Advertisement

Ruminal bacteria and protozoa composition, digestibility, and amino acid profile determined by multiple hydrolysis times

Open ArchivePublished:June 29, 2017DOI:https://doi.org/10.3168/jds.2016-12531

      ABSTRACT

      Microbial samples from 4 independent experiments in lactating dairy cattle were obtained and analyzed for nutrient composition, AA digestibility, and AA profile after multiple hydrolysis times ranging from 2 to 168 h. Similar bacterial and protozoal isolation techniques were used for all isolations. Omasal bacteria and protozoa samples were analyzed for AA digestibility using a new in vitro technique. Multiple time point hydrolysis and least squares nonlinear regression were used to determine the AA content of omasal bacteria and protozoa, and equivalency comparisons were made against single time point hydrolysis. Formalin was used in 1 experiment, which negatively affected AA digestibility and likely limited the complete release of AA during acid hydrolysis. The mean AA digestibility was 87.8 and 81.6% for non-formalin-treated bacteria and protozoa, respectively. Preservation of microbe samples in formalin likely decreased recovery of several individual AA. Results from the multiple time point hydrolysis indicated that Ile, Val, and Met hydrolyzed at a slower rate compared with other essential AA. Singe time point hydrolysis was found to be nonequivalent to multiple time point hydrolysis when considering biologically important changes in estimated microbial AA profiles. Several AA, including Met, Ile, and Val, were underpredicted using AA determination after a single 24-h hydrolysis. Models for predicting postruminal supply of AA might need to consider potential bias present in postruminal AA flow literature when AA determinations are performed after single time point hydrolysis and when using formalin as a preservative for microbial samples.

      Key words

      INTRODUCTION

      Nutrient supply and requirement models, such as the Cornel Net Carbohydrate and Protein System (CNCPS;
      • Higgs R.J.
      • Chase L.
      • Ross D.
      • Van Amburgh M.
      Updating the Cornell Net Carbohydrate and Protein System feed library and analyzing model sensitivity to feed inputs.
      ;
      • Van Amburgh M.E.
      • Collao-Saenz E.
      • Higgs R.
      • Ross D.
      • Recktenwald E.
      • Raffrenato E.
      • Chase L.
      • Overton T.
      • Mills J.
      • Foskolos A.
      The Cornell Net Carbohydrate and Protein System: Updates to the model and evaluation of version 6.5.
      ) and the
      • NRC
      , along with their derivative models predict postruminal flows of bacterial biomass. Bacterial protein flow is assigned an AA content, AA profile, and a digestibility of individual AA to calculate supply of metabolizable AA. These field-applicable models typically use the AA profile of bacteria obtained from the literature (
      • Storm E.
      • Brown D.S.
      • Ørskov E.R.
      The nutritive value of rumen micro-organisms in ruminants 3. The digestion of microbial amino and nucleic acids in, and losses of endogenous nitrogen from, the small intestine of sheep.
      ;
      • Clark J.H.
      • Klusmeyer T.H.
      • Cameron M.R.
      Microbial protein synthesis and flows of nitrogen fractions to the duodenum of dairy cows.
      ;
      • Volden H.
      • Harstad O.M.
      Amino acid composition of bacteria harvested from the rumen of dairy cows fed three diets differing in protein content and rumen protein degradability at two levels of intake.
      ), and few account for protozoal AA flows, which can contribute a substantial amount to total microbial AA flow (
      • Dijkstra J.
      • France J.
      • Tamminga S.
      Quantification of the recycling of microbial nitrogen in the rumen using a mechanistic model of rumen fermentation processes.
      ;
      • Fessenden S.W.
      ). A new, dynamic version of the CNCPS (v. 7;
      • Higgs R.
      ; R. J. Higgs, Cornell University, current address Ohaupo, New Zealand, unpublished data) uses a similar approach in a N-based model and mechanistically accounts for protozoa and endogenous AA contributions to total AA flow in addition to bacteria and feed. To improve this model, accurate representations of the AA content and digestibility of bacterial and protozoal AA were needed to understand sources of error in predictions of particular AA. On an N basis, the predictions of NAN in CNCPS version 7 were reasonably accurate and precise. However, predictions of individual AA, such as Lys, Met, Ile, and Val, were biased, likely due to a lack of information about the true content those AA in both microbes and feeds (
      • Higgs R.
      ; R. J. Higgs, Cornell University, current address Ohaupo, New Zealand, unpublished data).
      The methods used for isolation of microbial fractions and analysis of AA vary widely across the literature, and much of the data used for nutrition models still rely on older methods where more robust alternatives now exist. An example can be found in the isolation of protozoa, where differential centrifugation has historically been used to isolate microbial cells with significant contamination from bacteria and feed particles. A procedure developed for isolation of cultivatable mixed ruminal protozoa for quantitative PCR and competition studies (
      • Sylvester J.T.
      • Karnati S.K.
      • Yu Z.
      • Morrison M.
      • Firkins J.L.
      Development of an assay to quantify rumen ciliate protozoal biomass in cows using real-time PCR.
      ,
      • Sylvester J.T.
      • Karnati S.K.R.
      • Yu Z.
      • Newbold C.J.
      • Firkins J.L.
      Evaluation of a real-time PCR assay quantifying the ruminal pool size and duodenal flow of protozoal nitrogen.
      ;
      • Denton B.L.
      • Diese L.E.
      • Firkins J.L.
      • Hackmann T.J.
      Accumulation of reserve carbohydrate by rumen protozoa and bacteria in competition for glucose.
      ) can be used to isolate protozoa for more accurate nutrient analysis. In addition, several of the methods employ the use of formalin in the isolation and storage of microbial and specifically protozoal samples (
      • Martin C.
      • Williams A.G.
      • Michalet-Doreau B.
      Isolation and characteristics of the protozoal and bacterial fractions from bovine ruminal contents.
      ;
      • Rossi M.F.
      • Martinele I.
      • D'Agosto M.
      Quantitative and differential analysis of ciliate protozoa in rumen content samples filtered before and after fixation.
      ); this might be problematic, as formaldehyde is known to react with AA (
      • Barry T.N.
      The effectiveness of formaldehyde treatment in protecting dietary protein from rumen microbial degradation.
      ).
      Intestinal digestibility of individual microbial AA has been estimated by several different methods, including regression approaches (
      • Tas M.V.
      • Evans R.
      • Axford R.
      The digestibility of amino acids in the small intestine of the sheep.
      ;
      • Hvelplund T.
      • Hesselholt M.
      Digestibility of individual amino acids in rumen microbial protein and undegraded dietary protein in the small intestine of sheep.
      ), in vitro assays such as the modified 3-step assay (
      • Gargallo S.
      • Calsamiglia S.
      • Ferret A.
      Technical note: A modified three-step in vitro procedure to determine intestinal digestion of proteins.
      ), and in vivo assays utilizing the mobile bag technique. In vivo and in vitro procedures relying on retention in bags are largely inadequate, as both indigestible and digestible protein and AA can pass through the pores in the bags, resulting in inflated digestibility values. In vivo assays have clear advantages in providing realistic and applicable enzymatic hydrolysis conditions; however, they also require surgically altered animal models. The precision-fed cecectomized rooster bioassay was recently applied to ruminal bacteria (
      • Fonseca A.C.
      • Fredin S.
      • Ferraretto L.
      • Parsons C.
      • Utterback P.
      • Shaver R.
      Short communication: Intestinal digestibility of amino acids in fluid-and particle-associated rumen bacteria determined using a precision-fed cecectomized rooster bioassay.
      ); however, data are still lacking on protozoa AA digestibility. Additionally, avian species have different enzymes and pH conditions that might reduce the ability to make direct comparisons to ruminant digestion (
      • Keller P.J.
      Pancreatic proteolytic enzymes.
      ;
      • Guerino F.
      • Baumrucker C.R.
      Methionine and lysine uptake by cattle small intestine in vitro.
      ).
      • Ross D.A.
      • Gutierrez-Botero M.
      • Van Amburgh M.E.
      Development of an in vitro intestinal digestibility assay for ruminant feeds.
      developed an in vitro assay with ruminal fluid to determine intestinally unavailable N (uN) in ruminant feeds that addressed some issues of nonphysiologic or species-specific enzyme activities, poor retention of small particles, and extensive cannulation procedures. This assay might provide an adequate assessment of intestinal digestibility of ruminal bacteria and protozoa. Considering the ability of formaldehyde to reduce feed protein degradation, it is also possible that the use of formalin in preservation of microbial samples, as reported in
      • Reynal S.M.
      • Broderick G.A.
      • Ahvenjärvi S.
      • Huhtanen P.
      Effect of feeding protein supplements of differing degradability on omasal flow of microbial and undegraded protein.
      and others, might reduce the measured intestinal digestibility of microbial AA. An evaluation of the digestibility of unpreserved and formalin-preserved samples using the uN assay might simultaneously provide more information on formalin effects on individual microbial AA while testing the ability of the uN assay to detect differences in samples treated with a compound known to decreased digestibility.
      Amino acid content of feeds and microbes have historically been determined by single time point hydrolysis, as this represents a compromise between maximal release of AA from the matrix while minimizing the loss of acid labile AA (
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      ). Determination of AA at multiple time points followed by least squares nonlinear regression provides more accurate estimates of the AA profile (
      • Darragh A.J.
      • Moughan P.J.
      The effect of hydrolysis time on amino acid analysis.
      ). This approach has been used in purified protein (
      • Darragh A.J.
      • Garrick D.J.
      • Moughan P.J.
      • Hendriks W.H.
      Correction for amino acid loss during acid hydrolysis of a purified protein.
      ), milk protein (
      • Rutherfurd S.M.
      • Moughan P.J.
      • Lowry D.
      • Prosser C.G.
      Amino acid composition determined using multiple hydrolysis times for three goat milk formulations.
      ), and common animal feedstuffs (
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      ). Previous work in our laboratory indicated that, to obtain the greatest release of branched-chain AA in forages, hydrolysis times needed to be greater than 21 h, whereas Ile release was greatest at 70 h (
      • Ross D.A.
      ). To our knowledge, AA determination after multiple hydrolyses times has not been performed on rumen microbial biomass.
      Given the data from
      • Darragh A.J.
      • Moughan P.J.
      The effect of hydrolysis time on amino acid analysis.
      ,
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      , and the observations made in the data from R. J. Higgs (Cornell University, current address Ohaupo, New Zealand, unpublished), the hypothesis of the current study was that the standard method of determination of AA in ruminal bacteria and protozoa using single time point hydrolysis is not equivalent to AA determination after multiple time point hydrolysis and nonlinear least squares regression. Additional objectives of our study were to summarize the chemical composition and AA profile of ruminal bacteria and protozoa from high-producing lactating dairy cows and to evaluate intestinal digestibility of microbial AA using a newly developed in vitro assay.

