THE EFFECT OF SUPPLEMENTING NATIVE RUMEN MICROBES ON MILK PRODUCTION OF DAIRY CATTLE By Katelyn E. Goldsmith A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Animal Science - Master of Science 2022 1 ABSTRACT THE EFFECT OF SUPPLEMENTING NATIVE RUMEN MICROBES ON MILK PRODUCTION OF DAIRY CATTLE By Katelyn E. Goldsmith Dairy cows are commonly fed direct-fed microbials (DFM) to improve milk production and efficiency. Most DFM are not native to the dairy cow rumen which may limit their ability to interact with the native microbiome. We evaluated the effects of two DFM supplements containing 4 native rumen microorganisms on the production of dairy cows. Ninety Holstein cows (45 ± 10 kg milk/d, mean ± standard deviation; 40% primiparous) were fed a common diet. After 14 d, they were blocked by parity, days in milk, and energy corrected milk (ECM) per unit of metabolic body weight. Within block, cows were randomly assigned to treatments, which were top-dressed daily for the next 112 d. Treatments were 150 g of ground corn mixed with 1) no live DFM (CON), 2) 5 g of a live DFM (Galaxis 2.0; G2), and 3) 5 g of a live DFM (Galaxis 2.0 Plus; G2P). G2 and G2P were products of Native Microbials Inc. (San Diego, CA) and contained the same organisms but in different concentrations. Supplementation with DFM did not alter yield of total milk, protein, or fat, but slightly decreased body weight gain and body condition score gain with no difference between G2 and G2P. DFM tended to decrease dry matter intake (DMI) but did not significantly improve feed efficiency (ECM/DMI). DFM did not alter digestibility of fiber, starch, protein, or fat, and did not alter concentrations of glucose or non-esterified fatty acids but tended to decrease concentration of insulin in plasma. DFM decreased somatic cell counts in milk with no difference between G2 and G2P. In conclusion, supplementation with native DFM had little impact on milk production and efficiency. ACKNOWLEDGMENTS I want to extend my appreciation to all those who supported me during my time at MSU. I would like to thank Dr. Michael VandeHaar for his mentorship. Through his support and encouragement, I have grown immensely not only as a scientist but in confidence and knowledge. I am incredibly grateful to have had the opportunity to work with him. I also extend gratitude to my committee members, Dr. Bradford, Dr. Weber-Nielsen, and Dr. Abuelo. I appreciate their support of my research as well as investment in my education. Also, thanks to Jim Liesman for guiding me through the entire process of collecting, organizing, analyzing, and presenting data. Lynn Worden for being an incredible resource on the ins and outs of running experiments at the MSU Dairy farm. The MSU dairy farm staff for their dedication and hard work as well as understanding and guidance as I conducted this research. Their work was integral to the success of this research. The undergraduate students who worked on this research deserve a large portion of my gratitude. Without them, this research would not have been possible. I appreciate their reliability, flexibility, and investment in the success of this project. They made 9 months of early mornings and late nights at the farm incredibly fun. I want to thank my high school teacher and mentor, Cheryl Kluesner. The time spent in her classroom sparked my interest in the intersection of science and dairy farming. Her support and mentorship continue to resonate in my professional and personal life each day. Finally, words cannot describe the gratitude I feel for my family and friends who support me and my goals. I am incredibly grateful to Kristen Gallagher as well as Kirby and Sydney Krogstad who were the support group that helped me navigate beginning graduate school during iii a pandemic. Many thanks to all the other grad students for their encouragement, help, and entertaining experiences. Lastly, the sacrifices made by my parents and siblings to support and encourage me throughout life means more than I could ever explain. I am forever grateful to them. iv TABLE OF CONTENTS LIST OF TABLES……………………………………………………………………. vii LIST OF FIGURES…………………………………………………………………... viii KEY TO ABBREVIATIONS………………………………………………………… ix CHAPTER 1………………………………………………………………………….. 1 INTRODUCTION……………………………………………………………………. 1 CHAPTER 2………………………………………………………………………….. 3 LITERATURE REVIEW…………………………………………………………….. 3 The Rumen Microbiome……………………………………………………… 3 Rumen Microbiome Establishment…………………………………... 3 Rumen Microbiome Grouping………………………………………... 4 Energy Production by the Rumen Microbiome………………………. 5 Dietary Effects on Microbiome Composition………………………… 6 Microbiome and Rumen pH………………………………………….. 6 Rumen Microbiome Manipulation…………………………………… 7 Direct-Fed Microbials………………………………………………………… 8 Yeast and the Rumen…………………………………………………. 9 Species of Yeast as DFM……………………………………………... 10 Bacterial DFM………………………………………………………... 12 Lactic Acid Producing Bacteria (LAB) as DFM..……………………. 12 Lactic Acid Utilizing Bacteria (LUB) as DFM..……………………... 14 Combination DFM……………………………………………………. 16 Endomicrobial Supplement…………………………………………………... 18 Ruminococcus bovis………………………………………………….. 18 Butyrivibrio fibrisolvens……………………………………………… 19 Clostridium beijerinckii……………………………………………… 21 Pichia kudriavzevii…………………………………………………... 22 Conclusions…………………………………………………………………… 23 CHAPTER 3………………………………………………………………………….. 25 MATERIALS AND METHODS…………………………………………………….. 25 Design and Treatments……………………………………………………….. 25 Data Collection and Analysis………………………………………………… 27 Statistical Analysis……………………………………………………………. 29 CHAPTER 4………………………………………………………………………….. 30 RESULTS AND DISCUSSION……………………………………………………… 30 CHAPTER 5………………………………………………………………………….. 41 OVERALL CONCLUSIONS………………………………………………………… 41 v APPENDIX…………………………………………………………………………… 43 REFERENCES………………………………………………………………………... 53 vi LIST OF TABLES Table 3.1. Ingredient and nutrient composition of common ration………………. 26 Effects of DFM supplementation on milk yield, milk composition, Table 4.1. BW, BCS, and efficiency……………………………………………... 32 Table 4.2. Effect of DFM supplementation on nutrient digestibility…………….. 35 Table 4.3. Effect of DFM supplementation on plasma metabolite concentrations 39 Income over feed costs when milk income or feed costs change for Table A.1. control treatment……………………………………………………… 46 Income over feed costs when milk income or feed costs change for Table A.2. G2P treatment………………………………………………………… 46 Difference between G2P and CON income over feed costs when milk Table A.3. income or feed costs change1……………………………………….. 46 Comparison of production, efficiency, digestibility, and plasma Table A.4. metabolite concentrations between cohorts…………………………... 49 Comparison of response to DFM supplementation broken out by Table A.5. cohort1 and treatment2……………………………………………….. 51 Table A.6. Nutrient composition of common ration for each cohort……………... 52 vii LIST OF FIGURES Figure A.1. Daily minimum and maximum temperatures for duration of trial……. 