Can diet change the diversity of ruminal bacterial species in beef cattle?

Background The objective of this study was to assess the effects of diet on bacterial species in the solid fraction of the ruminal content using the gene sequences of the conserved 16S rDNA region steers fed one of the following diets: canola (C), cottonseed (A), sunower (G), soybean (SO), corn silage (S) and control diet (PD). Canola, cottonseed, sunower and soybean were fed as whole seeds. Six crossbred steers (Body weight = 416.33 ± 93.30 kg; mean ± SD), castrated male, and tted with ruminal cannula were used. The experimental design was a 6 × 6 Latin square design. Results Cellulolytic bacteria were predominant for all diets, with 47.75% of Operational Taxonomic Units (OTU) in animals fed the cottonseed diet. Amylolytic bacteria were identied for all diets, representing 62.51% OTU in animals consuming the sunower diet. Proteolytic bacteria were identied for all diets, corresponding to 65.96% OUT in animals fed the sunower diet. Lactic bacteria were identied for all diets. Megasphaera elsdenii bacterium was identied for all diets, with a greater diversity of this bacterium in steers fed the control diet. This bacterium may reduce the availability of hydrogen in the rumen due to propionate production and lactate utilization. Conclusion Oilseed in the diet showed a similarity of bacteria species with 47.5% of changing of the ruminal ora.


Background
Thousands of microorganisms belonging to the three domains [1] colonize the rumen environment: Bacteria, Archaea, and Eucarya (fungi and protozoa). Bacteria are highly diverse and abundant in the rumen representing approximately 95% of the total microbiota [2].
The main species of structural carbohydrate (i.e., cellulose and hemicellulose) bacteria, are Ruminococcus albus, Ruminococcus avefaciens and Fibrobacter succinogenes and the fermentation end-products are propionate, butyrate, succinate, formate, carbon dioxide, water, and especially acetate [3]. As Ruminococcus bacteria are hydrogen producers (via acetate production), their growth and consequently bre degradation can be inhibited by hydrogen accumulation [4,5].
The amylolytic bacteria Treponema bryantii and Treponema saccharophilum mainly ferment non-structural carbohydrates [6] uing malt oligosaccharides [7]. Ruminobacter amylophilus is predominant in the rumen of starch-fed cattle and the range of substrate used by this bacterium is limited to starch, maltose and maltodextrins [8]. Ruminobacter amylophilus has also a signi cant proteolytic activity [9].
The proteolytic bacteria Prevotella bryantii and Prevotella brevis are capable of using starch and polysaccharides such as xylenes and plant cell wall pectin. However, they are not able to degrade cellulose [7,10] and they have proteolytic activity. Lactic bacteria species are represented mainly by Lactobacillus. They are homofermentative bacteria and include Lactobacillus amylovorus, Lactobacillus fermentum, and Lactobacillus pentosus that play an important role as initiators of the ruminal fermentation [11,12].
Eubacterium pyruvativorans is a rapidly growing bacterium that uses pyruvate [13] although pyruvate does not have is not present in high concentration in ruminal uid [14]. Megasphaera elsdenii is considered as one of the main lactate fermenting bacteria converting lactate into propionate and. In the absence of lactate, M. elsdenii produces acetate and butyrate but does not ferment glucose into propionate [15,16].
The higher abundance of bacterial populations involved in propionate production is associated with reduced methane emissions compared with acetate production because more hydrogen is used for propionate production, thus reducing the availability for methane production [4].
The use of oilseeds in ruminant nutrition is an alternative to improve feed e ciency, reducing the methane energy losses the use of hydrogen for biohydrogenation of unsaturated fatty acids. The metagenomic approach is a tool allowing the evaluation of bacterial diversity by studying the total microbial DNA extracted directly from the environment [17]. Using DNA sequencing technology, the ruminal microbiota can be quickly investigated [18]. In order to identify and quantify non-cultured microorganisms, the characterization of the population of prokaryote microorganisms is carried out through the sequencing of the 16S rDNA gene [19].
The objective of the present work was to use the metagenomic approach (i.e., gene sequences of the conserved region 16S rDNA) to determine the bacterial diversity in the solid fraction of ruminal content in steers fed corn silage and different oilseeds (canola, cottonseed, sun ower, soybean) rich with polyunsaturated fatty acids (i.e., omega-3 and omega-6 fatty acids).

