Advertisement

Effect of 2-hydroxy-4-(methylthio) butanoate (HMTBa) supplementation on rumen bacterial populations in dairy cows when exposed to diets with risk for milk fat depression

Open ArchivePublished:December 18, 2019DOI:https://doi.org/10.3168/jds.2019-17389

      ABSTRACT

      Diet-induced milk fat depression (MFD) is a condition marked by a reduction in milk fat yield experimentally achieved by increasing dietary unsaturated fatty acids and fermentable carbohydrates. 2-Hydroxy-4-(methylthio) butanoate (HMTBa) is a methionine analog observed to reduce diet-induced MFD in dairy cows. We hypothesize that the reduction in diet-induced MFD by HMTBa is due to changes in the rumen microbiota. To test this, 22 high-producing cannulated Holstein dairy cows were placed into 2 groups using a randomized block design and assigned to either control or HMTBa supplementation (0.1% of diet dry matter). All cows were then exposed to 3 different diets with a low risk (32% neutral detergent fiber, no added oil; fed d 1 to 7), a moderate risk (29% neutral detergent fiber and 0.75% soybean oil; fed d 8 to 24), or a high risk (29% neutral detergent fiber and 1.5% soybean oil; fed d 25 to 28) for diet-induced MFD. Rumen samples were collected on d 0, 14, 24, and 28, extracted for DNA, PCR-amplified for the V1–V2 region of the 16S rRNA gene, sequenced on an Illumina MiSeq (Illumina, San Diego, CA), and subjected to bacterial diversity analysis using the QIIME pipeline. The α diversity estimates (species richness and Shannon diversity) were decreased in the control group compared with the HMTBa group. Bacterial community composition also differed between control and HMTBa groups based on both weighted UniFrac (relative abundance of commonly detected bacteria) and unweighted UniFrac (presence/absence) distances. Within the HMTBa group, no differences were observed in bacterial community composition between d 0 and d 14, 24, and 28; however, in the control group, d 0 samples were different from d 14, 24, and 28. Certain bacterial genera including Dialister, Megasphaera, Lachnospira, and Sharpea were increased in the control group compared with the HMTBa group. Interestingly, these genera were positively correlated with milk fat trans-10,cis-12 conjugated linoleic acid and trans-10 C18:1, fatty acid isomers associated with biohydrogenation-induced MFD. It can be concluded that diet-induced MFD is accompanied by significant alterations in the rumen bacterial community and that HMTBa supplementation reduces these microbial perturbations.

      Key words

      INTRODUCTION

      Milk fat concentration is a major determinant of the economic and nutritional value of milk and is sensitive to external stimuli, particularly components of a dairy cow's diet. Milk fat concentration below the genetic potential of the cow when feeding diets that alter rumen fermentation is referred to as diet-induced or biohydrogenation (BH)-induced milk fat depression (MFD), a condition that does not influence milk yield or other milk components, but reduces the value of milk (
      • Harvatine K.J.
      Managing milk fat depression.
      ).
      To better understand MFD, an experimental dietary model of lowering NDF and increasing starch and UFA was established to produce diet-induced MFD in dairy cows (
      • Bauman D.E.
      • Baumgard L.
      • Corl B.
      • Griinari D.J.
      Biosynthesis of conjugated linoleic acid in ruminants.
      ;
      • Rico D.E.
      • Harvatine K.J.
      Induction of and recovery from milk fat depression occurs progressively in dairy cows switched between diets that differ in fiber and oil concentration.
      ;
      • Rico D.E.
      • Ying Y.
      • Clarke A.
      • Harvatine K.
      The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
      ). The decrease in milk fat yield during MFD is caused by bioactive fatty acids (FA) produced in the rumen (
      • Bauman D.E.
      • Griinari J.M.
      Nutritional regulation of milk fat synthesis.
      ). In ruminants, dietary PUFA are converted to saturated fats by rumen microbes through BH. During diet-induced MFD, BH pathways are altered in the rumen resulting in the formation of specific trans FA that, upon reaching the mammary gland, reduce milk fat synthesis (
      • Harvatine K.J.
      • Boisclair Y.
      • Bauman D.
      Recent advances in the regulation of milk fat synthesis.
      ). The severity of MFD is dependent upon the rate and extent of BH and the specific intermediates formed, which in turn are influenced by dietary risk factors including the concentrations of PUFA and fermentable carbohydrates in the diet (
      • Jenkins T.C.
      • Wallace R.
      • Moate P.
      • Mosley E.
      Board-invited review: Recent advances in biohydrogenation of unsaturated fatty acids within the rumen microbial ecosystem.
      ;
      • Fuentes M.C.
      • Calsamiglia S.
      • Cardozo P.
      • Vlaeminck B.
      Effect of pH and level of concentrate in the diet on the production of biohydrogenation intermediates in a dual-flow continuous culture.
      ). Although MFD in dairy cows is the direct consequence of rumen microbial activity, the role played by microbes in MFD is not clear.
      2-Hydroxy-4-(methylthio) butanoate (HMTBa), an analog of methionine, is a feed additive that has the potential to modify milk fat yield (
      • Zanton G.I.
      • Bowman G.R.
      • Vázquez-Añón M.
      • Rode L.M.
      Meta-analysis of lactation performance in dairy cows receiving supplemental dietary methionine sources or postruminal infusion of methionine.
      ). 2-Hydroxy-4-(methylthio) butanoate is available in the rumen, where it is rapidly absorbed and converted to methionine, thus having a methionine-sparing effect in ruminants (
      • Lobley G.E.
      • Wester T.
      • Calder A.G.
      • Parker D.
      • Dibner J.
      • Vázquez-Añón M.
      Absorption of 2-hydroxy-4-methylthiobutyrate and conversion to methionine in lambs.
      ). Reports suggest that HMTBa may support fiber degradation and microbial protein synthesis and, ultimately, better fermentation in the rumen (
      • Lee C.
      • Oh J.
      • Hristov A.
      • Harvatine K.
      • Vazquez-Anon M.
      • Zanton G.
      Effect of 2-hydroxy-4-methylthio-butanoic acid on ruminal fermentation, bacterial distribution, digestibility, and performance of lactating dairy cows.
      ). 2-Hydroxy-4-(methylthio) butanoate may interact with cellulolytic bacteria directly and may indirectly influence non-cellulolytic bacteria in the rumen (
      • Martin C.
      • Mirande C.
      • Morgavi D.
      • Forano E.
      • Devillard E.
      • Mosoni P.
      Methionine analogues HMB and HMBi increase the abundance of cellulolytic bacterial representatives in the rumen of cattle with no direct effects on fibre degradation.
      ). Although the production responses of HMTBa have been attributed to alteration of the ruminal environment, knowledge of the effects of HMTBa on ruminal microbiota is sparse.
      Decreasing forage NDF and increasing starch and UFA are dietary factors well described to increase risk of diet-induced MFD as they lead to a switch to altered BH pathways in the rumen, resulting in the formation of bioactive FA that cause MFD.
      • Baldin M.
      • Zanton G.
      • Harvatine K.
      Effect of 2-hydroxy-4-(methylthio) butanoate (HMTBa) on risk of biohydrogenation-induced milk fat depression.
      and M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data) have used these dietary factors to expose dairy cows to low, moderate, and high levels of risk for MFD and have evaluated the effects of HMTBa supplementation on alleviating the condition. These authors found that HMTBa supplementation was effective in maintaining milk fat yield and concentration during moderate- and high-risk diet scenarios. Furthermore, they found that HMTBa supplementation maintained lower concentrations of the trans-10 C18:1 isomer in milk, indicating that the shift to altered BH in the rumen is inhibited by HMTBa supplementation. The current study was performed as an accompaniment to M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data) based on the hypothesis that HMTBa-mediated inhibition of BH-induced MFD is modulated by alterations in the ruminal microbiota. To this end, we analyzed changes in ruminal bacterial populations and their associations with milk FA in dairy cows exposed to dietary induction of MFD with and without HMTBa supplementation.

      MATERIALS AND METHODS

      Experimental Design and Treatments

      The design of the experiment and animal assignment to treatments were previously described by M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data) and were approved by the Pennsylvania State University Institutional Animal Care and Use Committee. Briefly, 22 ruminally cannulated multiparous Holstein cows were used in a randomized block design. All cows were maintained on a diet with low risk for MFD during a 7-d pretrial period. At the end of the pretrial period, cows were paired by milk production and were randomly assigned to 2 treatments: unsupplemented control or HMTBa supplemented group. Details of the experimental diets are presented (Supplemental Table S1; https://doi.org/10.3168/jds.2019-17389). The HMTBa (Alimet; Novus International Inc., St. Charles, MO) was provided at 0.1% of dietary DM in a corn carrier (10% HMTBa) and mixed into the TMR. The control group received an equal amount of the same ground corn carrier in their diets. During the course of the experimental period, animals were fed a diet with low risk of BH-induced MFD from d 1 to 7 (31.7% NDF, 27.2% starch, 2.18% UFA, no added oil), moderate risk from d 8 to 24 (29.1% NDF, 29.3% starch, 3.80% UFA), and high risk from d 25 to 28 (28.7% NDF, 29.6% starch, 4.45% UFA). Dietary UFA were increased using a combination of rapidly available FA from soybean oil and more slowly available FA from roasted soybeans. Diets were fed during pretrial and treatment periods as a TMR once daily at 0700 h at 110% of expected daily intake.

