Butanoic Acid Isopropyl Ester and Methionine

Home , Semantides

Evaluation of 2-Hydroxy-4-(methylthio) Butanoic Acid Isopropyl Ester and Methionine

Supplementation on Efficiency of Microbial Protein Synthesis and Rumen Bacterial

Populations

Thesis

Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in

the Graduate School of The Ohio State University

Colleen Fowler, B.S.

Graduate Program in Animal Science

The Ohio State University

2009

Thesis Committee:

Dr. Jeffrey Firkins, Advisor

Dr. Zhongtang Yu

Dr. Thaddeus Ezeji

Abstract

Because microorganisms that grow in the rumen flow to the abomasum and are digested, the primary source of protein for milk production is of microbial origin, so increasing microbial protein supply should either increase milk protein or decrease the need for expensive bypass protein. Moreover, improving the consistency of prediction of microbial protein supply decreases the amount of protein that must be fed to the dairy cow, reducing urinary N excretion and decreasing feed costs. Methionine is co-limiting with lysine in most dairy diets. 2-hydroxy-4-(methylthio) butanoic acid (HMBi) is a methionine precursor for bacteria and might be degraded more slowly and, therefore, might be sustained longer over the feeding cycle for bacterial uptake or absorption than methionine itself. HMBi has been shown to increase nitrogen efficiency, expressed as the proportion of ingested nitrogen secreted in milk, when supplemented to dairy cows, but its mode of action (rumen microbial protein synthesis, delayed absorption, or both) has not been ascertained. To show the effect of HMBi and methionine on bacterial population structure and protein synthesis, we used four continuous culture fermenters inoculated with rumen fluid from two lactating Holstein cows. We fed the cultures a normal dairy ration with a 50/50 ratio of concentrate and alfalfa. Our experimental design was a 4x4 Latin square. Our treatments were control (CON), 0.11% HMBi

(HMBi), 0.097% methionine (MET), and 0.055% HMBi plus 0.048% methionine (HMBi

+ MET). All doses were on an equivalent methionine molar basis. We dosed the stable

isotopes 1-13C-methionine, 3D-methionine (methyl hydrogens labeled with deuterium)

and U-13C-HMBi into the fermenters at six consecutive feedings to trace the

incorporation of the carbon skeletons into microbial protein and methylation of substrates

15 by methionine. We also infused NH3-N with the buffer during each period to measure

efficiency of microbial protein synthesis. We fed the fermenters thrice per day, with the

treatments and isotope dissolved in liquid and dosed directly into the fermenter contents

(the control received an equivalent amount of water). The data were compared using

three orthogonal contrasts, all methionine treatments versus control and the linear and

quadratic responses to methionine substitution by HMBi. Supplementation of methionine

and HMBi had no effect on digestibilities of ADF and true OM. NDF and hemicellulose

digestibility were linearly affected (P=0.04 and P<0.01, respectively) by supplementation. The flow of non-ammonia N, bacterial N, total N, and non-ammonia, non-bacterial N were not affected. Ammonia N flow tended (P=0.08) to be affected linearly. Concentration of NH3-N tended to be affected linearly (P=0.07) and

quadratically (P=0.08). Peptide N was affected linearly (P=0.04) and tended to be

affected quadratically (P=0.09). The proportion of bacterial N from NH3-N was affected linearly (P=0.02). Propionate production was affected linearly (P<0.01) and quadratically (P=0.05), and control was greater than the average of the three methionine

treatments (P=0.01). Isobutyrate was affected linearly (P=0.05). Isovalerate tended to be

iii

affected linearly (P=0.08). Valerate was affected linearly (P<0.01) and quadratically

(P=0.05), and control was less (P=0.01) than the average of the three methionine treatments. The production of total VFA was affected linearly (P=0.02). The acetate:propionate tended to be affected linearly (P=0.08) and quadratically (P=0.06).

The increase in hemicellulose digestibility due to methionine supplementation indicates a possible increase in activity of hemicellulose degrading bacteria. Though relative changes in population abundance were not detected by DGGE, these bacteria might increase activity to support bacterial growth that is aided by methionine. The changes in

VFA production also indicate a possible shift in activity of populations present in the rumen due to supplementation of methionine and HMBi.

This thesis is dedicated to my mother and husband who have always truly believed that I can do whatever it is that I want to do.

Acknowledgements

I would like to thank my family, who, through their support, made it possible for me to go to graduate school. I also want to thank my friends for listening to me through all the ups and downs.

Dr. Jeff Firkins gave me the opportunity to do something new. Without his trust that I could do it, I would never have tried. Dr. Zhongtang Yu was always patient while I worked through the problems that I encountered. I also want to thank Dr. Thaddeus Ezeji for agreeing to be on my advisory committee.

Without the help of Amanda Stalford, the best undergraduate intern ever, I may not have made it through the trial, and I really appreciate all her hard work. Also, without Dr. Sanjay Karnati’s help through the preparation, execution, and analysis of the data from the trial, I know that I would not have made it. The people in the microbiology lab, especially Mike Nelson and Jill Stiverson, helped me trudge through the molecular work when I did not think that I could ever do it. Lastly, I want to thank all my fellow graduate students for their support and help.

Vita

September 29, 1984 …………………………………………………Born - Springfield,

Ohio

August 2003 – May 2005……………………………………………Student

University of Kentucky

September 2005 – August 2007……………………………B.S. in Agriculture in Animal

Science

The Ohio State University

August 2007 – Present………………………………………… Graduate Research

Assistant

The Ohio State University

Field of Study

Major Field: Animal Science

Ruminant Nutrition

vii

Table of Contents

Abstract……………………………………………………………………………..ii

Dedication…………………………………………………………………………..v

Acknowledgements…………………………………………………………………vi

Vita………………………………………………………………………………….vii

List of Tables………………………………………………………………………...x

List of Figures………………………………………………………………………..xi

Chapter 1: Review of Literature……………………………………………………..1

Introduction………………………………………………………………….1

Biological Role of Methionine………………………………………………3

General Bacterial Biosynthesis and Uptake of Methionine………………....3

S-adenosylmethionine……………………………………………………….5

Bacterial Regulation of Methionine Biosynthesis and Uptake………………6

Metagenomics……………………………………………………………….10

Protein Nutrition in Dairy Cows……………………………………………..14

N recycling/Amino Acid Requirements…………………………………...... 15

Protein and Carbohydrate Effects on Growth for Rumen Bacteria………….17

AA Supplementation of Rumen Bacteria……………………………………19

Continuous Culture………………………………………………………….23

viii

Methionine and HMBi Supplementation to Dairy Cows……………….24

Chapter 2: Effects of 2-hydroxy-4-(methylthio) butanoic acid isopropyl ester and methionine supplementation on populations of rumen bacteria and efficiency of microbial protein synthesis………………………………………………………………..27

Introduction…………………………………………………………….27

Materials and Methods…………………………………………………29

Results and Discussion…………………………………………………43

Conclusions…………………………………………………………….51

References………………………………………………………………………53

List of Tables

Table 1. Ingredients of Diet…………………………………………………….36

Table 2. Nutrient Composition of Diet…………………………………………37

Table. 3 Nutrient digestibilities, nitrogen fluxes, and nitrogen partitioning for rumen bacteria in continuous cultures supplemented with HMBi,

DL-methionine or both………………………………………………………….38

Table 4. Volatile fatty acid production per day for rumen bacteria in continuous culture

supplemented with HMBi, DL-methionine or both…………………………….40

List of Figures

Figure 1. Comparison of the bacterial community structure from continuous culture samples………………………………………………………………………….42

List of Abbreviations

AA amino acid

ADF acid detergent fiber

CP crude protein

CsCl cesium chloride

CsTFA cesium trifloroacetic acid d day

DGGE denaturing gradient gel electrophoresis

DM dry matter

DMI dry matter intake

DNA deoxyribonucleic acid

EMPS efficiency of microbial protein synthesis h hour

HMB 2-hydroxy-4(methylthio) butanioc acid

HMBi 2-hydroxy-4(methylthio) butanioc acid isopropyl ester

MCP microbial crude protein

Met methionine

N nitrogen

NDF neutral detergent fiber

NAN non-ammonia nitrogen

xii

NANBN non-ammonia, non-bacterial N

OM organic matter

PCR polymerase chain reaction

RNA ribonucleic acid

SAS Statistical Analysis System

SEM standard error of the mean

VFA volatile fatty acid

xiii

Chapter 1: Review of Literature

Introduction

Methionine is an important molecule for eukaryotic and prokaryotic growth.

Bacteria use methionine for incorporation into protein and as a source of carbon and nitrogen. Also, because methionine contains sulfur, it can act to absorb the oxidation of free radicals through the production of sulfoxide when located on the exterior of proteins

(Brosnan et al., 2007). In eukaryotic and prokaryotic systems, methionine has an important role as S-adenosylmethionine, a highly active methylating agent, participating in the production of small molecules such as phosphatidylcholine and synthesis of polyamines (Chaing et al., 1996; Loenen, 2006).

Rumen bacteria are stimulated by amino acids and peptides (Atasoglu et al.,

1999), though the mechanisms have yet to be elucidated. Because methionine has functions beyond other amino acids as sources of nitrogen and carbon and protein structure, it might positively affect microbial growth through an increase in microbial protein production or efficiency. Bacteria can synthesize methionine through two main pathways using inorganic sulfur or cysteine and take up methionine through specific transporters or permeases (Rodionov et al., 2004; Sperandio et al., 2007). Methionine synthesis and uptake are highly regulated through riboswitches. Because the recycling of blood urea nitrogen protects the rumen environment from very low ammonia

concentrations (Firkins et al., 2007), rumen bacteria can produce all the necessary amino acids for milk production. The rumen bacteria can synthesize methionine specifically and incorporate it into bacterial protein (Or-Rashid et al., 2001). However, the supply might be insufficient because milk production from high producing dairy cows has been shown to be limited by methionine and lysine (Rulquin et al., 2006). Because methionine may stimulate growth of rumen bacteria and dairy cows might benefit from a postruminal supply, supplementation of methionine to dairy cows has been studied. In vivo and in vitro studies have shown the effects of methionine supplementation. In cows, methionine supplementation has increased milk protein yield (Rulquin and Delaby, 1997) and duodenal infusion increased milk fat and milk protein production (Socha et al., 2008). In batch culture, methionine increased bacterial growth (Gil et al., 1974b).

When methionine is degraded by rumen microorganisms for carbon and nitrogen, the

amount that escapes the rumen might not be adequate for milk production. Therefore, analogs

have been produced to decrease degradation, increasing the supply of methionine to the cow and

lengthening the supply of methionine to the rumen environment. 2-hydroxy-4-(methylthio)-

butanoic acid (HMB) and its isopropyl ester (HMBi) have been designed for this purpose. HMBi has been shown in cows to increase milk, milk fat, protein yields and true protein percentage when fed a diet deficient in methionine (St-Pierre and Sylvester, 2005; Rulquin et al., 2006). In both studies, HMB did not have any increases in milk production or components. In continuous culture, HMB was shown to decrease the proportion of microbial nitrogen originating from ammonia-N by about 13% (Noftsger et al., 2003), indicating that HMB is not simply a methionine precursor. HMBi, which has been shown to be 50% degraded in the rumen (Robert

et al., Schwab et al., 2001), might be degraded more slowly by the rumen bacteria and provide a

steadier supply of HMB to the bacterial populations to positively affect growth.

Biological Role of Methionine

Methionine is one of the two sulfur containing amino acids that are incorporated into

proteins. Methionine plays an important role in protein structure and function. Due to the

terminal methyl group, the side chain of methionine is hydrophobic (Brosnan et al., 2007).

Therefore, methionine is often in the hydrophobic interior of globular proteins. For prokaryotic

proteins, the methionine residues that are found on the exterior of a protein are susceptible to

oxidation of the sulfur atom to sulfoxide by reactive oxygen species. Some research has shown

that the exterior methionine residues may protect the catalytic site of proteins from oxidation.

Also, researchers have found evidence of an oxidation-reduction cycle through the action of

methionine sulfoxide reductase. Methionine also plays a very important role as the metabolite

N-formylmethionine, which serves as the initiating amino acid for protein synthesis in

prokaryotes. Methionine itself is the initiator in eukaryotic protein synthesis. The absence of

methionine has been shown to limit a ruminant’s growth and production, further illustrating its importance on a whole animal, production level (Or-Rashid et al., 2001).