52 viii KEY TO ABBREVIATIONS AA = Amino acid ADF = Acid detergent fiber ADG = Average daily gain ATP = Adenosine triphosphate ATTD = Apparent total tract digestibility BCS = Body condition score BHBA = Beta-hydroxybutyrate BW = Body weight CAPE = Captured energy CFU = Colony forming units CON = Control treatment CP = Crude protein DFM – Direct-fed microbial DIM = Days in milk DMI = Dry matter intake ECM = Energy corrected milk FEEDE = Feed energy G2 = Galaxis 2.0 treatment G2P = Galaxis 2.0 Plus treatment GIT = Gastrointestinal tract iNDF = Indigestible neutral detergent fiber IOFC = Income over feed cost ix LAB = Lactic acid producing bacteria LUB = Lactic acid utilizing bacteria MCP = Microbial protein MILKE = Milk energy MOS = Mannan-oligosaccharides MUN = Milk urea nitrogen NDF = Neutral detergent fiber NEFA = Non-esterified fatty acid OM = Organic matter SARA = Subacute ruminal acidosis SCC = Somatic cell count SD = Standard deviation SEM = Standard error of the mean TMR = Total mixed ration VFA = Volatile Fatty Acid x CHAPTER 1 INTRODUCTION The rumen microbiome is a robust community of microorganisms that plays an integral role in many functions of the ruminant. These microorganisms convert feed to usable energy, producing approximately 70% of the energy supplied to the animal, as well important precursors for substrates essential to milk production and composition (Bergman, 1990). Improving the function of the rumen microbiome by supplementing additional, beneficial organisms could improve its energy and substrate production. For these reasons, DFM are commonly used in the dairy industry to manipulate and improve the function of the rumen microbiome and therefore, overall milk production and efficiency. Many studies have evaluated the effect of DFM supplementation in ruminants (Yoon and Stern, 1995; Chaucheyras-Durand et al., 2008; Seo et al., 2010). Many modes of action have been proposed for DFM and include moderating pH, moderating redox potential, and improving nutrient digestibility (Chaucheyras-Durand et al., 2008; Yoon and Stern, 1995). DFM sometimes increase milk production (Boyd et al., 2011; Nocek et al., 2003) and lower rumen pH (Nocek et al., 2002), but results are inconsistent and vary by DFM species, strain, dosage, frequency, and animal physiological status (Chaucheyras-Durand et al., 2008). Most organisms in commercial DFM are not native to the rumen (Henderson et al., 2015). The strong resilience of the rumen microbiome may limit interaction with these non- native organisms (Weimer, 2015). For this reason, supplementing native rumen microorganisms may improve DFM effects. Research evaluating native rumen microorganism supplementation and its effects on production or efficiency of ruminants is limited. A study by Goetz et al. (2021) supplemented a DFM with two native rumen organisms to dairy cows and reported no significant effects. When completing a retrospective analysis, Goetz et al. (2021) found low and high 1 producing cows responded differently to native DFM. Therefore, we were interested in how DFM containing native microorganisms affect milk production and nutrient digestibility in dairy cattle. 2 CHAPTER 2 LITERATURE REVIEW The Rumen Microbiome The rumen microbiome is the community of prokaryotes (bacteria and archaea) and eukaryotes (fungi and protists) that exist in a symbiotic relationship with the ruminant. One mL of rumen fluid contains approximately 25 billion bacteria and archea, 10 million protozoa, and 10 thousand fungi (Leedle et al., 1982; Clarke et al., 1965; Joblin, 1981). When considering the total contents of the rumen, over three quadrillion (1x1015) microbes inhabit the environment (Jurgens, 2002). For many years, species identification was difficult as most rumen microorganisms are unculturable. With recent advances in genome analytics, researchers have identified over 5,000 unique microorganisms (Stewart et al., 2019). The rumen microbiome provides many advantages to the ruminant. These microorganisms enable the cow to ferment cellulosic material, which would be otherwise undigestible, and form volatile fatty acids (VFA). Microbes ferment other organic matter as well. Fermentation accounts for approximately 70% of the energy supplied to the animal (Bergman, 1990). VFA also serve as important precursors for substrates essential to milk synthesis. Rumen microbes also produce vitamins and amino acids through fermentation (Moran, 2005). Researchers and the industry are interested in understanding the roles, interactions, limitations, and potential for manipulation of the rumen microbiome. Rumen Microbiome Establishment The species that form the rumen microbiome community are established early in dairy cattle. Within hours of birth, bacteria and yeast are already colonizing the rumen (Rey et al., 2014; Abecia et al., 2014; Guzman et al., 2015). These microorganisms are introduced to the calf through their dam, other adult animals, and the surrounding environment (Ziolecki and 3 Briggs, 1961). Following initial colonization, the species and concentrations of the rumen community shift depending on diet, management, and unknown innate animal factors (Rey et al., 2014; Bryant and Small, 1960; Eadie, 1962; Malmuthuge et al., 2014). Within 3-4 weeks after birth, the microbial community reaches a state of stabilization where many of the present microorganisms form the core successional microbiome (Rey et al., 2014; Abecia et al., 2014; Guzman et al., 2015; Furman et al., 2020). The rumen core successional microbiome consists of species that maintain a high persistency throughout life. Although it may experience slight composition changes depending on age and diet, it is highly resistant to perturbations (Furman et al., 2020; Moraϊs and Mizrahi., 2019). Microorganisms that are early rumen colonizers (0 – 10 days after birth) are more likely to persist and be a part of the core successional microbiome (Furman et al., 2020). The species that compose the core successional microbiome may affect animal health, production, and efficiency throughout their life. Recent research suggest genetics may also be important in rumen colonization (Wallace et al., 2019). Rumen Microbiome Grouping With such a vast number of microorganisms, it can be difficult to describe their roles and interactions. Moraϊs and Mizrahi (2019) propose evaluating the rumen microbiome using functional grouping and community states. A functional group can consist of many different strains, species, or even kingdoms of microorganisms that serve the same function in the rumen. For example, all microbial species that utilize pyruvate to produce propionate compose one functional group. Together, functional groups form the community state. Community states are the stable microbiome composition that may alter slightly but primarily resists change unless drastically perturbed. Community states may have different compositions between animals but still carry out the same processes. Henderson et al. (2015) suggests that microbial interactions 4 within community states do not follow exact relationships, but instead demonstrate flexible associations between members of functional groups. This flexibility aids the ruminant in adapting to and utilizing various diets (Henderson et al., 2015). Functional and community groups help to simplify comparison of rumen microbial compositions that although different, still complete the same tasks. Energy Production by the Rumen Microbiome The primary function of the rumen microbiome is fermentative digestion of feed to harvest energy. Through fermentation, microorganisms also