      Sampling and Measurements

      Procedures for feed intake and milk responses were previously described by M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data). Feed intake was measured daily. Feed ingredients were sampled once per week and stored at −20°C, thawed at room temperature, dried at 55°C in a forced-air oven for 72 h, and ground in a Wiley mill through a 1-mm screen (A. H. Thomas, Philadelphia, PA). Feed samples were composited within dietary phase (equal dry weight basis). Feed samples were analyzed for CP, NDF, and ADF by wet chemistry procedures (Cumberland Valley Analytical Services Inc., Maugansville, MD). Briefly, CP was determined according to AOAC International (2000) using method 990.03, ADF was determined according to AOAC International (2000) method 973.18, and ash-free NDF according to
      • Van Soest P.J.
      • Robertson J.
      • Lewis B.
      Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition.
      using heat-stable amylase and sodium sulfite. Starch was determined by an enzymatic method (
      • Karkalas J.
      An improved enzymic method for the determination of native and modified starch.
      ; Hazyme, Centerchem, Norwalk, CT) after samples were gelatinized with sodium hydroxide. Total FA concentration and FA profile of feed samples were determined by GC after direct methylation (
      • Sukhija P.S.
      • Palmquist D.
      Rapid method for determination of total fatty acid content and composition of feedstuffs and feces.
      ), as described by
      • Rico D.E.
      • Harvatine K.J.
      Induction of and recovery from milk fat depression occurs progressively in dairy cows switched between diets that differ in fiber and oil concentration.
      .
      Cows were milked twice daily at 0600 and 1800 h and milk yield was determined by an integrated milk meter (AfiMilk, SAE Afikim, Afikim, Israel). Milk was sampled at both milkings on d 0, 7, 14, 24, and 28, and was composited based on yield at each milking. One milk sample from the composite was stored at 4°C with preservative (Bronolab-WII, Advanced Instruments Inc., Norwood, MA) until analysis for fat and protein content by Fourier transform infrared spectroscopy (Fossomatic 4000 Milko-Scan and 400 Fossomatic, Foss Electric, Hillerød, Denmark; at Dairy One Laboratory, Ithaca, NY). A second milk sample was immediately centrifuged at 3,000 × g for 15 min at 4°C and the fat cake was stored at −20°C before analysis of FA composition, as described by
      • Baldin M.
      • Zanton G.
      • Harvatine K.
      Effect of 2-hydroxy-4-(methylthio) butanoate (HMTBa) on risk of biohydrogenation-induced milk fat depression.
      . Briefly, FA were extracted in hexane isopropanol, base transmethylated with sodium methoxide, and quantified by GC with a fused-silica capillary column (SP-2560, 100 m × 0.25 mm i.d. with 0.2-μm film thickness; Supelco Inc., Bellefonte, PA) and a flame ionization detector.
      Whole rumen digesta was collected via cannula twice daily (approximately 1 h before and 6 h after feeding) on d 0, 7, 14, 24, and 28. Samples were collected from 5 different locations in the rumen (cranial dorsal, cranial ventral, central, caudal dorsal, and caudal ventral), mixed in a bucket, and a composite subsample of approximately 200 g was immediately placed on dry ice and archived at −80°C until analysis of microbial populations was performed.

      Genomic DNA Extraction and PCR Amplification

      The genomic DNA from whole digesta samples was extracted using bead-beating followed by precipitation with isopropanol and ethanol and finally extraction with a commercial kit (QIAamp DNA Stool Mini Kit; Qiagen Sciences, Germantown, MD) based on the procedure described in
      • Yu Z.
      • Morrison M.
      Comparisons of different hypervariable regions of rrs genes for use in fingerprinting of microbial communities by PCR-denaturing gradient gel electrophoresis.
      . For each extracted genomic DNA sample, the V1–V2 region of bacterial 16S rRNA gene was PCR-amplified in triplicate using the bacterial-specific primers F27 (5′-GAGTTTGATCCTGGCTCAG-3′) and R388 (5′-TGCTGCCTCCCGTAGGAGT-3′) barcoded with a unique 12-base error-correcting Golay code for multiplexing as described in
      • Song S.J.
      • Lauber C.
      • Costello E.K.
      • Lozupone C.A.
      • Humphrey G.
      • Berg-Lyons D.
      • Caporaso J.G.
      • Knights D.
      • Clemente J.C.
      • Nakielny S.
      • Gordon J.I.
      • Fierer N.
      • Knight R.
      Cohabiting family members share microbiota with one another and with their dogs.
      . Polymerase chain reaction was performed using the Accuprime Taq DNA Polymerase System (Invitrogen, Carlsbad, CA). The thermal cycling conditions involved an initial denaturing step at 95°C for 5 min followed by 20 cycles (denaturing at 95°C for 30 s, annealing at 56°C for 30 s, extension at 72°C for 90 s) and an extension step at 72°C for 8 min. The amplicons from each DNA sample were combined and then each library was added to a pool in equimolar concentration. The final pool was bead purified using Beckman Coulter Agencourt AMPure XP Beads (Beckman Coulter, Brea, CA). Multiplex DNA sequencing was performed with an Illumina MiSeq (Illumina, San Diego, CA) to obtain 2 × 250 bp paired-end reads. Raw sequencing information was deposited at National Center for Biotechnological Information within the Short Read Archive under the project accession number PRJNA580199.

      Sequencing, Data Analysis, and Statistical Analysis

      The 16S rDNA reads were analyzed using QIIME 1.8.0 pipeline (
      • Caporaso J.G.
      • Kuczynski J.
      • Stombaugh J.
      • Bittinger K.
      • Bushman F.D.
      • Costello E.K.
      • Fierer N.
      • Pena A.G.
      • Goodrich J.K.
      • Gordon J.I.
      • Huttley G.A.
      • Kelley S.T.
      • Knights D.
      • Koenig J.E.
      • Ley R.E.
      • Lozupone C.A.
      • McDonald D.
      • Muegge B.D.
      • Pirrung M.
      • Reeder J.
      • Sevinsky J.R.
      • Turnbaugh P.J.
      • Walters W.A.
      • Widmann J.
      • Yatsunenko T.
      • Zaneveld J.
      • Knight R.
      QIIME allows analysis of high-throughput community sequencing data.
      ). First, the forward and reverse Illumina reads were joined together using the join_paired_ends.py script. The merged sequences were de-multiplexed and quality filtered. The operational taxonomic units (OTU) were formed by clustering sequences based on 97% similarity threshold using the UCLUST algorithm (
      • Edgar R.C.
      Search and clustering orders of magnitude faster than BLAST.
      ). Representative sequences for each OTU were aligned with PyNast (
      • Caporaso J.G.
      • Bittinger K.
      • Bushman F.D.
      • DeSantis T.Z.
      • Andersen G.L.
      • Knight R.
      PyNAST: A flexible tool for aligning sequences to a template alignment.
      ). The resultant multiple sequence alignment was used to infer a phylogenetic tree with FastTree (
      • Price M.N.
      • Dehal P.S.
      • Arkin A.P.
      FastTree 2–approximately maximum-likelihood trees for large alignments.
      ). The taxonomy of each sequence was identified using uclust consensus taxonomy assigner by performing a search against GreenGenes taxonomy (12/10 release; 
      • McDonald D.
      • Price M.N.
      • Goodrich J.
      • Nawrocki E.P.
      • DeSantis T.Z.
      • Probst A.
      • Andersen G.L.
      • Knight R.
      • Hugenholtz P.
      An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea.
      ). The 16S sequence information from the before and after feeding samples was combined to form one representative sample. The OTU tables were rarefied at 48,536 sequences per sample for α and β diversity analysis. Package qiimer (
      • Bittinger K.
      qiimer: Work with QIIME output files in R. R package, Version 0.9 4.
      ) was used to open QIIME output files in R (version 3.1.1;
      • R Core Team
      R: A language and environment for statistical computing.
      ). Analyses of community similarity (β diversity) were performed by calculating pairwise distance using phylogenetic metric UniFrac. Two measures of α diversity were computed including Shannon diversity, an indicator of evenness in community structure, and richness, the number of OTU observed. The measured α diversity matrices were compared between the treatment groups using Wilcoxon rank sum test. A nonparametric permutational multivariate ANOVA test (
      • Anderson M.J.
      A new method for non-parametric multivariate analysis of variance.
      ), implemented in the vegan package for R, was used to test the effects of day of sampling and treatment within the day of sampling on overall community composition, as measured by weighted and unweighted UniFrac distance. The raw read counts from the 16S rDNA OTU abundance table were collapsed at taxonomic rank and compositionally normalized (relative abundance) such that each sample sums to 1. To test the differences in bacterial taxa between control and HMTBa treatment groups across all days of sampling, analysis of composition of microbiomes (ANCOM;
      • Mandal S.
      • Van Treuren W.
      • White R.A.
      • Eggesbø M.
      • Knight R.
      • Peddada S.D.
      Analysis of composition of microbiomes: A novel method for studying microbial composition.
      ) available in R was used. The significance of test is determined using the Benjamini-Hochberg procedure that controls for false discovery rate. Further, the taxa that differed significantly between the 2 groups using the ANCOM test were tested for significance at each day of sampling using Wilcoxon rank sum test. A P-value of 0.05 was used to define significance. The Spearman correlation coefficients (rs) were calculated to evaluate correlations between genera and FA profiles.