General Bacterial Biosynthesis and Uptake of Methionine

Bacteria have two main pathways to synthesize methionine (Rodionov et al., 2004). The transsulfuration pathway, elucidated in Escherichia coli, uses cysteine and cystathionine as a source of sulfur and an intermediate, respectively. The direct sulfhydrylation pathway, first found in Leptospira meyeri, bypasses cystathionine and uses inorganic sulfur. The industrially

utilized Corynebacterium glutamicum uses both transsulfuration and direct sulfhydrylation

pathways (Hwang et al., 2002). The carbon source for the aspartate family, to which methionine,

threonine, asparagine, lysine, and isoleucine belong, is oxaloacetate (OAA) (Walker et al., 2005).

Although anaerobic bacteria generally lack all the enzymes for a complete TCA cycle, OAA can

flux both forward and backward in reference to the TCA cycle for generation of important

metabolic precursors. Methionine biosynthesis begins with homoserine from aspartate (Hwang

et al., 2002). In the direct sulfhydrylation pathway, homoserine receives an acyl group from

homoserine O-acetyltransferase (metB in enterobacteria, firmicutes, cyanobacteria; metX in proteobacteria, actinobacteria, fungi, and metA in Streptococcus mutans) to make O- acetylhomoserine (Rodionov et al., 2004; Sperandio et al., 2007). With inorganic sulfur in the form of sulfide, O-acetylhomoserine sulfhydrylase, metY (cysD in S. mutans), forms homocysteine. In the transsulfuration pathway, O-acetylhomoserine with cysteine through cystathionine γ-synthase (metB, metI) forms cystathionine. Cystathionine forms homocysteine through cystathionine β-lyase (metC). In the last step, methionine is synthesized from homocysteine by methionine synthase (metE). E. coli has two methionine synthases, a B12- dependent (metH) and a B12-independent (metE). The product of metH synthesizes methionine

100-fold faster than the product of metE (Rodionov et al., 2004). Methionine also has a salvage

pathway that recycles methylthioadenosine from polyamine synthesis back to methionine

through methylthioribose. In Bacillus subtilis, this pathway is encoded by the gene cluster

mtnKSUVWXYZ. The signaling molecule AI-2 is also a metabolite of methionine (Sperandio et

al., 2007). AI-2 is directly synthesized from S-ribosylhomocysteine by S-adenosylhomocysteine

nucleosidase (Pfs) and S-ribosylhomocysteinase (LuxS) in S. mutans. AI-2 is involved in cell-to-

cell intraspecies communication and gene regulation in Gram positive and Gram negative

bacteria.

Methionine can also be transported into the bacterial cell. The review by Rodionov et al.

(2004) discusses methionine metabolism and methionine metabolite transport systems in Gram

positive bacteria. A new family of the ABC superfamily has been identified for methionine

transport, and it contains metNPQ, a permease that transports methionine sulfoxide and both D- and L-methionine (B. subtillis), which is regulated by S boxes (Bacillales and clostridia), methionine specific T boxes (lactobacilli), or MET boxes (streptococci). These regulation systems are discussed later. In enteric bacteria, metNIQ codes for another methionine ABC transporter. The metT protein is similar to the NhaC N:H antiporter superfamily and is predicted to transport methionine. The protein from the gene mtnABC is a transporter that is assumed to have methylthioribose specificity. In E. coli, metD codes for a high affinity methionine transport system (Rodionov et al., 2004). In S. mutans, atmBDE codes for a protein that is a member of the permease transporter family and thus transports thiosulfate, homocysteine, and methionine

(Sperandio et al., 2007).

SAM

S-adenosylmethionine (SAM) is a metabolically important methionine metabolite that participates in many diverse biochemical reactions, making it one of the most often used substrates in biological systems after ATP (Loenen, 2006). It is synthesized from methionine by methionine adenosyltransferase (metK) (Rodionov et al., 2004). The (S, S) enantiomer is the biologically active form (Chaing et al., 1996). The (R, S) enantiomer inhibits methylases. SAM acts as the source of methyl groups for many methylation reactions (Brosnan et al., 2007). The

positively charged sulfonium ion of SAM transfers its methyl group to a nucleophilic acceptor,

forming S-adenosylhomocysteine (SAH) and a methylated acceptor. These methylation

reactions occur through the action of methyltransferases. The methyltransferases are then

inhibited by SAH. SAH is recycled back to homocysteine by the action of two enzymes

(Rodionov et al., 2004; Sperandio et al., 2007). S-adenosylhomocysteine nucleosidase (mtn, pfs) forms S-ribosylhomocysteine (SRH) from SAH, and S-ribosylhomocysteinase (luxS) forms homocysteine from SRH.

The methylation reactions that involve SAM are many and diverse. Some functions are the biosynthesis of small molecules such as phosphatidylcholine through three successive methylations of phosphatidyl-N-monomethylethanolamine (Chaing et al., 1996) and the modification of DNA, RNA, and proteins (Brosnan et al., 2007). Protein methylation can alter steric orientation, charge, and hydrophobicity, regulating the protein’s function (Chaing et al.,

1996). Carboxymethylation and demethylation of glutamate residues in receptor-transducer proteins are the primary mechanisms for signal adaptation in bacterial chemotaxis. Also, in E. coli, methylation of elongation factor Tu corresponds with the growth phase. In the logarithmic phase, Tu is singly methylated at Lys56. In the stationary phase, Lys56 is doubly methylated.

Different areas of the SAM molecule are used for different biochemical reactions.

Researchers have identified fifteen superfamilies of SAM-binding proteins (Loenen, 2006).

Decarboxylated SAM also enters the polyamine pathway for spermidine and spermine synthesis.

SAM can bind to RNA and, therefore, act as a regulator.

Bacterial Regulation of Methionine Biosynthesis and Uptake

Structural changes in the mRNA transcripts often regulate gene expression (Nudler and

Mironov, 2004). Specific proteins have been found that bind RNA in response to metabolite concentration and change its conformation. These proteins can either stabilize the RNA for translation, or they can alter the structure of the RNA directly or indirectly when the Shine-

Delgarno sequence is sequestered, or causing transcription or translation to be terminated.

Riboswitches are RNA sequences that bind small molecules, such as metabolites, without need of any intermediate proteins and allow the mRNA to sense metabolite concentrations (Loenen,

2006). Some of the first research to show the existence of riboswitches was work on B. subtilis vitamin metabolism (Nudler and Mironov, 2004). The untranslated leader region of the riboflavin operon has a conserved regulatory element, rfn. Mutations in the rfn-box stopped repression by addition of flavins and led to riboflavin overproduction. The researchers found no protein candidate to mediate repression in the presence of flavins. Therefore, they proposed that the flavins themselves interacted with the rfn-box through direct binding, changing the conformation of rfn, promoting a structure, termed the terminator structure that leads to the disruption or termination of transcription or translation. Metabolites can also bind to the RNA to form an antiterminator structure, which would allow gene expression. A third structure, the anti- antiterminator, occurs when the native structure of the RNA allows gene expression (the antiterminator) and binding of the metabolite at a distance from the site of the terminator structure or the ribosome binding site would again stop transcription or translation.

Riboswitches can also function as a sequestor, antisequestor, or anti-antisequestor in the same way as the terminator conformations (Nudler and Mironov, 2004). The metabolite can bind to the riboswitch sequence and cause a conformational change that hides, or sequesters, the

Shine-Delgarno sequence. In another case, the metabolite may bind and cause a conformational

change that reveals, or antisequesters, the Shine-Delgarno sequence. The anti-antisequestor

conformation occurs when the native RNA conformation is the antisequestor and the metabolite

binds and hides the sequence.

Amino acid metabolism has also been found to be controlled by riboswitches (Rodionov

et al., 2004). The S-box regulon is made up of at least 60 transcriptional units from many different bacterial species. The S-box is a conserved regulatory leader sequence. B. subtilis has

11 operons under S-box control. Staphylococcus aureus and Clostridium acetobutylicum also have S-box sequences preceding transcriptional units between a putative promoter and a terminator sequence, suggesting that S-box regulation is widespread in Gram positive bacteria

(Grundy and Henkin, 1998). Many of these S-box sequences precede genes involved in methionine and SAM biosynthesis. The leader sequence has an intrinsic transcription terminator and competing antiterminator (Nudler and Mironov, 2004). The S-box is an anti-antiterminator.

Through studies that over expressed methionine adenosyltransferase, increasing the production of SAM from methionine, it has been shown that SAM, not methionine, is the effector for these

S-box controlled genes. When SAM binds to the S-box, a downstream hairpin forms, prematurely terminating transcription (Rodionov et al., 2004). The S-box regulon contains genes for methionine biosynthesis and transport and for methionine adenosyltransferase (metK). S-box regulation is found in Bacillales and Clostridiales, though not in Lactobacillales. In B. subtilis,

Bacillus cereus, and Oceanobacillus iheyensis, the S-box controls the expression of genes for cysteine biosynthesis and methionine salvage pathways. In the Lactobacillales, Enterococcus,

Streptococcus, and Lactococcus, Fuchs et al. (2006) found a riboswitch that is responsive to

SAM. This regulatory sequence has been designated the SMK box. Elevated concentrations of

SAM altered the leader RNA structure, sequestering the Shine-Delgarno sequence, thereby

terminating translation. SAM binds specifically to RNA that contains sequences from position

15 to 118 relative to the predicted transcription start site. Fuchs et al. (2007) further showed that the 30S ribosomal subunit binds to the SMK box in E. coli. The addition of SAM specifically inhibits this binding, preventing initiation of translation.

In other Gram positive bacteria, biosynthetic and transport genes for methionine are regulated by methionine-specific T-boxes (Rodionov et al., 2004). T-box regulation is used extensively in Lactobacillales, exclusively controlling methionine biosynthesis and transport. In

Bacillales and Clostridiales, T-box regulation is used for methionyl tRNA synthetases

(Vitreschak et al., 2008). The 14 nucleotide region, designated the T-box, was first discovered in

B. subtilis (Grundy and Henkin, 2003). The genes following the conserved leader regions

responded to the absence of the cognate amino acid and not to general amino acid starvation,

demonstrating amino acid specificity in the regulation of those genes. The primary sequences

and secondary structures were highly conserved and preceded the T-box and intrinsic terminator

sequences. The T-box sequence binds uncharged tRNA, which acts as the effector, inducing

formation of the antiterminator structure (Rodionov et al., 2004). The T-box sequence has a

‘specifier codon’ that interacts with the anticodon of the uncharged tRNA. Determining this

specifier codon allows the prediction of the amino acid specificity of the regulatory signal. The

methionine specific T-box sequences have an AUG specifier codon that interacts with the

anticodon on the methionyl tRNA. Rodionov et al. (2004) found 30 methionine specific T-boxes

in the Lactobacillales group. In Firmicutes, the branched-chain and aromatic aminoacyl tRNA

synthetases are regulated by T-boxes. Conversely in Bacillales, most amino acid aminoacyl

tRNA synthetases have T-box regulation (Vitreschak et al., 2008).

A third type of regulation of the methionine biosynthetic and transport genes is found in streptococci (Rodionov et al., 2004). A 17-base-pair palindromic sequence, designated the MET

box, binds the SAM responsive repressor, MetJ in Gram positive bacteria. Sperandio et al.

(2007) describe MetR in S. mutans, a DNA binding element that is a Lys-R type transcriptional regulator that also binds to MET box sequences. MetR activates transcription by binding to the

MET boxes in the promoter regions of two gene clusters that code for genes involved in

methionine biosynthesis and uptake in S. mutans. Homocysteine acts as a coeffector and

increases affinity of MetR for DNA, thereby inducing MetR-dependent transcription. In E. coli,

the repressor MetJ, along with SAM as a corepressor, negatively regulates the methionine

biosynthetic and transport genes. Some of the same genes are positively regulated by MetR

(Sperandio et al., 2007).

Metagenomics

Gene detection is difficult in the rumen consortia. A small percentage of the bacteria

present in the rumen can be cultured, and, therefore, studied in vitro. Metagenomics is a way to

circumvent culture-dependent methods through study of the genome of the entire communities.

The procedure involves extraction of all the DNA or RNA of the entire community, cloning the

nucleic acid material into a suitable vector, transforming the clones into a host bacterium, and

screening the resultant colonies for transformants (Handelsman, 2004). Sequence-based analysis

involves the direct sequencing of clones containing phylogenetic anchors that lead to the

taxonomic group or by random sequencing to find a gene of interest and then its phylogenetic

anchor. Functional analysis includes identifying a clone that produces a certain function. To

rely on this analysis, the vector and host bacteria strains must be chosen carefully to ensure

transcription, translation and secretion of the gene product. The concept of metagenomics can be easily applied to the rumen environment. Procedures have been developed to extract total DNA and RNA from rumen contents (Yu and Mohn, 1999). Firkins et al. (2007) describes metagenomics as new avenues to approach current problems with the rumen microbial ecosystem. These metagenomic analyses have led to construction of archival libraries containing genetic material for functional gene discovery. The ability to retrieve this genetic information combined with the increased availability of DNA sequencing is leading to new understanding of microbial community functions. Also, the databases of semantides, often the gene for the small subunit ribosomal RNA, allows researchers to assess diversity in the rumen microbial communities.

