The rumen bacterial community structure and phylogenetic diversity of fluid- and solidassociated bacteria (FAB and SAB, respectively) was investigated in Brahman-cross steers fed several diets based on low-quality pangola grass hay and molasses. Diets used in Experiment 1 (as fed basis) consisted of 100% pangola hay (0M), 75:25 hay/molasses (25M), 50:50 hay/molasses (50M), and 25:75 hay/molasses (75M). Diets used in Experiment 2 (as fed basis) consisted of molasses/hay (HM, 50:50), hay/molasses/urea (MU, 50:48:2), hay/molasses/casein (MCAS, 50:45:5), and hay/molasses/cottonseed meal (MCSM, 30:50:20). In both experiments, four rumen-cannulated steers were fed once daily at 0800 h. Steers were allocated to one of the four diets in a 4x4 Latin square design. Rumen fluid samples were taken during the initial period of molasses introduction (Experiment 1). Following adaptation to the diets, rumen fluid and solid phase samples were taken immediately prior to feeding and 8 h after feeding over two consecutive days (Experiments 1 and 2). Rumen FAB and SAB communities were analysed by amplification of the V2-V3 region of the 16S rDNA gene, followed by denaturing gradient gel electrophoresis (DGGE). Bacterial diversity was estimated by applying statistical ecology indices to DGGE data. Intake, digestibility, microbial protein production, fluid and solid fractional outflow rate, pH, NH3N and VFA content in the rumen were also measured.
Monitoring changes in FAB structure during the period of gradual introduction of molasses to the diets indicated that bacteria responded slowly to a dietary change. Soon after the start of the dietary change (days 1-8), the appearance of transient dominant bands (species) became evident as these bands were not present three weeks after the diet change. Comparison of morning and afternoon DGGE bacterial profiles from steers adapted to the diets fed in Experiments 1 and 2 (after 27 and 20 days on the new diet, respectively) revealed that the structure of the rumen bacterial community was particularly stable over at least two days (days 27 and 28, for Experiment 1, and days 20 and 21 Experiment 2). These results indicate that the sampling time had no influence on the assessment of the bacterial community structure and diversity once the animal had adapted to a diet. The inclusion of different amounts of molasses in the diet resulted in significantly different FAB and SAB population structures, in terms of DGG banding patterns (i.e., differences in position and intensity of bands DGGE in the profile).
The main finding in terms of estimated species diversity in samples of rumen contents from Experiment 1 was that steers fed hay-only (0M) or hay plus low levels of molasses (25M) displayed a more diverse FAB population when compared to steers consuming high levels of molasses (50M and 75M). Therefore, it appears that replacing slowly degradable low-quality fibre for readily available non-structural carbohydrate resulted in decreased FAB species diversity, as determined by diversity indexes calculated from the number and intensity of DGGE bands in each profile. FAB DGGE patterns were significantly different to SAB patterns for the 0M diet but not for the 75M diet.
Feeding a diet of pangola hay and molasses (50:50) with or without different nitrogen sources (Experiment 2) had no effect on FAB structure; however, differences in SAB structure were detected between the HM and MCAS. The response of the SAB population to casein addition, a highly degradable true protein source may be associated with a greater supply of branched-chain fatty acids that became available after degradation of branched-chain amino acids. Diversity indices for FAB and SAB samples collected in Experiment 2 showed no diet or steer effect. No differences were found in the structure of FAB and SAB between steers consuming the same diet. Phylogenetic diversity in samples from Experiment 1 was assessed by cloning, sequencing and phylogenetic analysis of excised DGGE bands. Between one and four bands were excised per profile. Dominant FAB and unique SAB bands were reamplified and cloned. A total of 269 partial 16S rDNA sequences were retrieved from FAB and SAB DGGE clone libraries.
Phylogenetic relationships between unique sequences and their closest culturable rumen bacteria were estimated. The majority of the phylotypes (47-67%) belonged to the Firmicutes phylum and this result was independent of the diet the steer was fed. Phylotypes belonging to Bacteroidetes represented between 33-44% of the total phylotypes. The following phyla were also represented: Fibrobacteria (6%; detected in 25M only), and the proposed OP11 phylogenetic division (3% in both 0M and 25M diets). The majority of the phylotypes identified in this study were unknown species. Some of the phylotypes identified, although unknown, had been reported to be present in the gastrointestinal system of native African herbivores, Norwegian reindeer, and cattle and sheep consuming diverse diets. Of the 90 phylotypes identified, four had sequences identical to those of known culturable bacteria. These were: Butyrivibrio fibrisolvens (detected in 0M, 25M and 75M diets), Clostridium polysaccharolyticum (25M and 75M diets), Ruminococcus albus (0M diet), and Fibrobacter succinogenes (25M diet). Phylogenetic analysis demonstrated that dominant bacterial populations in the rumen of steers fed hay and hay plus a small amount of molasses were different from those recovered from the rumen of steers fed high levels of molasses. In particular, phylotypes related to unclassified clostridia appeared to be present only when the proportion of hay in the diets was high; whereas Selenomonas-Quinella type bacteria were more prevalent in the rumen of steers fed high-molasses diets.
The presence of Quinella ovalis-like bacteria as dominant organisms in high-molasses diets was confirmed by real time PCR enumeration and microscopy. Q. ovalis-like bacteria tended to increase as the percentage of molasses in the diet increased while the total rumen bacterial populations remained relatively stable. Q. ovalis-like bacteria comprised between 1 and 61% of the total bacteria population (20% average) in 75M. DM intake and digestible organic matter increased quadratically while DM digestibility of the diet, MPS production and the efficiency of MPS increased linearly with increasing level of molasses in the diet. Molasses inclusion in the diet had no effect on rumen pH, ammonia and VFA concentration in the rumen fluid, urine pH or fluid and solid outflow rate. Significant correlations were found between rumen fluid pH, total and individual volatile fatty acids content, fractional outflow rate of the fluid and solid phases, microbial protein synthesis, Isotricha population and changes in abundance of certain rumen FAB phylotypes. Thus the populations of Prevotella ruminicola-like bacteria were negatively correlated to MPS. It is possible that these bacteria, which were abundant in the rumen of low quality hay-fed steers, had low growth rates. The positive correlation between the populations of Quinella ovalis-like bacteria (dominant in 50M and 75M diets) and the proportion of propionate in the rumen fluid suggested that these organisms were associated with propionate production. The populations of Ruminococcus albus and Fibrobacter succinogenes were linked to the abundance of the ciliate Isotricha.
The results from the present study have shown that the level of molasses influenced the structure of rumen bacterial populations. Phylogenetic shifts in dominant bacteria in contrasting diets were also identified. An increase in molasses resulted in significant changes in nutritional parameters, in particular, an increase in the supply of microbial protein, which the phylogenetic study has shown is comprised of different bacterial species and proportions thereof.