      RESULTS

      Milk Production Responses

      In the accompanying study by M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data), feeding a moderate-risk diet was found to be sufficient to cause diet-induced MFD, as evidenced by a 2.5-fold increase in trans-10 C18:1 and a 13% reduction in milk fat concentration in the control group between d 7 (end of low risk diet) and d 24 (end of moderate-risk diet; P < 0.01 for both variables). Milk fat yield was significantly higher in the HMTBa group at d 14 (midpoint of moderate diet; P < 0.05), d 24 (P < 0.05), and d 28 (end of high-risk diet; P < 0.05) as compared with the control (Supplemental Figure S1; https://doi.org/10.3168/jds.2019-17389). In addition, trans-10 C18:1 was significantly lower in the HMTBa group at d 14, 24, and 28 (P < 0.05 for all time points), indicating that HMTBa was successful in preventing the shift to the altered BH pathway. No significant change was seen between the 2 groups in trans-11 C18:1.

      Sequencing Details

      According to M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data), no significant differences were observed between control and HMTBa on the low-risk diet (d 7), and for this reason, the rumen samples from d 7 were not analyzed for bacterial information in this study. Almost 154 16S rRNA-based bacterial communities that contributed to approximately 1 million high quality reads were analyzed. These reads were binned into 137,563 OTU (minimum: 48,536; maximum: 103,131; mean: 72,392). The range for Good's coverage across all samples was 94 to 97%, indicating a good sequencing depth per sample.

      Within-Sample Variation (Alpha Diversity)

      The number of bacterial species (OTU) was estimated by species richness and the distribution of bacterial species within each community was estimated by Shannon diversity index (Figure 1a–f; Supplemental Table S2; https://doi.org/10.3168/jds.2019-17389). Both species richness (Figure 1a) and Shannon diversity (Figure 1d) were reduced (P < 0.05) in the control compared with HMTBa across all days with increasing risk of BH-induced MFD. No significant differences were noted between d 0 and d 14, 24, and 28 for observed species in either control or HMTBa (Figure 1b, c). However, within the control, compared with d 0, d 14, 24, and 28 showed greater variation with the most noticeable variation on d 28 during the highest risk phase (Figure 1b), whereas no such variation was noted between days of sampling with HMTBa (Figure 1c). For Shannon diversity, differences were noted between d 0 compared with d 14, 24, and 28 in the control group only (Figure 1e). Interestingly, no differences were noted between d 14, 24, and 28 in either control or HMTBa (Supplemental Table S2).
      Figure thumbnail gr1
      Figure 1Measurement of community diversity (α diversity) of the rumen bacterial communities fed control or 2-hydroxy-4-(methylthio) butanoate (HMTBa) in diets with increasing risk of biohydrogenation-induced milk fat depression (d 0 = covariant, d 14 = moderate risk, d 24 = moderate risk, and d 28 = high risk). Panels include (A) number of observed species between treatment groups, (B) number of observed species between days in the control, (C) number of observed species between days in the HMTBa treatment, (D) Shannon diversity between treatments, (E) Shannon diversity between days in the control, and (F) Shannon diversity between days in the HMTBa treatment. Boxes represent the interquartile range (IQR), and the lines inside represent the median. Whiskers denote the lowest and highest values within 1.5 times of the IQR; dots denote the diversity index observations for each sample.

      Between-Bacterial Community Comparison (Beta Diversity)

      The weighted and unweighted UniFrac distances were used to derive principal coordinates (Figure 2a and 2b). Bacterial communities differed (P < 0.05; Permanova test; Table 1) by dietary risk of MFD and treatment in both weighted (relative abundance of commonly present bacterial taxa) and unweighted (presence-absence information of bacterial taxa) distance matrices. In both weighted and unweighted matrices, in the control, bacterial community composition differed (P < 0.05) on d 0 as compared with d 14, 24, and 28, whereas no such differences were noted in the HMTBa group. In both control and HMTBa, bacterial communities did not differ (P > 0.05) between d 14 and d 24 and 28 (moderate and high risk of MFD).
      Figure thumbnail gr2
      Figure 2Comparison of bacterial community composition between control and 2-hydroxy-4-(methylthio) butanoate (HMTBa) supplementation and between days in which diets were fed with increasing risk of biohydrogenation-induced milk fat depression (d 0 = covariant, d 14 = moderate risk, d 24 = moderate risk, and d 28 = high risk) using principal coordinate analysis. (A) Weighted UniFrac distances based on relative abundance of bacterial operational taxonomic units (OTU) and (B) unweighted UniFrac distances based on presence or absence information of bacterial OTU (PC1 = principal coordinate 1; PC2 = principal coordinate 2).
      Table 1Permutational multivariate ANOVA for the effect of 2-hydroxy-4-(methylthio) butanoate (HMTBa) and day of sampling while feeding diets with increasing risk of biohydrogenation-induced milk fat depression (d 0 = covariant, d 14 = low risk, d 24 = moderate risk, and d 28 = high risk) on ruminal bacterial community composition based on weighted and unweighted UniFrac distances
      ItemWeightedUnweighted
      Treatment0.010.001
      Day0.0030.002
      Treatment × day0.4820.458
      Control group
       Day (overall)0.0210.028
       d 0 vs. d 140.0450.02
       d 0 vs. d 210.0160.022
       d 0 vs. d 280.0130.011
       d 14 vs. d 240.0890.176
       d 14 vs. d 280.2950.261
       d 24 vs. d 280.1230.269
      HMTBa group
       Day (overall)0.1280.118
       d 0 vs. d 140.1360.143
       d 0 vs. d 210.0640.076
       d 0 vs. d 280.1280.079
       d 14 vs. d 240.3490.693
       d 14 vs. d 280.1350.182
       d 24 vs. d 280.3490.388

      Phylogenetic Composition of Bacterial Communities in the Rumen

      Representative sequences from the OTU information were assigned to 21 bacterial phyla and 351 genera; however, only 83 genera were present across 75% of the samples (Supplemental Table S3; https://doi.org/10.3168/jds.2019-17389). The most dominant bacterial phyla were Firmicutes (47.84%) and Bacteroidetes (41.46%), which together constituted approximately 90% of total bacterial abundance. The other less abundant phyla included Proteobacteria (3.10%), Fibrobacteres (2.90%), Spirochaetes (1.83%), Cyanobacteria (0.77%), Tenericutes (0.77%), and Actinobacteria (0.52%). Among Firmicutes, Lachnospiraceae (10.72%), Clostridiales (10.22%), Butyrivibrio (6.25%), Ruminococcaceae (6.06%), Ruminococcus (5.77%), Coprococcus (3.55%), and Succiniclasticum (3.38%) were the most abundant bacterial lineages. Bacteroidetes was dominated by Prevotella (34.56%), Bacteroidales (5.96%), S24–7 (3.58%), Paraprevotellaceae (0.87%), and RF16 (0.86%). Proteobacteria was dominated by Succinivibrionaceae (4.13%), Fibrobacteres by Fibrobacter (2.90%), Spirochaetes by Treponema (3.34%), Cyanobacteria by YS2 (2.17%), Tenericutes by Anaeroplasma (1.99%) and RF39 (1.97%), and Actinobacteria by Coriobacteriaceae (2.08%) members.

      Changes in Ruminal Bacteria with MFD Dietary Risks and Effects of HMTBa

      To compare changes in bacterial populations between HMTBa and control, a new tool, the ANCOM test (
      • Mandal S.
      • Van Treuren W.
      • White R.A.
      • Eggesbø M.
      • Knight R.
      • Peddada S.D.
      Analysis of composition of microbiomes: A novel method for studying microbial composition.
      ), was used; this tool was selected because it includes rare bacterial populations. Using this tool, 7 genera were identified that differed between control and HMTBa across all days of sampling (Figure 3; P < 0.05). The d 0 sample was used as a covariate. The ANCOM analysis revealed an overall increase (P < 0.05) in the relative abundance of Dialister, Lachnospira, Megasphaera, and Sharpea in the control compared with HMTBa. Within the control group, these 4 genera were increased on d 14 (moderate-risk phase) compared with d 0. The genera Dialister and Megasphaera were reduced on d 24 (moderate-risk phase) compared with d 14 but increased on d 28 (high risk phase) compared with d 24 in the control. In HMTBa, these 2 genera were reduced on d 14, 24, and 28 compared with d 0. However, the magnitude of decrease was greater on d 14 and 24 compared with d 28. A different pattern was noted for the genera Lachnospira and Sharpea in the control group, where these 2 genera were increased on d 14 and then gradually reduced by d 28. In the HMTBa group, the relative abundance of these 2 genera was sustained from d 0 to 28. In contrast, SR1, L7A_E11 from Firmicutes, and F16 from TM7 were reduced in the control group compared with the HMTBa group (P < 0.05). Both L7A_E11 and F16 were reduced on d 14 compared with d 0 and remained low through d 24 and 28 (moderate- and high-risk phases). Genus SR1 was gradually decreased from d 0 to 28 in HMTBa but remained higher compared with the control.
      Figure thumbnail gr3
      Figure 3Significantly differentially abundant bacterial genera identified by analysis of composition of microbiomes (ANCOM;
      • Mandal S.
      • Van Treuren W.
      • White R.A.
      • Eggesbø M.
      • Knight R.
      • Peddada S.D.
      Analysis of composition of microbiomes: A novel method for studying microbial composition.
      ) between control and 2-hydroxy-4-(methylthio) butanoate (HMTBa) supplementation and between days in which diets were fed with increasing risk of biohydrogenation-induced milk fat depression (d 0 = covariant, d 14 = moderate risk, d 24 = moderate risk, and d 28 = high risk). The effect of HMTBa is identified within each day based on Wilcoxon rank sum test and shown on the plots (***P < 0.001; **P < 0.01; *P < 0.05; †P < 0.10). The error bars indicate SEM.