One metagenomic technique is stable isotope probing (SIP). SIP can identify populations

of microorganisms that utilize a certain substrate. SIP has been used successfully as a

cultivation-independent method with environmental microorganisms. Substrates are labeled with

a stable isotope, often 13C or 15N. A sample of the environment with the microbial community is

then incubated with the labeled substrate. After an identified period of time that allows the label

to be incorporated into the nucleic acids of the active populations, all the RNA or DNA from the

population is extracted. When SIP was first used, phospholipid fatty acids (PLFA) were

extracted from the sample and analyzed using isotope ratio-mass spectrometry (IRMS).

Microorganisms often have unique PLFA, and the labeled PFLA could identify the active

population. Now the extracted nucleic acids are separated in a density gradient into ‘light’ and

‘heavy’ bands. The ‘heavy’ bands are made up of the nucleic acids that have been labeled with

the stable isotope. Semantides in the labeled nucleic acids can identify the active populations.

Cesium trifloroacetate and cesium chloride are used for RNA and DNA separation, respectively.

The nucleic acids are applied to the gradients and run in an ultracentrifuge for 36 to 65 hours.

The separated bands can be visualized by adding a nucleic acid dye, such as ethidium bromide, to the gradient. If the bands are not visualized, fractions of the gradient are taken and run on an agarose gel. Also, libraries can be constructed from the labeled nucleic acids, possibly identifying functional genes (Dumont and Murrell, 2005; Neufeld et al., Whiteley et al., 2007).

Some considerations for the use of SIP include which nucleic acids to extract, what substrates are suitable for this application, and how long to incubate the samples with the labeled substrates. RNA-SIP is more sensitive in identifying active populations if the incubation time or

the extent of isotopic labeling of the substrate is minimal (Neufeld et al., 2007). For DNA-SIP,

the incubation period may need to exceed several generations to label the active populations to

an extent that they will be detectable. Using DNA is an attractive option due to the relative ease

of DNA extraction as compared to RNA and the availability of fingerprinting techniques, such as

denaturing gradient gel electrophoresis (DGGE) and terminal restriction fragment length polymorphism (T-RFLP), which can identify potential shifts in groups among the light and heavy bands.

When using DNA-SIP, another concern is the difference in G+C content of different bacterial species. Variable G+C content in the DNA from members of environmental communities changes the buoyant density of the DNA (Rolfe and Meselson, 1959). The guanidine and cytosine base pairs have greater nitrogen content than adenine and thymine base pairs, causing the G+C pairs to be more dense, thereby increasing the density of DNA that has a high G+C content. Buckley et al. (2007) showed the buoyant density of 15N-labeled DNA from

E. coli (51% G+C content) was similar to unlabeled DNA from Pseudomonas aeruginosa (67%

G+C content). When separated in a CsCl gradient, labeled E. coli DNA and unlabeled P.

aeruginosa DNA were in the same ‘heavy’ fractions. The authors overcame this obstacle when

they applied the ‘heavy’ DNA to another CsCl gradient containing the intercalcating agent bis-

benzimide, which causes a decrease in buoyancy that is inversely proportional to DNA G+C

content. The second centrifugation successfully separated the labeled from the unlabeled DNA.

Some suitable substrates that have been demonstrated to achieve nucleic acid labeling are

CH4, methanol, CO2, and glucose (Brinkmann et al., 2008; Egert et al., Neufeld et al., 2007).

These substrates have been successful because the utilizing species metabolize the substrates to a high degree, or the incubation periods are extended (Neufeld et al., 2007). The length of the incubation period must also be carefully considered. If the time is too short, then not enough label will be incorporated into the biomarker. If the incubation is too long, there could be cross feeding of the label due to predation of the primary utilizers. When Egert et al. (2007) studied the consortia in a model of the human gastrointestinal tract, they stabilized the bacteria in the model for 16 hours and then starved them for two hours. The starvation period ensured that the glucose utilizing bacteria would take up the U-13C-glucose when it was added to the culture. The researchers also took samples over time to ensure they had a sample that incorporate the label into the nucleic acids.

Fingerprinting methods can follow the stable isotope probing procedures to identify population changes in the labeled and unlabeled nucleic acids, or be used independently. DGGE has been used successfully to identify changes in microbial populations due to nutritional treatments. Karnati et al. (2007) compared DGGE and ribosomal intergenic spacer length polymorphism (RIS-LP) and found that DGGE was a reliable method for analyzing changes in the microbial communities in digesta samples from dairy cows. Karnati et al. (2009) also used

DGGE to resolve protozoa and methanogens and to identify changes resulting from defaunation

and microbial inhibitors in a continuous culture system.

Protein Nutrition in Dairy Cows

Protein requirements for dairy cows have been studied consistently as understanding of nutrient requirements, microbial protein synthesis and rumen degradability of feedstuffs has improved. Clark (1975) reviewed several studies from the late 1960’s and early 1970’s showing that abomasal infusion of casein increased total milk and milk protein yield. Implementation of the metabolizable protein system in place of the crude protein system in dairy cow diets has improved dairy nutrition (Firkins et al., 2007). Because microbial crude protein (MCP) flows out of the rumen and to the abomasum, it is a major source of protein for milk production. The amino acid profile of rumen bacteria closely resembles milk’s amino acid profile. The efficiency of microbial protein synthesis (EMPS) can have a major effect on MCP yield and, therefore, milk production. However, it is difficult to quantify the flow of nitrogen fractions in the rumen to MCP that then flows to the duodenum and omasum. These nitrogen fractions in the rumen originate from rumen degraded protein (RDP) and nitrogen recycling. It has also been difficult in the past to assess the effect of protein and carbohydrate sources on the rumen bacteria.

In 1974, Broderick et al. found that methionine, lysine, and valine were the most likely limiting amino acids for milk production by measuring accumulation of amino acids in plasma when cows were received elevated amounts of casein post-ruminally. Using postruminal infusions of amino acids, Rulquin et al. (2006) show that methionine and lysine are the first two limiting amino acids for milk production in dairy cows. Schwab et al. (1992) constructed a study to evaluate lysine and methionine limitations in four stages of lactation with a diet consisting of

corn silage, haycrop silage, corn meal, soybean meal, dried distillers grains with soluble, and

wheat middlings. At peak lactation (week 4) infusion of methionine and lysine together had the

same protein production (g/day) as casein infusion, and infusion of methionine alone was less.

For early lactation (weeks 8 to 12), the results were similar to peak for protein percentage and

protein production. In midlactation (week 17 to 21), casein infusion and infusion of methionine

and lysine together were similar and higher than infusion of methionine and lysine separately for

protein percentage and protein production. The results for late lactation (week 27 to 31) only

showed treatment differences for protein percentage. Infusion of methionine and lysine together

was higher than either infused separately. Casein infusion was not different than any treatment.

This study provides evidence that methionine and lysine are the first two limiting amino acids in

all stages of lactation in situations similar to the study parameters.

In a study designed to assess ruminally protected lysine and methionine, Robinson et al.

(1998) suggest that methionine may enhance production of milk components beyond the role of a limiting amino acid due to the level of methionine in the diet more than meeting the 1989 NRC requirement and still showing a numerical, though not statistically significant, increase in milk protein and milk fat. The authors also note that methionine deficiencies have been suggested to affect milk fat synthesis because SAM is a methyl donor in transmethylation reactions in lipid biosynthesis.

N recycling/Amino Acid Requirements

Due to blood urea nitrogen (BUN) returning to the rumen, the rumen environment seems to be protected against very low ammonia concentrations (Firkins et al., 2007). This cycling of

BUN also indicates that amino-N precursors might become more limiting than ammonia-N in

cattle fed high amounts of rumen degraded starch. Carbohydrate supply also influences the

amount of ammonia-N assimilated into microbial protein, and there is evidence for an optimum range of rumen degraded starch above and below which can have detrimental effects on efficiency and total microbial protein synthesis.

Because most bacteria can synthesize the amino acids necessary for growth from carbon skeletons, the paradigm of amino acid requirements applied to animals is most likely not appropriate for rumen bacteria. Rumen bacteria can grow with only nonprotein nitrogen sources, such as ammonia (Kajikawa et al., 2002). However, amino acids, along with propionate, are gluconeogenic precursors and can affect both total milk and milk protein production in the mammary gland (Clark, 1975). Rumen bacteria also need only a source of sulfur to produce the sulfur containing amino acids, methionine and cysteine (Van Soest, 1994). They can use most forms of sulfur because sulfate is reduced through sulfite to sulfide. Rumen sulfide most often exists as hydrogen sulfide because its first pK is 6.7, which is higher than most rumen pH measurements. As a mineral, sulfur has significant interactions with copper and molybdenum.

Sulfur requirements can vary in ruminants. Sheep that grow large amounts of wool would need more sulfur than a dairy cow because the keratin that comprises the wool has more sulfur AA than milk protein. The requirement for sulfur can be expressed as a N:S, since most sulfur is delivered to the rumen microorganisms as RDP. Though an increased supply of urea would increase the need to evaluate N:S as an expression of requirement. Bacterial protein has a N:S of

13:1, but bacteria have nonprotein nitrogen in nucleic acids and cell wall components, so estimating the N:S that flows to the cow is difficult. Lactation as well as heavy wool growth may require more sulfur than is in the MCP.

Protein and Carbohydrate Effects on Growth of Rumen Bacteria

Rumen bacterial growth is dependent on carbohydrates and RDP as major nutrients.

Dairy diets can be highly variable for sources of both (Hoover and Stokes, 1991). Microbial

growth has been shown in vitro to be dependent first on energy available in the form of readily

fermentable carbohydrates (Russell et al., 1983). Previous research shows that supplementation

of protein can further increase growth, but to a much less extent. Without the addition of

carbohydrates to mixed rumen bacteria in defined media, the addition of casein caused no

measurable increase in growth and ammonia accumulated in the media. In another experiment,

researchers found an interaction between casein and carbohydrate addition, as they varied both and measured MPS. The addition of casein to the media with carbohydrates did significantly

increase MPS.

In a review, Hoover and Stokes (1991) discuss factors in carbohydrate and protein digestion and their effects on yield of MCP. Because pH and turnover rate have such a large effect on microbial protein yield, the authors discussed in vitro research using batch and continuous culture and compared the results to in vivo studies. Production of bacterial proteolytic enzymes is constitutive (Wallace and Brammell, 1985), and serine, cysteine, and metalloproteinases are present in mixed rumen bacteria (Brock et al., 1982). Protozoa and fungi both produce some extracellular proteases to varying degrees with different rates of hydrolysis

(Hoover and Stokes, 1991). It has also been shown that the rate of proteolysis is greater than peptide hydrolysis. Therefore, peptides accumulate in the rumen after feeding. The N sources for the rumen bacteria are peptides, AA, and ammonia, because peptides are quickly hydrolyzed

to AA and ammonia.

The hydrolysis of plant polysaccharides involves both polysaccharidases and glycoside

hydrolases that are produced by protozoa, bacteria, and fungi (Williams, 1988). In adherent and

nonadherent microbes, these enzymes have different specific activities (Williams and Strachan,

1984) and change postprandially (Williams et al., 1989). Studies have shown that some large

peptides are taken up more rapidly and used more efficiently than most small peptides and AA

(Hoover and Stokes, 1991). For different bacterial species, the maximum length for uptake of

large peptides may differ. Also, hydrophobic peptides are hydrolyzed more slowly than

hydrophilic peptides.

Protozoa excrete AA when they ingest proteins from casein and heat treated soy (Hino

and Russell, 1987). The predation of bacteria by protozoa has also been shown to recycle AA

and peptides (Hoover and Stokes, 1991). Protozoa seem to preferentially use bacteria as a protein source so the AA and polypeptides they excrete are primarily the same as bacterial protein (Firkins et al, 2007). Also, hyperammonia producing bacteria, which rapidly deaminate

AA to produce ammonia, play a significant role in N recycling in the rumen. Eschenlauer et al.,

(2002) cites many studies by Russell and colleagues describing the hyperammonia producing

bacteria. Peptostreptococcus anaerobius, Clostridium sticklandii, and C. aminophilum were isolated and able to grow on Trypticase, producing ammonia at levels higher than other rumen bacteria, considered high enough to make a contribution to the general rumen environment, even though the numbers of the hyperammonia producing bacteria in the rumen were not very high.

The action of these hyperammonia producing bacteria along with protozoal predation provide additional sources of N, AA, and peptides to the rumen environment for bacterial growth.