Animals, experimental design
The experiment was carried out at the Dom Bosco Catholic University and at the Faculty of Veterinary Medicine and Animal Science of the Federal University of Mato Grosso do Sul, in Campo Grande, Mato Grosso do Sul, Brazil.
Six ruminally-cannulated male crossbred steers with and mean weight of 416 ± 93.3 kg at the beginning of the experiment were use in 6 × 6 Latin square design. Each experimental period consisted of 14 days of which 13 days for adaptation to the experimental diets and 1 day for ruminal sampling. The animals were vaccinated, dewormed and allocated in individual stalls (3 × 6m, 18 m 2 ) with cover and having free access to water. Animals were fed (8:00 A.M.) ad libitum (5% refusals on as-fed basis) one of the 6 experimental diets: Canola (C); cottonseed (A); sun ower (G); soybean (SO); control diet (PD) and corn silage (CS).
The oilseeds grains were included in the diets to achieve an ether extract (EE) concentration of 80 g/kg DM compared with an EE concentration of 50 and 30 g/kg DM, for control diet (PD) and corn silage exclusive diet, respectively.
The composition (chemical and composition) of the experimental treatments is shown in Table 1. The diets were formulated to meet nutrient requirements for an average daily gain of 1.25 kg/day [20].

Collection of ruminal content
Ruminal content was collected via the cannula on day 14 or each experimental period before feeding animals (i.e. 7:00 AM). Ten grams of the solid fraction of ruminal content were mixed with 10 ml of Tris-Borate-EDTA buffer (pH 7.0; Sigma Aldrich; add city and country please), shaken vigorously for 3 minutes (vortex; reference of the equipment) and ltered on 100-micron mesh tissue. The ltrate was centrifuged at 13,0000 RCF for 13 minutes at 4°C. The supernatant was discarded, and the remaining pellet was suspended in 0.8 ml of TE Tris-EDTA buffer (10X, pH 8.0; Sigma Aldrich; add city and country please). The suspended content was centrifuged at 15,000 RCF for 11 min at 4°C, the supernatant was discarded, and the precipitate was stored at -20°C. The total DNA extracted from the ruminal content was quanti ed using Qubit dsDNA HS Assay Kit (Thermo Fisher Scienti c ® ; city, country), and the quality of the genetic material was assessed by 0.8% agarose gel electrophoresis. The extracted DNA samples were stored at -20°C for further ampli cation by the Polymerase Chain Reaction (PCR).

16S rDNA gene Sequencing
The samples were sequenced on a large-scale DNA for GenOne Solutions in Biotechnology. The ampli cation, as well as the sequencing of the V3-V4 (466 bp) regions of the 16S rDNA were performed by the Illumina HiSeq platform (http://www.genone.com.br) (V3 341F CCTAYGGGRBGCASCAG, V4 806R GGACTACNNGGGTATCTAAT). After ampli cation and sequencing a genomic library, data processing, species composition and abundance were obtained. The complexity and complexity difference for each sample, and clustering of species composition for each sample and between samples were made.
Analysis of 16S rDNA gene data Subsequent data analysis was processed using the Quantitative Insights into Microbial Ecology (QIIME) program. The QIIME is a free program based on Phyton scripts that allows the classi cation of 16S rDNA sequences into Operational Taxonomic Units (OTU) and using them as a basis for building phylogenetic trees, plot taxonomic graphs, build interaction networks, alpha and beta diversity, among others [24]. Thus, OTU were de ned by clustering at 97%, using as reference, the most recent OTUs database of Greengenes using the Uclust method [25].
The general structure of the bacterial community of phylum and genus were analysed using relative abundance plot. The alpha diversity analysed by a rarefaction curve and OTU observations. The beta-diversity analysis measured by the UniFrac distance matrix, which was used to demonstrate similarity or dissimilarity among the analyzed samples [26].
Statistical tests on the taxonomic differences between the samples were calculated using the STAMP software using Fisher's exact test with multiple Bonferroni correction (P <0.01) (nominal coverage of 95%) [27].