produce waste products including VFA (France and Dijkstra, 2005). VFA serve as a crucial source of energy for the ruminant. The rumen microbiome produces approximately 70% of the energy needed by the host, mainly through VFA production (Bergman, 1990). Once produced, VFA are absorbed through the rumen epithelium into the bloodstream and transported to tissues (Stevens, 1970; Dijkstra et al., 1993). Absorbed propionate is primarily used in the liver for gluconeogenesis, providing almost all the glucose for the ruminant. Acetate and butyrate undergo oxidation via the citric acid cycle in tissues for energy utilization (France and Dijkstra, 2005). Additionally, acetate serves as a primary substrate for lipogenesis. The pattern of VFA produced in the rumen is directly linked to the rumen microbiome composition (Dijkstra et al., 1994). The rumen microbiome also serves as a valuable source of protein to the animal. Rumen microbes break down ingested protein to amino acids and ammonia to form microbial protein (MCP). Rumen microbes regularly leave the rumen via attachment to feed particles or the ruminal fluid (Moran, 2005). In the small intestine, MCP is broken down to AA and absorbed for use by the animal, accounting for approximately two thirds of the ruminant’s total absorbable 5 protein (Tas et al., 1981; Storm et al., 1983). Absorbed amino acids are partitioned to tissues for use including milk protein synthesis, growth, and maintenance (Henry and Morrison, 1928). Overall, the rumen microbiome plays a critical role in energy and protein status of the dairy cow. Dietary Effects on Microbiome Composition Diet is integral to the diversity and richness of the rumen microbial community. Although the rumen core microbiome remains stable through perturbations, the concentrations of member species maintain some plasticity (Belanche et al., 2019). When diets shift, plasticity allows the microbiome to adjust according to the nutrient substrates available. For example, dairy cows consuming fiber rich diets had increased concentrations of cellulolytic protozoa, fungi, and methanogens when compared to consuming starch rich diets (Belanche et al., 2012). Compared to eating a pasture diet, sheep consuming a starch rich diet had a less diverse and dense rumen microbiome but produced more VFA. When transitioned to a pasture diet, the sheep’s microbiome adjusted to the newly available nutrients and was more diverse and denser (Belanche, 2019). Diet contributes greatly to the composition, concentration, and complexity of the rumen microbiome. Microbiome and Rumen pH The microbiome is an important regulator of rumen pH. Maintaining the typical pH of the rumen (5.6-7) is critical for the survival and efficacy of the rumen microorganisms (Grünberg and Constable, 2009). When VFA accumulate in the rumen faster than they can be absorbed by the rumen epithelium, typically from ingesting large quantities of highly fermentable carbohydrates, rumen pH begins to drop (Britton and Stock, 1991). When pH falls between 5.2- 5.6 for an extended period, cows experience subacute rumen acidosis (SARA). SARA decreases microbial community richness, diversity, fibrolytic activity, buffering capacity, and rumen 6 motility (Stewart, 1977; Lorenz, 2016; McCann et al., 2016). Low rumen pH stimulates the growth of lactic acid producing bacteria, generating lactate. Lactate has a lower pKa than VFA (pKa 3.9 vs. 4.9) making it more difficult for the rumen epithelium to absorb (Nagaraja and Titgemeyer, 2006). If pH is not stabilized at this point, a spiraling effect may occur. When pH drops below 5.2, described as acute acidosis, inflammation, erosion, and ulceration of the ruminal epithelium can occur (Mutsvangwa et al., 2002; Lorenz, 2016). Understanding the relationship between diet, rumen microorganisms, and rumen pH is important to maintain healthy conditions for animal and microbial success. Rumen Microbiome Manipulation Across cows, the core microbiome may be similar, but each is unique in its concentration and composition of species. Most microbes composing the rumen core microbiome are early colonizers, almost all introduced during the days following birth (Furman et al., 2020). This suggests that early intervention can have lasting effects on the core rumen microbiome. Research has shown that manipulating feed management in early life formed unique microbiomes that persisted later in life (Yáñez-Ruiz et al., 2015; Abecia et al., 2014). Furman et al. (2020) demonstrated that different birth delivery methods (cesarean section compared to vaginal) affected rumen microbial composition that endured through life. Research is still needed to determine best timeframes, methods, and microorganisms for effective and persistent intervention in calves. Once established, the core microbiome is difficult to significantly alter. In a study by Weimer et al. (2010), 2 pairs of cows whose bacterial community composition were significantly different underwent exchange of 95% of their rumen contents. These cows reestablished their pre-exchange pH and VFA profiles within 24 hours and bacterial community composition within 7 14-61 days. Zhou et al. (2018) conducted complete rumen content exchange between 8 pairs of beef steers of varying feed efficiencies. Like Weimer et al. (2010), Zhou et al. (2018) found the animal’s rumen microbiome reestablished to similar compositions as prior to exchange. This research suggests that mature ruminants maintain a strong specificity for their core microbiome and even drastic perturbations cannot completely alter its composition. This strong specificity of the rumen microbiome must be considered when considering DFM supplementation. Direct-Fed Microbials Direct-fed microbials (DFM) have been researched and used in the dairy industry at length. DFM have been defined as products containing viable bacteria and/or yeast by the United States Food and Drug Administration (1995). DFM were originally conceived as a route to improve performance and health by manipulating the post-ruminal microbiome (Fuller, 1989). As the rumen microbiome’s role in energy production and overall health was better understood, DFM effects on rumen function became a major interest to researchers and farmers. DFM products have been researched to evaluate their modes of action, rumen effects, and animal production responses. These studies have generated inconsistent results. Additionally, the ability to compare studies is difficult due to many confounding variables influencing the effect of DFM. Many studies use different microorganisms, dosage, frequency, delivery form, diets, and animals of various physiological states. These factors can all affect the rumen microbiome and therefore, the effectiveness of DFM (Ban and Guan, 2021). This makes drawing clear conclusions difficult and therefore when evaluating the literature, readers should be aware of these factors. Regardless, results from DFM supplementation studies suggest potential to benefit rumen health, performance, and efficiency. 