied (P<0.05) between species from 95 (P. thorningii) to 220 mg/g DM (F. villosa). The content of CT varied (P<0.05) from 53.2 (F. villosa) to 98.3 mg/g DM in P. thorningii. High proportion of CT was bound to protein (40-51.4%) compared to soluble (19.5-33.1 %) and fibre-bound (22.7-27.3%) CT fractions. Characterization of PAs in leaves revealed presence of flavan-3-ol and flavan-3,4-diols flavonoids. The increase in GP due to PEG supplementation varied (P<0.05) from 44.4 ml/g OM (14.5%) to 132.3 ml/g OM (69.4%) at 16 h, and from 34.9 ml/g OM (10.3%) to 132.2 ml/g OM (57.2%) at 24 h, in H. abyssinica and D. cinerea, respectively. Variable responses on gas production between species’ leaves due to PEG supplementation demonstrate adverse effects of tannins on depressed feed digestibility. Improved gas production and digestibility were due to the ability of PEG to bind and complex tannins, and recover feed nutrients bound by tannins. 194 EFFECT OF TANNINS ON BROWSE LEAVES UTILIZATION 195 RUBANZA C.D.K. ET AL. Therefore, utilization of these browse species’ leaves as protein supplements to ruminants could be optimized through reduction of levels of phenolic and tannin anti-nutritional factors (ANFs).


INTRODUCTION
Browse tree foliages could be used as source of protein in the tropical region for improved ruminant nutrition and production.Livestock production in arid and semi-arid regions of the world is limited by inadequate feed nitrogen (N) supply from low protein content of basal feeds (hay, straw and stover) (Leng, 1990).In north-western Tanzania during dry seasons ruminants are grazed on standing hay forages that have too low crude protein (CP) contents to meet the normal protein requirement of 80 g/kg DM of CP needed for optimal rumen microbial function (Annison and Bryden, 1998).Browse fodder from semi-arid regions of Tanzania indicate high CP values (100-200 g/kg DM) (Topps, 1997;Rubanza et al., 2003).However, utilization of browse legume fodder as protein supplements to ruminants fed on low-quality roughages could be limited by their contents of phenolic and tannin feed anti-nutritional factors (ANFs) due to their effects on reduced feed digestibility and nutrient utilization (Mangan, 1988;Makkar, 1989).
Tannins are phenolic plant secondary compounds that possess enough hydroxyl groups to form complexes with proteins and carbohydrates (Mangan, 1988;Makkar, 1989).Tannins are classified into hydrolysable tannins (HT) that represent the polyesters of gallic acids (gallotannins) and hexahydroxydiphenic acids (ellagitannins) (Haslam, 1998), and condensed tannins (CT) or proanthocyanidins (PA), which refer to complex of oligomers and polymers of flavonoid units (Haslam, 1998;Schofield et al., 2001).There is scanty information from the literature on nutritive potential and content of tannin in tropical browse fodder.Similarly, levels of proanthocyanidins (PA), and proportions of soluble CT fraction and CT bound to dietary protein and fibre have rarely been reported.Also, little is known on forages' proanthocyanidins composition, tannin structure and tannin structure-biological activity relationship in ruminants (Waterman, 2000).
Protein analogies such as polyethylene glycol (PEG) could alleviate adverse effects of tannins on feed digestibility due to their ability to neutralize tannins'antinutritive activity (Makkar et al., 1995).PEG contains sufficient oxygen molecules capable of forming strong hydrogen bonds with the phenolic and hydroxyl groups in tannins; and to precipitate tannin from solutions (Silanikove et al., 2001a).This property of binding tannins using tannin binding agents has been used for assessment of adverse effects of tannins on nutrients availability in vitro (Makkar et al., 1995;Getachew et al., 2000), in sacco (Silanikove et al., 1996) and in vivo (Silanikove et al., 2001b).Despite wide application of PEG in nutritional studies, there is little information from the literature on the use of PEG to neutralize tannin in different tropical forages.Available data are limited to those of Makkar et al. (1995) and Makkar and Becker (1996) on Acacia bateri, A. saligna and Dichrostachys cinerea; Getachew et al. (2000) on Dichrostachys spp.; and Rubanza et al. (2003) on leaves and pods of Acacia nilotica, A. polyacantha and A. tortilis.
A study was therefore, conducted to screen four browse fodder species' leaves from D. cinerea, F. villosa, H. abyssinica and P. thorningii indigenous to Tanzania, in order to: 1. quantify levels of phenolic and tannin ANFs; 2. determine soluble, protein-bound and fibre-bound CT fractions; 3. elucidate proanthocyanidins' structure and composition, and 4. assess adverse effects of tannins' anti-nutritive activity on in vitro gas production using PEG (MW 6000) tannin bioassay.