Most sources of oligosaccharides, such as starch, sucrose, cellobiose, xylans, and pectin, along with glucose, xylose, and maltose, provide hexoses and pentoses that are readily

fermentable and affect EMPS similarly (Hoover and Stokes, 1991). The inclusion of

nonstructural carbohydrates (NSC) as compared to structural carbohydrates in a diet is less

important than the overall rate of carbohydrate digestion. Increasing the amount of NSC increases ruminal carbohydrate digestion, supplying the rumen bacteria with more energy. When

Hoover and Stokes (1991) reviewed lactating cow studies, they calculated, with sufficient energy, the conversion of ruminally degradable protein to MCP as high as 88%. In another study in the review, inclusion of ruminally degradable protein increased carbohydrate digestion.

Traditional views are that a lack of synchrony of energy and nitrogen release will

decrease EMPS. Kim et al (1999) found infusion of maltodextrin with diets that have excess

RDP decreased ammonia-N, and continuous infusion decreased it more than the other two

infusion treatments, in which both infused the maltodextrin for two periods of 6 hours per day,

one synchronized with feeding and one not. All infusion treatments increased butyrate in the

rumen, and the continuous treatment increased valerate. Urinary-N excretion and BUN were

decreased by the infusion treatments. Calculated MPS was increased by continuous and

synchronized infusion. However, citing other in vivo work, Firkins et al. (2006) and others argue

that the improvement by synchronized feedings may be equivocal and show that BUN and

utilization of stored intracellular polysaccharide in both bacteria and protozoa can supply N and energy for microbial growth between feedings.

AA Supplementation

Optimizing bacterial growth through the supplementation of amino acids is most likely due to a decrease in energy use from catabolizing carbon sources and synthesizing the appropriate amino acids. In a review, Firkins et al. (2007) discuss rumen bacterial amino acid

synthesis. Guliye et al. (2005) report that no single amino acid deletion decreased bacterial

EMPS. Also, the rumen bacterial amino acid pool size remained similar in cattle fed different

diets (Ives et al., 2002), but the amino acids synthesized from carbon skeletons varied by diet

(Bach et al., 2004). Some amino acids are preferentially metabolized and others are

preferentially incorporated into microbial protein by the rumen bacteria (Atasoglu et al., 2004).

Carbon catabolite repression (CCR) is a regulatory mechanism that stops the use of a carbon source when the preferred carbon source is available through attenuation of genes that

code for proteins involved in the metabolism of the nonpreferred carbon source (Stulke and

Hillen, 1999). In Firmicutes, HPr, a bifunctional enzyme with kinase and phosphorylase activity,

is the central regulator of carbon metabolism. The uptake of the preferred carbon source changes

the concentration of metabolites. These changes stimulate kinase activity in HPr, which then

interacts with the substrate transporters. CCR is mediated by global transcription regulators,

catabolite control protein A, CcpA, in Firmicutes and cAMP receptor protein, Crp, in

Enterobacteriaceae (Deutscher, 2008). Amino acid de novo synthesis pulls intermediates from

the TCA cycle. Coordinating amino acid flux with protein synthesis and synthesis of nonprotein

cell components could theoretically increase EMPS (Firkins et al., 2007).

Peptide and amino acid supplementation has been shown to have different effects on

rumen bacteria in vitro. Peptide supplementation, in the form of pancreatic casein hydrolysate,

with NH4Cl decreased the proportion of particulate nitrogen derived from ammonia by 63%, as

compared to an NH4Cl -only treatment (Atasoglu et al., 1999). Amino acid supplementation

decreased the proportion by 45%. Both peptides and amino acids increased the rate of

carbohydrate fermentation, as measured by gas production. Xylose, cellobiose and soluble starch

were supplied in the media, so these effects were most likely on noncellulolytic bacteria. The

authors concluded that MCP from ammonia-N is not fixed, but dependent on ammonia N:total N

available. When the researchers measured de novo production of individual amino acids,

methionine had the lowest proportion derived from ammonia, except proline, for the peptide and

amino acid supplementations, 0.16 and 0.24, respectively (Atasoglu et al., 1999). Therefore,

preformed methionine can be highly incorporated into MCP, decreasing the need for de novo

synthesis.

Atasoglu et al. (2004) showed that supplying the 20 amino acids to rumen bacteria

increased bacterial protein synthesis. The researchers also discussed amino acid and microbial

protein turnover as synthesis compared to outflow. Nolan and Stachiw (1979) showed that

synthesis, measured by incorporation of N into microbial cells, was almost twice the outflow of

microbial N from the rumen in Merino sheep on low quality roughage, indicating that microbial

protein breakdown and resynthesis occurs to a great extent. Supplementation of specific amino

acids may have differential effects on microbial growth. Another study showed that

supplementation of the 20 amino acids increased growth rate by 46 % and efficiency by 15%

(Kajikawa et al., 2002). Atasoglu et al. (2004) showed that glutamate and aspartate were the

most abundant amino acids in microbial protein, isoleucine, phenylalanine, lysine, and leucine

were often incorporated into microbial protein, the glutamate family was often catabolized, and lysine synthesis may limit microbial growth. Amino acids also stimulate fibrolytic bacteria

(Firkins et al., 2007). In a study to determine which, if any, single amino acid could stimulate growth as compared to ammonia alone and all 20 amino acids, only glutamine and glutamate significantly increased growth as compared to ammonia (Kajikawa et al., 2002). Growth was depressed in the cysteine, isoleucine, leucine, phenylalanine, and threonine treatments by more

than 10%. The rumen bacteria supplemented with methionine grew similarly to those with

ammonia only. The study also showed that the removal of leucine, tryptophan, tyrosine,

glutamate, methionine, phenylalanine, and valine from the 20 amino acid supplement

significantly decreased the growth stimulation.

Supplying preformed amino acids to rumen bacteria can trigger feedback inhibition,

utilized to spare carbon for ATP, limiting synthesis of entire families of amino acids (Kajikawa

et al, 2005). Threonine, isoleucine, and phenylalanine, specifically, suppress production of other amino acids. This inhibition, however, may be overcome by the addition of other amino acids that are antagonistic to those inhibitors. Kajikawa et al. (2005) showed that the addition of leucine and valine relieved isoleucine inhibition, tryptophan and tyrosine relieved phenylalanine inhibition, and glutamate, serine, valine, alanine, and glutamine relieved threonine inhibition.

These studies show that proper balance of preformed amino acid supplementation can prevent feedback inhibition and may increase MPS.

In most bacteria, methionine can be synthesized from precursor molecules. Or-Rashid et

al. (2001) showed that rumen bacteria can synthesize methionine from homocysteine, and

subsequently incorporated the methionine into microbial protein. The rumen bacteria could also

produce methionine from cystathionine. Homocysteine was found in the incubation media,

suggesting that the rumen bacteria first convert cystathionine to homocysteine and then

homocysteine to methionine, using the transsulfuration pathway. To further support the use of

the transsulfuration pathway, rumen bacteria incubated with homoserine and cysteine also

produced methionine, through cystathionine, which was found in the incubation media. The

researchers gave a hypothetical scheme for methionine biosynthesis. Irreversibly, homoserine +

cysteine Æ cystathionine Æ homocysteine Æ methionine.

Amino acid synthesis in rumen bacteria is related to volatile fatty acid (VFA) synthesis.

Oxaloacetate from pyruvate can be used to synthesize propionate (Van Soest, 1994). In

concentrate-fed animals, bacteria, especially Megasphaera elsdenii, produce much higher

amounts of propionate than in forage fed animals. Also, amino acids can be degraded for VFA

synthesis. Valine, leucine, and isoleucine can be deaminated to isobutyrate, isovalerate, and 2-

methylbutyrate, respectively. These VFA are growth factors for cellulolytic microorganisms,

and other bacteria use them for long chain fatty acid synthesis.

Continuous Culture

Passage rate and substrate supply may have independent effects on microbial protein

synthesis (Firkins et al., 2007). Increasing the amount of substrate can stimulate growth by

diluting maintenance energy requirements for the population. In dual-flow continuous culture

systems, the effects of substrate supply and passage rate can be separated.

Studies have shown that increasing passage rate increases EMPS. Firkins et al. (2007) suggested that the increase in efficiency is due to an increase in gene expression in those

populations that are highly competitive for the substrate available, or that have another strategy

to outcompete the other bacteria present.

A dual-flow continuous culture system allows escape of metabolites at a constant rate so

pool size is not perturbed. Samples of the microbial populations from the fermentation vessels

would be less prone to sampling error due to the extremely large size of a rumen compared to the

sample size. Also, the samples would be more homogenous due to constant, equal agitation by

the paddles. Continuous culture also allows the use of isotopically labeled substrates in small

amounts. Large amounts could elicit animal responses not linked to microbial changes and

would be cost prohibitive.

Methionine and HMBi Supplementation to Dairy Cows

Postruminal infusion of methionine increased milk protein production (Pisulewski et al.,

1996). This effect of methionine supplementation indicates that increasing methionine flow out of the rumen may increase production. The same effect could be achieved by increasing total

MCP flow from the rumen. Methionine supplementation also increased milk protein yield, even when total milk yield was reduced (Rulquin and Delaby, 1997). Socha et al. (2008) tried to determine the methionine requirement for high producing dairy cows in peak, early, and mid- lactation when fed diets that were formulated to meet of slightly exceed nutrient requirements.

Duodenal infusion of 16g/day of methionine during peak lactation increased milk true protein production linearly (P=0.034). However, having no quadratic effect indicates that the methionine requirement was not met for peak lactation. During early lactation, infusion of

16g/day tended to increase 3.5% fat corrected milk linearly (P=0.052), increased milk true protein production linearly (P=0.013) and increased milk fat production linearly (P=0.030).

These results also indicate that the methionine requirement was not met. In mid-lactation, the increase in milk true protein percent had a quadratic effect (P=0.043). The calculated curve reached plateau at 12.4g/day of methionine supplementation. The effect of methionine supplementation also increased milk production linearly (P=0.037). These results indicate that methionine is limiting in peak, early and mid-lactation for cows fed diets that are formulated to meet nutrient requirements and methionine supplementation may increase production parameters.

Because methionine is used by rumen microorganisms for protein synthesis and absorbed

by the animal through the rumen wall, analogs have been developed to provide methionine to the

animal and rumen microorganisms. The methionine and methionine hydroxyl analogs (MHA)

were shown to have ruminal effects because they did not change serum methionine:valine (Papas

et al., 1974). Older studies have shown the comparative degradation rates of methionine and its

analogs in vitro. Patterson and Kung (1988) showed the disappearance of the MHA to be

significantly less than methionine when incubated with rumen bacteria. Alimet, a liquid form of

MHA, disappeared at a significantly slower rate than MHA. Methyl and ethyl esters of MHA

were rapidly converted to MHA. However, bolus dosing, often at non-physiological amounts, precludes an accurate assessment of MHA degradation in vivo under normal conditions.

2-Hydroxy-4-(methylthio) butanoic acid (HMB) is a methionine hydroxy analog that is currently being fed to dairy cows. HMB was shown in continuous culture by Noftsger et al.

(2003) to decrease the proportion of microbial nitrogen originating from ammonia-N by about

13%. If HMB was simply a methionine precursor, the amount of nitrogen in MCP from ammonia should not decrease. Therefore, HMB probably has another effect on MPS. HMBi is the isopropyl ester of HMB and might be degraded more slowly by the rumen bacteria and provide a steadier supply of methionine to the bacterial populations to positively affect growth.

The effects of amino acid supplementation on growth most likely involve modulation of gene expression in the affected bacterial species. Noftsger et al. (2005) cites two abstracts showing that HMBi is 50% undegraded in the rumen through blood and milk true protein analyses

(Robert et al., Schwab et al., 2001). The undegraded fraction is absorbed through the rumen wall where the isopropyl group is removed and HMB goes into the blood stream The other 50% is decomposed to HMB and isopropanol. This was further supported by the measured passages

rates of HMB from HMBi and HMB from HMB through the omasum of 2.3% and 5.3%,

respectively (Noftsger et al., 2005). These results do not support other work showing large

amounts of HMB passing from the rumen in the liquid to supply methionine to the cow

postruminally.

Rulquin et al. (2006) compared the effects of HMB and HMBi when fed to lactating dairy

cows. Supplementation of a methionine deficient diet with HMBi increased milk protein yield

and true protein percentage. HMB had no effect on milk production or milk components. St-

Pierre and Sylvester (2005) also showed that HMBi increased milk, milk fat, and protein yields

when fed to cows in early lactation that are fed a diet deficient in methionine, but not in lysine.