8 Yeast and the Rumen Yeast is commonly included in DFM because of their role as “oxygen scavengers”. Yeasts are facultative anaerobes, surviving in environments with or without oxygen present. The rumen is a primarily anaerobic environment, but dissolved oxygen can be introduced to the rumen fluid through eating, drinking, salivating, or rumination (Newbold et al., 1996). Dissolved oxygen can severally inhibit the growth and performance of anaerobic ruminal bacteria (Newbold et al., 1996). Yeast can scavenge the rumen for oxygen and utilize it for respiration. By eliminating dissolved oxygen, yeast promotes the anerobic rumen environment and subsequently, the success of bacteria, particularly fibrolytic species (Chaucheyras-Durand et al., 2008). A study by Newbold et al. (1996) exhibited this relationship in fermentation simulators by supplementing mutants of the yeast species S. cerevisiae that were unable to metabolize oxygen. Mutant S. cerevisiae were unable to promote bacterial populations or O2 uptake whereas strains with normal oxygen consuming ability increased bacterial abundance and O2 uptake. As a facultative anaerobe, yeasts can play a unique role in promoting rumen microbial success. Yeast’s uptake of oxygen is not only important for preserving anerobic conditions, but also maintaining the rumen redox potential. Rumen bacteria, particularly fibrolytic species, require low redox potential to complete metabolic functions of fermentation (Huang et al., 2018). Julien et al. (2020) demonstrated that anerobic ruminal bacteria diversity and richness is closely related to the rumen redox potential. Redox potential in the rumen is directly linked to, and increases, with the presence of dissolved oxygen. Therefore, yeast’s consumption of dissolved oxygen may help to maintain low redox potential and promote bacterial growth. Live yeast also produce growth factors such as vitamins and organic acids that are utilized by the bacteria, 9 further decreasing redox potential (Jouany, 2006; Chaucheyras-Durand et al., 2008). In vitro studies have demonstrated that feeding live yeast decreases the redox potential in the rumen of lambs (Chaucheyras-Durand and Fonty, 2002; Mathieu et al., 1996) and dairy cows (Křížová et al., 2011). This in vitro research suggests DFMs containing yeast can be used to maintain a low rumen redox potential and effectively promote microbial performance. Yeast may also stabilize rumen pH by promoting viable bacterial growth. Yeast produce growth factors such as amino acids and peptides that promote growth of rumen lactate-utilizing bacteria (LUB; Chaucheyras-Durand et al., 2008). LUB can ferment lactic acid, a major contributor to acute acidosis, and increase rumen pH. Additionally, yeast can stimulate protozoa that utilize lactate, compete with amylolytic bacteria, and ferment starch at a slower rate (Brassard et. al., 2006; Mendoza et al., 1993; Williams and Coleman, 1997). Studies have reported that feeding supplemental live yeast increased average, minimum, and maximum rumen pH (Křížová et al., 2011; Bach et al., 2007; Nocek et al., 2003; Thrune et al., 2009) whereas others (Yoon and Stern, 1996; Nocek et al., 2002) saw no difference. Bach et al. (2007) reported that cows supplemented with live yeast had shorter bouts of rumen pH below both 5.6 and 6.0. Stabilizing rumen pH and stimulating bacterial growth can promote optimum animal health, digestion, and performance. Overall, yeast may improve rumen conditions by maintaining an adequate anaerobic environment, low redox potential, and moderating pH. Species of Yeast as DFM The most studied supplemental yeast species for dairy cattle is Saccharomyces cerevisiae. S. cerevisiae is a common yeast used in industrial processes from baking to beer making and is not native to the rumen. Supplementation of live S. cerevisiae to cattle has garnered inconsistent results (Yoon and Stern, 1996; Doreau and Jouany, 1998; Tesfaye and Hailu, 2019). 10 Supplemental S. cerevisiae has increased milk yield in goats (Stella et al., 2007) and decreased milk yield for dairy cows (Rossow et al., 2018). Both studies also reported decreased milk fat concentrations. In comparison, similar studies found no difference in milk yield (Maamouri and Ben Salem, 2021; Hossain et al., 2014) but some reported increased milk protein concentrations (Rossow et al., 2018; Hossain et al., 2014). When evaluating S. cerevisiae’s effect on digestion, several researchers found no change in ATTD of NDF, OM or ADF (Yoon and Stern, 1996; Chung et al., 2011; Doreau and Jouany, 1998) but Chung et al. reported lower ATTD of CP (2011). Chung et al. (2011) also detected no effect on animal’s BW or BCS change. Research reporting the effects of S. cerevisiae on cow efficiency was not found at the time of this literature review. The inconsistent results of supplementary S. cerevisiae may be tied to several factors. Dosing of DFM is influential on its effect. Nocek et al. found that various doses of the same strain of S. cerevisiae had differing effects on rumen pH and nutrient digestion (2002). Additionally, many studies feed different strains of S. cerevisiae, which may act differently. Chung et al. (2011) reported two strains of supplemental S. cerevisiae had different effects on rumen pH and VFA concentration in nonlactating dairy cows. Other contributing factors that have not been explicitly evaluated include stage of lactation, parity, and health status of supplemented animals. Research into the yeast species Pichia kudriavzevii as a DFM is limited. Current research is primarily focused on P. kudriavzevii’s role as an aflatoxin detoxifier. When fed TMR contaminated with aflatoxin, dairy cows supplemented with P. kudriavzevii had increased DMI, but did not differ in BW gain or digestibility of feed (Intanoo et al., 2020). In chickens, supplemental P. kudriavzevii did not affect ADG in the basal diet but did improve ADG of 11 chickens fed aflatoxin contaminated feed (Magnoli et al., 2017). A study by Geotz et al. (2021) found supplemental P. kudriavzevii and C. beijerinckii (a bacteria) could have beneficial milk yield responses based on cow’s production level. Additional research of supplementary P. kudriavzevii for ruminants is needed to better understand its effectiveness as a direct-fed microbial. Overall, supplementation of yeast to the diet of dairy cattle has inconsistent but promising results. Additional research is needed to best understand the interactions of yeast strains, dosage, and frequency for dairy cattle DFM. Bacterial DFM Bacterial DFMs have been used extensively in the dairy industry. The bacterial species used in DFM vary greatly but are commonly divided into three groups: lactic acid producing bacteria, lactic acid utilizing bacteria, or other species. It is hypothesized that these bacteria may moderate the rumen environment, microbiome, and energy substrate production. Recent research also suggests bacterial DFM may participate in competitive exclusion and immunomodulation (McAllister et al., 2011). Overall, bacteria appear to be an important consideration in DFM formulation. Lactic Acid Producing Bacteria (LAB) as DFM Lactic acid producing and utilizing bacteria are commonly included in DFM together due to the symbiotic relationship. Through the fermentation of feedstuffs, LAB produce lactic acid. This lactic acid can then be used by lactic acid utilizing bacteria (LUB) to produce energy substrates such as VFA. Additionally, increasing the concentration of lactic acid in the rumen stimulates LUB growth (Yoon and Stern, 1996). An increased LUB population can moderate lactate concentration, stabilizing the rumen pH and rumen microflora. 