Study area
This study was conducted in Shinyanga region located at 1100-1300 m above sea level in north-western Tanzania (2-3 o S; 31-31.5 o E).Average annual rainfall ranges from 600-800 mm that rains from November through mid April.Dry season starts from June to October.Minimum and maximum temperatures vary from 15 to 18.3 o C and 27.6 to 30.2 o C, respectively.Common vegetation includes short grasses and scattered shrubs and trees.

Forage sample collection and processing
Browse leaves (soft twigs less than 5 mm diameter) samples were hand plucked from 8 to 10 trees selected at random in each of four sub-plots (70 × 70 m) in four rangelands in five administrative districts (Bariadi, Kahama, Meatu, Shinyanga Urban and Shinyanga Rural) in the study area.The samples were harvested at an advanced maturity stage in late rainy season between mid April and mid May in 2002.The samples were pooled for each individual tree species, and dried at 50 o C in a forced air oven for 48 h to constant weight.Foliage leaf samples for chemical analysis were ground to pass through a 2.00 mm-sieve and sub-sampled into 12 bulk samples (three samples for each species) for further laboratory analyses and nutritional studies.Foliage leaf samples for phenolic and tannin assays and for in vitro gas production PEG tannin bioassay were further ground to pass through 1.0 mm-sieve.Chemical compositions including phenolics and tannin assays were determined in duplicates.In vitro gas production and tannin PEG bioassay were conducted in triplicates.

Phenolics and tannin assays
Extraction and determination of total phenolics and tannins.Extraction of phenolics and tannins from leaf samples was carried out using the method of Makkar (2000).The TEP and TET were assayed according to the method of Jolkumen-Tiito (1985) using Folin-Ciocalteu's reagent (Sigma-Aldrich Chemie, Steinhein, Germany) and tannic acid standard (Sigma-Aldrich Chemie, Steinhein, Germany) as described by Makkar (2000).TET was estimated gravimetrically as a difference of phenolics remaining from total phenolics after binding tannins with polyvinyl polypyrrolidone (PVPP; Sigma-Aldrich Chemie, Steinhein, Germany) (Makkar et al., 1993).The contents of TEP and TET were expressed as tannic acid equivalent.

