Sainfoin (Onobrychis viciifolia): a forgotten crop for dairy cows with future potential SUMMARY The world population growth and rising incomes are expected to increase the consumption of animal-derived foods such as meat, eggs and milk. However, livestock production should not only be directed towards increasing productivity but should also incorporate environmental, food safety and animal welfare aspects. Therefore, farm businesses have to respond to the high environment impact of their activities, by using low-input systems including the use of forage legumes. Recent studies have demonstrated that forage legumes with moderate levels of condensed tannins (CT) are beneficial for animal nutrition and animal health. Sainfoin (Onobrychis viciifolia Scop.) is a tanniniferous forage legume containing CT that has potential nutritional and health benefits, i.e. preventing bloating, reducing nematode larval establishment, improving nitrogen (N) utilization and reducing greenhouse gas and N emissions (Chapter 1). However, the use of sainfoin as a fodder crop in dairy cow rations in northwestern Europe is still rather unknown. This thesis investigated the potential of sainfoin in the dairy cow diets and the effect of CT structural properties on rumen fermentation and biohydrogenation (BH). Chapter 2 reports a study where the effect of sainfoin silage on nutrient digestibility, animal performance, energy and N utilization and methane (CH4) production in dairy cows was investigated. Six rumen cannulated, lactating dairy cows were randomly assigned to either a control (CON) or sainfoin based (SAIN) diet. The CON diet was a mixture of grass silage, corn silage, concentrate and linseed. In the SAIN diet, 50% of the grass silage DM in the CON diet was exchanged by sainfoin silage. Total daily dry matter (DM), organic matter (OM) and neutral detergent fiber (NDF) intake did not differ between the two diets. The apparent digestibility of DM, OM, NDF and acid detergent fiber (ADF) were respectively, 5.7, 4.0, 15.7 and 14.8% lower for the SAIN diet. Methane production per kg DM intake was lowest for the SAIN diet and CH4 production as a percentage of gross energy intakes tended to be lower while milk yield was greater for the SAIN diet. Nitrogen intake, N retention and energy retained in body protein were greater for the SAIN than the CON diet. Nitrogen retention as a percentage of N intake tended to be greater for the SAIN diet. These results showed that inclusion of sainfoin silage at the expense of grass silage in dairy cow rations reduced CH4 per kg DM intake. Although nutrient digestibility was decreased, sainfoin silage improved milk production and redirected metabolism towards body protein accretion at the expense of body fat. In Chapter 3, reticular fatty acid (FA) flow and ruminal BH of C18:3n-3 is reported using the reticular sampling technique (Cr-EDTA and Yb-acetate as digesta flow markers) in the lactating cows fed the SAIN and CON diet in Chapter 2. The reticular flows of DM, OM and N were not affected by dietary treatment. However, NDF flow was higher (1.87 vs. 1.40 kg/d) where the cows were fed the SAIN diet. A higher mono-unsaturated FA flow was caused by the higher trans-9-C18:1 and cis-9-C18:1 flow for the SAIN compared to the CON fed cows. The flows of trans-9,trans-12-C18:2 and cis-12,trans-10 C18:2 were higher in the SAIN diet fed cows, but total poly-unsaturated FA flow was not affected by the different diet treatments. The SAIN diet fed cows had a significant lower ruminal BH of cis-9-C18:1 and C18:3n-3, compared to the CON fed cows and tended to a lower ruminal BH in case of cis-9,cis-12-C18:2. These results show that inclusion of sainfoin silage at the expense of grass silage in dairy cow rations reduces ruminal BH of dietary cis-9-C18:1 and C18:3n-3. The effects of replacing grass silage by sainfoin silage in a TMR on milk production and FA in milk fat of the dairy cows in Chapter 2 is reported in Chapter 4. Milk yield reported in Chapter 4 was highest for the SAIN diet with every kg of OM digested of the SAIN diet resulting, on average, in 0.2 kg more milk production. The SAIN diet fed cows had a higher C18:3n-3 and cis-9,cis-12-C18:2 proportion in milk fat compared to the CON diet fed cows. A higher proportion of total trans-C18:1 was found in the cows fed the SAIN diet. There were no differences in proportion of total saturated and unsaturated FA in milk fat between the two diets. Our results showed that replacing grass silage by sainfoin silage improved milk yield and milk FA profile of dairy cows. Effects of the structural properties of CT, i.e. average polymer size (or mean degree of polymerization, mDP); percentage of cis flavan-3-ols (%cis) and percentage of prodelphinidins (%PD) in CT extracts on CH4 production and fermentation characteristics of rumen fluid using an in vitro gas production technique was investigated in Chapter 5. Extracts of CT from eight plants; black currant leaves, goat willow leaves, goat willow twigs, pine bark, red currant leaves, sainfoin plants, weeping willow catkins and white clover flowers were extracted, in order to obtain CT with a wide range in mDP, %PD and %cis. All CT extracts reduced CH4 concentration, decreased the maximum rate of fermentation for CH4 production and rate of substrate degradation. The correlation between CT structure on the one hand and CH4 production and fermentation characteristics on the other hand showed that the %PD within CT had the largest effect on fermentation characteristics, followed by mDP and %cis. Chapter 6 reports results of an in vitro study to investigate the effects of the structural properties CT (mDP, %cis and %PD) on rumen fermentation and BH end-products. The total volatile FA (VFA), ammonia concentration and the proportion of branched chain VFA was reduced in all CT extracts, compared to the control. The proportion of cis-9-C18:1; cis-9,cis-12-C18:2; cis-9,cis-12,cis-15-C18:3 were numerically higher in all CT sources, while the proportion of C18:0 and fractional rate of BH of C18:3n-3 were numerically lower in all CT sources, compared to the control. The correlation between CT structural properties on the one hand and fermentation and BH end-products on the other hand showed that the CT with a high %PD and smaller mDP had the largest effect on fermentation end-products. However, mDP was found to be the most important factor affecting rumen BH. Chapter 7 provides a general synthesis on the major findings of the studies presented in the preceding chapters. In addition, results are reported of a further in vitro as well as an in situ study in which I investigated the mechanisms of CT action in the rumen, in the post-rumen compartments and digestive tract. In the in situ study, fresh sainfoin (Esparcette) was incubated in the rumen and in the abomasum before digested during passage through the digestive tract. For the in vitro study, sainfoin (Ambra) was incubated with rumen fluid buffer for 1, 2, 4, 8, 12, 24 hours. After incubation in situ and in vitro, the incubated material was analyzed for tannin content by butanol-HCl assay. The results showed that the soluble CT dramatically reduced upon introduction in the digestive tract. Additional analyses showed that CT had bound to the fiber and protein (diet and microbes) fractions in the digestive tract. The present work showed that sainfoin silage can be used in dairy cow rations to improve milk production and N utilization and reduce CH4 emissions per kg DM intake. Moreover, sainfoin silage, when replacing part of the grass silage in a TMR of dairy cows, increases ruminal unsaturated FA flow into the reticulum and reduces ruminal BH of dietary cis-9-C18:1 and C18:3n-3. Cows fed sainfoin silage at the expense of grass silage in a TMR increase the proportion of unsaturated FA in milk fat. In terms of condensed tannin structure, mDP and %PD appear to be the most important properties of CT that affect fermentation and BH end-products. Condensed tannins with a mDP ranging from 5 to 10 monomeric units and a %PD > 70.0% seem to have the highest biological activity in the rumen.