      Correlation Analysis Between Bacteria Associated with MFD Dietary Risks and Milk FA Profiles

      Associations between bacterial taxa and production responses were analyzed using Spearman correlation (Supplemental Table S4; https://doi.org/10.3168/jds.2019-17389). Bacterial genera that were significantly positively or negatively correlated with a coefficient value of >0.3 were considered (Figure 4). Coefficient values >0.5 were considered strong, whereas those between 0.3 and 0.5 were considered weak. The analysis showed that very few bacterial populations at the genus level had positive correlations with DMI and milk yield, with the notable exceptions of Coriobacteriaceae, Dialister, and Veillonellaceae. It is evident that most bacterial lineages from Firmicutes were positively correlated with milk fat yield and concentration. The representatives from this phylum, including Blautia, Anaerovibrio, Selenomonas, Christenellaceae, and Clostridiales, showed a high positive correlation (P < 0.05) with milk fat concentration and a weak positive correlation (P < 0.05) with milk fat yield. None of the bacterial populations had strong positive correlations with milk protein percentage, whereas a few populations, including Dialister, Eubacterium, and Acidaminococcus, had weak negative correlations. No interesting patterns were noted between protein yield and bacterial populations.
      Figure thumbnail gr4
      Figure 4Association between the relative sequence abundance of bacterial taxa and production responses in cows fed control or 2-hydroxy-4-(methylthio) butanoate (HMTBa) in diets with differing risk for biohydrogenation-induced milk fat depression. The color code indicates the direction of the correlations (blue indicates positive and red indicates negative), and the shade represents the strength of the correlation. OBCFA = odd- and branched-chain fatty acids.
      For the FA isomers, very few bacterial genera showed a weak positive correlation with cis-9,trans-11 CLA, including Dialister and Sharpea, whereas BS11, Christensenellaceae, and Paludibacter had a negative correlation with this isomer. None of the bacterial populations had either strong positive or negative associations with trans-11 C18:1, the main intermediate of the normal BH pathway. The bacterial genera that had a significant (P < 0.05) positive correlation to both trans-10,cis-12 CLA and trans-10 C18:1 isomers were Coriobacteriaceae, Prevotella, Acidaminococcus, Dialister, Lachnospira, Megasphaera, Pseudoramibacter-Eubacterium, Sharpea, Shuttleworthia, and Veillonellaceae. Interestingly, Dialister, Lachnospira, Megasphaera, and Sharpea, all of which increased in the control, were positively correlated (P < 0.05) not only with the 2 trans-10 isomers of the alternate BH pathway, but also with milk yield. These genera were negatively correlated (P < 0.05) with milk fat and protein concentration.

      DISCUSSION

      Dietary Models to Investigate MFD

      Diet-induced MFD continues to be a common problem in dairy herds despite improved nutritional management (
      • Baldin M.
      • De Souza J.
      • Ticiani E.
      • Sandri E.
      • Dresch R.
      • Batistel F.
      • Oliveira D.
      Milk fat response to calcium salts of palm or soybean in a normal or milk fat depression scenario in dairy ewes.
      ). Several factors, including the concentrations and availability of FA in dairy cow rations, fermentability of diets, feeding management, rumen modifiers, and individual cow-to-cow factors, predispose dairy cows to MFD (
      • Harvatine K.J.
      Managing milk fat depression.
      ). Experimental induction and recovery models have been useful in investigation of the pathophysiology of diet-induced MFD (
      • Rico D.E.
      • Harvatine K.J.
      Induction of and recovery from milk fat depression occurs progressively in dairy cows switched between diets that differ in fiber and oil concentration.
      ;
      • Rico D.E.
      • Ying Y.
      • Clarke A.
      • Harvatine K.
      The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
      ). Briefly, in this model, MFD is consistently achieved within 10 d by reducing forage NDF and increasing both starch and PUFA and rescued within 10 to 18 d of returning to a recovery diet. Increasing levels of dietary starch are known to reduce ruminal pH and consequently alter microbial populations (
      • Fuentes M.C.
      • Calsamiglia S.
      • Cardozo P.
      • Vlaeminck B.
      Effect of pH and level of concentrate in the diet on the production of biohydrogenation intermediates in a dual-flow continuous culture.
      ), while increasing levels of dietary PUFA are known to inhibit the growth of microbial populations (
      • Maia M.R.
      • Chaudhary L.C.
      • Bestwick C.S.
      • Richardson A.J.
      • McKain N.
      • Larson T.R.
      • Graham I.A.
      • Wallace R.J.
      Toxicity of unsaturated fatty acids to the biohydrogenating ruminal bacterium, Butyrivibrio fibrisolvens.
      ). Because these dietary modifications significantly alter the ruminal environment, including the microbiota, they lead to a progressive buildup of trans-10,cis-12 CLA and trans-10 isomers in the rumen, which are associated with diet-induced MFD (
      • Rico D.E.
      • Ying Y.
      • Clarke A.
      • Harvatine K.
      The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
      ,
      • Rico D.E.
      • Preston S.
      • Risser J.
      • Harvatine K.
      Rapid changes in key ruminal microbial populations during the induction of and recovery from diet-induced milk fat depression in dairy cows.
      ). Using an induction-recovery model of MFD,
      • Rico D.E.
      • Preston S.
      • Risser J.
      • Harvatine K.
      Rapid changes in key ruminal microbial populations during the induction of and recovery from diet-induced milk fat depression in dairy cows.
      and
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      have previously demonstrated significant alterations in the ruminal microbial populations by d 10 of the induction phase and restoration to the original state by d 14 of the recovery phase. Collectively, these studies established the link between dietary risk factors, rumen microbial composition, and milk FA and provided insights into the physiology behind diet-induced MFD.
      A challenge model has been useful in testing the effect of feed or supplements on induction of diet-induced MFD.
      • Baldin M.
      • Zanton G.
      • Harvatine K.
      Effect of 2-hydroxy-4-(methylthio) butanoate (HMTBa) on risk of biohydrogenation-induced milk fat depression.
      ,
      • Baldin M.
      • Tucker H.A.
      • Harvatine K.J.
      Milk fat response and milk fat and urine biomarkers of microbial nitrogen flow during supplementation with 2-hydroxy-4-(methylthio) butanoate.
      ), and M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data) investigated the effect of HMTBa on risk for MFD in 3 separate experiments with designs that sequentially increased dietary MFD risk factors to represent low-, moderate-, and high-risk diets. This sequential increase in dietary MFD risk allows the evaluation of the effects of HMTBa in preventing the shift to altered BH while reducing the need for washout periods. The experiments differed in the degree of challenge and length of feeding challenge diets. Similar to previous investigation of the effect of HMTBa in the challenge model, HMTBa had no effect on the low risk diet, but reduced the shift to trans-10 pathway and induction of MFD on the moderate- and high-risk phases [M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine, unpublished data].

      Changes in Ruminal Microbiota with Increasing Dietary MFD Risks

      Because M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data) found no significant change in FA isomers between control and HMTBa groups at d 7 (low-risk phase), those ruminal samples were not analyzed in the current study with the presumption that microbial populations are acclimating to the low-risk diet and it may be too early to find differences in bacterial communities at this stage. However, the bacterial community composition in ruminal samples from dairy cows on moderate-risk (d 14 and 24) and high-risk (d 28) MFD diets were found to be different from those in d 0 samples in the control group, indicating that MFD dietary risks induce changes in ruminal bacterial communities. This finding is similar to our previous report (
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      ), which found that the MFD induction diet is associated with altered bacterial communities when compared with recovered bacterial communities. The moderate- and high-risk diets had the same concentration of NDF and starch, but the high-risk diet had a higher PUFA concentration. Interestingly, no differences in the bacterial community composition between moderate- and high-risk diets were identified in this study, indicating that both fermentable carbohydrates and PUFA concentrations drive alterations in ruminal bacteria. However, the high-risk period was maintained for only 5 d, and further studies may be required to investigate the effect of further increasing PUFA concentrations alone for longer periods (>5 d) on ruminal bacteria.
      During the course of the MFD induction period,
      • Rico D.E.
      • Ying Y.
      • Clarke A.
      • Harvatine K.
      The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
      reported a gradual reduction of cis-9,trans-11 CLA isomer and a progressive buildup of trans-10,cis-12 CLA isomer.
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      , in an accompanied study to
      • Rico D.E.
      • Ying Y.
      • Clarke A.
      • Harvatine K.
      The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
      , identified that members of Firmicutes and Actinobacteria were increased whereas members of Bacteroidetes were decreased at d 10 of the induction period. In the current study, Firmicutes remained stable whereas Actinobacteria doubled in both moderate and high MFD dietary risk samples compared with d 0. It is possible that differences in NDF concentrations [26.1% in the induction diet in
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      and
      • Rico D.E.
      • Ying Y.
      • Clarke A.
      • Harvatine K.
      The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
      vs. 29.1 and 28.7% in moderate- and high-risk diets in the current study and M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data)] may have resulted in differences in the relative abundance of Firmicutes between the 2 studies. Increased Actinobacteria, represented mostly by Coriobacteriaceae, as observed in this study, has been associated with dyslipidemic phenotypes (Martinez et al., 2013) and is known to produce trans-10 isomer (

      Zened, A., S. Combes, L. Cauquil, C. Rousseau, C. Klopp, A. Troegeler-Meynadier, and F. Enjalbert. 2011. The ruminal level of trans-10 fatty acids of dairy cows is linked to the composition of bacterial community. In 4th Congress of European Microbiologists, Geneva, Switzerland.