Soluble and bound condensed tannins assay
Extraction and determination of soluble and bound condensed tannin fractions.The soluble or extractable, protein-bound and fibre-bound CT fractions were estimated using the modified butanol/HCl technique (Porter et al., 1986) by extracting the samples with 1 % aqueous sodium deodocyl sulphate (SDS) and 5% 2-β-mercaptoethanol solution as described by Jackson et al. (1996) and Stewart et al. (2000).A sample of approximately 0.1 g DM (1.0 mm) was extracted (in eppendorff tubes) three times with a mixture of 4 ml of acetone/water (7:3 v/v), containing 1 g/l ascorbic acid, and 2 ml of dichloromethane.The tubes were vortexed, then centrifuged at 3 000 rpm for 15 min while transferring the upper (aqueous) layer to a 10 ml volumetric flask, and the lower (dichloromethane and acetone) was discarded.The combined aqueous fractions containing soluble or extractable CT were made up to 10 ml with distilled water.
Extraction and determination of protein-bound condensed tannins.The solid residues following extraction of free-bound CT were dried for few minutes in a stream of nitrogen at room temperature to drive off any residual volatile solvents.A-3 ml of 10 g/l aqueous sodium deodocyl sulphate (SDS) containing 50 g/l 2-β-mercaptoethanol (SDS) solution was added to the residues and heated at 95 o C for 45 min as described by Stewart et al. (2000).The mixture was centrifuged for 15 min at 3 000 rpm and the supernatant solution transferred to a 10 ml volumetric flask and repeated for three times.The combined supernatant solution was made to 10 ml with SDS solution.Contents of CT in these fractions were assayed by butanol/HC/Fe 3+ (Porter et al., 1986).A mixture of aqueous extract (0.25 ml) and SDS-solution (0.5 ml) was added to 6 ml of butanol/HCl (95:5 v/v); vortexed and heated at 95 o C on a metal heat block for 1 h, with further vortexing after 30 min.Absorbance of the red anthocyanidin products (condensed tannins) was measured at 550 nm.
Extraction and determination of fibre-bound condensed tannins.Fibre-bound CTs were estimated using same procedure as for the protein-bound, except that 6 ml of butanol/HCl and 0.5 ml of SDS-solution were added directly to the solid residue during the extraction process as described by Stewart et al. (2000).The mixture was heated at 100 o C on a metal heat block for 1 h, cooled, centrifuged at 3 000 rpm for 15 min before the absorbance of the supernatant was measured at 550 nm.Absorbance values for extractable, protein-bound and fibre-bound CT fractions for each of four browse species were compared with the respective species' blank values.Blank samples constituted plant extracts that were extracted in a similar procedure with butanol/H 2 O (95:5 v/v) replacing the butanol/HCl as described by Jackson et al. (1996) and Stewart et al. (2000).Absorbance values for soluble, protein-bound and fibre-bound CT fractions were converted to condensed tannin concentrations by including external tannin standard purified from Acacia nilotica at known concentrations in each of the three runs.The concentration of CT in the browse samples were converted to mg A. nilotica tannin equivalent/g DM from A. nilotica tannin regression equation: where Y is absorbance at 550 nm, and X is CT concentration expressed as A. nilotica tannin.Purification of tannin from A. nilotica leaves for the standard was achieved by back extraction with diethyl ether to remove low molecular weight phenolics and pigments as detailed by Terrill et al. (1992).

Effect of polyethyethylene glycol supplementation on in vitro gas production
Animals and management.Rumen fluid for in vitro tannin bioassay was obtained from three healthy mature Japanese Corriedale female sheep fitted with permanent rumen cannulae (70 mm).The fistulated animals were kept on standard daily ration of 800 g timothy hay and 200 g concentrates divided into two equal meals fed at 8.00 and 16.00 h daily.The animals were supplemented with minerals and had free access to water throughout the experimental period.Rumen fluid was obtained from the three sheep early in the morning before feeding, and filtered through three layers of cheese-cloth.
The mixture was kept stirred under CO 2 flushing at 39 o C using a magnetic stirrer fitted with a hot plate.
Polyethylene glycol (PEG) tannin bioassay.Adverse effects of tannin ANFs on in vitro gas production, and thus digestibility, were assessed by incubation of approximately 500 mg DM test feed samples (in triplicates) with or without 1.0 g PEG molecular weight (MW, 6 000) (Wako Pure Chemicals Industries Ltd., Osaka, Japan).PEG (MW 6 000) was preferred due to its higher binding capacity to tannins at near neutral pH that approaches true biological value.The feed samples were incubated in 100 ml glass syringes based on Menke and Steingass (1988) standard procedure.A portion (40.0±0.5 ml) of the rumen fluid/buffer medium was transferred into graduated syringes (100 ml) and incubated into water bath at 39±0.1 o C as described by Blümmel and Ørskov (1993).Vaseline oil was applied to the piston to ease movement and prevent escape of gas.The syringes were pre-warmed (39 o C) for 1 h, before addition of 40±0.5 ml of rumen fluid/buffer mixture into each syringe, and incubated in a water bath maintained at 39±0.1 o C. The syringes were gently shaken every hour during the first 8 h of incubation.Gas production readings (ml) were recorded after 2, 4, 6, 8, 12, 16 and 24 h for both PEG treated samples and blank samples.

Statistical analyses
Data on chemical compositions, phenolics, condensed tannins and CT fractions in forages were subjected to the analysis of variance (ANOVA) using the General Linear Model Procedure (GLMP) (SAS/Statview, 1999), and were analysed based on statistical model: where: Y ij is the general observation on chemical composition, phenolics, condensed tannins, CT fractions and proanthocyanidins' composition; µ ij is the general mean common for each parameter under investigation; S i is the i th effect of browse fodder species on the observed parameters; e i is the standard error term.Data on in vitro gas production were subjected to the analysis of variance (ANOVA) using the General Linear Model Procedure (GLMP) (SAS/ Statview, 1999), and were analysed based on statistical model: where: Y ij is the general observation on gas production, µ ij is the general mean common for each observation; S i is the i th effect of browse fodder species on the observed parameters; P j is the j th effect of PEG supplementation on in vitro gas production; (SxP) ij refer to i th and j th interaction effects of fodder species and PEG supplementation on in vitro gas production, e i is the standard error term.