      MATERIALS AND METHODS

      All cannulated cows used as rumen or omasal fluid donors for the microbial isolations in this experiment were cared for according to the guidelines of the Institutional Animal Care and Use committee appropriate for the university responsible for their care. The committees reviewed and approved the experiment and all procedures carried out in the study.

      Microbial Isolation Procedures

      Microbial samples from several independent experiments in lactating dairy cattle were obtained and analyzed for nutrient composition, AA content, and intestinal nutrient and AA digestibility. Bacteria and protozoa included in the analysis were from the following experiments. Trial A was an omasal sampling trial with 8 cows in a 2-treatment switchback design investigating effects of a commercial by-product feed on omasal nutrient flow (
      • Fessenden S.W.
      ). Trial B was an omasal sampling trial with 12 cows in a 3-treatment Latin square design investigating the effect of rapidly degradable starch on omasal nutrient flow (A. Foskolos, unpublished data). Trial C was a ruminal N balance and recycling trial with 12 cows in a 3-treatment randomized complete block design investigating ruminal N- or MP-deficient diets (
      • Recktenwald E.B.
      ;
      • Recktenwald E.B.
      • Ross D.A.
      • Fessenden S.W.
      • Wall C.J.
      • Van Amburgh M.E.
      Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
      ). One additional protozoal sample was obtained from T. Hackmann at the University of Florida from repeated isolations from the rumen of a lactating dairy cow at the Ohio State University (Columbus) campus (trial D). For trials A to C, equal parts DM were combined within microbial type, resulting in a composited sample of bacteria and protozoa from each experiment. Therefore, the possible effects of treatments from trials A to C are not represented in this data set. Information regarding the chemical composition of the average diet fed to cows in each experiment, along with the number of individual collections and isolations represented by each composited sample, are reported in Table 1. Due to limited amount of sample for some trials (D and C), not all analysis were performed on all samples as noted throughout the text.
      Table 1Donor cow diet ingredient and chemical composition, intake, and milk production
      ItemTrial
      Trial A: Fessenden (2016); Trial B: A. Foskolos (unpublished data); Trial C: Recktenwald (2010) and Recktenwald et al. (2014); Trial D: T. Hackmann, unpublished data.
      ABCD
      Number of isolations in each sample72108NA
      NA = not available.
      8
      Diet ingredient composition, % of DM
       Corn silage44.028.645.645.3
       Haycrop silage12.022.913.8
       Wheat straw2.1
       Corn meal12.028.611.112.5
       Barley grain, ground6.2
       Soybean meal4.58.6
       Canola meal9.2
       Rumen-protected soybean meal8.23.82.1
       Commercial fermentation byproduct1.5
       Corn distillers3.8
       Cottonseed8.46.4
       Wheat middlings3.42.8
       Soybean hulls5.85.5
       Citrus pulp3.31.07.3
       Sugar2.2
       Molasses0.9
       Fatty acid supplement1.20.61.0
       Blood meal1.71.41.1
       Minerals, vitamins, and additives6.22.27.73.7
      Diet chemical composition
       OM, % of DM93.992.692.591.8
       CP, % of DM16.016.914.816.3
       Soluble protein, % of CP35.743.731.934.1
       RDP,
      Chemical composition estimated using CNCPS v. 6.55 (Van Amburgh et al., 2015) using diet ingredient composition.
      % of CP
      51.959.854.362.0
       aNDFom,
      aNDFom = the NDF content after the addition of alpha-amylase, sodium sulfite, and ash correction.
      % of DM
      31.130.633.632.1
       ADF, % of DM19.820.5NA20.2
       ADL, % of DM3.03.41.93.1
       Sugars, % of DM5.43.55.42.8
       Starch, % of DM27.627.925.426.7
       Ether extract, % of DM4.93.94.55.6
       ME,
      Metabolizable energy predicted using CNCPS v. 6.55 (Van Amburgh et al., 2015).
      Mcal/kg
      2.52.62.72.6
      Cattle intake and production
       DMI, kg/d27.326.123.8NA
       Milk production, kg/d41.741.630.9NA
      1 Trial A:
      • Fessenden S.W.
      ; Trial B: A. Foskolos (unpublished data); Trial C:
      • Recktenwald E.B.
      and
      • Recktenwald E.B.
      • Ross D.A.
      • Fessenden S.W.
      • Wall C.J.
      • Van Amburgh M.E.
      Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
      ; Trial D: T. Hackmann, unpublished data.
      2 NA = not available.
      3 Chemical composition estimated using CNCPS v. 6.55 (
      • Van Amburgh M.E.
      • Collao-Saenz E.
      • Higgs R.
      • Ross D.
      • Recktenwald E.
      • Raffrenato E.
      • Chase L.
      • Overton T.
      • Mills J.
      • Foskolos A.
      The Cornell Net Carbohydrate and Protein System: Updates to the model and evaluation of version 6.5.
      ) using diet ingredient composition.
      4 aNDFom = the NDF content after the addition of alpha-amylase, sodium sulfite, and ash correction.
      5 Metabolizable energy predicted using CNCPS v. 6.55 (
      • Van Amburgh M.E.
      • Collao-Saenz E.
      • Higgs R.
      • Ross D.
      • Recktenwald E.
      • Raffrenato E.
      • Chase L.
      • Overton T.
      • Mills J.
      • Foskolos A.
      The Cornell Net Carbohydrate and Protein System: Updates to the model and evaluation of version 6.5.
      ).
      For trials A and B, microbial samples were obtained using the omasal sampling technique developed by
      • Huhtanen P.
      • Brotz P.G.
      • Satter L.D.
      Omasal sampling technique for assessing fermentative digestion in the forestomach of dairy cows.
      and adapted by
      • Reynal S.M.
      • Broderick G.A.
      Effect of dietary level of rumen-degraded protein on production and nitrogen metabolism in lactating dairy cows.
      . Samples of whole omasal contents were collected from the omasal canal every 2 h during three 8-h intervals. Details of sampling for trial A are described in
      • Fessenden S.W.
      . Trial B sampling occurred per a very similar sampling schedule as trial A by the same researchers (A. Foskolos and S. Fessenden). Trial C collection methods are described in
      • Recktenwald E.B.
      and
      • Recktenwald E.B.
      • Ross D.A.
      • Fessenden S.W.
      • Wall C.J.
      • Van Amburgh M.E.
      Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
      . Trial D protozoa were collected on 4 separate days with 2 separate aliquots filtered per day for 8 aliquots total (T. Hackmann, unpublished data).
      Bacterial isolations for trial A and C were performed according to
      • Whitehouse N.L.
      • Olson V.
      • Schwab C.
      • Chesbrot W.
      • Cunningham K.
      • Lykos T.
      Improved techniques for dissociating particle-associated mixed ruminal microorganisms from ruminal digesta solids.
      with modifications. Briefly, whole omasal contents were filtered through 4 layers of cheesecloth, solids were rinsed once with saline, and the filtrate (I) was treated with formalin (0.1% vol/vol in final solution) and stored at 4°C. The solids retained on the cheesecloth were incubated for 1 h at 39°C in a 0.1% methylcellulose solution, mixed for 1 min at low speed (Omni Mixer, Omni International, Kennesaw, GA) to detach solids-associated bacteria, and held at 4°C for 24 h. The contents were then squeezed through 4 layers of cheesecloth and the filtrate (II) was treated with formalin (0.1% vol/vol in final solution). Filtrates I and II were then combined and centrifuged at 1,000 × g for 5 min at 4°C to remove small feed particles and protozoa. The supernatant was centrifuged at 15,000 × g for 20 min at 4°C and the bacterial pellet, representing both solid- and liquid-associated bacteria, was collected and stored at −20°C until lyophilization and later analysis. Bacterial isolation for trial B followed the same procedure as described above, but formalin was not used.
      Protozoa from trials A and B were isolated from whole contents using the same procedure as described by
      • Denton B.L.
      • Diese L.E.
      • Firkins J.L.
      • Hackmann T.J.
      Accumulation of reserve carbohydrate by rumen protozoa and bacteria in competition for glucose.
      and modified as reported in
      • Fessenden S.W.
      ; Figure 1). The only difference between trials was the omission of formalin and centrifugation in trial B. The isolations were performed by the same researcher for both trials. Strained omasal fluid (250 mL) was combined 1:1 with prewarmed, anaerobically prepared with an N-free buffer (Simplex type, pH 6.8;
      • Williams A.
      • Coleman G.
      ) and added to a prewarmed separatory funnel. Plant particles were removed by aspiration after 1 h of incubation at 39°C. Funnel contents were then preserved with formalin (0.1% vol/vol in final solution) and stored for <4 d at 4°C. Preserved contents were centrifuged at 1,000 × g for 5 min at 4°C, the pellet was resuspended in saline, and protozoa were isolated on a nylon cloth with a 20 μm pore size (14% open area, Sefar, Buffalo, NY). The protozoa isolate was washed several times with saline (500 mL) to reduce bacterial contamination. After isolation, protozoa were stored at −20°C before lyophilization. Protozoa from trial C were isolated from strained ruminal fluid by flocculation to remove feed particles followed by preservation with formalin and centrifugation at 500 × g for 5 min at 10°C. The pellet obtained after centrifugation was assumed to be representative of the ruminal protozoa (
      • Recktenwald E.B.
      ). Protozoa isolation from trial D was performed at the Ohio State University according to
      • Denton B.L.
      • Diese L.E.
      • Firkins J.L.
      • Hackmann T.J.
      Accumulation of reserve carbohydrate by rumen protozoa and bacteria in competition for glucose.
      , except 25 mL instead of 30 mL of clarified fluid was filtered per isolation.
      Figure thumbnail gr1
      Figure 1Flowchart for preparation of protozoa isolates in trials A and B. Fractions discarded are crossed out. Trial B isolation differed only in the omission of the formalin preservation step.