Results
From the Venn diagram ( Fig. 1), 2,495 Operating Taxonomic Units (OTU) were identi ed of which 1,188 OTUs were shared among all diets, corresponding to a similarity rate of 47.61%. The standard and canola diets shared 86 shared OUT while the standard and sun ower diets shared 42 OTU were. The standard and cottonseed diets shared 69 OTU whereas the canola and cottonseed diets shared 74 OTU. Canola and sun ower diets shared 62 OTU, and 41 OTU were shared between sun ower and cottonseed diets. The bacterial community changed slightly with the experimental diet fed to animals (Fig. 1).
The standard and canola diets exclusively shared 3.45% of the bacterial species; while 1.68% of the bacterial species were shared exclusively between the standard and sun ower diets shared. The control diets and cottonseed exclusively shared 2.76% of the bacterial species whereas the canola and cottonseed diets exclusively shared 2.96% of the bacterial species. Canola and sun ower diets exclusively shared 2.48% of the bacterial species and 1.64% of the bacterial species were exclusively shared between the sun ower and cottonseed diets. Thus, changing the experimental diet slightly affected bacterial species little changed due to the lipid level in the diets change more likely because of the chemical composition between the diets. i.e. lipid source from grains or fatty acid pro le of them.
The Venn diagram (Fig. 2) revealed that there is an overlap of bacteria between the experimental diets. Corn silage and the control diets shared 1481 OTU, suggesting that the bacterial community present in this overlap probably does not change with the type of the diet fed to animals. Corn silage and control diets shared 66.71% of bacteria; suggesting that regardless of the type of diet, change in averaged 33.28% across all diets.
Comparing the control diet to corn silage diet, a total of 2220 OTU were observed, of which 20% and 13.28% were related to the corn silage diet and the control diet, respectively.
One hundred and fty strains of bacteria, corresponding to 10.84% of OTU were identi ed while other bacteria (i.e., 89.16%) were not identi ed by the database. The identi ed bacteria, we classi ed according to the speci city activity (i.e., amylolytic, proteolytic and lactic acid) of each bacteria ( Table 2).
In addition to these bacteria, others with unknown functions were reported so far in the literature. Other bacteria not belonging to the rumen environment were also found and the reason of their presence in the rumen environment is not yet known. Among them: pathogenic bacteria (Propionibacterium granulosum; Ralstonia pickettii; and Serratia marcescens); bacteria of human disease (Rhodococcus hoagie; Aeromonas caviae; Klebsiella pneumonia; Gallibacterium salpingitidis; and Pseudomonas monteilli); bacteria found in water (Paladibaculum fermentans; Pseudoclavibacter caeni; Bacterium enrichment culture clone R4-41B; and Methylobacterium aquaticum); bacteria found in plants (Pseudoxanthomonas suwonensis and Sphingomonas melonis); and bacteria of ruminant disease (Dietzia maris; Micrococcus luteus; Prevotella heparinolytica; and Bibersteinia trehalosi).

