Yeast β-glucan-mannan combinations in broiler diets: roles of processing and β-glucan solubility in growth...

Yeast-derived β-glucan- and mannan-rich materials are used as functional feed additives, but their effects may depend on processing and composition. This study evaluated crude, fractionated, and recombined yeast polysaccharide preparations in 1008 one-day-old male Ross 308 broilers assigned to eight dietary treatments for 38 d (seven pens per treatment; 18 birds per pen), with the pen as the experimental unit. Growth performance, organ weights, intestinal morphology, caecal metabolites, and microbial composition were assessed. No supplemented treatment differed from the non-supplemented control in growth performance. Compared with the crude cell wall preparation, the insoluble-type β-glucan-mannan preparation with a relatively low β-glucan level had a higher caecal propionate concentration (11.78 vs. 7.23 μmol/g; P = 0.005) and a higher jejunal villus height-to-crypt depth ratio (P < 0.05). However, it also increased finisher and overall feed conversion ratios and decreased relative full caecum weight (all P < 0.05). This preparation was also characterised by Faecalibacterium, the Ruminococcus torques group, and Erysipelatoclostridium. The corresponding soluble-type preparation did not differ from the crude preparation for these outcomes but was characterised by Clostridium sensu stricto 1. In addition, soluble-type preparations had higher caecal butyrate concentrations than insoluble-type preparations (P = 0.002), whereas the higher β-glucan level increased relative full and empty caecum weights (both P < 0.05). Preparation-specific microbial patterns were also evident at the higher level, with Marvinbryantia characterising the soluble-type preparation and Romboutsia and Turicibacter characterising the insoluble-type preparation. Combining soluble- and insoluble-type fractions produced non-additive responses in relative ileum weight and caecal metabolites, without yielding higher metabolite concentrations than the corresponding single-fraction preparations. Overall, responses depended on processing-related preparation characteristics, β-glucan level, and component balance, with clearer effects on caecal fermentation and microbial composition than on growth performance.