      Chemical Analysis and Hydrolysis Times

      All samples were analyzed for DM after 6 h at 105°C and ash according to
      • AOAC International
      . Total N was determined using a combustion assay (Leco FP-528 N Analyzer, Leco Corp., St. Joseph, MI). Amino acid content of all samples was determined by HPLC following hydrolysis for 24 h at 110°C in a block heater (
      • Gehrke C.W.
      • Wall Sr., L.
      • Absheer J.
      • Kaiser F.
      • Zumwalt R.
      Sample preparation for chromatography of amino acids: Acid hydrolysis of proteins.
      ). Trial B bacteria and protozoa were also hydrolyzed for 2, 4, 6, 12, 18, 21, 24, 30, 48, 72, 120, and 168 h to evaluate the rate of release of each AA. The time points chosen were based on a similar analysis performed on milk proteins (
      • Rutherfurd S.M.
      • Moughan P.J.
      • Lowry D.
      • Prosser C.G.
      Amino acid composition determined using multiple hydrolysis times for three goat milk formulations.
      ). The entire time course was performed twice for each sample, and the reported values are the mean of the 2 determinations. Insufficient sample amount from trials A, C, and D precluded the multiple time point hydrolysis of samples for AA determination.
      For all AA, excluding Met, Cys, and Trp, a sample containing 2 mg of N was weighed into hydrolysis tubes with 25 μL of 250 mM norleucine as an internal standard. Samples were then hydrolyzed as described above with high-purity 6 M HCl (5 mL) after flushing with N2 gas (
      • Mason V.C.
      • Bech-Andersen S.
      • Rudemo M.
      Hydrolysate preparation for amino acid determinations in feed constituents.
      ). For Met and Cys, additional aliquots containing 2 mg of N and the internal standard were preoxidized with 1 mL of performic acid (0.9 mL of 88% formic acid, 0.1 mL of 30% H2O2, and 5 mg of phenol) for 16 h at 4°C before acid hydrolysis (
      • Elkin R.G.
      • Griffith J.
      Hydrolysate preparation for analysis of amino acids in sorghum grains: Effect of oxidative pretreatment.
      ). After hydrolysis, tube contents were filtered through Whatman 541 filter paper (GE Healthcare Life Sciences, Marlborough, MA) and filtrate was diluted to 50 mL in a volumetric flask with HPLC-grade H2O. Aliquots (0.5 mL) were evaporated at 60°C under constant N2 flushing, with 3 rinses and re-evaporations with HPLC-grade H2O to remove acid residues. After final evaporation, the hydrolysate was dissolved in 1 mL of Na diluent (Na220, Pickering Laboratories, Mountain View, CA).
      Individual AA hydrolysates were separated using an Agilent 1100 series HPLC (Agilent Technologies, Santa Clara, CA) fitted with a sodium cation exchange column (Cat. no 1154110T, Pickering Laboratories) using a 4-buffer step gradient and column temperature gradient. Detection of separated AA was performed at 560 nm following postcolumn ninhydrin derivation. Standards (250 nmol/mL) for the individual AA were prepared by diluting a pure standard in sample buffer. The volume of sample and standards loaded onto the column was 10 μL.
      For Trp determination, a separate aliquot of sample containing 2 mg of N was hydrolyzed with 1.2 g of Ba(OH)2 at 110°C for the same time course as other AA on a block heater according to the method of
      • Landry J.
      • Delhaye S.
      Simplified procedure for the determination of tryptophan of foods and feedstuffs from barytic hydrolysis.
      . Included in the hydrolysis was 125 μL of 5-Methyl-Trp (5 mM) as an internal standard. After cooling to precipitate barium ions, an aliquot (3 μL) of the hydrolysate was added to 1 mL of acetate buffer (0.07 M sodium acetate) an analyzed using fluorescence detection (excitation = 285 nm, emission = 345 nm) after HPLC separation.

      In Vitro Digestibility of N and AA

      Microbial samples from trials A and B were analyzed for intestinal digestibility of N and AA according to the assay described by
      • Ross D.A.
      with minor modifications. The ruminal incubation step was omitted because microbial samples were isolated from the omasum. For each sample, 150 mg of DM was weighed in duplicate into 125-mL Erlenmeyer flasks and 40 mL of prewarmed ruminal buffer was added (
      • Van Soest P.J.
      ). Samples were then acidified to a pH of 2 with 3 M HCl followed by addition of 2 mL of pepsin solution (282 U/mL). After 1 h of incubation at 39°C in a shaking water bath, contents of the flask were neutralized with 2 mL of 2 M NaOH. Ten milliliters of enzyme mixture containing trypsin (24 mg/mL), chymotrypsin (20 mg/mL), amylase (50 mg/mL), and lipase (4 mg/mL) was then added to the flasks, followed by 24 h of incubation at 39°C in a shaking water bath. After incubation, flask contents were filtered on previously tared Whatman 934AH filters under vacuum. Samples were allowed to air dry, followed by drying and storage in a desiccator. Filter plus residue weight was then recorded, and DM remaining on the filter was corrected for a blank carried throughout the process. Each filter was cut in half and weighed; with one half used for determination of residual N, whereas the other half was used for AA analysis of the residual material. Determination of residual AA, except Trp, was performed after 24 h of hydrolysis with preoxidation of Met and Cys, as described previously. Insufficient sample N on the filters precluded the determination of Trp on the residues.

      Calculations and Statistical Analysis

      Digestibility of DM, N, and individual AA was calculated as the disappearance of DM, OM, N, or AA after enzymatic hydrolysis, corrected for the procedure blank. Determination of the AA concentration of microbes after multiple hydrolysis times was performed using a method similar to
      • Rutherfurd S.M.
      • Moughan P.J.
      • Lowry D.
      • Prosser C.G.
      Amino acid composition determined using multiple hydrolysis times for three goat milk formulations.
      and
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      . Amino acid concentration (mg/g of DM) was plotted against hydrolysis time and a nonlinear equation was used to fit the curves to each plot:
      B(t)=Aoh(elheht)hl,


      where B(t) is the AA concentration at time t, h is the hydrolysis rate (proportion of bound AA hydrolyzed per hour), l is the loss rate (proportion of bound AA destroyed per hour), and Ao is the actual AA content of the protein within the sample; Aoh and l for each sample were derived from each AA using least squares nonlinear regression with the constraints that Ao > 0, and h > 0, and e is a mathematical constant whose natural logarithm is equal to 1 and is approximately equal to 2.71828.
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      used an additional term to account for free AA content (analyzed as AA determined before hydrolysis); this was not included in the current model, as free AA in bacterial and protozoal samples was considered to be negligible.
      The AA profile of trial B bacteria and protozoa determined using the different hydrolysis methods was compared using 2 one-sided paired t-tests (TOST option in the TTEST procedure of SAS version 9.3; SAS Institute, Cary, NC). This procedure, common in bio-equivalence testing, allows the researcher to specify biologically relevant differences in means determined by competing methods. In our study, an AA profile difference that would theoretically alter the calculated flow of any individual AA by ≥6 g/d was considered biologically relevant because (1) commercial AA products fed at 10 g/d (a common minimum feeding rate) provide approximately 6 g of metabolizable AA (
      • Whitehouse N.L.
      ), and (2) with a commonly limiting AA, such as Met, a 6-g shift represents a 10% difference in the total supply of 60 g/d. To convert this 6 g/d of metabolizable AA value into a relevant change in bacterial or protozoal AA profile (measured in g/100 g of AA) the following calculations were made. The microbial AA flowing out of the rumen of a lactating dairy cow was assumed to be approximately 1,900 g/d (
      • Reynal S.M.
      • Ipharraguerre I.R.
      • Lineiro M.
      • Brito A.F.
      • Broderick G.A.
      • Clark J.H.
      Omasal flow of soluble proteins, peptides, and free amino acids in dairy cows fed diets supplemented with proteins of varying ruminal degradabilities.
      ;
      • Fessenden S.W.
      ). Assuming 80% of the flow is bacteria AA and 20% is protozoa AA (
      • Dijkstra J.
      • France J.
      • Tamminga S.
      Quantification of the recycling of microbial nitrogen in the rumen using a mechanistic model of rumen fermentation processes.
      ;
      • Fessenden S.W.
      ), this corresponds to 1,520 and 380 g/d of AA for bacteria and protozoa, respectively. Therefore, a 6-g/d change in supply corresponds to 0.4 (6 g/1,520 g) and 1.5 percentage units (6 g/380 g) change in any individual AA in the AA profile of bacteria and protozoa, respectively. In a nutrition modeling context,
      • Fessenden S.W.
      saw significant improvements in CNCPS predictions of AA flow when similar magnitude changes in AA profile of microbial fractions were evaluated. The bio-equivalency testing framework requires thoughtful interpretation of the results. Emphasis is placed on the comparison of the 90% confidence interval of the mean difference and its comparison the predetermined biologically relevant ranges defined previously in this section.
      For AA digestibility, a 2-sample t-test was used to compare the digestibility estimates between trials A and B. Only means from trials A and B were analyzed, as a limited amount of sample precluded digestibility and multiple hydrolysis time analysis for trials C and D. For all analysis, n = 2 for each comparison.