Again, HMB did not have any increases in milk production or components. Estimated urinary N

excretion was decreased in the cows on the HMBi supplemented diet by 17.5 g/day, so more N

was recovered in the milk. Therefore, gross N efficiency (proportion of ingested N secreted in

milk) was increased by HMBi. The HMBi treatment numerically increased plasma methionine,

but the increase was not significant (p=0.07). Noftsger et al. (2005) compared the production

effects of 0.10% HMB, 0.13% HMBi, and 0.088% DL-methionine. HMBi and methionine increased milk protein percentage, and HMBi caused a higher protein production than methionine. Apparent rumen digestibility of organic matter was increased in all three supplemented diets.

Chapter 2: Effects of 2-hydroxy-4-(methylthio) butanoic acid isopropyl ester and methionine supplementation on populations of rumen bacteria and efficiency of microbial protein synthesis

Introduction

Many dairy producers are currently trying to decrease fecal and urinary nitrogen excretion from high producing dairy cows due to increased federal regulations and public interest. To achieve this end, many producers are decreasing the amount of protein

(crude protein and rumen degraded protein (RDP)) in diets. Much of the protein in diets fed to dairy cows is degraded and utilized by the rumen microorganisms. The amino acid profile of rumen bacteria closely reflects that of milk protein, so utilization of bacteria protein for milk protein is efficient for the cow. Because the majority of the protein that the cow absorbs is of microbial origin, increasing microbial protein synthesis (MPS), more specifically, increasing the efficiency of microbial protein synthesis (EMPS) would greatly benefit dairy producers by decreasing the amount of nitrogen that is excreted into the environment by decreasing the amount of protein that must be in the diet.

Supplementation of amino acids that are limiting for rumen bacterial growth or total milk yield and production of components might also decrease the amount of rumen degraded protein (RDP) needed in the diet. Even though amino acids can be synthesize

de novo by the rumen bacteria, the concentrations might not be sufficient for optimized

MPS, or leading to variable MPS relative to the ability of models to predict MPS.

Because of this variable prediction, dietary crude protein (CP) is often formulated with a safety factor. If the diet contained a sustained supply of preformed amino acids, synthesis of microbial crude protein (MCP) might be increased because growth would be synchronized with carbohydrate availability (energy) from dietary starch and fiber.

Because supplying the 20 amino acids to rumen bacteria can increase bacterial protein synthesis (Atasoglu et al., 2004), supplementation of specific amino acids might have differential effects on microbial growth. Therefore, research to advance our knowledge in microbial conversion of RDP to MCP could profoundly lead to improved efficiency of dietary conversion of CP to milk protein without increasing the risk of depressed milk production.

Postruminal supplementation of methionine has been shown to improve milk protein yield (Rulquin and Delaby, 1997). However, oral supplementation of methionine might not cause these production effects because methionine should be quickly degraded to its carbon skeleton and ammonia by the rumen bacteria or incorporated into bacterial protein. For this reason, methionine analogs were created to supply methionine postruminally. The methionine analog 2-hydroxy-4-(methylthio) butanoic acid (HMB) and its isopropyl ester (HMBi) did not increase plasma methionine concentrations

(Noftsger et al., 2005), indicating that they are primarily rumen degraded. The positive production effects of HMBi supplementation include increased total milk production, fat- corrected milk production and true milk protein (St-Pierre and Sylvester, 2005). These effects might be due to increased EMPS, which would increase the supply of MCP to the

cow for total milk and milk protein production or changes in the microbial populations in the rumen that support milk fat and protein synthesis. For example, acetate supplies the carbon for synthesis of short chain fatty acids for milk fat in the mammary gland (Van

Soest, 1994). Therefore, an increase in the cellulolytic and amylolytic bacteria that produce acetate (Baldwin and Allison, 1983) might support an increase in milk fat production. Also, because propionate is a major substrate for hepatic gluconeogenesis from ruminal fermentation (Seal and Reynolds, 1993), an increase in propionate production by bacteria such as Megasphaera elsdenii in animals fed high concentrate diets (Van Soest, 1994) could increase milk production. For the current study, the objectives were to determine if HMBi caused changes in the rumen microbial populations and determine its effects on EMPS.

Materials and Methods

Experimental Design

A dual flow continuous culture fermentation system designed to model the ruminal environment for microorganism growth was used in a 4x4 Latin square (Noftsger et al., 2003). The four periods consisted of 12 days, with the first eight days for adaptation. The cultures were fed 100 g DM for the average fermenter volume. The diets were 50:50 forage:concentrate and fed in three equal meals per day at 0000, 0800, and 1600. The diet composition is shown in Table 1 and was formulated without soybean meal to be minimal to complement RDP in alfalfa to prevent excessive RDP from

potential responses to methionine supplementation. The treatments contained no

supplement (CON), 0.11% HMBi (HMBi), 0.097% DL-methionine (MET), or 0.055%

HMBi + 0.048% DL-methionine (HMBi+MET). The amounts of supplementation of

HMBi and DL-methionine were chosen from the previous continuous culture study of

Noftsger et al. (2003). The supplements and control were added as 3.6 mL of liquid

directly into the fermentation vessels at each feeding. For each treatment, six doses of

isotopically labeled HMBi and DL-methionine were added to each feeding, starting on day

9 at 0800. During dosing, 25% of the HMBi supplement was replaced with U-13C-

HMBi; 60% of the MET supplement was replaced with 30% 1-13C-methionine and 30%

2 methyl- H3-methionine; 55% of the HMBi+MET supplement was replaced by 25% U-

13C-HMBi and 30% 1-13C-methionine. After the sixth dose, the labeled supplements were switched back to their 100% unlabeled doses. When supplements were added, all outflow was stopped by turning off the filtrate pumps and physically blocking the overflow tube for 30 seconds while the contents of the fermentation vessels mixed.

Continuous Culture Operation

The dual flow continuous culture system is based on the system described by

Hoover et al. (1976). The inoculum for the fermenters originated from two multiparous lactating Holstein cows receiving a normal lactation diet. Four liters of rumen contents were collected through a rumen cannula from each cow. The contents were then mixed and squeezed through eight layers of cheesecloth. The resulting inoculum was divided into the four fermentation vessels. NaCO3 was added to buffer the inoculum during the initial hours. The pH was monitored multiple times per day and was maintained between

6.0 and 7.0 by continual infusion of buffer and manual adjustment with 6 N HCl and 5 N

NaOH. Temperature was kept constant at 39ºC. The cultures were continually agitated

and purged of oxygen by N2 to maintain anaerobiosis. The DMI (averaging 100.0 g/day, ranging from 94.3 to 108.2 g/day) for each culture was adjusted for the volume of the fermentation vessel. The fermentation vessels varied in volume from 1600 to 1835 mL, with agitation. The liquid and solid dilution rates were 12.5 and 5.5% per hour, respectively. These rates were maintained by regulated filtrate removal and buffer input.

Solid and liquid output was weighed and adjusted daily to maintain these flow rates.

Sample Collection and Analysis

On days 10, 11, and 12, a 15% sample of the total effluent (liquid and solid) was taken and composited. The three-day composite sample was lyophilized and analyzed for

N using the Kjeldahl method (AOAC, 1990), for NDF in the presence of heat-stable alpha amylase and sodium sulfite, and ADF (Van Soest, 1991). Hemicellulose was calculated as NDF – ADF. For each of the three days, a 47-mL aliquot of the effluent sample was acidified with 3 mL of 6 N HCl to stop fermentation, and all the samples were composited. The composited sample was analyzed for VFA using gas chromatography (Harvatine et al., 2002) and ammonia-N by the colorimetric method of

Chaney and Marbach (1962). Bacteria were pelleted from a total effluent (liquid and solid) sample of the 3-day composite. Nitrogen was determined for lyophilized bacteria by micro-Kjeldahl (AOAC, 1990). Organic matter was determined for effluent and bacteria by heating at 600°C for 24 hours.

A pilot study with the isotopically labeled HMBi showed that the HMBi enrichment of the fermentation medium plateaued after the second dose. For the secondary pool of the isotopically labeled methionine to plateau as well, samples were taken after the forth dose. Three 15-mL samples of total effluent were taken and composited. For period 1, the samples were taken at 31.5, 34.5, and 36.5 h; for period 2, the samples were taken at 32, 39.5, and 41 h; for period 3, the samples were taken at 28,

31, and 39 h; for period 4, the samples were taken at 32, 39, and 41 h. These samples were stored at -80°C for nucleic acid extraction.

Bacteria and effluent samples were taken during and after the isotopes were dosed into the cultures. For the first period, samples were taken before the first dose

(background), 0.5, 8, 8.5, 10, 12, 24.5, 32.5, 48.5, 52, 55, 58, 60, 64, 72.5, 79, 81, 85, 87,

90, 93, and 96 h. The samples were ~10 mL for analysis of the effluent and ~30 mL for extraction of bacteria. The samples for 79 through 96 h were for bacteria only. After the initial analyses of label incorporation into the bacteria protein, the number of samples were decreased for periods 2, 3, and 4, to facilitate sampling efficiency: before the first dose (background), 0.5, 8, 8.5, 12, 24.5, 32.5, 40.5, 44, 47, 50, 52, 56, 64.5, 71, 79, and

85 h. The samples at 71, 79, and 85h were bacteria only. The samples taken on the hour

(e.g., 8 h) were taken before feeding. The samples on the half hour (e.g., 8.5 h) were taken after feeding. The bacteria were pelleted by differential centrifugation.

Approximately 30 mL of effluent was centrifuged at 4°C, 500 x g for 15 minutes to remove feed particles. The supernatant was centrifuged at 4°C, 23,300 x g for 15 minutes. The pellet was washed and resuspended in 0.9% saline and centrifuged again at

4°C, 23,300 x g for 15 minutes. The resulting pellet was frozen and lyophilized.

DNA extraction

The frozen effluent samples were thawed and 1.5 mL was added to the bead-

beating tube as described by Yu and Morrison (2004). The samples were centrifuged at

16000 x g for 15 minutes. The extraction continued as described, with two rounds of

bead-beating to break the cells, nucleic acid precipitation, and treatment with RNase A and proteinase K. The DNA was then column-purified using the DNA Mini Stool Kit

(QIAGEN, Valencia, CA) according to the manufacturer’s recommendations.

DGGE

The V3 region of 16S rRNA genes was amplified using primers 357f-GC

(synthesized with a 40-bp GC clamp at its 5’ end) and 519r (Larue et al., 2005). The amplification reactions were comprised of 100 pmol of each primer, 250 µM dNTP, 2 mM magnesium chloride, 0.05% bovine serum albumin, 1X PCR buffer, and 5.0 units of

Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA) in 50µL. A hot start

thermocycling program was used as follows: initial denaturation at 94°C for 5 min; 10

touch down cycles at 94°C for 30 sec, then the annealing temperature decreased from

61°C to 56°C by 0.5°C/cycle, followed by an extension step at 72°C for 30 sec; then 25

cycles with denaturation at 94°C for 30 sec, annealing at 56°C for 30 sec, and extension

at 72°C for 1 min; and the program ended with a final 72°C extension for 30 min

(Karnati, 2006). The PCR reaction included negative and positive controls. The PCR

product was approximately 200 bp, as verified by electrophoresis in a 1% agarose gel.

The PCR products were resolved on a 7.5% polyacrylamide gel (37.5:1) with a 40 to 70%

denaturing gradient for 1600 volt-hours. The gel was stained with SYBR green and visualized with a Gel Logic 1500 imager (Kodak, New Haven, CT).

The TIFF images were imported from the Kodak molecular imaging software v.

4.0 (New Haven, CT) into the database in the BioNumerics software (BioSystematica,

Devon, United Kingdom). Bands were automatically identified using the band-searching algorithm and manually refined to find and mark uncertain bands. To adjust for any possible migration differences, external reference markers and internal reference bands were identified. The bands were compared by calculating cluster analysis using the

Jaccard method, which accounts for band intensity as well as number of bands in each lane for calculation of similarity coefficients.

RNA Extraction and Gradient Separation

The RNA was extracted using an amended version of the Killing Two Birds with

One Stone method by Yu and Mohn (1999). Effluent sample (2 mL) was centrifuged at

16000 x g for 15 minutes at 4°C. The pellet was stored in RNAlater (Ambion, Austin,

TX) at -80°C. The pellet was transferred to the bead beating tube. The nucleic acids were extracted. The RNA was column-purified with the RNeasy kit (QIAGEN,

Valencia, CA) following the manufacturer’s instructions. The RNA was visualized on a

1% agarose gel, with 20% iodoacetic acid.

The RNA was run in a cesium trifluoroacetate (CsTFA) gradient to separate the light and heavy (isotope labeled) bands. The gradient was comprised of 4 mL 2.0 g/mL

CsTFA, 170 mL formamide, 765 µL dH2O, 20 µL SYBR green, and 45 µL RNA solution

(Whiteley et al., 2007) for each sample. The samples were put in 5.1 mL, 13 x 51 mm

polyallomer Quick-Seal centrifuge tubes (Beckmann Coulter, Palo Alto, CA). The tubes were sealed with a heat sealer, and balanced within 10 mg. The tubes were ultracentrifuged at 128,000 x g for 65 hours at 20°C.

