Milk products are an important component of human diets, with beneficial effects for human health, but also one of the major sources of nutritionally undesirable saturated fatty acids (SFA). Recent discoveries showing the importance of the rumen microbiome on dairy cattle health, metabolism and performance highlight that milk composition, and potentially milk SFA content, may also be associated with microorganisms, their genes and their activities. Understanding these mechanisms can be used for the development of cost-effective strategies for the production of milk with less SFA. This work aimed to compare the rumen microbiome between cows producing milk with contrasting FA profile and identify potentially responsible metabolic-related microbial mechanisms. Forty eight Holstein dairy cows were fed the same total mixed ration under the same housing conditions. Milk and rumen fluid samples were collected from all cows for the analysis of fatty acid profiles (by gas chromatography), the abundances of rumen microbiome communities and genes (by whole-genome-shotgun metagenomics), and rumen metabolome (using 500 MHz nuclear magnetic resonance). The following groups: (i) 24 High-SFA (66.9–74.4% total FA) vs. 24 Low-SFA (60.2–66.6%% total FA) cows, and (ii) 8 extreme High-SFA (69.9–74.4% total FA) vs. 8 extreme Low-SFA (60.2–64.0% total FA) were compared. Rumen of cows producing milk with more SFA were characterized by higher abundances of the lactic acid bacteria Lactobacillus, Leuconostoc, and Weissella, the acetogenic Proteobacteria Acetobacter and Kozakia, Mycobacterium, two fungi (Cutaneotrichosporon and Cyphellophora), and at a lesser extent Methanobrevibacter and the protist Nannochloropsis. Cows carrying genes correlated with milk FA also had higher concentrations of butyrate, propionate and tyrosine and lower concentrations of xanthine and hypoxanthine in the rumen. Abundances of rumen microbial genes were able to explain between 76 and 94% on the variation of the most abundant milk FA. Metagenomics and metabolomics analyses highlighted that cows producing milk with contrasting FA profile under the same diet, also differ in their rumen metabolic activities in relation to adaptation to reduced rumen pH, carbohydrate fermentation, and protein synthesis and metabolism.

An important limitation of continuous-culture fermenters is their inability of maintaining microbial populations similar to those observed in the rumen, especially protozoa numbers, which usually decrease markedly or even disappear. Two approaches (a polyurethane-sponge (SP) and a filter system (FIL) for additionally retaining protozoa were tested in continuous culture system already designed to retain protozoa (Muetzel et al., 2009), and their effects on microbial populations, fermentation parameters and microbial biomass were assessed. Two 14-day incubation runs were carried out with 6 fermenters, and in each run two fermenters were randomly assigned to each of the experimental treatments (control, SP and FIL). Total protozoa numbers assessed by microscopic counting were 1.7 and 2.1 times greater in SP and FIL fermenters than in control ones on day 14, although differences did not reach the significance level (P = 0.855). Protozoal DNA concentration on day 14 were 1.6 and 1.4 times greater in SP and FIL fermenters, respectively, than in control ones, but differences were not significant (P = 0.524). Results from protozoal DNA concentrations determined in each fermenter on the different sampling days (n= 48) were positively correlated (P < 0.001) with the numbers of total (r = 0.826), entodiniomorphid (r =0.824) and holotrich (r = 0.675) protozoa determined by microscopic counting, indicating that both methods are valid to assess protozoa populations. The proportion of holotrich in FILfermenters was relatively constant over the incubation period (ranging from 10.5 to 13.3% of total protozoa), but decreased with time in control (from 10.9% at day 2 to 6.7% at day 14) and SP (from 10.9 to 6.9%) fermenters. Neither the bacterial DNA concentration nor the relative abundance of fungal and archaeal DNA were influenced by any of the modifications tested, but values changed over the sampling period (days 10-14). Bacterial DNA concentration increased (P < 0.001) from day 10 to 14 in all fermenters, whereas the relative abundance of fungal and archaeal DNA decreased (P < 0.001). The tested modifications did not affect (P > 0.05) fermentation parameters, which reached a steady-state after 6 days of incubation. Values of microbial biomass determined using purine bases as a microbial marker were significantly correlated with the amount of bacterial plus protozoal DNA (r = 0.794; P = 0.002; n =12) in each Abbreviations: ADFom, acid detergent fiber expressed exclusive of residual ash; aNDFom, neutral detergent fiber with heat-stable amylase and expressed exclusive of residual ash; CON, control treatment (fermenters without modification); CP, crude protein; EE, ether extract; FIL, fermenters provided with a filter system; PCA, principal component analysis; PB, purine bases; SP, fermenters provided with a polyurethane-sponge; VFA, volatile fatty acids fermenter. In conclusion, the two tested modifications increased the protozoa numbers in continuous-culture fermenters, and the FIL allowed maintaining a proportion of holotrich pr...