      RESULTS AND DISCUSSION

      Microbial Chemical Composition and Digestibility

      All donor cows were fed diets with similar chemical and nutrient composition (Table 1). Diets were typical of the northeastern and Midwestern United States, with corn silage and alfalfa silage as the principal forages. Organic matter content in bacteria and protozoa was similar to values obtained previously from ruminal and omasal isolates (
      • Brito A.F.
      • Broderick G.A.
      • Reynal S.M.
      Effect of varying dietary ratios of alfalfa silage to corn silage on omasal flow and microbial protein synthesis in dairy cows.
      ,
      • Brito A.F.
      • Broderick G.A.
      • Reynal S.M.
      Effects of different protein supplements on omasal nutrient flow and microbial protein synthesis in lactating dairy cows.
      ), although OM content is strongly influenced by the isolation procedures used (
      • Martin C.
      • Williams A.G.
      • Michalet-Doreau B.
      Isolation and characteristics of the protozoal and bacterial fractions from bovine ruminal contents.
      ). Trial A bacteria and protozoa had numerically decreased DM, OM, and N digestibility compared with trial B (Table 2); this is likely due to the use of formalin in trial A versus trial B, as formaldehyde readily reacts with proteins to form products resistant to digestion (
      • Barry T.N.
      The effectiveness of formaldehyde treatment in protecting dietary protein from rumen microbial degradation.
      ). Whereas diet and animal differences between trials might also have contributed to observed differences, the direction and magnitude of the difference between trial A and B for OM, DM, and N digestibility suggests formalin treatment is a likely cause of the differences observed. Bacteria isolates had similar AA as a percent of DM, whereas the lower N content of bacteria from trial C (7.5%) resulted in AA N contributing more to total N compared with bacteria from trials A and B. Protozoa from trial D had the highest AA N as a percent of N. Bacterial AA N values for trials A and B were on the low end of the range (54.9–86.7% of total N) reported by
      • Clark J.H.
      • Klusmeyer T.H.
      • Cameron M.R.
      Microbial protein synthesis and flows of nitrogen fractions to the duodenum of dairy cows.
      . This could be related to the site of sampling, as microbes in trials A and B were isolated from omasal contents whereas trials C and D were isolated from ruminal contents, which may contribute to different amounts of NAN.
      • Volden H.
      • Mydland L.T.
      • Harstad O.M.
      Chemical composition of protozoal and bacterial fractions isolated from ruminal contents of dairy cows fed diets differing in nitrogen supplementation.
      reported diaminopimelic acid and purines were affected by diet and differed between protozoa and bacterial fractions in the rumen; similarly,
      • Illg D.
      • Stern M.
      In vitro and in vivo comparisons of diaminopimelic acid and purines for estimating protein synthesis in the rumen.
      , noted wide ranges in nonamino N concentration between duodenal and ruminal samples.
      Table 2Chemical composition and intestinal digestibility of bacteria and protozoa isolates
      ItemTrial
      Trial A: Fessenden (2016); Trial B: A. Foskolos (unpublished data); Trial C: Recktenwald (2010) and Recktenwald et al. (2014); Trial D: T. Hackmann, unpublished data.
      ABCD
      Bacteria
       OM, % of DM83.992.190.1
       N, % of DM8.18.97.5
       uN,
      Intestinally unavailable N (uN) as determined by the procedure of Ross et al. (2013).
      % of total N
      36.415.1
       AA, % of DM28.234.332.5
       AA N, % of N47.351.857.8
       DM digestibility, %53.669.0
       OM digestibility, %68.281.9
       N digestibility, %63.684.9
      Protozoa
       OM, % of DM87.085.090.893.8
       N, % of DM8.38.28.18.4
       uN,
      Intestinally unavailable N (uN) as determined by the procedure of Ross et al. (2013).
      % total N
      46.724.7
       AA, % of DM28.431.737.445.3
       AA N, % of N53.253.765.971.3
       DM digestibility, %48.761.9
       OM digestibility, %65.988.4
       N digestibility, %53.375.3
      1 Trial A:
      • Fessenden S.W.
      ; Trial B: A. Foskolos (unpublished data); Trial C:
      • Recktenwald E.B.
      and
      • Recktenwald E.B.
      • Ross D.A.
      • Fessenden S.W.
      • Wall C.J.
      • Van Amburgh M.E.
      Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
      ; Trial D: T. Hackmann, unpublished data.
      2 Intestinally unavailable N (uN) as determined by the procedure of
      • Ross D.A.
      • Gutierrez-Botero M.
      • Van Amburgh M.E.
      Development of an in vitro intestinal digestibility assay for ruminant feeds.
      .
      Microbial isolates averaged 50.6 and 49% EAA as a percent of total AA for bacteria and protozoa, respectively (Table 3). Bacterial isolates from trial B had numerically increased concentrations of Lys and Met, whereas trials B and D protozoa also demonstrated increased Lys and Met concentrations. For NEAA, bacteria and protozoa isolations were similar among trials, with the exception of Tyr, which was reduced in trials A and C. Again, the differences in microbial isolates from trial A and C for individual AA are likely due to formalin treatment.
      • Volden H.
      • Mydland L.T.
      • Harstad O.M.
      Chemical composition of protozoal and bacterial fractions isolated from ruminal contents of dairy cows fed diets differing in nitrogen supplementation.
      reported decreased recoveries of Lys, Met, and Tyr with versus without formaldehyde treatment in solid- and liquid-associated bacteria, and
      • Whitehouse N.L.
      • Olson V.
      • Schwab C.
      • Chesbrot W.
      • Cunningham K.
      • Lykos T.
      Improved techniques for dissociating particle-associated mixed ruminal microorganisms from ruminal digesta solids.
      reported approximately 20% less Tyr in microbial samples after treatment with formaldehyde. Beyond these differences, likely due to formaldehyde treatment, the AA profile of bacteria and protozoa agreed fairly well with literature reports, with some exceptions. Methionine averaged 3.2% of total AA among all samples and was at the high end of the range reported by
      • Clark J.H.
      • Klusmeyer T.H.
      • Cameron M.R.
      Microbial protein synthesis and flows of nitrogen fractions to the duodenum of dairy cows.
      . The variability of reported AA composition is likely related more to the isolation techniques rather than true differences among microbial populations. Protozoa AA composition has been shown to remain fairly constant among sampling times (
      • Martin C.
      • Bernard L.
      • Michalet-Doreau B.
      Influence of sampling time and diet on amino acid composition of protozoal and bacterial fractions from bovine ruminal contents.
      ;
      • Volden H.
      • Harstad O.M.
      • Mydland L.T.
      Amino acid content and profile of protozoal and bacterial fractions isolated from ruminal contents of lactating dairy cows fed diets differing in nitrogen supplementation.
      ). Differences among microbial fractions (solid-associated bacteria, liquid-associated bacteria, and protozoa) have been well documented (
      • Chiquette J.
      • Benchaar C.
      Effect of diet and probiotic addition on chemical composition of free or particle-associated bacterial populations of the rumen.
      ;
      • Korhonen M.
      • Ahvenjärvi S.
      • Vanhatalo A.
      • Huhtanen P.
      Supplementing barley or rapeseed meal to dairy cows fed grass-red clover silage: II. Amino acid profile of microbial fractions.
      ); however, reasons for the differences are not clear. Procedures used to detach microbes report recoveries ranging from 20 (
      • Martín-Orúe S.M.
      • Balcells J.
      • Zakraoui F.
      • Castrillo C.
      Quantification and chemical composition of mixed bacteria harvested from solid fractions of rumen digesta: effect of detachment procedure.
      ) to 80% (
      • Whitehouse N.L.
      • Olson V.
      • Schwab C.
      • Chesbrot W.
      • Cunningham K.
      • Lykos T.
      Improved techniques for dissociating particle-associated mixed ruminal microorganisms from ruminal digesta solids.
      ); this might call into question the true ability of recovered bacteria to represent the particle associated bacteria (
      • Korhonen M.
      • Ahvenjärvi S.
      • Vanhatalo A.
      • Huhtanen P.
      Supplementing barley or rapeseed meal to dairy cows fed grass-red clover silage: II. Amino acid profile of microbial fractions.
      ). Ultimately, it is likely that differences in isolation methods are responsible for much of the reported ranges of AA composition, whereas a smaller portion of the variation can be considered a true difference in AA composition (
      • Fonseca A.C.
      • Fredin S.
      • Ferraretto L.
      • Parsons C.
      • Utterback P.