12 In the small intestine, LAB may benefit the ruminant through competitive exclusion with pathogenic organisms. Some pathogenic organisms attach to the epithelial wall of the small intestine, promoting their growth and decreasing peristaltic removal (Jones and Rutter, 1972). LAB can compete for sites of adherence to the intestinal epithelial cells or produce products that limit other organisms’ ability to adhere. This decreases the prevalence and success of pathogenic organisms colonizing the small intestine. Strains of Lactobacillus have been found to inhibit colonization of epithelium cells by E. coli and Salmonella strains through competitive adhesion or producing adhesion limiting products (Sherman et al., 2005; Chen et al., 2007). Mature biofilms formed in the GIT by well established organisms can limit bacterial DFM competition. Thus, McAllister et al. (2011) suggests that DFM may have a greater competitive potential in the GIT of young animals who have not yet formed complex biofilms. Overall, through competitive exclusion with pathogenic organisms, LAB can beneficially influence GIT health. LAB can impact the ruminant’s immune system through both a direct antibacterial effect and indirectly enhancing immune responses. All major bacterial groups produce bacteriocins, small peptides and proteins that inhibit competitor bacteria’s growth (McAllister et al., 2011). Bacteriocins typically target bacteria similar to the producing strain through varying mechanisms including membrane disruption or nucleic acid degradation (Riley and Wertz, 2002). Due to their prevalence, bacteriocins are hypothesized to play a role in rumen microbial interactions but little research exists investigating this. Bacteriocins produced by strains of E. faecium and S. bovis have been reported to inhibit some bacterial species’ growth (Mantovani et al., 2002). Research by Lee et al. (2002) and Lima et al. (2009) found bacteriocins may be as effective as an ionophore in limiting rumen methane production or AA degradation. Similar to treatment with other antimicrobials, bacteria may become resistant to bacteriocins. This warrants consideration 13 when formulating and feeding DFM. Additional research of bacteriocins and their role in the ruminant may improve the selection of bacterial species for DFM. The last believed mode of action of LAB is immunomodulation in the GIT. Research in poultry and cell culturing has indicated DFM may enhance GIT immune response. These indicators include enhanced innate, humoral, or cellular responses of the immune system (Erickson and Hubbard, 2000; Isolauri et al., 2001; Miettinen et al., 1996; Tejada-Simon and Pestka, 1999; Haller et al., 2001). There is little research in ruminants regarding DFM’s influence on the immune system, but species, dose, and other factors likely contribute (McAllister et al., 2011). Investigation into the relationship between supplemental bacteria and the ruminant’s immune system is an area of needed research. Lactic acid producing bacteria play several beneficial roles in the ruminant suggesting they may be useful in DFM. Through their symbiotic relationship with LUB, LAB may promote rumen bacterial growth. Additionally, LAB may improve the host’s health both directly and indirectly through competitive exclusion, bacteriocins, and immunomodulation. These aspects present reasonable arguments for researching and utilizing LAB in DFM for ruminants. Lactic Acid Utilizing Bacteria (LUB) as DFM Lactic acid utilizing bacteria are common in the rumen. These species ferment lactic acid and convert it to various substrates, primarily propionate. In ruminants, propionate serves as the major precursor for glucose production in the liver (Huntington, 1989). Therefore, feeding DFM with LUB may be a route to improve energy status of dairy cows. Some studies (Stein et al., 2006; Weiss et al., 2008) saw increased ruminal propionate molar concentrations when supplementing Propionibacterium (a LUB) in dairy cattle, whereas Kenney (2013) and Ghorbani et al. (2002) saw no effect in feedlot cattle. DFM dosage likely plays an important role in 14 propionate response. Ghorbani et al. (2002) saw no effect when supplementing 1010 cfu/d Propionibacterium, but Stein et al. (2006) saw a positive response with 6×1011 cfu/d Propionibacterium. Despite inconsistent effects on propionate production, no studies have detected increased plasma glucose levels in cows supplemented with LUB (West and Bernard, 2011; Boyd et al., 2011). Fermentation of lactic acid to VFA by LUB may help to moderate rumen pH and mitigate bouts of acute acidosis. Acute acidosis is commonly characterized as below normal ruminal pH and high concentrations of lactic acid (Hernández et al., 2014). Acute acidosis is not very common in dairy cattle but may occur during substantial shifts in dietary starch content such as during the transition period. Research into DFM containing LUB has shown inconsistent outcomes on moderating rumen pH. Lawrence et al. (2021) found that L. animalis and P. fruendenrichii increased rumen pH during a dramatic shift in dietary starch content. When supplementing M. elsdenii in a highly fermentable diet, Kung and Hession (1995) saw increased pH in vitro whereas Aikman et al. (2011) saw no significant effect in vivo. These results may indicate that rumen pH moderation by LUB is dependent on bacterial species, diet, or research model. Lactic acid utilizing bacteria are typically supplemented to ruminants due to their ability to consume and convert lactic acid to more favorable products. Fermentation of lactic acid may decrease lactate and increase propionate concentrations, therefore increasing rumen pH. These reactions may improve the ruminant’s energy status and mitigate development of acute acidosis. Further research investigating the extent, mechanism, or bacterial interactions through which LUB affects ruminant performance may be warranted. 15 Combination DFM Direct-fed microbials commonly include a combination of multiple strains of bacteria and/or yeast. This approach aims to optimize the benefits yeast and bacteria can provide individually, and together. A prevalent combination is the yeast species S. cerevisiae with the LAB species E. faciem. Two transition cow studies reported increased milk and component concentrations when supplementing these combined species (Nocek et al., 2003; Nocek and Kautz, 2006). Additionally, these studies reported cows fed the DFM had increased plasma glucose and insulin concentrations as well as decreased BHBA and NEFA (Nocek et al., 2003; Nocek and Kautz, 2006). This may suggest supplemental species provided more substrate for gluconeogenesis, increasing available precursors for milk production and tissue use. In contrast, a similar transition cow study by AlZahal et al. (2014) saw no effect on milk production or plasma metabolites. When fed S. cerevisiae and E. faciem, starch ATTD was increased in lactating dairy cows (AlZahal et al., 2014), but not feedlot cattle (Beauchemin et al., 2003). DFM did not affect ATTD of organic matter, NDF, ADF, or protein in feedlot cattle (Beauchemin et al., 2003). Despite possible