Phenolics and tannins compositions
Total extractable phenolics and tannins.The selected browse forages had significantly (P<0.05)variable levels of TEP, TET and CT among fodder species (Table 2).Flagea villosa had (P<0.05)highest TEP content (234 mg/g DM) compared to P. thorningii that had (P<0.05)lowest total phenolics (112 mg/g DM).There was no difference (P>0.05) in TEP content between D. cinerea and P. thorningii (114 vs 112 mg/g DM, respectively).Flagea villosa had (P<0.05)higher TET content (220 mg/g DM) than H. abyssinica (139 mg/g DM), while P. thorningii and D. cinerea had the least TET contents of 95 and 96 mg/g DM, respectively.The content of CT varied (P<0.05) from 53.2 to 98.3 mg/g DM in F. villosa and P. thorningii, respectively.There was no (P>0.05)difference in CT between F. villosa and H. abyssinica.
Soluble and bound-condensed tannin fractions.The forages had variable (P<0.05)CT fractions (Table 2) with D. cinerea having higher (P<0.05)soluble CT fraction of 22.9 mg/g DM compared to H. abyssinica that had the least soluble CT fraction (14.5 mg/g DM).Piliostigma thorningii had high (P<0.05)protein-bound and fibre-bound CT contents of 50.5 and 28.6 mg/g DM, respectively, compared to F. villosa and D. cinerea, which contained (P<0.05)lowest contents (22.2 and 13.4 mg/g DM, respectively).The forages had most of their CT in the protein-  a,b,c,d different letters in the same column indicate significant differences (P<0.05)*** P<0.001 1 total extractable phenolics (TEP) and total extractable tannins (TET) were expressed as mg/g tannic acid equivalent 2 total condensed tannins (TCT), protein-bound CT and fibre-bound CT were expressed as mg/g Acacia nilotica CT tannin equivalent bound fraction that constituted 40.0-51.4% of the total CT in H. abyssinica and P. thorningii, respectively.Soluble CT fraction constituted 19.5-33.1% of the total CT in P. thorningii and F. villosa, respectively.The browse species had relatively high fibre-bound CT fraction that constituted 22.7-29.1% of the total CT in D. cinerea and P. thorningii, respectively.

Effect of polyethylene glycol treatment on in vitro gas production
Effect of incubation of browse leaf samples with polyethylene glycol (PEG) on in vitro gas production (GP) is shown in Figure 1.Incubation of the fodder species in vitro with PEG resulted in increase (P<0.05) in gas production with variable responses among fodder species.Dichrostachys cinerea showed the highest response on the increase in gas production of 132.3 ml/g OM (69.4%) to PEG treatment at 16 h incubation interval compared to H. abyssinica, which showed the lowest increase of 44.4 ml/g OM (14.5%).The increase in gas production due to PEG treatment at 24 h incubation interval varied from 34.9 ml/g OM (10.3%) to 132.2 ml/g OM (57.2 %) in H. abyssinica and D. cinerea, respectively.The response on the increase in gas production from D. cinerea due to PEG supplementation was slightly higher than P. thorningii.Flagea villosa showed the relatively lower responses than D. cinerea and P. thorningii.