Statistical Analysis

Data were analyzed using Proc MIXED of SAS (2004) according to the following model:

Yijk = µ + fi + Pj + Tk + eijk

where:

Yijkl is the dependent, continuous variable,

µ is the overall population mean,

fi is the random effect of the ith fermenter (i = 1, 2, 3, 4)

Pj is the fixed effect of the jth period (j = 1, 2, 3, 4),

Tk is the fixed effect of the kth treatment (k = 1, 2, 3, 4),

2 eijkl is the residual error, assumed independent and ~ N(0, σ e).

The first preplanned orthogonal contrast compared the average of all the methionine treatments vs. the control (CON). Two contrasts were used to determine linear or quadratic response to HMBi, HMBi+MET, and MET.

Table 1. Ingredients of Diet

Ingredient Base Diet

% DM of Diet

Alfalfa pellets 50

Pelleted concentrate 50

Corn grain, ground 31 Dry distillers grains 6.5

Blood meal, ring-dried 0.5

Soybean hulls 10.8

CaPO4 0.33

Limestone 0.065

MgO 0.13

Trace mineralized salt1 0.5

Vit A 0.012

Vit D 0.033

Vit E 0.059

1Contained 0.10% Mg, 38.08% Na, 58.0% Cl, 0.04% S, 5,000 mg/kg of Fe, 7,500 mg/kg of Zn, 2,500 mg/ kg of Cu, 6,000 mg/kg of Mn, 100 mg/kg of I, 60 mg/kg of Se, and 50 mg/kg of Co.

Table 2. Nutrient Composition of Diet

Nutrient % DM

NDF 34.9

ADF 23.6

CP1 14.2

RDP2 7.8

RUP2 6.4

1 Found experimentally (Kjeldahl method)

2 Calculated from book values NRC, 2001

Table 3. Nutrient digestibilities, nitrogen fluxes, and nitrogen partitioning for rumen bacteria in continuous cultures supplemented with HMBi1, DL-methionine or both Treatment Contrasts2 CON vs Con Met HMBi+Met HMBi SEM all Linear Quadratic Digestibility (%) NDF 42.0 46.3 46.0 38.8 2.6 NS3 0.04 NS ADF 48.4 46.6 51.8 47.3 3.6 NS NS NS Hemicellulose 28.9 44.4 33.8 20.7 5.0 NS 0.01 NS True OM 50.4 51.3 50.6 49.2 2.0 NS NS NS

Nitrogen flows (g/day) Ammonia N 0.36 0.39 0.39 0.32 0.03 NS 0.08 NS NAN4 2.50 2.38 2.45 2.38 0.15 NS NS NS

38 Bacterial N 1.57 1.62 1.57 1.50 0.10 NS NS NS Total N 3.15 3.10 3.18 2.97 0.15 NS NS NS NANBN5 0.98 0.82 0.92 0.94 0.11 NS NS NS

Nitrogen Partitioning Ammonia N (mg/dl) 6.22 6.80 7.20 5.65 0.37 NS 0.07 0.08 TCA-soluble N (mg/dl) 11.4 12.1 12.3 11.8 0.37 0.06 NS NS Peptide N6 (mg/dl) 5.17 5.28 5.14 6.13 0.27 NS 0.04 0.09 Bacterial N derived from NH3- N, % 86.2 81.8 88.5 89.1 1.80 NS 0.02 NS Bacterial N efficiency7 28.9 32.0 31.7 30.9 1.95 NS NS NS

1HMBi = 2-hydroxy-4-(methylthio) butanoic acid.

2All supplementation (HMBi, HMBi+MET, MET) vs control, and linear and quadratic responses to the three supplements.

3NS= Not significant; P>0.10.

4NAN= Non-ammonia nitrogen.

5NANBN= Non-ammonia non-bacterial nitrogen = NAN – bacterial N

6Nitrogen soluble in 10% (vol/vol) trichloroacetic acid. 39 7Grams microbial N produced/kilogram OM truly digested.

Table 4. Volatile fatty acid production per day for rumen bacteria in continuous culture supplemented with HMBi1, DL-methionine or both

Treatment Contrasts2

Con Met HMBi+Met HMBi SEM Con vs all Linear Quadratic Production (mmol/day)

Acetate 119 118 116 111 5.11 NS3 0.06 NS

Propionate 35.7 36.1 32.4 31.9 1.24 0.01 <0.01 0.05 Isobutyrate 0.56 0.66 0.66 0.52 0.05 NS 0.05 NS 40

Butyrate 19.9 20.6 18.4 18.5 0.84 NS NS NS Isovalerate 3.33 3.37 4.90 4.80 0.57 NS 0.08 NS

Valerate 4.06 3.97 4.81 4.90 0.25 0.01 <0.01 0.05

BCVFA4 3.88 4.02 5.56 5.32 0.6 NS NS NS Total VFA 183 183 177 172 7 NS 0.02 NS Acetate:Propionate 3.33 3.28 3.58 3.48 0.09 NS 0.08 0.06

1HMBi = 2-hydroxy-4-(methylthio) butanoic acid.

2All supplementation (HMBi, HMBi+MET, MET) vs control.

3NS= Not significant; P>0.10.

4BCVFA= Branched chain VFA, which includes isobutyrate and isovalerate.

Figure 1. Comparison of the bacterial community structure from continuous culture samples; dendrogram shows cluster analysis that was performed using the Jaccard method; P = Period 1 to 4; HMBi = 2-hydroxy-4-(methylthio) butanoic acid isopropyl ester. MET= DL-methionine. The length of scale depicts the percent similarity between different lanes.

Results and Discussion

Digestibility of Nutrients

Treatment did not affect (P>0.10) ADF or true OM digestibilities (Table 3). NDF

(P=0.04) and hemicellulose digestibilities (P=0.01) were affected linearly. Numerically digestibilities were higher with Met and lower with HMBi, which is difficult to explain.

Because NDF includes hemicellulose and ADF digestibility was not affected, the effect for NDF digestibility is likely due to the affect for hemicellulose digestibility.

Methionine and HMB supplementation have yielded inconsistent effects on nutrient digestibilities. Supplementation of HMB at 0.055 and 0.110% in continuous culture had no significant effect on digestibilities of hemicellulose, NDF, or true OM

(Noftsger et al., 2003), although our hypothesis was that the continuous feeding of the cultures might have negated the effects of HMB supplementation compared with dairy cows fed once or twice daily. With 0.20, 0.77, and 1.43% supplemental HMB given twice daily in continuous culture, Vazquez-Anon et al. (2001) found no changes in true

OM, CP, NDF, ADF, or NSC digestibilities. The amount of supplementation greater than

0.20% might provide an unphysiological excess of HMB to the bacteria, though.

Other studies have shown benefits in fiber digestibility with supplementation of methionine sources. Gil et al. (1973a) found an increase in cellulose digestibility in vitro when HMB or DL-methionine was 0.8% of the media for mixed rumen bacteria and cellulose was the only substrate. The cellulolytic bacteria might have been favored because the inoculum originated from steers fed Bermuda grass hay ad libitum. In batch

culture with inoculum from cows fed a purified ration of cellulose, corn starch, and urea,

cellulose digestion was greater with methionine than HMB (Salsbury et al., 1971).

Cellulose digestion sharply decreased on day 5, followed by a sharp increase. The

authors suggested that this sudden dip in cellulose digestibility might be indicative of a

population shift. In a study with a 2x2 factorial arrangement of 35 and 43% RUP and

high or low dietary methionine, Bach and Stern (1999) found high dietary methionine increased apparent NFC digestibility. This increase in digestibility might result from the requirement of some amylolytic and saccharolytic bacteria for amino acids (Baldwin and

Allison, 1983). Also, saccharolytic Selenomonas ruminantium requires methionine specifically. The increase in NDF and hemicellulose digestibility in the current study could be due to a shift in microbial populations from methionine supplementation.

Hemicellulose consists mostly of linear xylose chains with varying amounts of arabinose, uronic acids, and galactose (Baldwin and Allison, 1983). Cellulolytic bacteria and oligotrich protozoa have been shown to hydrolyze hemicellulose. However, protozoa washed out of the continuous culture system, indicating that the increase in hemicellulose degradation could be due to an increase in the action of cellulolytic bacteria. Griswold et al. (2003) suggest that hemicellulose-degrading bacteria might use amino acids when N is limiting. The formulation of the experimental diet in the current study was for minimal to slightly limiting RDP, so the hemicellulose-degrading bacteria might have used methionine as a N source during periods when NH3 N dropped below optimum levels of

2 to 5 mg/dL (Hoover et al., 1986). Griswold et al. (2003) has similar RDP values for a

low RDP treatment (8.85%). Although in a pilot study we determined that the diet and

buffer would provide NH3 N to maintain concentrations at that optimum, it is possible

that during periods between feedings the concentrations dropped to benefit the bacteria

receiving an NH3 N supply from methionine.

Nitrogen Fluxes and Partitioning

The control diet for this experiment was formulated to be minimal to limiting in

RDP to express possible microbial responses to methionine supplementation that could occur in cows fed diets formulated to decrease N excretion. The flows of non-ammonia

N (NAN), bacterial N, total N, and non-ammonia non-bacterial N (NANBN) in the effluent were not different by treatment (Table 3). NH3 N flow tended (P=0.08) to be

linearly affected by treatment, decreasing as HMBi replaced MET. Previously,

methionine supplementation caused an increase in ammonia N flow (Bach and Stern,

1999).

The concentration of NH3 N tended to be affected linearly (P=0.07) and

quadratically (P=0.08), mostly because HMBi decreased NH3 N concentration compared with MET or HMBi+MET. Ammonia concentration has been affected by HMB and methionine supplementation in various ways. In continuous culture, supplementation of

HMB caused no differences in NH3 N concentration (Vazquez-Anon et al., 2001). Also,

NH3 N concentration was decreased in vitro by methionine and HMB supplementation

(Gil et al., 1973a). Ammonia N might be increased by methionine supplementation due

to an increase in deamination, but net uptake and increased growth would decrease NH3

N concentration. A decrease in NH3 N from HMBi supplementation might be caused by

an increase on NH3 utilization for amination of HMBi to methionine. An increase in

amination of HMBi could also explain the increase in bacterial N derived from ammonia

N for the HMBi and HMBi+Met treatments. Previously, HMB in continuous culture

decreased the proportion of bacterial N from ammonia N (Nofsger et al., 2003), possibly

increasing proteolysis if HMB was used simply as a carbon source, or that it stimulates

uptake of preformed amino acids. That study had much higher RDP than did our study.

The concentration of NH3 N tended to be affected linearly (P=0.07) and

quadratically (P=0.08), mostly because HMBi decreased NH3 N concentration compared with MET or HMBi+MET. For the concentration of TCA-soluble N, which includes ammonia and peptides, the control tended (P=0.06) to be less than the average of the methionine treatments. However, peptide N, which excludes NH3 N, was affected linearly (P=0.04) and tended to be affected quadratically (P=0.09). These contrasts and the lack of difference compared with control are explained by peptide concentration being greatest for HMBi. The concentrations of peptide N for all treatments were greater than those reported by Griswold et al. (2003), which used similar procedures suggesting that peptide N supply was adequate in the current study. The percentage of bacterial N derived from NH3 N was increased (P=0.02) as HMBi replaced Met. The corresponding

increase in peptide N and decreasing NH3 N concentrations with increasing HMBi supplementation seems to support decreased preference for preformed amino N, which is contrary to our original hypothesis. The percentage of bacterial N from NH3-N was

increased in the HMBi treatment as compared to the control, possibly indicating that

HMBi was aminated to methionine and the methionine was incorporated into bacterial

protein; however, because methionine is only a small component of the protein in

bacteria, our data would indicate that other amino acids were synthesized de novo, also.

In the methionine treatment, the percentage of bacterial N from NH3-N was decreased as

compared to the control indicating that the bacteria were using the available preformed amino acids. Together, these results indicate that rumen bacteria incorporate methionine into bacterial protein whether it is provided as methionine or HMBi, but the complexity of control of protein synthesis requires analysis of other amino acids in addition to methionine. Or-Rashid et al. (2001) showed that rumen bacteria could synthesize methionine from different precursors, such as homocysteine and cystathionine, and incorporate it into microbial protein. However, tracer studies document complex interactions among amino acids that await further clarification (Walker er al., 2005).