influences on nutrient digestibility or blood metabolites, this species combination did not affect BW or BCS in either dairy cows or feedlot steers (AlZahal et al., 2014; Nocek and Kautz, 2006; Beauchemin et al., 2003). Currently, no studies have reported on how this species may affect efficiency in ruminants. Another common combination is L. acidophilus and P. freudenreichii, LAB and LUB species respectively. This combination increased milk yield and component concentrations in some studies (Boyd et al., 2011; West and Bernard, 2011), but not all (Raeth-Knight et al., 2007; Ferraretto and Shaver, 2015). West and Bernard (2011) reported L. acidophilus and E. 16 freudenreichii increased efficiency (ECM/DMI), indicating the species may have improved diet digestibility or energy partitioning. Boyd et al. (2011) reported supplemented cows had increased NDF and CP ATTD, but other studies did not detect improvements in feed digestibility or efficiency (Ferraretto and Shaver, 2015; Boyd et al., 2011, Raeth-Knight et al., 2007). Studies found no effects of L. acidophilus and P. freudenreichii on BW, BCS, or plasma metabolite concentrations (Boyd et al., 2011; Ferraretto and Shaver, 2015; West and Bernard, 2011). The inconsistent results across these studies can likely be attributed to each using differing dosages or strains of L. acidophilus and P. freudenreichii. Research of other species for DFM is much more limited. A study by Goetz et al. (2021) fed C. beijerinckii in combination with P. kudriavzevii to lactating dairy cattle. In this study, no effect was detected when considering all experimental animals but in a retrospective analysis, treatment effect differed based on original milk production. Treated animals who were low producers (<53kg/d) had increased ECM production whereas high producing cows (≥53kg/d) had decreased milk production. Xu et al. (2017) reported that a combination of L. casei and L. planatarum increased milk yield and decreased SCC after 15 days of supplementation but did not alter milk composition. Propionibacterium with L. planatarum or L. rhamnosus had no effects on milk yield or composition in both high and low starch diets (Philippeau et al., 2017). Lastly, in feedlot steers, M. elsdenii and R. bromii did not improve average daily gain or BCS change (Klieve et al., 2012). Due to the limited research of these microorganisms for DFM, little is known about their effects in ruminants. Overall, DFM utilizing combinations of LAB, LUB, and yeast could benefit ruminants by leveraging the benefits of each species with their symbiotic interactions. Benefits may include 17 improved VFA production, blood metabolite concentrations, milk yield and composition, and overall animal efficiency. Endomicrobial Supplement Many DFM contain species of bacteria and yeasts that are not native to the rumen of dairy cattle. This limits the organisms’ ability to compete or interact with the established, stable rumen microbiome (Moraïs and Mizrahi, 2019). Weimer (2015) summarized several studies where rumen fibrolytic bacteria of donor species were supplemented to other, different ruminant species. In these studies, supplemented bacteria showed weak or no persistence in the recipient’s rumen. This has generated interest in the efficacy of supplementing rumen species that are native to the recipient animal. The term “endomicrobial” has been used to describe DFM products containing microorganisms naturally occurring within the recipient animal species. The supplement studied in our research was an endomicrobial containing four, native rumen species of yeast and bacteria. These species of Ruminococcus bovis, Butyrivibrios fibrisolvens, Clostridium beijerinckii, and Pichia kudriavzevii have been proposed as promising microorganisms for endomicrobial supplementation. Ruminococcus bovis Ruminococcus bovis is a novel bacteria species first isolated from the ruminal contents of a dairy cow and named in 2021 (Gaffney et al.). A part of the Firmicutes phylum, R. bovis is a strict anaerobic coccoid bacterium that grows in chains (Schoch et al., 2020; Gaffney et al., 2021). R. bovis can survive the rumen environmental conditions, but its optimal pH for growth is 7.0-7.5. Acidic conditions can hinder the bacteria’s growth when between 6-6.5 pH and the species cannot grow when pH is lower than 5.5 (Gaffney et al., 2021). 18 R. bovis belongs to the Ruminococcus genus which comprises 3.9% of the total rumen bacterial community of cattle (Henderson et al., 2015). This genus is important for starch degradation, constituting 70-80% of starch hydrolyzing bacteria in barley fed cattle (Xia et al., 2015). R. bovis ferments starch, glucose, and glycogen, but not other sugar sources including glycerol or cellobiose (Gaffney et al., 2021). Fermentation by R. bovis produces acetate as a major product, and ethanol and glycerol as minor products (Gaffney et al., 2021). These fermentation characteristics indicate that supplemental R. bovis could promote digestion of starch and production of acetate to improve milk yield or composition. Due to its recent identification, there is currently no published research regarding R. bovis as a DFM for ruminants. Butyrivibrios fibrisolvens First isolated from the rumen of dairy cattle in the 1950s, B. fibrisolvens is a strictly anaerobic bacteria (Bryant and Small, 1955). This species is an active fiber, starch, and sugar fermenter but has shown little to no cellulose degradation capabilities (Sechovcová et al., 2019). The primary fermentation output of B. fibrisolvens is butyrate and it is considered the major butyrate producer of the rumen. (Stewart et al., 1997). In addition to carbohydrate fermentation, this species conducts polyunsaturated fatty acid biohydrogenation, particularly of linoleic acid (Kepler et al., 1966; Polan et al., 1964). Biohydrogenation by B. fibrisolvens produces the intermediate conjugated linoleic acid, a fatty acid desired for its potential health properties (Kritchevsky, 2000). With these characteristics in mind, suggested benefits of supplemental B. fibrisolvens are enhanced fiber digestion, increased energy status due to greater butyrate production, and improved fatty acid profile of milk. B. fibrisolvens may also enhance the cow’s immune system. To digest β-mannans present in hemicellulose, B. fibrisolvens produces the fibrolytic enzyme β-mannanase (Nakai et 19 al.,1993). Through the hydrolysis of β-mannan by this enzyme, mannan-oligosaccharides are released (MOS, Franco et al., 2004). There is little known about MOS exact function in animals but research by Ibuki et al. (2014) found MOS may play an important role in the immune system, particularly in the GIT. Ibuki et al. (2014) reported that pigs with intestinal inflammation fed supplemental MOS had decreased expression in the colon of pro-inflammatory factors TNF-α, IL-1β, and IL-17. In dairy cattle, somatic cell count is positively related to immune expression of TNF-α, IL-1β, and IL-17. If MOS moderated these factors it dairy cattle, it may decrease SCC (Oviedo-Boyso et al., 2007; Rainard et al., 2015, 2016). Tewoldebrhan et al. (2017) reported that low doses of supplemented β-mannanase (0.1% of DM) decreased SCC in lactating dairy cows with no effect on milk or component yield. Besides SCC, previous