DISCUSSION
The current findings on CP were higher than the previously reported CP content of standing hay forages (30 g/kg DM) (Rubanza, 1999) in semi-arid western Tanzania in the dry season.Following high CP contents, browse leaves could be used as ruminants' feed supplements for N in the tropics especially during dry seasons.Variations in chemical compositions between fodder species could be partly explained by genotypic differences in factors that control synthesis and accumulation of nutrients in plant tissues; and partly by forages' stage of growth and plant proportion that was harvested for analysis.Minson (1990) reported decrease in cell contents (CP, sugars, vitamins, neutral detergent soluble contents) and digestibility with forage maturity at the expense of increased fibre content and lignification.Similarly, Topps (1997) reported a decrease in CP and increase in fibre fractions in Acacia spp.due to advanced leaf maturity.Despite browse fodder nutritive superiority, high levels of phenolic and tannin ANFs could limit utilization of N from browse legume foliages.
Adverse effects due to high levels of tannins on digestibility could be through inhibition of digestive enzymes and complex formation with dietary nutrients and toxic effects on rumen microorganism.Variation of phenolics and tannin concentration among browse fodder species could be due to factors which control synthesis and accumulation of polyphenolic compounds in plants, mainly plant genotype, environmental factors, genotype-environment interaction (Wong, 1973) and foliages' stage of growth.
The reported total condensed tannins (CT) contents were comparable to CT contents reported earlier in Acacia spp.and Dichrostachys spp.leave and pods.Rubanza et al. (2003) reported CT content (butanol/HCl) of 46, 51, 55 and 74 mg/g for A. polyacantha, A. tortilis, A. nilotica and Dichrostachys spp.leaves, respectively.In the current work, F. villosa had higher CT content than Acacia spp.leaves.The CT content in D. cinerea was comparable to Dichrostachys spp. reported previously (74 vs 75 mg/g DM) (Rubanza et al., 2003).Comparison of the presented results on the total condensed tannins and CT fractions is limited by lack of information from the literature on CT composition in tropical browse foliages.However, results presented in the current work on total CT contents were higher than values of CT fractions reported by Terrill et al. (1992) in grasses and herbaceous legumes, and even higher than those reported by Jackson et al. (1996) and Mupangwa et al. (2000).Mupangwa et al. (2000) reported lower total CT of 29.5, 16.9 and 12.4 mg/g DM in Cassia rotundifolia, Lablab purpureum and Macroptilium atropurpureum, respectively, compared to the current findings of 53, 54, 75 and 98 mg/g DM in F. villosa, H. abyssinica, D. cinerea and P. thorningii, respectively.High levels of CT in browse species' leaves in this study could cause depressed N and dry DM digestibility (Aerts et al., 1999) through formation of tannin-protein complexes (Mangan, 1988).
The four browse species' leaves in the current study were lower (19.5-33.1% of total CT) in the soluble fraction than those reported by Jackson et al. (1996) for tropical legumes Arachis pintoi, Centrosema latidens and Desmodium ovalifolium (70-95% of total CT).Higher contents of CT fractions of browse species' leaves in this study than values reported by Terrill et al. (1992) in herbage forages were due to the fact that the latter forages contained negligible CT contents.Reported results on the distribution of CT fractions compare slightly with values of 12-44% of total CT (butanol/HCl) for soluble CT fraction in C. rotundifolia, L. purpureum, and M. atropurpureum (Mupangwa et al., 2000).High proportion of protein-bound CT fraction (40.0-51.4%)could be associated with depressed browse fodder digestibility.The fibre-bound CT fraction constituted a significant proportion (22.7-29.1%) of the total CT.High proportion of bound CT reduce rumen fibre digestion possibly due to a reduction caused by inactivation of hemicellulase and cellulase enzymes or inhibition of the attachment of the rumen microbes (Muhammed et al., 1994).The latter author reported inhibitory effect of tannins on microbial endoglucocanase activity and the attachment of fungi to cellulose in vitro, which result to reduced degradation of cellulose.Min et al. (2003) further noted reduced rumen proteolytic bacterial growth, and some specific bacterial populations measured in vivo due to CT anti-nutritive activity.
Variation in flavonoid classes between studied browse fodder species could be due to differences in genetic and biochemical processes that control synthesis and accumulation of flavonoids in plant tissues, environmental factors (light intensity, mineral deficiency and moisture stress) and interactions between the environment and genetic factors (Wong, 1973;Haslam, 1998;Aerts et al., 1999).Some important genetically controlled biochemical processes include overall flavonoid production, specific flavonoid synthesis and distribution of flavonoids in different plant tissues (Wong, 1973).Variable levels of pelargonidin, delphinidin and cyanidin flavonoids composition between foliage species could be associated to tannin reactivity and thus biological activity due to the chemical structure or stereochemistry of the polymerized proanthocyanidins.The detected proanthocyanidins in D. cinerea, F. villosa, H. abyssinica and P. thorningii browse fodder species could have different tannin anti-nutritive activity and thus variable biological activity due to their variable flavonoid structure (Haslam, 1998).