      ).
      In this study, certain bacterial genera such as Dialister, Lachnospira, Megasphaera, and Sharpea were elevated on both moderate- and high-risk MFD diets and were positively correlated (P < 0.05) with trans-10,cis-12 CLA and trans-10 isomers. These findings corroborate with
      • Dewanckele L.
      • Vlaeminck B.
      • Hernandez-Sanabria E.
      • Ruiz-González A.
      • Debruyne S.
      • Jeyanathan J.
      • Fievez V.
      Rumen biohydrogenation and microbial community changes upon early life supplementation of 22:6n-3 enriched microalgae to goats.
      , where these 4 genera, along with Bifidobacterium, Coriobacteriaceae, Lactobacillus, and Eubacterium, were positively associated with trans-10 isomers. It has also been previously shown (
      • Latham M.J.
      • Storry J.E.
      • Sharpe M.E.
      Effect of low-roughage diets on the microflora and lipid metabolism in the rumen.
      ) that 4 lactating dairy cows, when fed a high-roughage diet (8 kg of hay and 10 kg of concentrate) and gradually transitioned to a milk fat-depressing low-roughage diet (until milk fat % reduced to <2%), had an increase in Megasphaera elsdenii, Lactobacillus, Bifidobacterium, and Lachnospira multiparus on milk fat-depressing diets, which is in agreement with findings in this study. An increase in the relative population size of M. elsdenii in cows with MFD has been previously reported (
      • Palmonari A.
      • Stevenson D.M.
      • Mertens D.R.
      • Cruywagen C.W.
      • Weimer P.J.
      pH dynamics and bacterial community composition in the rumen of lactating dairy cows.
      ;
      • Weimer P.J.
      • Stevenson D.M.
      • Mertens D.R.
      Shifts in bacterial community composition in the rumen of lactating dairy cows under milk fat-depressing conditions.
      ;
      • Mohammed R.
      • Stevenson D.M.
      • Beauchemin K.A.
      • Muck R.E.
      • Weimer P.J.
      Changes in ruminal bacterial community composition following feeding of alfalfa ensiled with a lactic acid bacterial inoculant.
      ). While certain strains of M. elsdenii, such as YJ-4, have been shown to produce trans-10,cis-12 CLA isomer (
      • Kim Y.J.
      • Liu R.H.
      • Rychlik J.L.
      • Russell J.B.
      The enrichment of a ruminal bacterium (Megasphaera elsdenii YJ-4) that produces the trans-10, cis-12 isomer of conjugated linoleic acid.
      ), several other strains of M. elsdenii were unable to produce the same isomer in in vitro models (
      • Wallace R.J.
      • McKain N.
      • Shingfield K.J.
      • Devillard E.
      Isomers of conjugated linoleic acids are synthesized via different mechanisms in ruminal digesta and bacteria.
      ;
      • Maia M.R.G.
      • Chaudhary L.C.
      • Figueres L.
      • Wallace R.J.
      Metabolism of polyunsaturated fatty acids and their toxicity to the microflora of the rumen.
      ). Although the specific species of Megasphaera was not identified in this study, the relative abundance of this genus increased in cows with diet-induced MFD. Further studies on identification of the specific bacteria involved in the production of trans-10,cis-12 CLA isomers will be useful to develop strategies for reducing MFD in dairy cows. Application of metagenomic and metatranscriptomic approaches to analyze rumen samples may help to identify the bacteria as well as describe their functional role involved in the production of isomers that lead to MFD.
      Some uncultured bacterial lineages, such as those within Lachnospiraceae, have been reported to participate in ruminal BH (
      • Huws S.A.
      • Kim E.J.
      • Lee M.R.
      • Scott M.B.
      • Tweed J.K.
      • Pinloche E.
      • Wallace R.J.
      • Scollan N.D.
      As yet uncultured bacteria phylogenetically classified as Prevotella, Lachnospiraceae incertae sedis and unclassified Bacteroidales, Clostridiales and Ruminococcaceae may play a predominant role in ruminal biohydrogenation.
      ), although the specific mechanism is not well understood. In our previous study (
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      ), we have reported the association of unclassified Lachnospiraceae with trans-10 isomer in the rumen of dairy cows with diet-induced MFD. Interestingly, under a similar dietary regimen as that used in this study, we found Lachnospira, a member of Lachnospiraceae, to be associated with trans-10 isomer. All 4 of the genera discussed above were increased on d 14 after 7 d on the moderate-risk diet, but declined toward the end of d 24, suggesting a possible adaptation of the dairy cows' ruminal microbiota to the moderate-risk diet. Upon switching to a high-risk diet on d 24, Dialister and Megasphaera increased by d 28, suggesting these genera may have a role in the production of trans-10 isomers; however, such observations require further investigations to demonstrate the cause-effect relationship. The other 2 genera, Lachnospira and Sharpea, did not increase with higher UFA, suggesting that their increased abundance may be associated with changes in the ruminal environment induced by interactions between simultaneously increasing starch and fat concentrations.
      • Kamke J.
      • Kittelmann S.
      • Soni P.
      • Li Y.
      • Tavendale M.
      • Ganesh S.
      • Janssen P.H.
      • Shi W.
      • Froula J.
      • Rubin E.M.
      • Attwood G.T.
      Rumen metagenome and metatranscriptome analyses of low methane yield sheep reveals a Sharpea-enriched microbiome characterised by lactic acid formation and utilisation.
      reported that in sheep with lower methane-yield phenotype, a rapid heterofermentation is supported by an enrichment of Sharpea and Megasphaera spp. in the rumen, leading to lower hydrogen production, thus naturally controlling for methane formation. Collectively, these data support that rapid fermentation in the rumen induced by higher starch concentrations is conducive for specific bacterial phylotypes such as Sharpea, Megasphaera, and Lachnospiraceae. These changes in ruminal conditions, when paired with higher UFA as noted in this study, may lead to significant alterations in the rumen including shifting BH pathways.