research in calves and dry cows found that supplementing MOS improved overall animal health, performance, and immune response to a viral challenge (Heinrichs et. al., 2003; Franklin et al., 2005). Considering MOS and β-mannanase’s effect on animal health, research is needed regarding how supplementation of B. fibrisolvens may affect the expression of immune factors and the subsequent animal response. The Butyrivibrios genus compose 4.1% of the total rumen microbial community, one of the most prevalent genera in cattle (Henderson et al., 2015). Butyrivibrios is classified as a main genus in dairy cattle, indicating the genus is heritable and a part of the core microbiome (Gonzalez-Recio et al., 2018). B. fibrisolvens concentrations can also be affected by diet composition. In a study conducted with beef cattle, when animals were switched from a diet containing a forage to concentration ratio of 100:0 to 60:40, B. fibrisolvens concentrations decreased 10-fold (Fernando et al., 2010). This research supports previous findings of a small study with dairy cattle by Mrázek et al. (2006), highlighting the important relationship the 20 bacterial species shares with dietary fiber content. Although a common species identified in the rumen, research could not be found investigating B. fibrisolvens as a DFM. Clostridium beijerinckii Clostridium beijerinckii is a common bacteria isolated from many sources beyond the rumen including feces and soil (Boone et al., 2001). The species is commonly used for industrial biofuel production due to its efficient butanol production from various biomass sources. C. beijerinckii can utilize sucrose, cellobiose, and starch for growth, butyric acid, and acetic acid production (Leschine, 2005). Increased production of the VFA by the rumen can improve the ruminant’s energy status as well as milk yield and composition. Beyond VFA production, C. beijerinckii may benefit the rumen environment and microflora through its efficient production of H2 from carbohydrate degradation (Hoang et al., 2018). Hydrogen as either an ion (H+) or dissolved gas (H2) is a central regulator of rumen fermentation and end products through modulation of the rumen redox potential (Czerkawski, 1986). Increased concentrations of H2 decreases rumen redox potential, promoting fermentation and production of propionate and methane. When partial pressure of rumen H2 decreases, fermentation shifts and favors acetate production over other end products (Hegarty and Gerdes, 1999). The fermentation characteristics of C. beijerinckii suggests supplementation of the species may moderate the rumen’s redox potential encouraging fermentation and VFA production. Clostridium beijerinckii as a DFM has been scarcely researched. Recently, Goetz et al. (2021) fed a combination of C. beijerinckii and P kudriavzevii to lactating dairy cattle. In this study, no effect was detected when considering all experimental animals for milk yield and composition, DMI, efficiency or body weight. When conducting a retrospective analysis, 21 treatment had differing responses based on a cow’s milk production levels. Cows fed DFM who were low producers (<53kg/d) experienced improved ECM production. In contrast, high producing cows (≥53kg/d) decreased in milk yield. Feeding the same species combination, Lefler et al. (2020) reported a treatment by week interaction with improved daily milk, ECM, and milk fat yield. These studies provide evidence that C. beijerinckii may be a beneficial DFM, but more research is needed to better understand its effects. Pichia kudriavzevii Pichia kudriavzevii, previously known as Candida krusei (Douglass et al., 2018), is the most abundant yeast in the rumen of dairy cattle (Fernandes et al., 2019). Named in 1965 by Boidin, Pignal, and Besson (1965), this species is widely abundant in nature and isolated from many sources including berries, feces, and eggs (Kurtzman et al., 2011). P. kudriavzevii is a very acid tolerant species, growing in conditions as low as 2 pH (Qvirist et al., 2017). In vitro experiments show that this yeast ferments a wide range of carbohydrates including glucose, sucrose, lactate, hemicellulose, cellulose, and starch (Kurtzman et al., 2011; Elahi and Rehman, 2018; Natalicia Mendes de Almeida et al., 2012). Due to its ability to break down cellulose and hemicellulose (Elahi and Rehman, 2018), P. kudriavzevii could improve rumen fiber digestibility. An in vitro study by Fernandes et al. (2019) found strains of P. kudriavzevii improved NDF digestibility at 12 and 48h of fermentation in rumen cultures. In contrast, Santos et al. (2015) reported a different strain of P. kudriavzevii decreased NDF digestibility at 12 and 24h of in vitro incubation. Together, these results indicate that P. kudriavzevii may benefit rumen fiber digestion, but responses may be strain dependent. Interest in supplementing this yeast species to ruminants is relatively new and therefore, in vivo research is limited. As previously mentioned, research by Goetz et. al. (2021) found that 22 in combination with C. beijerinckii, treatment of these microorganisms increased ECM production in low producing cows (<54kg/d) but decreased milk production of high producing cows (>54kg/d). A study conducted by Intanoo et al. (2020) evaluated the detoxification effects of P. kudriavzevii supplementation in cows fed aflatoxin B1 contaminated feeds. Results showed P. kudriavzevii had the capacity to detoxify aflatoxin B1 and subsequently improve DMI and ECM yield in treated animals. Lastly, Suntara et al (2021) found that cows fed rice straw ensiled with P. kudriavzevii compared to S. cerevisiae had increased rumen microbe population 4h after feeding and improved DM digestibility. There was no difference between treatments on intake, milk production, or feed efficiency. A common yeast in the rumen, P. kudriavzevii has recently piqued the interest of researchers as a potential species to use in DFM. With some promising results reported, more research is needed to properly understand the potential effects of Pichia kudriavzevii supplementation. Conclusions The species colonizing the rumen are a diverse, abundant, and resilient community essential to animal performance and health. By supplementing combinations of bacteria and yeast, farmers may be able to improve rumen microbial fermentation, VFA production, and pH. Improvement of these factors can lead to enhanced cow energy status, health, efficiency, milk production and composition. Inconsistent results in research have indicated that dosage, species, strains, and diet are important considerations when developing DFM for dairy cattle. The objective of our research was to investigate the combination of native rumen species R. bovis, B fibrisolvens, C. beijerinckii, and P. kudriavzevii’s effects on milk production and 23 composition, efficiency, blood metabolites, and nutrient digestion. These species have minimal reported research but display many beneficial functions, suggesting they may be good candidates for a DFM. This research will advance our knowledge of this novel combination of species, guide future research of best species and dosage for DFM, and determine the benefits this DFM may provide dairy farmers. 