      • Shaver R.
      Short communication: Intestinal digestibility of amino acids in fluid-and particle-associated rumen bacteria determined using a precision-fed cecectomized rooster bioassay.
      ).
      Table 3Amino acid profile (% of total AA) of microbial isolates using a 24-h hydrolysis time
      ItemBacteria AA
      Trial A: Fessenden (2016); Trial B: A. Foskolos (unpublished data); Trial C: Recktenwald (2010) and Recktenwald et al. (2014); Trial D: T. Hackmann, unpublished data.
      Protozoa AA
      Trial A: Fessenden (2016); Trial B: A. Foskolos (unpublished data); Trial C: Recktenwald (2010) and Recktenwald et al. (2014); Trial D: T. Hackmann, unpublished data.
      ABCDABCD
      EAA
       Arg5.45.05.05.55.45.44.7
       His2.12.12.02.92.52.32.0
       Ile5.04.04.94.73.85.75.5
       Leu4.85.66.85.56.14.64.2
       Lys4.77.54.85.78.85.310.2
       Met3.34.52.62.73.12.13.8
       Phe6.66.06.87.46.57.37.6
       Thr6.35.75.45.44.86.14.7
       Trp5.75.55.34.64.53.11.4
       Val6.75.95.65.74.74.74.7
       Total EAA50.851.949.350.150.246.548.8
      NEAA
       Ala7.96.96.75.95.14.74.0
       Asp12.211.18.811.810.812.111.6
       Cys1.41.51.12.02.11.72.2
       Glu12.411.313.014.013.615.213.6
       Gly5.74.95.24.74.54.34.0
       Pro2.12.06.92.73.07.95.6
       Ser5.54.65.45.35.45.54.2
       Tyr2.06.03.53.55.42.06.1
       Total NEAA49.248.150.749.949.853.551.2
      1 Trial A:
      • Fessenden S.W.
      ; Trial B: A. Foskolos (unpublished data); Trial C:
      • Recktenwald E.B.
      and
      • Recktenwald E.B.
      • Ross D.A.
      • Fessenden S.W.
      • Wall C.J.
      • Van Amburgh M.E.
      Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
      ; Trial D: T. Hackmann, unpublished data.
      Bacteria and protozoa from trial A demonstrated decreased digestibility (Table 4) for most AA, which again is likely related directly to formalin treatment. Total bacterial EAA digestibility averaged 74.9 and 88.0% for trials A and B, respectively (P = 0.01). Protozoa AA digestibility was also likely affected by formalin treatment in trial A; however, not all AA were significantly different between trials. Arginine, Leu, Val, and Glu all demonstrated decreased digestibility with formalin treatment. The ability of the uN assay to detect differences in digestibility due to formalin treatment indicates the assay might be a useful evaluation tool for other protein-containing feedstuffs, especially rumen-protected protein supplements. Other techniques, such as the mobile bag technique (
      • Hvelplund T.
      • Weisbjerg M.R.
      • Andersen L.S.
      Estimation of the true digestibility of rumen undegraded dietary protein in the small intestine of ruminants by the mobile bag technique.
      ) and the modified 3-step assay (
      • Gargallo S.
      • Calsamiglia S.
      • Ferret A.
      Technical note: A modified three-step in vitro procedure to determine intestinal digestion of proteins.
      ), rely on retention of all undigested proteins in bags, and, as such, estimates of microbial digestibility of AA from those assays are of limited value due to potential for loss from the bag. The in vivo nature of the mobile bag technique does ensure exposure to idealized enzymatic hydrolysis conditions; however, this requires extensively cannulated animals and a significant amount of sample to use the method, something not easily done with isolated omasal microbial samples. Heat-damaged blood meal, although relatively soluble and able to pass through pores, can have a very low digestible fraction (
      • Ross D.A.
      • Gutierrez-Botero M.
      • Van Amburgh M.E.
      Development of an in vitro intestinal digestibility assay for ruminant feeds.
      ). Alternative in vivo techniques, such as the precision-fed cecectomized rooster bioassay, have been used in ruminant feeds (
      • Titgemeyer E.
      • Merchen N.
      • Han Y.
      • Parsons C.
      • Baker D.
      Assessment of intestinal amino acid availability in cattle by use of the precision-fed cecectomized rooster assay.
      ,
      • Boucher S.E.
      • Calsamiglia S.
      • Parsons C.M.
      • Stein H.H.
      • Stern M.D.
      • Erickson P.S.
      • Utterback P.L.
      • Schwab C.G.
      Intestinal digestibility of amino acids in rumen undegradable protein estimated using a precision-fed cecectomized rooster bioassay: I. Soybean meal and SoyPlus.
      ), and
      • Fonseca A.C.
      • Fredin S.
      • Ferraretto L.
      • Parsons C.
      • Utterback P.
      • Shaver R.
      Short communication: Intestinal digestibility of amino acids in fluid-and particle-associated rumen bacteria determined using a precision-fed cecectomized rooster bioassay.
      recently applied the technique to ruminal bacteria isolates. Total AA digestibility reported by
      • Fonseca A.C.
      • Fredin S.
      • Ferraretto L.
      • Parsons C.
      • Utterback P.
      • Shaver R.
      Short communication: Intestinal digestibility of amino acids in fluid-and particle-associated rumen bacteria determined using a precision-fed cecectomized rooster bioassay.
      averaged 76%, with a range of 62 (Cys) to 82% (Met). The mean AA digestibility in non-formalin-treated bacteria in the current trial was 88%, with a range of 84 (Tyr) to 95% (Cys). Differences between the studies are likely related to the different enzyme activities and digestive processes between these 2 methods and the enzyme and pH differences between avian and ruminant digestion (
      • Guerino F.
      • Baumrucker C.R.
      Methionine and lysine uptake by cattle small intestine in vitro.
      ;
      • Ross D.A.
      • Gutierrez-Botero M.
      • Van Amburgh M.E.
      Development of an in vitro intestinal digestibility assay for ruminant feeds.
      ).
      • Storm E.
      • Brown D.S.
      • Ørskov E.R.
      The nutritive value of rumen micro-organisms in ruminants 3. The digestion of microbial amino and nucleic acids in, and losses of endogenous nitrogen from, the small intestine of sheep.
      calculated a mean intestinal digestibility of 85% (range of 80 to 88%) in sheep maintained on VFA, minerals, and isolated ruminal microorganisms. Mathematical techniques have also been used to estimate digestibility, and
      • Tas M.V.
      • Evans R.
      • Axford R.
      The digestibility of amino acids in the small intestine of the sheep.
      used a regression approach to estimate true digestibility of rumen bacteria at 87%, whereas
      • Hvelplund T.
      • Hesselholt M.
      Digestibility of individual amino acids in rumen microbial protein and undegraded dietary protein in the small intestine of sheep.
      reported true AA digestibilities between 80 and 91% for most AA using a similar approach. General agreement between the previously used techniques in ruminants and the current application of the assay developed by
      • Ross D.A.
      • Gutierrez-Botero M.
      • Van Amburgh M.E.
      Development of an in vitro intestinal digestibility assay for ruminant feeds.
      suggest that in vitro uN determination might be useful for future studies of AA digestibility in diverse supplemental protein sources of metabolizable AA.
      Table 4Intestinal digestibility (% of AA) of AA of omasal bacteria and protozoa
      ItemBacteriaProtozoa
      Trial
      Trial A: Fessenden (2016); Trial B: A. Foskolos (unpublished data); Trial C: Recktenwald (2010) and Recktenwald et al. (2014); Trial D: T. Hackmann, unpublished data. n = 2 for each comparison.
      SEMP-valueTrial
      Trial A: Fessenden (2016); Trial B: A. Foskolos (unpublished data); Trial C: Recktenwald (2010) and Recktenwald et al. (2014); Trial D: T. Hackmann, unpublished data. n = 2 for each comparison.
      SEMP-value
      ABAB
      EAA
       Arg74.088.20.90.0369.089.00.40.01
       His78.490.71.90.0369.770.911.90.93
       Ile76.888.90.60.0075.385.73.00.08
       Leu78.592.21.50.0174.392.51.10.01
       Lys75.591.01.80.0268.277.512.60.54
       Met80.788.81.00.0778.490.10.90.17
       Phe69.083.21.20.0167.579.72.70.06
       Thr76.389.81.10.0173.579.29.30.61
       Trp
       Val69.988.04.00.0563.283.33.80.04
       Total EAA74.988.01.40.0170.682.84.40.11
      NEAA
       Ala72.587.22.20.0267.580.16.60.20
       Asp77.590.81.20.0175.886.35.00.18
       Cys82.194.80.70.0589.393.61.70.15
       Glu68.585.23.60.0467.587.70.80.01
       Gly69.485.83.70.0563.572.813.40.56
       Pro77.988.41.40.1078.180.74.70.66
       Ser75.089.31.30.0171.956.733.60.70
       Tyr34.084.30.40.0055.973.84.50.06
       Total NEAA71.687.51.90.0270.280.27.60.32
      Total AA73.387.81.70.0170.481.66.00.21
      1 Trial A:
      • Fessenden S.W.
      ; Trial B: A. Foskolos (unpublished data); Trial C:
      • Recktenwald E.B.
      and
      • Recktenwald E.B.
      • Ross D.A.
      • Fessenden S.W.
      • Wall C.J.
      • Van Amburgh M.E.
      Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
      ; Trial D: T. Hackmann, unpublished data. n = 2 for each comparison.