      Effect of HMTBa on Reducing Perturbations in Ruminal Microbiota

      Several reports indicate that dairy cows supplemented with methionine hydroxy analogs have demonstrated improvements in bacterial protein synthesis (
      • Arambel M.
      • Bartley E.
      • Camac J.
      • Dennis S.
      • Nagaraja T.
      • Dayton A.
      Rumen degradability and intestinal availability of a protected methionine product and its effects on rumen fermentation, and passage rate of nutrients.
      ), protozoal counts (
      • Lundquist R.G.
      • Stern M.
      • Otterby D.
      • Linn J.
      Influence of methionine hydroxy analog and DL-methionine on rumen protozoa and volatile fatty acids.
      ), ruminal lipid synthesis (
      • Patton R.A.
      • McCarthy R.
      • Griel Jr., L.C.
      Observations on rumen fluid, blood serum, and milk lipids of cows fed methionine hydroxy analog.
      ), fiber digestion (
      • Salsbury R.L.
      • Marvil D.
      • Woodmansee C.
      • Haenlein G.
      Utilization of methionine and methionine hydroxy analog by rumen microorganisms in vitro.
      ;
      • de Vuyst A.
      • Vanbelle M.
      • Joassart J.
      • Baguette A.
      The effect of methionine hydroxyanalog supplementation of the diet on the concentration of ciliate protozoa in the rumen of sheep.
      ), and shifts in VFA patterns (
      • Lundquist R.G.
      • Stern M.
      • Otterby D.
      • Linn J.
      Influence of methionine hydroxy analog and DL-methionine on rumen protozoa and volatile fatty acids.
      ). One such methionine analog, HMTBa, is used as a source of methionine for dairy cows (
      • Chen Z.H.
      • Broderick G.
      • Luchini N.
      • Sloan B.
      • Devillard E.
      Effect of feeding different sources of rumen-protected methionine on milk production and N-utilization in lactating dairy cows.
      ). It has been shown that HMTBa is highly available in the rumen (
      • Zanton G.I.
      • Bowman G.R.
      • Vázquez-Añón M.
      • Rode L.M.
      Meta-analysis of lactation performance in dairy cows receiving supplemental dietary methionine sources or postruminal infusion of methionine.
      ;
      • Koenig K.M.
      • Rode L.M.
      • Knight C.D.
      • Vázquez-Añón M.
      Rumen degradation and availability of various amounts of liquid methionine hydroxy analog in lactating dairy cows.
      ;
      • Jones B.A.
      • Mohamed O.E.
      • Prange R.W.
      • Satter L.D.
      Degradation of methionine hydroxy analog in the rumen of lactating cows.
      ), but the mechanism of its action in the rumen is not completely understood. It has been reported that HMTBa may provide specific AA and peptides for cellulolytic microbes (
      • Van Kessel J.S.
      • Russell J.
      The effect of amino nitrogen on the energetics of ruminal bacteria and its impact on energy spilling.
      ) or may indirectly influence several non-cellulolytic species in the rumen.
      In the current study, supplementing HMTBa to dairy cows exposed to moderate and high MFD dietary risks had a positive effect on the ruminal bacterial populations in their entirety as revealed by α diversity (species richness and diversity) and β diversity (principal coordinate analysis). First, the overall bacterial species richness and diversity, which were reduced with MFD dietary risks, were mitigated with HMTBa supplementation. Reductions in richness and diversity metrics of microbial communities in the rumen have been associated with higher production efficiency in dairy cattle (
      • Shabat S.K.B.
      • Sasson G.
      • Doron-Faigenboim A.
      • Durman T.
      • Yaacoby S.
      • Miller M.E.B.
      • White B.A.
      • Shterzer N.
      • Mizrahi I.
      Specific microbiome-dependent mechanisms underlie the energy harvest efficiency of ruminants.
      ) as well as dietary changes, although dietary changes have a significant effect on overall bacterial community and structure (
      • Tajima K.
      • Aminov R.I.
      • Nagamine T.
      • Matsui H.
      • Nakamura M.
      • Benno Y.
      Diet-dependent shifts in the bacterial population of the rumen revealed with real-time PCR.
      ;
      • Henderson G.
      • Cox F.
      • Ganesh S.
      • Jonker A.
      • Young W.
      • Janssen P.H.
      Global Rumen Census Collaborators
      Rumen microbial community composition varies with diet and host, but a core microbiome is found across a wide geographical range.
      ). For example, a switch to a high-grain diet (
      • AlZahal O.
      • Li F.
      • Guan L.L.
      • Walker N.D.
      • McBride B.W.
      Factors influencing ruminal bacterial community diversity and composition and microbial fibrolytic enzyme abundance in lactating dairy cows with a focus on the role of active dry yeast.
      ) and a high-fat diet (
      • Huws S.A.
      • Kim E.J.
      • Cameron S.J.
      • Girdwood S.E.
      • Davies L.
      • Tweed J.
      • Vallin H.
      • Scollan N.D.
      Characterization of the rumen lipidome and microbiome of steers fed a diet supplemented with flax and echium oil.
      ) have been shown to reduce overall richness and diversity. In agreement with these reports, a change in dietary composition in the current study resulted in a reduction in species richness and diversity. Although changes in bacterial species and their distribution can be influenced by several factors, the extent of variability in species richness and diversity between samples is mostly associated with significant alterations in the microbiota (
      • Fecteau M.E.
      • Pitta D.W.
      • Vecchiarelli B.
      • Indugu N.
      • Kumar S.
      • Gallagher S.C.
      • Fyock T.L.
      • Sweeney R.W.
      Dysbiosis of the fecal microbiota in cattle infected with Mycobacterium avium ssp. paratuberculosis.
      ;
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      ;
      • Stewart H.L.
      • Southwood L.L.
      • Indugu N.
      • Vecchiarelli B.
      • Engiles J.B.
      • Pitta D.
      Differences in the equine faecal microbiota between horses presenting to a tertiary referral hospital for colic compared with an elective surgical procedure.
      ). Indeed, we found the variability between samples for species richness and diversity was greater in d 14, 24, and 28 compared with d 0 in the control group, indicating a gradual shift in bacterial communities with increasing dietary MFD risks. Interestingly, in the HMTBa group, the distribution of samples for species richness and Shannon diversity was consistent from d 0 through d 28, irrespective of the level of MFD dietary risk, indicating that HMTBa prevented aberrations in microbial community structure that are otherwise altered due to MFD dietary risks.
      Second, the findings of this study support the idea that HMTBa prevents the shift to altered BH pathways in response to MFD dietary risks. In the study performed by M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data), the authors reported (1) the significant increase of trans-10 isomer, (2) no change in the cis-9,trans-11 CLA and trans-11 isomers, and (3) an increase in the concentrations of odd- and branched-chain fatty acids including C15:0, C17:0 isomers in the milk FA profiles of cows on the moderate- and high-risk diets without HMTBa supplementation. As described above, Dialister, Sharpea, Megasphaera, and Lachnospira were positively correlated with the trans-10 isomer, a biomarker for the altered BH pathway. These genera were increased in the control group but were reduced (P < 0.05) in the HMTBa-supplemented group on moderate- and high-risk diets compared with d 0 samples, indicating that HMTBa supplementation influences specific bacterial populations, particularly bacteria that are associated with the altered BH pathway.
      Third, in agreement with the findings by
      • Baldin M.
      • Tucker H.A.
      • Harvatine K.J.
      Milk fat response and milk fat and urine biomarkers of microbial nitrogen flow during supplementation with 2-hydroxy-4-(methylthio) butanoate.
      and M. Baldin, G. I. Zanton (USDA, Madison, WI), and K. J. Harvatine (unpublished data) that no differences were noted in cis-9,trans-11 CLA isomers between control and HMTBa groups, no bacterial populations revealed any significant positive or negative associations with this isomer. For example, the genus Butyrivibrio, known for its role in BH (
      • Lourenço M.
      • Ramos-Morales E.
      • Wallace R.
      The role of microbes in rumen lipolysis and biohydrogenation and their manipulation.
      ), stayed consistent between treatment groups and did not have positive or negative correlations with FA isomers. Whereas our previous report (
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      ) revealed positive correlations between Butyrivibrio and both cis-9,trans-11 CLA and trans-10,cis-12 CLA in an induction recovery model, such patterns were not observed in this study. The relative abundance of Butyrivibrio in
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      was about 5% on the induction diet and was reduced by half to about 2.5% on the recovery diet. In this study, the relative abundance of Butyrivibrio was about 4% on d 0 and remained at 5% on all MFD-dietary risk phases in both the control and HMTBa treatment groups, indicating that HMTBa did not influence Butyrivibrio. Higher concentrations of UFA may elevate Butyrivibrio concentrations as observed in both control and HMTBa groups to increase conversion of UFA to CLA and ultimately to vaccenic acid, as a significant percentage of the reductase required to convert CLA to vaccenic acid comes from Butyrivibrio (
      • Hughes P.E.
      • Hunter W.J.
      • Tove S.
      Biohydrogenation of unsaturated fatty acids. Purification and properties of cis-9, trans-11-octadecadienoate reductase.
      ). However, it is not clear as to why HMTBa did not alter the relative abundance of Butyrivibrio in this study. Further studies are needed to understand the role of different species of Butyrivibrio in BH. Finally, we found no associations between bacterial populations and total or individuals OBCFA.
      Other Firmicutes members including Veillonellaceae, Acidoaminococcus, and Shuttleworthia were positively correlated with trans-10 isomer and trans-10,cis-12 CLA isomer but were not significantly influenced by HMTBa. Actinobacteria, predominated by genus Coriobacteriaceae, showed positive correlations with trans-10,cis-12 CLA and trans-10 isomer, similar to our previous findings (
      • Pitta D.W.
      • Indugu N.
      • Vecchiarelli B.
      • Rico D.
      • Harvatine K.
      Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
      ). However, the relative abundance of Coriobacteriaceae was not statistically different between the 2 treatment groups. These findings suggest that HMTBa does not alter the entire community but may alter only the few bacteria involved in the altered pathway or those that produce the trans-10 isomer. These findings on the role of HMTBa on specific bacteria and the production of the trans-10 isomer require further research.
      Notably, genera F16 (Actinobacteria), SR1 (candidate division), and L7A_E11 (Firmicutes) remained higher in the HMTBa group relative to the control group; however, their role in feed digestion has not been well described in the literature. The F16 is positively associated with gross feed efficiency (
      • Jewell K.A.
      • McCormick C.A.
      • Odt C.L.
      • Weimer P.J.
      • Suen G.
      Ruminal bacterial community composition in dairy cows is dynamic over the course of two lactations and correlates with feed efficiency.
      ), and SR1 members are known to be actively involved in sulfur metabolism (
      • Davis J.P.
      • Youssef N.H.
      • Elshahed M.S.
      Assessment of the diversity, abundance, and ecological distribution of members of candidate division SR1 reveals a high level of phylogenetic diversity but limited morphotypic diversity.
      ). Although these genera decreased significantly in the control group on both moderate- and high-risk diets compared with d 0 samples, the magnitude of reduction was lower in the HMTBa group. The role of these genera in the rumen and particularly in MFD requires further investigation.
      In conclusion, exposing cows gradually to increasing MFD dietary risks provided the platform to test the ability of HMTBa to alleviate signs of MFD. Supplementation with HMTBa prevented the shift to the altered BH pathway by modulating the ruminal bacteria in dairy cows. The MFD dietary risk factors altered the ruminal bacterial populations and increased bacteria that were positively associated with trans-10 isomer of FA, which is known to induce MFD. Supplementing HMTBa not only reduced the relative abundance of bacteria that were involved in trans-10 isomer production, but also prevented the overall shift in bacterial community profiles to those typical of cows with MFD. Further studies are needed to understand and decipher the interactions between HMTBa and rumen microbes.

