ABSTRACT
Effective degradation (ED) of crude protein (CP) was estimated in vitro at 0.02, 0.05 and 0.08 h−1 assumed ruminal passage rates for a total of 40 feedstuffs, for which in situ ED was available and used as reference degradation values. For this, the Streptomyces griseus protease test was used. The differences between in vitro CP degradation and the in situ CP degradation values were lowest in legume grains and highest in cereal by-products and barley. The differences between in situ and in vitro ED were expressed using a degradation quotient (degQ), where degQ = (EDin vitro − EDin situ)/EDin situ. Among the tested feedstuffs, eight specific clusters were identified according to degQ for the assumed passage rates. The feedstuffs clustered in an unspecific way, i.e., feedstuffs of different nutrient composition, origin or treatment did not necessarily group together. Formaldehyde−treated rapeseed meal, soybean meal, wheat, a treated lupin, sunflower meal and barley could not be assigned to any of the clusters. Groupwise degradation (range of degQ for assumed passage rates are given in brackets) was detected in grass silages (−0.17, −0.11), cereal by-products together with sugar beet pulp (−0.47, −0.35) and partly in legume grains (−0.14, 0.14). The clustering probably based on different specific nutrient composition and matrix effects that influence the solubility of feed protein and limit the performance of the protease. The matrix can be affected by treatment (chemically, thermally or mechanically), changing the chemical and physical structure of the protein within the plant. The S. griseus protease test had reliable sensitivity to reflect differences between native feedstuffs and treatments (thermally or chemically) that were found in situ. The in situ results, however, are mostly underestimated. The clustering results do not allow a clear conclusion on the groupwise or feed-specific use of carbohydrate-degrading enzymes as pre- or co-inoculants as part of the S. griseus protease test and need to be tested for its potential to make this test more conform with in situ data.
ABSTRACT
The objective of this study was to estimate the amount of post-rumen crude protein (prCP), a precursor to metabolisable protein for ruminants, of feeds and feed mixtures using an in situ and in vitro method, and to contrast the results of both approaches. For this, 34 samples were examined: 9 feeds, 13 feed mixtures made thereof and 12 feed mixtures provided by commercial dairy farms. As a minor aspect, additivity of protein values from feeds was evaluated by comparing measured and calculated values of feed mixtures. Effective prCP was calculated for assumed rumen passage rates (k) of 0.05 h-1 (prCP5) and 0.08 h-1 (prCP8) by in situ measurements of rumen undegraded CP and fermented organic matter (FOM) assuming an efficiency of 181 g microbial CP/kg FOM. Additionally, effective prCP in vitro was estimated using the modified Hohenheim gas test (modHGT) through incubation in rumen-fluid buffer solution for 8 and 24 h followed by ammonia distillation. In vitro estimations were highly correlated with in situ values for both passage rates (k = 0.05 h-1: p < 0.001, R2 = 0.68; k = 0.08 h-1: p < 0.001, R2 = 0.76). The in vitro method yielded higher values for effective prCP than the in situ approach with 29 g/kg OM (k = 0.05 h-1) and 37 g/kg OM (k = 0.08 h-1) on average for all samples. Small positive associative effects - reflecting non-additivity - were found, averaging at 2.2 g prCP8/kg OM (p < 0.05) for the in situ and 10.7 g prCP8/kg OM (p < 0.001) for the in vitro approach. Due to the need of an assumption of a certain value for microbial efficiency in situ, effective prCP might be more accurately estimated in vitro, accounting for nutrient-specific efficiencies as well as interactions between carbohydrate and protein degradation by rumen microbes. Furthermore, the modHGT highlighted associative effects more pronounced and seems suitable as a routine method due to the comparably low effort and high sample throughput. The potential of the modHGT to determine the protein value of feeds could be demonstrated by our study.