Effect of tannin on in vitro digestibility
Increase in gas production due to in vitro supplementation with PEG in browse leaves in this study shows the negative role of tannin on lowered fermentation and digestibility of the feed organic matter (OM).Therefore, condensed tannins in D. cinerea, F. villosa, H. abyssinica and P. thorningii leaves could limit digestibility of browse fodder and thus lower their nutritive potential.Variable responses on the increase in in vitro gas production shown by tannin-PEG bioassay among browse species in this study could be associated with depressed feed OM digestibility potential probably due to content of tannin and variable tannin anti-nutritive activity among browse leaves (Aerts et al., 1999).Tannins bind and complex feed proteins and carbohydrates (Mangan, 1988;Makkar, 1989), which result to lowered feed digestibility (Makkar et al., 1995).
Improved gas production and probably OM digestibility in tannin-rich browse could be due to high affinity of PEG to tannins (Makkar and Becker, 1996), and due to its ability to bind and inactivate tannins by forming tannin-PEG complexes (Makkar et al., 1995) and the release of feed nutrients, for example, protein, from tannin-protein complexes (Getachew et al., 2000).Improved in vitro gas production in the current study concur to findings reported elsewhere (Makkar et al., 1995;Makkar and Becker, 1996;Rubanza et al., 2003) in different browse fodder species.Makkar et al. (1995) and Makkar and Becker (1996) reported improved gas production due to PEG treatment in Acacia barteri, A. saligna, Dichrostachys cinerea leaves incubated with PEG.
Higher increase in gas production due fermentation of D. cinerea with PEG than other species suggests different tannin anti-nutritive activity among browse species' leaves (Aerts et al., 1999).Results on variable improved gas production between browse species' leaves due to addition of PEG were consistent to previous findings reported earlier (Rubanza et al., 2003) in Acacia nilotica, A. polyacantha, A. tortilis and Dichrostachys spp.leaves and pods.Low digestibility of both D. cinerea and P. thorningii could also be attributed to their high proportions of free CT, protein-bound CT and fibre-bound CT fractions due to their effects on depressed protein and fibre digestibility (Muhammed et al., 1994), and probably due to their toxic effects on rumen microbes (Odenyo et al., 1997).Lower in vitro gas production potential and probably OM digestibility of D. cinerea compared to F. villosa, H. abyssinica and P. thorningii leaves could also be associated with its relatively higher delphinidin/cyanidin ratio than in F. villosa, H. abyssinica and P. thorningii (Table 3).
Variable in vitro gas production potential and probably their digestibility between these browse species could also be partly explained by type of fibre and extent of fibre lignifications (Fonseca et al., 1998).Low gas production and probably low digestibility of D. cinerea and P. thorningii leaves could also be partly accounted for by their relatively higher lignin contents compared to F. villosa and H. abyssinica, which contained relatively lower lignin contents.

CONCLUSIONS
It was concluded that browse foliage leaves harvested from D. cinerea, F. villosa, H. abyssinica and P. thorningii could be used as N supplements to ruminants fed on low-quality roughages due to their high crude protein content.However, utilization of these browse fodder supplements to ruminants could be limited by their high contents of phenolics and tannins.Browse species' leaves in this study contained higher contents of phenolic and tannin ANFs than the lower beneficial level of tannin (50 mg/g DM) in ruminant diets.High proportion of total condensed tannins (CT) were bound to protein compared to soluble CT and fibrebound CT fractions.The species' leaves contained detectable proanthocyanidins that constituted flavan-3-ols and flavan-3,4-diols that elucidate variable flavonoids' stereochemistry.Improved in vitro gas production due to fermentation of the browse species' leaves with PEG were mainly due to the deactivation and reversed tannin anti-nutritive activity by PEG binding tannins; and the release of feed nutrients bound by tannins.
Further studies are needed to investigate effect of PEG and PEG analogies supplementation on effects of tannins in vivo in ruminants fed on tannin-rich browse foliages, and effect of browse foliages supplementation on animal productivity.

Figure 1 .
Figure1.Effect of polyethylene glycol (PEG) treatment on cumulative in vitro gas production (ml/g OM) from selected browse tree leaves at different incubation intervals, h

TABLE 1
,b,c,d different letters in the same column indicate significant differences (P<0.05)*** P<0.001 RUBANZA C.D.K. ET AL. a