      ACKNOWLEDGMENTS

      We are grateful to the Biomedical Research Core Facilities, University of Pennsylvania (Philadelphia), for sequencing services. The project was partially supported by Penn State University (State College) including the USDA National Institute of Food and Agriculture Federal Appropriations under project number PEN04539 and accession number 1000803, and the animal experiment was supported by Novus International (St. Charles, MO)

      REFERENCES

        • AlZahal O.
        • Li F.
        • Guan L.L.
        • Walker N.D.
        • McBride B.W.
        Factors influencing ruminal bacterial community diversity and composition and microbial fibrolytic enzyme abundance in lactating dairy cows with a focus on the role of active dry yeast.
        J. Dairy Sci. 2017; 100 (28390722): 4377-4393
        • Anderson M.J.
        A new method for non-parametric multivariate analysis of variance.
        Austral Ecol. 2001; 26: 32-46
        • Arambel M.
        • Bartley E.
        • Camac J.
        • Dennis S.
        • Nagaraja T.
        • Dayton A.
        Rumen degradability and intestinal availability of a protected methionine product and its effects on rumen fermentation, and passage rate of nutrients.
        Nutr. Rep. Int. 1987; 35: 661-672
        • Baldin M.
        • De Souza J.
        • Ticiani E.
        • Sandri E.
        • Dresch R.
        • Batistel F.
        • Oliveira D.
        Milk fat response to calcium salts of palm or soybean in a normal or milk fat depression scenario in dairy ewes.
        Livest. Sci. 2017; 206: 109-112
        • Baldin M.
        • Tucker H.A.
        • Harvatine K.J.
        Milk fat response and milk fat and urine biomarkers of microbial nitrogen flow during supplementation with 2-hydroxy-4-(methylthio) butanoate.
        J. Dairy Sci. 2019; 102 (31056326): 6157-6166
        • Baldin M.
        • Zanton G.
        • Harvatine K.
        Effect of 2-hydroxy-4-(methylthio) butanoate (HMTBa) on risk of biohydrogenation-induced milk fat depression.
        J. Dairy Sci. 2018; 101 (29128224): 376-385
        • Bauman D.E.
        • Baumgard L.
        • Corl B.
        • Griinari D.J.
        Biosynthesis of conjugated linoleic acid in ruminants.
        Proc. Am. Soc. Anim. Sci. 1999; 77: 1-14
        • Bauman D.E.
        • Griinari J.M.
        Nutritional regulation of milk fat synthesis.
        Annu. Rev. Nutr. 2003; 23 (12626693): 203-227
        • Bittinger K.
        qiimer: Work with QIIME output files in R. R package, Version 0.9 4.
        • Caporaso J.G.
        • Bittinger K.
        • Bushman F.D.
        • DeSantis T.Z.
        • Andersen G.L.
        • Knight R.
        PyNAST: A flexible tool for aligning sequences to a template alignment.
        Bioinformatics. 2010; 26 (19914921): 266-267
        • Caporaso J.G.
        • Kuczynski J.
        • Stombaugh J.
        • Bittinger K.
        • Bushman F.D.
        • Costello E.K.
        • Fierer N.
        • Pena A.G.
        • Goodrich J.K.
        • Gordon J.I.
        • Huttley G.A.
        • Kelley S.T.
        • Knights D.
        • Koenig J.E.
        • Ley R.E.
        • Lozupone C.A.
        • McDonald D.
        • Muegge B.D.
        • Pirrung M.
        • Reeder J.
        • Sevinsky J.R.
        • Turnbaugh P.J.
        • Walters W.A.
        • Widmann J.
        • Yatsunenko T.
        • Zaneveld J.
        • Knight R.
        QIIME allows analysis of high-throughput community sequencing data.
        Nat. Methods. 2010; 7 (20383131): 335-336
        • Chen Z.H.
        • Broderick G.
        • Luchini N.
        • Sloan B.
        • Devillard E.
        Effect of feeding different sources of rumen-protected methionine on milk production and N-utilization in lactating dairy cows.
        J. Dairy Sci. 2011; 94 (21426989): 1978-1988
        • Davis J.P.
        • Youssef N.H.
        • Elshahed M.S.
        Assessment of the diversity, abundance, and ecological distribution of members of candidate division SR1 reveals a high level of phylogenetic diversity but limited morphotypic diversity.
        Appl. Environ. Microbiol. 2009; 75 (19395567): 4139-4148
        • de Vuyst A.
        • Vanbelle M.
        • Joassart J.
        • Baguette A.
        The effect of methionine hydroxyanalog supplementation of the diet on the concentration of ciliate protozoa in the rumen of sheep.
        Z. Tierphysiol. Tierernahr. Futtermittelkd. 1975; 35 (813418): 316-321
        • Dewanckele L.
        • Vlaeminck B.
        • Hernandez-Sanabria E.
        • Ruiz-González A.
        • Debruyne S.
        • Jeyanathan J.
        • Fievez V.
        Rumen biohydrogenation and microbial community changes upon early life supplementation of 22:6n-3 enriched microalgae to goats.
        Front. Microbiol. 2018; 9 (29636742): 573
        • Edgar R.C.
        Search and clustering orders of magnitude faster than BLAST.
        Bioinformatics. 2010; 26 (20709691): 2460-2461
        • Fecteau M.E.
        • Pitta D.W.
        • Vecchiarelli B.
        • Indugu N.
        • Kumar S.
        • Gallagher S.C.
        • Fyock T.L.
        • Sweeney R.W.
        Dysbiosis of the fecal microbiota in cattle infected with Mycobacterium avium ssp. paratuberculosis.
        PLoS One. 2016; 11 (27494144)e0160353
        • Fuentes M.C.
        • Calsamiglia S.
        • Cardozo P.
        • Vlaeminck B.
        Effect of pH and level of concentrate in the diet on the production of biohydrogenation intermediates in a dual-flow continuous culture.
        J. Dairy Sci. 2009; 92 (19700707): 4456-4466
        • Harvatine K.J.
        Managing milk fat depression.
        in: Proc. Florida Ruminant Nutr. Symp. 2016: 64-77
        • Harvatine K.J.
        • Boisclair Y.
        • Bauman D.
        Recent advances in the regulation of milk fat synthesis.
        Animal. 2009; 3 (22444171): 40-54
        • Henderson G.
        • Cox F.
        • Ganesh S.
        • Jonker A.
        • Young W.
        • Janssen P.H.
        • Global Rumen Census Collaborators
        Rumen microbial community composition varies with diet and host, but a core microbiome is found across a wide geographical range.
        Sci. Rep. 2015; 514567
        • Hughes P.E.
        • Hunter W.J.
        • Tove S.
        Biohydrogenation of unsaturated fatty acids. Purification and properties of cis-9, trans-11-octadecadienoate reductase.
        J. Biol. Chem. 1982; 257 (7061501): 3643-3649
        • Huws S.A.
        • Kim E.J.
        • Cameron S.J.
        • Girdwood S.E.
        • Davies L.
        • Tweed J.
        • Vallin H.
        • Scollan N.D.
        Characterization of the rumen lipidome and microbiome of steers fed a diet supplemented with flax and echium oil.
        Microb. Biotechnol. 2015; 8 (25223749): 331-341
        • Huws S.A.
        • Kim E.J.
        • Lee M.R.
        • Scott M.B.
        • Tweed J.K.
        • Pinloche E.
        • Wallace R.J.
        • Scollan N.D.
        As yet uncultured bacteria phylogenetically classified as Prevotella, Lachnospiraceae incertae sedis and unclassified Bacteroidales, Clostridiales and Ruminococcaceae may play a predominant role in ruminal biohydrogenation.
        Environ. Microbiol. 2011; 13 (21418494): 1500-1512
        • Jenkins T.C.
        • Wallace R.
        • Moate P.
        • Mosley E.
        Board-invited review: Recent advances in biohydrogenation of unsaturated fatty acids within the rumen microbial ecosystem.
        J. Anim. Sci. 2008; 86 (18042812): 397-412
        • Jewell K.A.
        • McCormick C.A.
        • Odt C.L.
        • Weimer P.J.
        • Suen G.
        Ruminal bacterial community composition in dairy cows is dynamic over the course of two lactations and correlates with feed efficiency.
        Appl. Environ. Microbiol. 2015; 81 (25934629): 4697-4710
        • Jones B.A.
        • Mohamed O.E.
        • Prange R.W.
        • Satter L.D.
        Degradation of methionine hydroxy analog in the rumen of lactating cows.
        J. Dairy Sci. 1988; 71 (3379179): 525-529
        • Kamke J.
        • Kittelmann S.
        • Soni P.
        • Li Y.
        • Tavendale M.
        • Ganesh S.
        • Janssen P.H.
        • Shi W.
        • Froula J.
        • Rubin E.M.
        • Attwood G.T.
        Rumen metagenome and metatranscriptome analyses of low methane yield sheep reveals a Sharpea-enriched microbiome characterised by lactic acid formation and utilisation.
        Microbiome. 2016; 4 (27760570): 56
        • Karkalas J.
        An improved enzymic method for the determination of native and modified starch.
        J. Sci. Food Agric. 1985; 36: 1019-1027
        • Kim Y.J.
        • Liu R.H.
        • Rychlik J.L.
        • Russell J.B.
        The enrichment of a ruminal bacterium (Megasphaera elsdenii YJ-4) that produces the trans-10, cis-12 isomer of conjugated linoleic acid.