      AA Determination from Multiple Hydrolysis Times

      The release of individual AA in trial B bacteria and protozoa are in Figures 2, 3, and 4. Extraction of Ile, Met, and Val demonstrated greater release over time and thus positive slopes at time points greater than 24 h and hydrolysis rate (Table 5) were lowest for these AA. Of the NEAA of the protozoa, Ala, Cys, and Pro demonstrated increasing concentrations of AA as hydrolysis time increased. Serine concentrations of bacteria and protozoa decreased markedly after 24 h of hydrolysis as indicated by a relatively high loss rate (Table 5). Overall, total AA were hydrolyzed from the sample matrix at a rate of 0.415 and 0.357 mg/h for bacteria and protozoa, respectively. The same least squares nonlinear regression approach has been previously employed in the analysis of other AA-containing compounds, including lysozyme (
      • Darragh A.J.
      • Garrick D.J.
      • Moughan P.J.
      • Hendriks W.H.
      Correction for amino acid loss during acid hydrolysis of a purified protein.
      ), cat hair (
      • Hendriks W.
      • Tarttelin M.
      • Moughan P.
      The amino acid composition of cat (Felis catus) hair.
      ), human milk (
      • Darragh A.J.
      • Moughan P.J.
      The amino acid composition of human milk corrected for amino acid digestibility.
      ), and some common feedstuffs (
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      ). To our knowledge, no previous work has reported rumen microbial AA content after multiple hydrolysis times.
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      reported similarly low hydrolysis rate for Ile and Val, whereas Ser was reported to have the highest loss rate of any AA.
      Figure thumbnail gr2
      Figure 2Effect of hydrolysis time (h) on content of EAA (mg/g of DM) from freeze-dried isolations of omasal bacteria (•) and protozoa (○) from trial B. The mean of each time point (2 replicates each) is plotted against the least squared regression line for each respective data set.
      Figure thumbnail gr3
      Figure 3Effect of hydrolysis time (h) on content of NEAA (mg/g of DM) from freeze-dried isolations of omasal bacteria (•) and protozoa (○) from trial B. The mean of each time point (2 replicates each) is plotted against the least squared regression line for each respective data set.
      Figure thumbnail gr4
      Figure 4Effect of hydrolysis time (h) on content of EAA, NEAA, and total AA (mg/g of DM) from freeze-dried isolations of omasal bacteria (•) and protozoa (○) from trial B. The mean of each time point (2 replicates each) is plotted against the least squared regression line for each respective data set.
      Table 5Rate of hydrolysis (h)
      Proportion of bound AA hydrolyzed per hour.
      and loss (l)
      Proportion of bound AA destroyed per hour.
      for individual and total AA from omasal bacteria and protozoa isolates from trial B [values are presented as mean (SEM)]
      ItemBacteriaProtozoa
      h (h−1)l (h−1)h (h−1)l (h−1)
      EAA
       Arg0.414 (0.045)0.00039 (0.00078)0.303 (0.027)−0.00025 (0.00001)
       His0.577 (0.222)−0.00027 (0.00048)0.409 (0.110)0.00026 (0.00038)
       Ile0.107 (0.037)−0.00054 (0.00080)0.106 (0.007)−0.00122 (0.00034)
       Leu0.323 (0.019)−0.00037 (0.00086)0.217 (0.002)−0.00162 (0.00060)
       Lys0.421 (0.017)−0.00012 (0.00034)0.333 (0.026)−0.00016 (0.00051)
       Met0.234 (0.079)−0.00097 (0.00008)0.198 (0.100)−0.00112 (0.00041)
       Phe0.782 (0.193)−0.00050 (0.00108)0.529 (0.286)−0.00052 (0.00086)
       Thr0.323 (0.056)0.00077 (0.00007)0.318 (0.030)0.00015 (0.00042)
       Trp0.283 (0.196)0.00054 (0.00090)0.332 (0.303)−0.00072 (0.00122)
       Val0.312 (0.054)−0.00054 (0.00055)0.169 (0.021)−0.00153 (0.00003)
       Total EAA0.303 (0.009)−0.00038 (0.00022)0.287 (0.002)−0.00084 (0.00033)
      NEAA
       Ala0.571 (0.183)0.00001 (0.00063)0.658 (0.132)−0.00156 (0.00100)
       Asp0.523 (0.057)0.00014 (0.00039)0.548 (0.168)−0.00044 (0.00077)
       Cys0.377 (0.035)−0.00035 (0.00005)0.429 (0.125)−0.00070 (0.00023)
       Glu0.550 (0.027)−0.00019 (0.00005)0.397 (0.060)−0.00048 (0.00118)
       Gly0.599 (0.174)−0.00049 (0.00020)0.541 (0.030)−0.00081 (0.00009)
       Pro0.324 (0.115)0.00018 (0.00021)0.304 (0.014)−0.00081 (0.00045)
       Ser0.622 (0.097)0.00218 (0.00049)0.418 (0.011)0.00198 (0.00079)
       Tyr0.804 (0.254)0.00036 (0.00024)0.328 (0.476)−0.00046 (0.00038)
       Total NEAA0.568 (0.112)0.00014 (0.00024)0.447 (0.107)−0.00045 (0.00075)
      Total AA0.415 (0.048)−0.00018 (0.00021)0.357 (0.037)−0.00067 (0.00054)
      1 Proportion of bound AA hydrolyzed per hour.
      2 Proportion of bound AA destroyed per hour.
      The use of multiple hydrolysis times provides some insight into the appropriateness of single time point hydrolysis for AA in rumen microbial samples. Although both techniques are simply estimates of the theoretical unknown true AA composition, the regression method has been shown to more accurately estimate the true AA profile in purified proteins (
      • Darragh A.J.
      • Garrick D.J.
      • Moughan P.J.
      • Hendriks W.H.
      Correction for amino acid loss during acid hydrolysis of a purified protein.
      ). The AA profile determined from the regression compared with the value determined at 24 h was used to establish the equivalency of the 2 methods in relation to biologically relevant ranges (Table 6). This alternative framework of hypothesis testing requires thoughtful interpretation of the results. Whereas some AA may exhibit negligible mean differences between analysis method, such as His and Thr, the interpretation of the 90% confidence interval indicates that they are not equivalent, as the confidence interval lies outside the predetermined rage of biologically relevant differences. Of the bacterial AA, the 24-h time point method was determined to be not equivalent to the multiple time point hydrolysis method for every AA except Gly. The 90% confidence interval of the mean difference was greater than ± 1 g/100 g of AA for Ile, Leu, Met, and Val (Figure 5). The relatively large underestimation of Ile, Met, and Val results in an overestimation of approximately 5% for the rapidly hydrolyzed AA, such as Arg, Leu, and Lys. This is similar to the results of
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      , where soybean meal Ile content was underestimated by 8.4%, followed by Val (7.0%), Ser (4.6%), and Thr (4.3%). The relatively low range in acceptable equivalence (mean difference of −0.4 to 0.4 g/100 g of AA for bacteria) serves to emphasize the importance of the AA profile of bacteria on AA supply determinations.
      Table 6Comparison of the AA composition (g/100 g of AA) of omasal bacteria from trial B
      Trial B: A. Foskolos (unpublished data). n = 2 for all comparisons.
      determined using multiple hydrolysis time point or single hydrolysis time point methods
      AAMethodSingle − multipleSED
      SED = standard error of the difference.
      90% CIEQ
      Equivalence determined from 2 one-sided paired t-tests. Methods deemed to be equivalent if 90% CI falls within defined equivalency of −0.4 to 0.4 g/100 g of AA.
      SingleMultipleLowerUpper
      EAA
       Arg5.004.730.270.030.060.48No
       His2.122.110.010.14−0.850.86No
       Ile4.054.62−0.580.46−3.462.31No
       Leu5.605.320.280.26−1.351.91No
       Lys7.547.170.370.040.110.63No
       Met4.494.63−0.140.36−2.412.13No
       Phe6.005.770.230.09−0.310.77No
       Thr5.495.53−0.040.10−0.690.60No
       Trp5.975.770.200.03−0.010.41No
       Val5.926.32−0.410.32−2.411.60No
      NEAA
       Ala6.886.94−0.060.21−1.361.24No
       Asp11.0610.790.280.030.050.50No
       Cys1.451.400.050.07−0.400.51No
       Glu11.2611.060.210.58−3.453.86No
       Gly4.864.830.020.02−0.100.15Yes
       Pro2.031.910.120.16−0.891.12No
       Ser4.584.89−0.310.44−3.062.44No
       Tyr5.715.650.060.13−0.760.89No
      1 Trial B: A. Foskolos (unpublished data). n = 2 for all comparisons.
      2 SED = standard error of the difference.
      3 Equivalence determined from 2 one-sided paired t-tests. Methods deemed to be equivalent if 90% CI falls within defined equivalency of −0.4 to 0.4 g/100 g of AA.
      Figure thumbnail gr5
      Figure 5Equivalency chart for difference in AA composition (g/100 g of AA) of trial B bacteria determined using multiple versus single time point hydrolysis methods. Open circles () represent the mean difference and error bars represent the 90% CI around the mean difference. Shaded region represents equivalency defined as −0.4 to 0.4 g/100 g of bacterial AA.
      Protozoa AA determinations between methods showed more general agreement between hydrolysis methods, largely due to the greater range in equivalence limits (mean difference of −1.5 to 1.5 g/100 g of AA for protozoa). Six of the 10 EAA and 6 of the 8 NEAA were deemed equivalent between methods (Table 7). Similar to the bacterial results, Ile and Met were underestimated (13.4 and 16.5%, respectively; Figure 6) when determined with a single time point hydrolysis, resulting in over estimation of several other AA, namely Lys and Asp.
      Table 7Comparison of the AA composition (g/100 g of AA) of omasal protozoa from trial B
      Trial B: A. Foskolos (unpublished data). n = 2 for all comparisons.
      determined using multiple versus single time point hydrolysis methods
      AAMethodSingle − multipleSED
      SED = standard error of the difference.
      90% CIEQ
      Equivalence determined from 2 one-sided paired t-tests. Methods deemed to be equivalent if 90% CI falls within defined equivalency of −1.5 to 1.5 g/100 g of AA.
      SingleMultipleLowerUpper
      EAA
       Arg5.355.260.090.15−0.841.03Yes
       His2.532.520.010.01−0.030.05Yes
       Ile3.804.39−0.590.06−0.94−0.24Yes
       Leu6.116.25−0.140.41−2.702.42No
       Lys8.818.550.260.06−0.100.62Yes
       Met3.143.77−0.620.47−3.582.34No
       Phe6.496.58−0.080.24−1.611.45No
       Thr5.415.340.070.03−0.130.26Yes
       Trp4.764.95−0.190.27−1.901.52No
       Val4.654.75−0.100.04−0.380.18Yes
      NEAA
       Ala5.145.030.110.07−0.310.53Yes
       Asp10.8510.140.710.47−2.273.69No
       Cys2.062.16−0.090.01−0.15−0.04Yes
       Glu13.5613.030.520.040.300.74Yes
       Gly4.484.410.070.010.030.12Yes
       Pro2.992.890.100.10−0.520.71Yes
       Ser5.365.280.080.03−0.140.29Yes
       Tyr4.504.70−0.200.34−2.321.92No
      1 Trial B: A. Foskolos (unpublished data). n = 2 for all comparisons.
      2 SED = standard error of the difference.
      3 Equivalence determined from 2 one-sided paired t-tests. Methods deemed to be equivalent if 90% CI falls within defined equivalency of −1.5 to 1.5 g/100 g of AA.
      Figure thumbnail gr6
      Figure 6Equivalency chart for difference in AA composition (g/100 g of AA) of trial B protozoa determined using multiple versus single time point hydrolysis methods. Open circles () represent the mean difference and error bars represent the 90% CI around the mean difference. Shaded region represents equivalency defined as −1.5 to 1.5 g/100 g of protozoal AA.