In the current study, efficiency of bacterial protein synthesis was not affected by treatment. Supplementation of 0.20 and 0.77% HMB increased bacterial protein synthesis and efficiency numerically, although the effects were not statistically significant (P= 0.19 and P= 0.18, respectively) (Vazquez-Anon et al., 2001). Bacterial N flow and EMPS were not affected by methionine (Bach and Stern, 1999) or HMB

(Noftsger et al., 2003) supplementation. In contrast, bacterial N concentration has been increased by methionine and HMB supplementation with mixed batch cultures fed glucose and urea (Gil et al., 1973a), and HMB increased glucose utilization by rumen bacteria in vitro (Gil et al., 1973b), indicating an effect on growth.

The effect of HMB and methionine on growth and protein synthesis of rumen bacteria has been difficult to elucidate, possibly due to different populations present in the inoculum of continuous and batch cultures from different animals fed different diets.

These studies were performed across location: Gil et al. (1973a, b) in Florida, Bach and

Stern (1999) in Minnesota, Vazquez-Anon et al. (2001) in Pennsylvania, and Noftsger et al. (2003) in Ohio. In a single study using cows from the same herd, Salsbury et al.

(1971) found differences in the effect of methionine and HMB on cellulose digestibility by the rumen bacteria from two cows.

VFA

Volatile fatty acid (VFA) production is shown in Table 4. Butyrate and BCVFA were not affected by treatment. Acetate tended (P=0.06) to decrease as HMBi replaced

MET. Propionate was lower (P=0.01) for the average of the three treatments compared with control and was affected linearly (P<0.01) and quadratically (P=0.05). These contrasts demonstrate that both amounts of HMBi supplementation decreased propionate production. Isobutyrate was linearly (P=0.05) decreased with increasing HMBi, although numerically isobutyrate was greater for Met and HMBi+Met than control. Isovalerate tended (P=0.08) to increase linearly with increasing HMBi. All treatments were greater

(P=0.01) than control for valerate which was linearly (P<0.01) and quadratically

(P=0.05) affected by treatment. These contrasts indicate that valerate was increased for both HMBi+Met and HMBi. Total VFA production was decreased (P=0.02) with increasing HMBi replacement of Met. The acetate to propionate ratio tended to be affected linearly (P=0.08) and quadratically (P=0.06), indicating increasing ratios for both MBi+Met and HMBi than for Met.

These changes in VFA production due to supplementation of methionine as

HMBi or Met, like the changes in the in vivo and in vitro studies discussed below, might indicate a relative change in abundance or activity of the present bacterial populations as some populations incorporate preformed amino acids into bacterial protein, and others use amino acids as N and carbon sources. HMBi appeared to shift metabolism or benefit

those populations that readily aminate carbon skeletons and utilize the resulting

methionine. Also, the decrease in total VFA production without a concurrent decrease in

organic matter digestibility indicates a shift in the flux of the carbon skeletons. The

carbon that is not going to VFA production may be aminated and incorporated into

bacterial protein. In continuous culture, 0.13 (Windschitl and Stern, 1988), or 0.20, 0.77,

and 1.43% (Vazquez-Anon et al., 2001) HMB supplementation caused no changes in

concentration of total VFA, acetate, propionate, butyrate, or acetate:propionate.

Methionine supplementation in continuous culture has also been shown to have no effect

on total VFA, acetate, propionate, butyrate, and BCVFA concentrations (Bach and Stern,

1999). Methionine supplementation increased isobutyrate and butyrate concentrations in

vivo (Lundquist et al., 1985) and increased concentration of total VFA and molar

percentage of propionate and decreased molar percentage of isovalerate, valerate, and

acetate:propionate in vitro (Chung et al., 2006). An in vivo study comparing

supplementation of HMB, HMBi, and DL-methionine, showed no differences in VFA

concentration (Noftsger et al., 2005).

Because cellulolytic and amylolytic rumen bacteria produce relatively more

acetate (Baldwin and Allison, 1983), the tendency for acetate production to decrease with

HMBi supplementation might indicate a decrease in the action of amylolytic bacteria,

because there was an increase in hemicellulose digestibility. Propionate production

decreased in the HMBi treatments, indicating an affect of methionine supplementation as

HMBi which was not seen when methionine, itself, was supplemented. Fibrobacter succinogenes produces succinate (Scheifinger and Wolin, 1973). Succinate is not often found in the rumen because it is quickly decarboxylated to propionate by bacteria such as

S. ruminantium. Although HMBi decreased both acetate and propionate, propionate was decreased more, resulting in an increase in the acetate:propionate. The decrease in propionate could be from a decrease in activity of amylolytic bacteria or from a shift in glucose for anabolic purposes; with the decreasing acetate production from glucose, there could be a proportionate increase in the need for disposal of reducing equivalents through propionate production.

Production of isobutyrate increased with methionine, and isovalerate tended to increase with HMBi. These BCVFA are required for the growth of Ruminicoccus flavefaciens (Allison et al., 1962). Also, R. albus requires isobutyrate for growth (Miura et al., 1980). R. flavefaciens and Megasphaera elsdenii (Allison and Peel, 1971) incorporate isovalerate into bacterial protein as leucine. M. elsdenii and Prevotella ruminicola also incorporate isobutyrate into cell protein as valine. F. succinogenes requires BCVFA and valerate (Bryant and Doetsch, 1955). These BCVFA are produced by P. ruminicola and M. elsdenii (Allison, 1978). Because some proteolytic bacteria, such as Ruminobacter amylophilus, have no amino acid or VFA requirements in the presence of a protein source (Miura et al., 1980), the decrease in total VFA production with HMBi supplementation might not affect protein degradation and might indicate that other VFA utilizers have increased growth or activity. The decrease in total VFA production due to HMBi supplementation, along with a decrease in NH3-N concentration,

indicates that HMBi decreased deamination of feed amino acids or more likely increased

the synthesis of amino acids from carbon skeletons and NH3.

DGGE

DGGE analysis was performed on total DNA extracted from the microbial communities present in the fermentation vessels after ten days. The dendrogram (Figure

1) shows the banding patterns. The bands cluster primarily by period. Even though relative changes in abundance of the present populations were not seen, DGGE only assesses abundant populations (Karnati et al., 2009) and would presumably preclude the detection of the low-abundance but high-activity deaminating populations of bacteria.

Moverover, the activities of anabolic and catabolic metabolism of amino acids could be affected by HMBi and methionine supplementation without affecting population structure. Further analysis using microarray would help determine if HMBi replacement for methionine affected the community of the hyperammonia producers or if the major differences in N metabolism were a result of shifting metabolism of amino acids.

Conclusions

Methionine supplementation might increase the activity of hemicellulose- degrading bacteria. HMBi appeared to be aminated to methionine by rumen bacteria and incorporated into bacterial protein. HMBi supplementation increased the flow of peptides but decreased the net production of total VFA, thus indicating that carbon was being used for synthesis of amino acids rather than using preformed amino acids.

Because of the changes in the production of VFA, supplementation of methionine as

HMBi or methionine itself might change the activities of populations of rumen bacteria.

Further Work

To further elucidate the fate and effect of HMBi, the rate of production of methionine from HMBi and the rate of incorporation of methionine from methionine and

HMBi into bacterial protein should be measured. Moreover, more extensive analyses of microbial communities would verify that it was a change in metabolism rather than a change in key populations of bacteria, particularly for hyperammonia producing bacteria for which their low abundance would apparently preclude their detection by DGGE.

References

Allison, M. J., M. P. Bryant, and R. N. Doetsch. 1962. Studies on the metabolic function of branched-chain volatile fatty acids, growth factors for ruminococci. J. Bacteriol. 83:523-532.

Allison, M. J. and J. L. Peel. 1971. The biosynthesis of valine from isobutyrate by Peptostreptococcus elsdenii and Bacteroides ruminicola. Biochem. J. 121:431-437.

Allison, M. J. 1978. Production of branched-chain volatile fatty acids by certain anaerobic bacteria. Appl. Environ. Microbiol. 35:872-877.

Atasoglu, C., C. Valdes, C. J. Newbold, and R. J. Wallace. 1999. Influence of peptides and amino acids on fermentation rate and de novo synthesis of amino acids by mixed micro-organisms from the sheep rumen. Br. J. Nutr. 81:307-314.

Atasoglu, C., A. Y. Guliye, and R. J. Wallace. 2004. Use of stable isotopes to measure de novo synthesis and turnover of amino acid-C and -N in mixed micro-organisms from the sheep rumen in vitro. Br. J. Nutr. 91:253-261.

Bach, A., and M. D. Stern. 1999. Effects of different levels of methionine and ruminally undegradable protein on the amino acid profile of effluent from continuous culture fermenters. J. Anim. Sci. 77:3377-3384.

Bach, A., S. Calsamiglia, and M. D. Stern. 2004. Nitrogen metabolism in the rumen. J. Dairy Sci. 88(E Suppl.):E9-E21.

Baldwin, R. L., and M. J. Allison. 1983. Rumen metabolism. J. Anim. Sci. 57(Suppl.):461-477.

Brinkmann, N., R. Martens, C. C. Tebbe. 2008. Origin and diversity of metabolically active gut bacteria from laboratory-bred larvae of Manduca sexta (Sphingidae, Lepidoptera, Insecta). Appl. Environ. Microbiol. 74:7189-7196.

Brock, F. M., C. W. Forsberg, and J. G. Buchanan-Smith. 1982. Proteolytic activity of rumen microorganisms and effects of proteinase inhibitors. Appl. Environ. Microbiol. 44:561-569.

Broderick, G. A., L. D. Satter, and A. E. Harper. 1974. Use of plasma amino acid concentration to identify limiting amino acids for milk production. J. Dairy Sci. 57:1015- 1023.

Brosnan, J. T., M. E. Brosnan, R. F. P. Bertolo, and J. A. Brunton. 2007. Methionine: A metabolically unique amino acid. Livestock Sci. 112:2-7.

Bryant, M. P., and R. N. Doetsch. 1955. Factors necessary for the groeth of Bacteroides succinogenes in the volatile fatty acid fraction of rumen fluid. J. Dairy Sci. 38:340-350.

Buckley, D. H., V. Huangyutitham, S.-F. Hsu, and T. A. Nelson. 2007. Stable isotope probing with 15N achieved by disentangling the effects of genome G+C content and isotope enrichment on DNA density. Appl. Environ. Microbiol. 73: 3189-3195.

Chaney, A. L., and E. P. Marbach. 1962. Modified reagents for determination of urea and ammonia. Clin. Chem. 8:130-132.

Chiang, P. K., R. K. Gordon, J. Tal, G. C. Zeng, B. P. Doctor, K. Pardhasaradhi, and P. P. McCann. 1996. S-adenosylmethionine and methylation. FASEM J. 10:471-480.

Chung, Y. H., H. G. Bateman, II, C. C. Williams, C. C. Stanley, D. T. Gantt, T. W. Braud, L. L. Southern, J. D. Ward, P. G. Hoyt, and G. A. Sod. 2006. Effects of methionine and lysine on fermentation in vitro and in vivo, nutrient flow to the intestine, and milk production. J. Dairy Sci. 89:1613-1620.

Clark, J. H. 1975. Lactational responses to postruminal administration of protein and amino acids. J. Dairy Sci. 58:1178-1197.

Cotta, M. A., and J. B. Russell. 1982. Effect of peptides and amino acids on efficiency of rumen bacteria protein synthesis in continuous culture. J. Dairy Sci. 65:226-234.

Deutscher, J. 2008. The mechanisms of carbon catabolite repression in bacteria. Curr. Opin. Microbiol. 11:87-93.

Dumont, M. G., and J. C. Murrell. 2005. Stable isotope probing – linking microbial identity to function. Nat. Rev. Microbiol. 3:499-504.

Eschenlauer, S. C. P., N. McKain, N. D. Walker, N. R. McEwan, C. J. Newbold, and R. J. Wallace. 2002. Ammonia production by ruminal microorganisms and enumeration, isolation, and characterization of bacteria capable of growth on peptides and amino acids from the sheep rumen. Appl. Environ. Microbiol. 68:4925-4931.

Egert, M., A. A. Graaf, A. Maathuis, P. de Waard, C. M. Plugge, H. Smidt, N. E. P. Deutz, C. Dijkema, W. M. de Vos, and K. Venema. 2007. Identification of glucose-

fermenting bacteria present in an in vitro model of the human intestine by RNA-stable isotope probing. FEMS Microbiol. Ecol. 60:126-135.

Firkins, J. L., A. N. Hristov, M. B. Hall, G. A. Varga, and N. R. St-Pierre. 2006. Integration of ruminal metabolism in dairy cattle. J. Dairy Sci. 89(E Suppl.):E31-E51.

Firkins, J. L., Z. Yu, and M. Morrison. 2007. Ruminal nitrogen metabolism: Perspectives for integration of microbiology and nutrition for dairy. J. Dairy Sci. 90(E. Suppl.):E1- E16.