        J. Appl. Microbiol. 2002; 92 (11972704): 976-982
        • Koenig K.M.
        • Rode L.M.
        • Knight C.D.
        • Vázquez-Añón M.
        Rumen degradation and availability of various amounts of liquid methionine hydroxy analog in lactating dairy cows.
        J. Dairy Sci. 2002; 85 (12018438): 930-938
        • Latham M.J.
        • Storry J.E.
        • Sharpe M.E.
        Effect of low-roughage diets on the microflora and lipid metabolism in the rumen.
        Appl. Microbiol. 1972; 24 (16349951): 871-877
        • Lee C.
        • Oh J.
        • Hristov A.
        • Harvatine K.
        • Vazquez-Anon M.
        • Zanton G.
        Effect of 2-hydroxy-4-methylthio-butanoic acid on ruminal fermentation, bacterial distribution, digestibility, and performance of lactating dairy cows.
        J. Dairy Sci. 2015; 98 (25434334): 1234-1247
        • Lobley G.E.
        • Wester T.
        • Calder A.G.
        • Parker D.
        • Dibner J.
        • Vázquez-Añón M.
        Absorption of 2-hydroxy-4-methylthiobutyrate and conversion to methionine in lambs.
        J. Dairy Sci. 2006; 89 (16507704): 1072-1080
        • Lourenço M.
        • Ramos-Morales E.
        • Wallace R.
        The role of microbes in rumen lipolysis and biohydrogenation and their manipulation.
        Animal. 2010; 4 (22444606): 1008-1023
        • Lundquist R.G.
        • Stern M.
        • Otterby D.
        • Linn J.
        Influence of methionine hydroxy analog and DL-methionine on rumen protozoa and volatile fatty acids.
        J. Dairy Sci. 1985; 68 (3935701): 3055-3058
        • Maia M.R.
        • Chaudhary L.C.
        • Bestwick C.S.
        • Richardson A.J.
        • McKain N.
        • Larson T.R.
        • Graham I.A.
        • Wallace R.J.
        Toxicity of unsaturated fatty acids to the biohydrogenating ruminal bacterium, Butyrivibrio fibrisolvens.
        BMC Microbiol. 2010; 10 (20167098): 52
        • Maia M.R.G.
        • Chaudhary L.C.
        • Figueres L.
        • Wallace R.J.
        Metabolism of polyunsaturated fatty acids and their toxicity to the microflora of the rumen.
        Antonie van Leeuwenhoek. 2007; 91: 303-314
        • Mandal S.
        • Van Treuren W.
        • White R.A.
        • Eggesbø M.
        • Knight R.
        • Peddada S.D.
        Analysis of composition of microbiomes: A novel method for studying microbial composition.
        Microb. Ecol. Health Dis. 2015; 26 (26028277)27663
        • Martin C.
        • Mirande C.
        • Morgavi D.
        • Forano E.
        • Devillard E.
        • Mosoni P.
        Methionine analogues HMB and HMBi increase the abundance of cellulolytic bacterial representatives in the rumen of cattle with no direct effects on fibre degradation.
        Anim. Feed Sci. Technol. 2013; 182: 16-24
        • McDonald D.
        • Price M.N.
        • Goodrich J.
        • Nawrocki E.P.
        • DeSantis T.Z.
        • Probst A.
        • Andersen G.L.
        • Knight R.
        • Hugenholtz P.
        An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea.
        ISME J. 2012; 6 (22134646): 610-618
        • Mohammed R.
        • Stevenson D.M.
        • Beauchemin K.A.
        • Muck R.E.
        • Weimer P.J.
        Changes in ruminal bacterial community composition following feeding of alfalfa ensiled with a lactic acid bacterial inoculant.
        J. Dairy Sci. 2012; 95 (22192212): 328-339
        • Palmonari A.
        • Stevenson D.M.
        • Mertens D.R.
        • Cruywagen C.W.
        • Weimer P.J.
        pH dynamics and bacterial community composition in the rumen of lactating dairy cows.
        J. Dairy Sci. 2010; 93 (20059926): 279-287
        • Patton R.A.
        • McCarthy R.
        • Griel Jr., L.C.
        Observations on rumen fluid, blood serum, and milk lipids of cows fed methionine hydroxy analog.
        J. Dairy Sci. 1970; 53 (5463230): 776-780
        • Pitta D.W.
        • Indugu N.
        • Vecchiarelli B.
        • Rico D.
        • Harvatine K.
        Alterations in ruminal bacterial populations at induction and recovery from diet-induced milk fat depression in dairy cows.
        J. Dairy Sci. 2018; 101 (29103706): 295-309
        • Price M.N.
        • Dehal P.S.
        • Arkin A.P.
        FastTree 2–approximately maximum-likelihood trees for large alignments.
        PLoS One. 2010; 5 (20224823)e9490
        • R Core Team
        R: A language and environment for statistical computing.
        • Rico D.E.
        • Harvatine K.J.
        Induction of and recovery from milk fat depression occurs progressively in dairy cows switched between diets that differ in fiber and oil concentration.
        J. Dairy Sci. 2013; 96 (23958016): 6621-6630
        • Rico D.E.
        • Preston S.
        • Risser J.
        • Harvatine K.
        Rapid changes in key ruminal microbial populations during the induction of and recovery from diet-induced milk fat depression in dairy cows.
        Br. J. Nutr. 2015; 114 (26123320): 358-367
        • Rico D.E.
        • Ying Y.
        • Clarke A.
        • Harvatine K.
        The effect of rumen digesta inoculation on the time course of recovery from classical diet-induced milk fat depression in dairy cows.
        J. Dairy Sci. 2014; 97 (24746126): 3752-3760
        • Salsbury R.L.
        • Marvil D.
        • Woodmansee C.
        • Haenlein G.
        Utilization of methionine and methionine hydroxy analog by rumen microorganisms in vitro.
        J. Dairy Sci. 1971; 54 (5106928): 390-396
        • Shabat S.K.B.
        • Sasson G.
        • Doron-Faigenboim A.
        • Durman T.
        • Yaacoby S.
        • Miller M.E.B.
        • White B.A.
        • Shterzer N.
        • Mizrahi I.
        Specific microbiome-dependent mechanisms underlie the energy harvest efficiency of ruminants.
        ISME J. 2016; 10: 2958-2972
        • Song S.J.
        • Lauber C.
        • Costello E.K.
        • Lozupone C.A.
        • Humphrey G.
        • Berg-Lyons D.
        • Caporaso J.G.
        • Knights D.
        • Clemente J.C.
        • Nakielny S.
        • Gordon J.I.
        • Fierer N.
        • Knight R.
        Cohabiting family members share microbiota with one another and with their dogs.
        eLife. 2013; 2 (23599893)e00458
        • Stewart H.L.
        • Southwood L.L.
        • Indugu N.
        • Vecchiarelli B.
        • Engiles J.B.
        • Pitta D.
        Differences in the equine faecal microbiota between horses presenting to a tertiary referral hospital for colic compared with an elective surgical procedure.
        Equine Vet. J. 2019; 51 (30153353): 336-342
        • Sukhija P.S.
        • Palmquist D.
        Rapid method for determination of total fatty acid content and composition of feedstuffs and feces.
        J. Agric. Food Chem. 1988; 36: 1202-1206
        • Tajima K.
        • Aminov R.I.
        • Nagamine T.
        • Matsui H.
        • Nakamura M.
        • Benno Y.
        Diet-dependent shifts in the bacterial population of the rumen revealed with real-time PCR.
        Appl. Environ. Microbiol. 2001; 67 (11375193): 2766-2774
        • Van Kessel J.S.
        • Russell J.
        The effect of amino nitrogen on the energetics of ruminal bacteria and its impact on energy spilling.
        J. Dairy Sci. 1996; 79 (8872717): 1237-1243
        • Van Soest P.J.
        • Robertson J.
        • Lewis B.
        Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition.
        J. Dairy Sci. 1991; 74: 3583-3597
        • Wallace R.J.
        • McKain N.
        • Shingfield K.J.
        • Devillard E.
        Isomers of conjugated linoleic acids are synthesized via different mechanisms in ruminal digesta and bacteria.
        J. Lipid Res. 2007; 48 (17644775): 2247-2254
        • Weimer P.J.
        • Stevenson D.M.
        • Mertens D.R.
        Shifts in bacterial community composition in the rumen of lactating dairy cows under milk fat-depressing conditions.
        J. Dairy Sci. 2010; 93 (20059925): 265-278
        • Yu Z.
        • Morrison M.
        Comparisons of different hypervariable regions of rrs genes for use in fingerprinting of microbial communities by PCR-denaturing gradient gel electrophoresis.
        Appl. Environ. Microbiol. 2004; 70: 4800-4806
        • Zanton G.I.
        • Bowman G.R.
        • Vázquez-Añón M.
        • Rode L.M.
        Meta-analysis of lactation performance in dairy cows receiving supplemental dietary methionine sources or postruminal infusion of methionine.
        J. Dairy Sci. 2014; 97 (25242429): 7085-7101
      1. Zened, A., S. Combes, L. Cauquil, C. Rousseau, C. Klopp, A. Troegeler-Meynadier, and F. Enjalbert. 2011. The ruminal level of trans-10 fatty acids of dairy cows is linked to the composition of bacterial community. In 4th Congress of European Microbiologists, Geneva, Switzerland.