      Implications for AA Predictions in Mathematical Nutritional Models

      The nonequivalence of the determination methods are important to consider when developing models that rely on AA profiles of microbial protein and feedstuffs. The results from our study and the
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      data indicate that specific AA, especially Ile, Leu, Met, and Val, could be underestimated in many postruminal AA flow studies when utilizing single time point hydrolysis between 21 and 24 h. This consideration should recognized when literature values for AA are used in development and evaluation of nutritional models that seek to accurately predict AA supply, especially those that use mechanistic postabsorptive submodels. For example, in this analysis Met was determined to contribute more to total AA than has previously been reported. Currently, the CNCPS v.6.55 uses a profile that corresponds to approximately 1.2% of microbial AA as Met (
      • Higgs R.J.
      • Chase L.
      • Ross D.
      • Van Amburgh M.
      Updating the Cornell Net Carbohydrate and Protein System feed library and analyzing model sensitivity to feed inputs.
      ;
      • Van Amburgh M.E.
      • Collao-Saenz E.
      • Higgs R.
      • Ross D.
      • Recktenwald E.
      • Raffrenato E.
      • Chase L.
      • Overton T.
      • Mills J.
      • Foskolos A.
      The Cornell Net Carbohydrate and Protein System: Updates to the model and evaluation of version 6.5.
      ). Compared with the current analysis (4.7% of total AA), predictions of AA supply from the model would be expected to increase more than 2 fold (assuming microbial AA accounts for 50% of total AA). Adoption of these values will likely result in a re-evaluation of many common ratios and relationships currently used to balance EAA for lactating cattle. Given the data presented here and by
      • Rutherfurd S.M.
      Accurate determination of the amino acid content of selected feedstuffs.
      , this might also be true for many of the EAA. The current data, especially regarding the branched-chain AA, would help explain the prediction bias for those AA observed in CNCPS v.7 despite the relatively good prediction of NAN (
      • Higgs R.
      ; R. J. Higgs, Cornell University, current address Ohaupo, New Zealand, unpublished data). Overall, this analysis illustrates how sensitive nutritional models that rely on microbial AA profiles could be to errors in AA analysis, especially when a single profile accounts for a large portion of the predicted AA supply. Additionally, future studies should evaluate the use of formalin as a microbial preservative if AA analysis or digestibility is considered as an outcome. Model developers should not include any data from procedures that use formalin as a microbial preservative, as it will likely lead to biases and poor model evaluation.

      CONCLUSIONS

      Microbial composition and digestibility of individual AA are very important for the accurate predictions in many nutrition models used to feed dairy cattle. Previous literature reports have used techniques that have known limitations, whereas new procedures developed to address these limitations might provide better estimations of key parameters needed to properly characterize metabolizable AA supply. Multiple time point hydrolysis has been shown to improve the determination of AA in feeds, and the technique has been applied here. Digestibility of AA in microbial isolations were measured using a new in vitro technique. Although in vivo techniques might better represent the bio-physical condition of the ruminant animal, they can be expensive and rely on surgically altered animals. In vitro digestibility estimates were similar to those reported using previous techniques, and the assay was sensitive to formalin treatment, a process known to reduce bioavailability of proteins. Methionine and branched-chain AA concentration of microbial isolates was higher than previously reported, and this might have implications for predicting AA supply in cattle.

      ACKNOWLEDGMENTS

      We gratefully acknowledge and appreciate the financial support of Arm & Hammer Animal Nutrition, Princeton, New Jersey, for Sam Fessenden's doctoral program, and especially Elliot Block.