Fuchs, R. T., F. J. Grundy, and T. M. Henkin. 2006. Smk box is a new SAM-binding RNA for translational regulation of SAM synthetase. Nat. Struct. Mol. Biol. 13:226-233.

Fuchs, R.T., F. J. Grundy, T. M. Henkin. 2007. S-adenosylmethionine directly inhibits binding of 30S ribosomal subunits to the SMK box translational riboswitch RNA. Proc. Natl. Acad. Sci. 104:4876-4880.

Gil, L. A., R. L. Shirley, and J. E. Moore. 1973a. Effect of methionine hydroxy analog on bacterial protein synthesis from urea and glucose, starch or cellulose by rumen microbes in vitro. J. Anim. Sci. 37:159-163.

Gil, L. A., R. L. Shirley, and J. E. Moore. 1973b. Effect of methionine hydroxy analog on growth, amino acid content, and catabolic products of glucolytic rumen bacteria in vitro. J. Dairy Sci. 56:757-762.

Griswold, K. E., G. A. Apgar, J. Bouton, and J. L. Firkins. 2003. Effects of urea infusion and ruminal degradable protein concentration on microbial growth, digestibility, and fermentation in continuous culture. J. Anim. Sci. 81:329-336.

Grundy, F. J. and T. M. Henkin. 1998. The S-box regulon: a new global transcription termination control system for methionine and cysteine biosynthesis genes in Gram- positive bacteria. Molecular Microbiol. 30:737-749.

Grundy, F. J. and T. M. Henkin. 2003. The T box and S box transcription termination control systems. Frontiers in Bioscience. 8:D20-D31.

Guliye, A. Y., G. Atasoglu, and R. J. Wallace. 2005. Assessment of amino acid requirements for optimum fermentation of xylan by mixed micro-organisms from the sheep rumen. Anim. Sci. 80:353-360.

Handelsman, J. 2004. Metagenomics: Application of genomics to uncultured microorganisms. Microbiol. Molecular Biol. Rev. 68:669-685

Hino, T., and J. B. Russell. 1987. Relative contributions of ruminal bacteria and protozoa to the degradation of protein in vitro. J. Anim. Sci. 64:261-270.

Hoover, W. H., and S. R. Stokes. 1991. Balancing carbohydrate and protein for optimum rumen microbial yield. J. Dairy Sci. 74:3630-3644.

Hwang, B.-J., H.-J. Yeom, Y. Kim, and H.-S. Lee. 2002. Corynebacterium glutamicum utilizes both transsulfuration and sulfhydrylation pathways for methionine biosynthesis. J. Bacteriol. 184:1277-1286.

Ives, S. E., E. C. Titgemeyer, and T. G. Nagaraja. 2002. Technical note: Effect of removal of microbial cells by centrifugation on peptide and -amino nitrogen concentrations in ruminal fluid. J. Dairy Sci. 85:3059-3061.

Kajikawa, H., M. Mitsumori, and S. Ohmomo. 2002. Stimulatory and inhibitory effects of protein amino acids on growth rate and efficiency of mixed ruminal bacteria. J. Dairy Sci. 85:2015-2022.

Kajikawa, H., M. Mitsumori, K. Tajima, and M. Kurihara. 2005. Amino acids antagonistic to the amino acids inhibitory for growth rate of mixed ruminal bacteria. J. Dairy Sci. 88:2601-2603.

Karnati, S. K. R., Z. Yu, and J. L. Firkins. 2009. Investigation unsaturated fat, monensin, or bromoethanesulfonate in continuous cultures retaining ruminal protozoa. II. Interaction of treatment and presence of protozoa on prokaryotic communities. J. Dairy Sci. 92:3861-3873.

Karnati, S. K. R., J. T. Sylvester, S. M. Noftsger, Z. Yu, N. R. St-Pierre, and J. L. Firkins. 2007. Assessment of ruminal bacterial populations and protozoal generation time in cows fed different methionine sources. J. Dairy Sci. 90:798-809.

Kim, H. K., J.-J. Choung, and D. G. Chamberlain. 1999. Effects of varying the degree of synchrony of energy and nitrogen release in the rumen on the synthesis of microbial protein in lactating dairy cows consuming a diet of grass silage and a cereal-based concentrate. J. Sci. Food Agric. 79:1441-1447.

Loenen, W. A. M. 2006. S-adenosylmethionine: jack of all trades and master of everything? Biochem. Soc. Trans. 34:330-333.

Lundquist, R. G., M. D. Stern, D. E. Otterby, and J. G. Linn. 1985. Influence of methionine hydroxy analog and dl-Methionine on rumen protozoa and volatile fatty acids. J. Dairy Sci. 68:3055-3058.

Miura, H., M. Horiguchi, and T. Matsumoto. 1980. Nutritional interdependence among rumen bacteria, Bacteroides amylophilus, Megasphaera elsdenii, and Ruminococcus albus. Appl. Environ. Microbiol. 40:294-300.

Neufeld, J. D., M. G. Dumont, J. Vohra, and J. C. Murrell. 2007. Methodological considerations for the use of stable isotope probing in microbial ecology. Microbial Ecology. 53:435-442.

Noftsger, S. M., N. R. St-Pierre, S. K. R. Karnati, and J. L. Firkins. 2003. Effects of 2- hydroxy-4-(methylthio) butanoic acid (HMB) on microbial growth in continuous culture. J. Dairy Sci. 86:2629-2636.

Noftsger, S., N. R. St-Pierre, and J. T. Sylvester. 2005. Determination of rumen degradability and ruminal effects of three sources of methionine in lactating cows. J. Dairy Sci. 88:223-237.

Nolan, J. V., and S. Stachiw. 1979. Fermentation and nitrogen dynamics in merino sheep given a low-quality-roughage diet. Br. J. Nutr. 42:63-80.

NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Nat. Acad. Sci., Washington, DC.

Nudler, E., and A. S. Mironov. 2004. The riboswitch control of bacterial metabolism. Trends Biochem. Sci. 29:11-17.

Or-Rashid, M. M., R. Onodera, and S. Ward. 2001. Biosynthesis of methionine from homocysteine, cystathionine and homoserine plus cysteine by mixed rumen microorganisms in vitro. Appl. Microbiol. Biotechnol. 55:758-764.

Papas, A., G. A. B. Hall, E. E. Hatfield, and F. N. Owens. 1974. Response of lambs to oral or abomasal supplementation of methionine hydroxy analog or methionine. J. Nutr. 104:653-659.

Patterson, J. A., and L. Kung Jr. 1988. Metabolism of DL-methionine analogs by rumen microorganisms. J. Dairy Sci. 71:3292-3301.

Pisulewski, P. M., H. Rulquin, J. L. Peyraud, and R. Vérité. 1996. Lactational and systemic responses of dairy cows to postruminal infusions of increasing amounts of methionine. J. Dairy Sci. 79:1781-1791.

Robert, J. C., C. Richard, T. d’Alfonso, N. Ballet, and E. Depres. 2001. Investigation of the site of absorption and metabolism of a novel source of metabolizable methionine: 2- hydroxy-4(methylthio) butanoic acid isopropyl ester (HMBi). J. Dairy Sci. 84(Suppl. 1):35. (Abstr.)

Rodionov D. A., A. G. Vitreschak, A. A. Mironov, M. S. Gelfand. 2004. Comparative genomics of the methionine metabolism in Gram positive bacteria: a variety of regulatory systems. Nucleic Acids Res. 32:3340-3353.

Rolfe, R., and M. S. Meselson. 1959. The relative homogeneity of microbial DNA. Proc. Natl. Acad. Sci. USA 45:1039-1043.

Rulquin, A., and L. Delaby. 1997. Effects of the energy balance of dairy cows on lactational responses to rumen-protected methionine. J. Dairy Sci. 80:2513-2522.

Rulquin, H., B. Graulet, L. Delaby, and J. C. Robert. 2006. Effect of different forms of methionine on lactational performance of dairy cows. J. Dairy Sci. 89:4387-4394.

Russell, J. B., C. J. Sniffen, and P. J. Van Soest. 1983. Effect of carbohydrate limitation on degradation and utilization of casein by mixed rumen bacteria. J. Dairy Sci. 66:763– 775.

Salsbury, R. L., D. K. Marvil, C. W. Woodmansee, and G. F. Haenlein. 1971. Utilization of methionine and methionine hydroxyl analog by rumen microorganisms. J. Dairy Sci. 54:390-396.

Scheifinger, C. C., and M. J. Wolin. 1973. Propionate formation from cellulose and soluble sugars by combined cultures of Bacteroides succinogenes and Selenomonas ruminantium. Appl. Microbiol. 26:789-795.

Schwab, C. G., C. K. Bozak, N. L. Whitehouse, and M. M. A. Mesbah. 1992. Amino acid limitation and flow to duodenum at four stages of lactation. 1. Sequence of lysine and methionine limitation. J. Dairy Sci. 75:3486-3502.

Schwab, C. G., N. L. Whitehouse, A. M. McLaughlin, R. K. Kadariya, N. R. St-Pierre, B. K. Sloan, R. M. Gill, and J.-C. Robert. 2001. Use of milk protein concentration to estimate the methionine availability of two forms of 2-hydroxy-4-methylthio butanoic acid (HMB) for lactating cows. J. Dairy Sci. 84(Suppl. 1):35. (Abstr.)

Seal, C. J., and C. K. Reynolds. 1993. Nutritional implications of gastrointestinal and liver metabolism in ruminants. Nutr. Res. Rev. 6:185–208.

Socha, M. T., C. G. Schwab, D. E. Putnam, N. L. Whitehouse, B. D. Garthwaite, and G. A. Ducharme. 2008. Extent of methionine limitation in peak-, early-, and mid-lactation dairy cows. J. Dairy Sci. 91:1996-2010.

Sperandio, B., C. Gautier, S. McGovern, D. S. Ehrlich, P. Renault, I. Martin-Verstraete, and E. Guedon. 2007. Control of methionine synthesis and uptake by MetR and homocysteine in Streptococcus mutans. J. Bacteriol. 189:7032-7044.

St-Pierre, N. R., and J. T. Sylvester. 2005. Effects of 2-hydroxy-4-(methylthio) butanoic acid (HMB) and its isopropyl ester on milk production and composition by Holstein cows. J. Dairy Sci. 88:2487-2497.

Stulke, J., W. Hillen. 1999. Carbon catabolite repression in bacteria. . Curr. Opin. Microbiol. 2:195-201.

Van Soest, P. J. 1994. Nutritional ecology of the ruminant, 2nd ed. Cornell University Press, Ithaca, NY.

Vazquez-Anon, M., T. Cassidy, P. McCullough, and G. A. Varga. 2001. Effects of Alimet on nutrient digestibility, bacterial protein synthesis, and ruminal disappearance during continuous culture. J. Dairy Sci. 84:159-166.

Vitreschak, A. G., A. A. Mironov, V. A. Lyubetsky, and M. S. Gelfand. 2008. Comparative genomic analysis of T-box regulatory systems in bacteria. RNA. 14:717- 735.

Wallace, R. J., and M. L. Brammell. 1985. The role of different species of bacteria in the hydrolysis of protein in the rumen. J. Gen. Microbiol. 131: 821-832.

Walker, N. D., C. J. Newbold, and R. J. Wallace. 2005. Nitrogen metabolism in the rumen. Pages 71–115 in Nitrogen and Phosphorus Nutrition of Cattle. E. Pfeffer and A. Hristov, ed. CABI Publishing, Cambridge, MA.

Whiteley, A. S., B. Thompson, T. Leuders, and M. Manefield. 2007. RNA stable-isotope probing. Nature Protocols. 2:838-844.

Williams, A. G., and N. H. Strachan. 1984. Polysaccharide degrading enzymes in microbial populations from the liquid and solid fractions of bovine rumen digesta. Can. J. Anim. Sci. 64(Suppl):58-59.

Williams, A. G. 1988. Factors affecting the formation of polysaccharide-degrading enzymes by rumen micro-organisms. Anim. Feed Sci. Technol. 21:191-195.

Williams, A. G., S. E. Withers, and N. H. Strachan. 1989. Postprandial variations in the activity of polysaccharide-degrading enzymes in microbial populations from the digesta solids and liquor fractions of rumen contents. J. Appl. Bacteriol. 66:15-26.

Windschitl, P. M., and M. D. Stern. 1988. Influence of methionine derivatives on effluent flow of methionine from continuous culture of ruminal bacteria. J. Anim. Sci. 66:2937- 2947.

Yu, Z. and W. W. Mohn. 1999. Killing two birds with one stone: simultaneous extraction of DNA and RNA from activated sludge biomass. Can. J. Microbiol. 45:269-272.