Discussion
Diversity indices showed that the ruminal bacterial microbiota is quite rich and diverse, re ecting the reliability of the metagenomic technique associated with 16S rDNA gene sequencing to describe the bacterial community, which would not have been revealed by traditional bacterial culture techniques in microbiology. Using the diversity analysis, it was possible to identify that 47.61% (on average) of changes in bacteria was due to the type of diet fed to animals and this diversity was related to cellulolytic bacteria.
Diet is probably the most important factor in uencing the number and relative proportion of different species of ruminal microorganisms [28]. The change in population is a result of the change in ruminant diet [29], and the change in diet imposes on the animal a transition period in the microbial rumen population, with changes in the proportions between different species to give a new balance and promote a better adaptation to the new diet [30]. The adaptation period used in the present study was 13 days as recommended by the literature [31].
The change in the diversity of rumen microorganisms also varies according to the phase of microbial growth and nutrient availability [32]. This variation also occurs according to the time at which sampling is performed [33]. Variations in the composition of ruminal bacteria can still be attributed to differences between isolation and bacterial composition determination techniques, and signi cant differences in the composition of isolated bacteria can occur in animals fed different diets [34] For the cottonseed diet, 47.75% OTU of the were identi ed as cellulolytic bacteria and was probably due to NDF content of the cottonseed diet (Table 1). For the sun ower diet, 62.51% OTU were identi ed as amylolytic bacteria and this was more likely due to the starch content of the sun ower diet ( Table 2). The bacterium Ruminobacter amylophilus was identi ed with 47.34% of total bacteria, this bacterium has proteolytic and amylolytic activities.
Proteolytic bacteria were identi ed in 65.96% in the sun ower diet and the bacteria Ruminobacter amylophilus was identi ed with 47.34% of total of bacteria in the sun ower diet and these bacteria possess proteolytic and amylolytic activities.
Lactic bacteria were identi ed in 0.92% with the silage diet. This proportion was three times higher than for the other diets, probably due to the lactate levels present in the diet, mainly in corn silage and the corn meal levels used in the experimental diets.
The bacteria Megasphaera elsdenii was identi ed in all of diets being 0.1% in the corn silage diet; 0.06% in control diet; 0.1% in cottonseed diet; 0.3% in canola diet; 0.02% in sun ower diet, and 0.03% in the soybean diet.
These values suggest that the greater diversity of this bacterium in the ruminal uid of animals may reduce the availability of rumen hydrogen due to propionate production and lactate utilization.
Cellulolytic bacteria were identi ed in high percentage in ruminal content of animals fed the cottonseed diet.
Amylolytic and proteolytic bacteria were identi ed in great proportion in animals fed the sun ower diet. Lactic bacteria were identi ed in great proportion in ruminal content fed the corn silage diet.

Consent for publication
The authors declare that they agree for publication.

Competing interests
The authors declare that they have no competing interests. Luis Carlos Vinhas Ítavo 1* , coordinator, conception and design of the work; analysis, and interpretation of data; and substantively revised the paper, read and approved the nal manuscript; Alinne Pereira de Castro 2 , conception and design of the work and revised the paper, read and approved the nal manuscript; Camila Celeste Brandão Ferreira Ítavo 1 , conception and design of the work and revised the paper, read and approved the nal manuscript; Alexandre Menezes Dias 1 , conception and design of the work and revised the paper, read and approved the nal manuscript; Gelson dos Santos Difante 1 , conception and design of the work revised the paper, read and approved the nal manuscript; Geraldo Tadeu dos Santos 1 , conception and design of the work and revised the paper, read and approved the nal manuscript; Marcus Vinicius Garcia Niwa 1 , acquisition, analysis, and interpretation of data, read and approved the nal manuscript; Gabriella Jorgetti de Moraes 1 , acquisition, analysis, and interpretation of data, read and approved the nal manuscript; Alysson Martins Wanderley 1 , acquisition, analysis, and interpretation of data, read and approved the nal manuscript; Antonio Leandro Chaves Gurgel 1 , acquisition, analysis, and interpretation of data, read and approved the nal manuscript; Rodrigo Gonçalves Mateus 2 , revised the paper, read and approved the nal manuscript; Chaouki Benchaar 3, have drafted the work and substantively revised the paper, read and approved the nal manuscript;