      REFERENCES

        • AOAC International
        Official Methods of Analysis. 18th ed. AOAC International, Gaithersburg, MD2005
        • Barry T.N.
        The effectiveness of formaldehyde treatment in protecting dietary protein from rumen microbial degradation.
        Proc. Nutr. Soc. 1976; 35 (987587): 221-229
        • Boucher S.E.
        • Calsamiglia S.
        • Parsons C.M.
        • Stein H.H.
        • Stern M.D.
        • Erickson P.S.
        • Utterback P.L.
        • Schwab C.G.
        Intestinal digestibility of amino acids in rumen undegradable protein estimated using a precision-fed cecectomized rooster bioassay: I. Soybean meal and SoyPlus.
        J. Dairy Sci. 2009; 92 (19700710): 4489-4498
        • Brito A.F.
        • Broderick G.A.
        • Reynal S.M.
        Effect of varying dietary ratios of alfalfa silage to corn silage on omasal flow and microbial protein synthesis in dairy cows.
        J. Dairy Sci. 2006; 89 (16960069): 3939-3953
        • Brito A.F.
        • Broderick G.A.
        • Reynal S.M.
        Effects of different protein supplements on omasal nutrient flow and microbial protein synthesis in lactating dairy cows.
        J. Dairy Sci. 2007; 90: 1828-1841
        • Chiquette J.
        • Benchaar C.
        Effect of diet and probiotic addition on chemical composition of free or particle-associated bacterial populations of the rumen.
        Can. J. Anim. Sci. 1998; 78: 115-120
        • Clark J.H.
        • Klusmeyer T.H.
        • Cameron M.R.
        Microbial protein synthesis and flows of nitrogen fractions to the duodenum of dairy cows.
        J. Dairy Sci. 1992; 75 (1401380): 2304-2323
        • Darragh A.J.
        • Garrick D.J.
        • Moughan P.J.
        • Hendriks W.H.
        Correction for amino acid loss during acid hydrolysis of a purified protein.
        Anal. Biochem. 1996; 236 (8660495): 199-207
        • Darragh A.J.
        • Moughan P.J.
        The amino acid composition of human milk corrected for amino acid digestibility.
        Br. J. Nutr. 1998; 80 (9797640): 25-34
        • Darragh A.J.
        • Moughan P.J.
        The effect of hydrolysis time on amino acid analysis.
        J. AOAC Int. 2005; 88 (16001867): 888-893
        • Denton B.L.
        • Diese L.E.
        • Firkins J.L.
        • Hackmann T.J.
        Accumulation of reserve carbohydrate by rumen protozoa and bacteria in competition for glucose.
        Appl. Environ. Microbiol. 2015; 81 (25548053): 1832-1838
        • Dijkstra J.
        • France J.
        • Tamminga S.
        Quantification of the recycling of microbial nitrogen in the rumen using a mechanistic model of rumen fermentation processes.
        J. Agric. Sci. 1998; 130: 81-94
        • Elkin R.G.
        • Griffith J.
        Hydrolysate preparation for analysis of amino acids in sorghum grains: Effect of oxidative pretreatment.
        J. Assoc. Off. Anal. Chem. 1985; 68 (4086434): 1117-1121
        • Fessenden S.W.
        Amino acid supply in dairy cattle. PhD Dissertation. Animal Science Department, Cornell Univ., Ithaca, NY2016
        • Fonseca A.C.
        • Fredin S.
        • Ferraretto L.
        • Parsons C.
        • Utterback P.
        • Shaver R.
        Short communication: Intestinal digestibility of amino acids in fluid-and particle-associated rumen bacteria determined using a precision-fed cecectomized rooster bioassay.
        J. Dairy Sci. 2014; 97 (24746133): 3855-3859
        • Gargallo S.
        • Calsamiglia S.
        • Ferret A.
        Technical note: A modified three-step in vitro procedure to determine intestinal digestion of proteins.
        J. Anim. Sci. 2006; 84 (16864878): 2163-2167
        • Gehrke C.W.
        • Wall Sr., L.
        • Absheer J.
        • Kaiser F.
        • Zumwalt R.
        Sample preparation for chromatography of amino acids: Acid hydrolysis of proteins.
        J. Assoc. Off. Anal. Chem. 1985; 68: 811-821
        • Guerino F.
        • Baumrucker C.R.
        Methionine and lysine uptake by cattle small intestine in vitro.
        J. Anim. Sci. 1987; 65 (3114208): 619-629
        • Hendriks W.
        • Tarttelin M.
        • Moughan P.
        The amino acid composition of cat (Felis catus) hair.
        Anim. Sci. 1998; 67: 165-170
        • Higgs R.
        Development of a dynamic rumen and gastro-intestinal model in the Cornell Net Carbohydrate and Protein System to predict the nutrient supply and requirements of dairy cattle. PhD Dissertation. Department of Animal Science, Cornell Univ., Ithaca, NY2014
        • Higgs R.J.
        • Chase L.
        • Ross D.
        • Van Amburgh M.
        Updating the Cornell Net Carbohydrate and Protein System feed library and analyzing model sensitivity to feed inputs.
        J. Dairy Sci. 2015; 98 (26142848): 6340-6360
        • Huhtanen P.
        • Brotz P.G.
        • Satter L.D.
        Omasal sampling technique for assessing fermentative digestion in the forestomach of dairy cows.
        J. Anim. Sci. 1997; 75 (9159288): 1380-1392
        • Hvelplund T.
        • Hesselholt M.
        Digestibility of individual amino acids in rumen microbial protein and undegraded dietary protein in the small intestine of sheep.
        Acta Agric. Scand. 1987; 37: 469-477
        • Hvelplund T.
        • Weisbjerg M.R.
        • Andersen L.S.
        Estimation of the true digestibility of rumen undegraded dietary protein in the small intestine of ruminants by the mobile bag technique.
        Acta Agric. Scand. Anim. Sci. 1992; 42: 34-39
        • Illg D.
        • Stern M.
        In vitro and in vivo comparisons of diaminopimelic acid and purines for estimating protein synthesis in the rumen.
        Anim. Feed Sci. Technol. 1994; 48: 49-55
        • Keller P.J.
        Pancreatic proteolytic enzymes.
        in: Handbook of Physiology. Vol. 122. American Physiology Society, Washington, DC1968: 2605-2628
        • Korhonen M.
        • Ahvenjärvi S.
        • Vanhatalo A.
        • Huhtanen P.
        Supplementing barley or rapeseed meal to dairy cows fed grass-red clover silage: II. Amino acid profile of microbial fractions.
        J. Anim. Sci. 2002; 80 (12211389): 2188-2196
        • Landry J.
        • Delhaye S.
        Simplified procedure for the determination of tryptophan of foods and feedstuffs from barytic hydrolysis.
        J. Agric. Food Chem. 1992; 40: 776-779
        • Martin C.
        • Bernard L.
        • Michalet-Doreau B.
        Influence of sampling time and diet on amino acid composition of protozoal and bacterial fractions from bovine ruminal contents.
        J. Anim. Sci. 1996; 74 (8726749): 1157-1163
        • Martin C.
        • Williams A.G.
        • Michalet-Doreau B.
        Isolation and characteristics of the protozoal and bacterial fractions from bovine ruminal contents.
        J. Anim. Sci. 1994; 72 (7730192): 2962-2968
        • Martín-Orúe S.M.
        • Balcells J.
        • Zakraoui F.
        • Castrillo C.
        Quantification and chemical composition of mixed bacteria harvested from solid fractions of rumen digesta: effect of detachment procedure.
        Anim. Feed Sci. Technol. 1998; 71: 269-282
        • Mason V.C.
        • Bech-Andersen S.
        • Rudemo M.
        Hydrolysate preparation for amino acid determinations in feed constituents.
        Z. Tierphysiol. Tierernahr. Futtermittelkd. 1980; 43 (7386051): 35-48
        • NRC
        Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC2001
        • Recktenwald E.B.
        Urea N recycling and its utilization by ruminal microbial populations in lactating dairy cattle. PhD Dissertation. Department of Animal Science, Cornell Univ., Ithaca, NY2010
        • Recktenwald E.B.
        • Ross D.A.
        • Fessenden S.W.
        • Wall C.J.
        • Van Amburgh M.E.
        Urea-N recycling in lactating dairy cows fed diets with 2 different levels of dietary crude protein and starch with or without monensin.
        J. Dairy Sci. 2014; 97 (24377801): 1611-1622
        • Reynal S.M.
        • Broderick G.A.
        Effect of dietary level of rumen-degraded protein on production and nitrogen metabolism in lactating dairy cows.
        J. Dairy Sci. 2005; 88 (16230710): 4045-4064
        • Reynal S.M.
        • Broderick G.A.
        • Ahvenjärvi S.
        • Huhtanen P.
        Effect of feeding protein supplements of differing degradability on omasal flow of microbial and undegraded protein.
        J. Dairy Sci. 2003; 86 (12741554): 1292-1305
        • Reynal S.M.
        • Ipharraguerre I.R.
        • Lineiro M.
        • Brito A.F.
        • Broderick G.A.
        • Clark J.H.
        Omasal flow of soluble proteins, peptides, and free amino acids in dairy cows fed diets supplemented with proteins of varying ruminal degradabilities.
        J. Dairy Sci. 2007; 90 (17369230): 1887-1903
        • Ross D.A.
        Amino acid composition of ruminant feeds and feed fractions and evaluation of the methods used to obtain the insoluble and true precipitate protein fractions of feedstuffs. MS Thesis. Department of Animal Science, Cornell University, Ithaca, NY2004
        • Ross D.A.
        Methods to analyze feeds for nitrogen fractions and digestibility for ruminants with application for the CNCPS. PhD dissertation. Animal Science Department, Cornell University, Ithaca, NY2013
        • Ross D.A.
        • Gutierrez-Botero M.
        • Van Amburgh M.E.
        Development of an in vitro intestinal digestibility assay for ruminant feeds.
        in: Proc. Cornell Nutrition Conference, Syracuse, NY. Cornell University, Ithaca, NY2013: 190-202
        • Rossi M.F.
        • Martinele I.
        • D'Agosto M.
        Quantitative and differential analysis of ciliate protozoa in rumen content samples filtered before and after fixation.
        Rev. Bras. Zootec. 2013; 42: 831-834
        • Rutherfurd S.M.
        Accurate determination of the amino acid content of selected feedstuffs.
        Int. J. Food Sci. Nutr. 2009; 60 (18946798): 53-62
        • Rutherfurd S.M.
        • Moughan P.J.
        • Lowry D.
        • Prosser C.G.
        Amino acid composition determined using multiple hydrolysis times for three goat milk formulations.
        Int. J. Food Sci. Nutr. 2008; 59 (18608544): 679-690
        • Storm E.
        • Brown D.S.
        • Ørskov E.R.
        The nutritive value of rumen micro-organisms in ruminants 3. The digestion of microbial amino and nucleic acids in, and losses of endogenous nitrogen from, the small intestine of sheep.
        Br. J. Nutr. 1983; 50 (6193806): 479-485
        • Sylvester J.T.
        • Karnati S.K.
        • Yu Z.
        • Morrison M.
        • Firkins J.L.
        Development of an assay to quantify rumen ciliate protozoal biomass in cows using real-time PCR.
        J. Nutr. 2004; 134 (15570040): 3378-3384
        • Sylvester J.T.
        • Karnati S.K.R.
        • Yu Z.
        • Newbold C.J.
        • Firkins J.L.
        Evaluation of a real-time PCR assay quantifying the ruminal pool size and duodenal flow of protozoal nitrogen.
        J. Dairy Sci. 2005; 88 (15905439): 2083-2095
        • Tas M.V.
        • Evans R.
        • Axford R.
        The digestibility of amino acids in the small intestine of the sheep.
        Br. J. Nutr. 1981; 45 (7470432): 167-174
        • Titgemeyer E.
        • Merchen N.
        • Han Y.
        • Parsons C.
        • Baker D.
        Assessment of intestinal amino acid availability in cattle by use of the precision-fed cecectomized rooster assay.
        J. Dairy Sci. 1990; 73 (2341644): 690-693
        • Van Amburgh M.E.
        • Collao-Saenz E.
        • Higgs R.
        • Ross D.
        • Recktenwald E.
        • Raffrenato E.
        • Chase L.
        • Overton T.
        • Mills J.
        • Foskolos A.
        The Cornell Net Carbohydrate and Protein System: Updates to the model and evaluation of version 6.5.
        J. Dairy Sci. 2015; 98 (26142847): 6361-6380
        • Van Soest P.J.
        The Detergent System for Analysis of Foods and Feeds. Cornell University, Ithaca, NY2015
        • Volden H.
        • Harstad O.M.
        Amino acid composition of bacteria harvested from the rumen of dairy cows fed three diets differing in protein content and rumen protein degradability at two levels of intake.
        Acta Agric. Scand. Anim. Sci. 1998; 48: 210-215
        • Volden H.
        • Harstad O.M.
        • Mydland L.T.
        Amino acid content and profile of protozoal and bacterial fractions isolated from ruminal contents of lactating dairy cows fed diets differing in nitrogen supplementation.
        Acta Agric. Scand. Anim. Sci. 1999; 49: 245-250
        • Volden H.
        • Mydland L.T.
        • Harstad O.M.
        Chemical composition of protozoal and bacterial fractions isolated from ruminal contents of dairy cows fed diets differing in nitrogen supplementation.
        Acta Agric. Scand. Anim. Sci. 1999; 49: 235-244
        • Whitehouse N.L.
        • Olson V.
        • Schwab C.
        • Chesbrot W.
        • Cunningham K.
        • Lykos T.
        Improved techniques for dissociating particle-associated mixed ruminal microorganisms from ruminal digesta solids.
        J. Anim. Sci. 1994; 72 (8056682): 1335-1343
        • Whitehouse N.L.
        Using the plasma free amino acid dose response method to determine metabolizable protein concentrations of lysine and methionine in rumen protected supplements. PhD dissertation. Department of Animal and Nutritional Sciences, Univ. New Hampshire, 2016
        • Williams A.
        • Coleman G.
        The Rumen Protozoa. Springer-Verlag, New York, NY1992