Influence of protease inhibitors on nitrogen distribution in ensiled perennial ryegrass and the utilisation of silage nitrogen for growth by rumen bacteria in vitro

Protease inhibitors were added to perennial ryegrass (Lolium perenne; dry matter (DM), 188 g kg ^-1 fresh weight (FW); water soluble carbohydrates, 175 g kg ^-1 DM) at ensilage and their influence on silage nitrogen (N) distribution and the ability of soluble N fractions prepared from these silages to support the growth of Megasphaera elsdenii, Prevotella bryantii and Selenomonas ruminantium in vitro were investigated. Control silages were well fermented (pH 3.9), ammonia N and peptide N concentrations were 0.09 and 0.08 of total N, respectively, and the silage had undergone extensive proteolysis (soluble non-protein N, (NPN) 0.68 of total N). Formic acid treatment (5.4 g kg ^-1 FW) restricted silage fermentation, decreased soluble NPN (0.49 of total N) and ammonia-N concentrations and increased peptide N concentrations to 0.21 of total N. Two metallo-protease inhibitors, 8-hydroxyquinoline (5 g kg ^-1 FW) and 1,10-phenanthroline (2.4 g kg ^-1 FW) restricted silage fermentation and increased peptide-N concentrations to values similar to those obtained with formic acid treatment. Both of these additives decreased ammonia-N concentrations and, in addition, 1,10-phenanthroline decreased soluble NPN concentrations to 0.5 of total N. A third metallo-protease inhibitor, bestatin had no effect. Inhibitors of chymotrypsin-like enzymes had little effect on silage fermentation but markedly different effects on N distribution: N-acetyl-L-tyrosine ethyl ester (20 mg kg ^-1 FW) was without effect, whereas tosyl phenylalanyl chloromethyl ketone (TPCK, 42 mg kg ^-1 FW) decreased the concentrations of soluble NPN (0.59 of total N) and ammonia N and increased peptide-N to 0.22 of total N. The cysteine-protease inhibitor, 1-trans epoxysuccinyl-leucylamido-(4-guanidino) butane (22.5 mg kg ^-1 FW), decreased total soluble NPN (0.45 of total N) and ammonia-N concentrations with no effect on peptide-N concentrations. Silage extracts and N fractions purified from silage extracts were supplied as sources of N and glucose added to growth media for rumen bacteria and specific growth rates were determined turbidimetrically. M. elsdenii grew more rapidly in medium containing silage extracts than in complete medium, whereas P. bryantii and S. ruminantium grew more slowly. None of the protease inhibitors resulted in silage extracts which supported significantly more rapid growth of any of the bacteria compared to the control silage. However, 8-hydroxyquinoline and 1,10-phenanthroline were toxic to the bacteria when silage juices were added to the medium. The results suggest that grass protein is degraded by metallo- and cysteine proteases during ensilage and that silage peptide N concentrations may be manipulated by the use of formic acid, chelators and TPCK. Rumen bacteria grew well on silage extracts and N fractions prepared from silage extracts, but there was no indication that increasing the concentration of silage peptide N beyond 17 μg peptide N ml ^-1 supported higher growth rates.