24 CHAPTER 3 MATERIALS AND METHODS Design and Treatments Ninety lactating Holstein cows at the Michigan State University Dairy Cattle Teaching and Research Center were used in two cohorts (cohort 1, 11/2020 to 03/2021, n=39, 53% primiparous; and cohort 2, 01/2021 to 06/2021, n=51, 29% primiparous) in a randomized complete block design. Mean DIM, milk yield, and BW for all cows at the beginning of the study (mean ± SD) were 92 ± 23 d, 45.4 ± 10.3 kg/d, 659 ± 86 kg, respectively. Within cohort, cows were blocked by preliminary period parity, DIM, and ECM/BW0.75, and within block, randomly assigned to treatment. All experimental procedures were approved by the Michigan State University Institutional Animal Care and Use Committee. Cows were fed a common diet for a 14-d preliminary period, then treatments were topdressed daily on the common diets for 112 d. Treatments were 150 g of ground corn mixed with 1) no DFM (CON), 2) 5g of a live DFM (G2) containing Clostridium beijerinckii at 1.0×107 cfu and Pichia kudriavzevii, Ruminococcus bovis, and Butyrivibrios fibrisolvens at 1.0×108 cfu (Galaxis 2.0; Native Microbials Inc., San Diego, CA) or 3) 5g of a live DFM (G2P) that was similar to G2 but contained more C. beijerinckii (7.5×107 cfu) and P. kudriavzevii (1.0×109 cfu; Galaxis 2.0 Plus; Native Microbials Inc., San Diego, CA). Treatments were mixed into the top 15 cm of each cow’s feed daily before she had access to it. Treatments were obtained from the manufacturer every 3 mo in individual, airtight daily packets, stored at 2°C, and mixed with the ground corn before feeding. The ingredient and nutrient composition of the diets fed as TMR are presented in Table 3.1. Forage dry matter content was determined twice weekly, and diets were adjusted accordingly. Cows were milked three times daily at 730 h, 1530 h, and 2330 h (cohort 1, before 25 61 d of treatment) or 530 h, 1330 h, and 2130 h (cohort 1, after 61 d of treatment). Cohort 2 was milked at 400 h, 1200 h, and 2000 h. All cows were housed in tiestalls throughout the experiment and had access only to their own feed. Stalls were bedded with sawdust and cleaned three times daily. Orts were recorded daily prior to feeding. Feed was offered at 115% of expected intake once daily at 1030 h (cohort 1) or 800 h (cohort 2). Water was available ad libitum in each stall. Table 3.1. Ingredient and nutrient composition of common ration. Ingredient % of DM Corn Silage 29.3 Alfalfa Silage 13.6 Ground Corn 22.1 Cottonseed, Whole 6.8 Soybean Meal 8.5 Soybean Hulls 10.2 Base Vitamin/Mineral Mix1 2.0 High Cow Supplement Mix2 7.6 Nutrient Composition DM3 53.6 NDF 29.1 Forage NDF 17.7 Starch 27.3 CP 16.8 RUP4 33.9 1 Vitamin and mineral mix contained 22.0% fine ground corn grain, 20.5% MIN-AD (MIN-AD Inc., Winnemucca, NV), 20.0% calcium carbonate, 19.1% calcium phosphate di, 10.0% sodium chloride, 4.6% sodium sesquinate, 2.0% selenium, and <1% of each of the following: bleachable fancy tallow, Intellibond VITAL 5 (Micronutrients USA LLC, Indianapolis, IN), vitamin E, vitamin A, and Vit D3 500 (Baltivet, Dubingai, Lithuania). 2 High cow supplement mix contained 39.5% Amino Plus (Ag Processing Inc., Omaha, NE), 18.4% Caledonia Pass (Caledonia Farmers Elevator, Caledonia, MI), 15.8% sodium sesquicarbonate, 12.8% calcium carbonate, 8.7% fine ground corn grain, 2.7% urea, and 1.1% Smartamine M (Adisseo, Alpharetta, GA). 3 Expressed as percent of as fed. 4 Expressed as percent of CP based on 2021 NRC. 26 Data Collection and Analysis Body weight was measured for each cow 3 d per week. Body condition was scored by 3 trained investigators on a 5-point scale in 0.25 increments (Wildman et al., 1982) at -14, 0, 28, 56, 84, and 112 d of treatment. Daily milk yield was automatically recorded at each milking. Milk samples were collected for 6 consecutive milkings per week for component analysis. Samples were stored with preservative (Bronolab W-II liquid, Advanced Instruments, Norwood, MA) at 4°C until analysis. Individual milk samples were analyzed by CentralStar Cooperative, Inc (Grand Ledge, MI) for fat, true protein, lactose, MUN, and SCC concentrations by mid- infrared spectroscopy (AOAC, 1990, method 972.160). Milk yield and component concentrations for each milking were summed for a daily total and to calculate ECM and milk component yields. Energy-corrected milk was calculated as: ECM = [(0.324 x kg milk) + (12.816 x kg milk fat) + (7.129 x kg milk protein)]. Samples of all diet ingredients and TMR (~0.5 kg) were collected once weekly and stored at -20°C until composited by month and dried. Apparent total tract digestibility (ATTD) was determined on 10 blocks (30 cows) in each cohort. Samples of all diet ingredients (~0.5 kg) and orts from each cow (~1.0 kg) were collected daily over 5 d (d 29 to 34 for cohort 1 and d 35 to 40 for cohort 2). Within cohorts, samples of diet ingredients were composited for the 5 d and orts were composited by cow for the 5 d. Feces (~400 g) was sampled from the rectum of each cow or during defecation every 15 h resulting in 8 samples/cow representing every 3 h over a day. Feces were stored at -20°C until dried and composited on an equal DM basis for each cow. Diet ingredients, orts, and fecal samples were dried at 55°C for 72 h in a forced-air oven to determine DM. Dried samples were ground with a Wiley mill (5-mm screen; Arthur H. Thomas, Philadelphia, PA). Samples of diet ingredients, orts, and feces were analyzed by Cumberland 27 Valley Analytical Services (Waynesboro, PA) for CP (method 990.03; AOAC, 2000), starch (Hall, 2009), and ether extract (method 2003.05; AOAC, 2006). Ash was determined according to AOAC method 942.05 (2000) modified to ash a 1.5 g sample for 4 h. NDF and indigestible NDF were determined according to Van Soest et al. (1991) modified to use Whatman 934-AH glass micro-fiber filters with 1.5-μm particle retention (Cytiva, Marlborough, MA). Indigestible NDF, estimated as NDF residue after 240 h in vitro fermentation, was used as an internal marker to predict fecal output (Cochran et al., 1986). Blood was sampled (~15 mL) at every fecal collection and on 93 ± 1 d at -1, +2, and +6 h after feeding into EDTA- and NaF-coated tubes. Plasma was harvested after centrifugation at 2,000 × g for 15 min at 4°C and stored at -20°C until composited by cow and analyzed. Plasma non-esterified fatty acid (NEFA), glucose, and insulin concentrations were analyzed using commercially available kits according to the manufacturer’s instructions (NEFA: Sekishui Diagnostics, Burlington, MA; glucose: Sigma-Aldrich, St. Louis, MO; insulin: Mercodia, Uppsala, Sweden). Absorbance was measured with a micro-plate reader (SpectraMax 190; Molecular Devices Corp., Sunnyvale, CA). In addition to ECM and ECM/DMI, other efficiency measurements were calculated for this study. Milk energy (MilkE), captured energy (CapE), and feed energy (FeedE), and IOFC were calculated and analyzed. Milk energy (Mcal/kg), an estimate of energy secreted in milk from fat, protein, and lactose, was calculated as: MilkE = [(9.29 x kg fat/kg milk) + (5.63 x kg true protein/kg milk) + (3.95 x kg lactose/kg milk)] (NASEM, 2021; equation 3-14b). Captured energy (Mcal/kg), an estimate of change in body tissue energy and energy secreted in milk, was calculated as: CapE = [MilkE + 6.3 x BW change/d] (NASEM, 2021). Energy content of diet (GE, Mcal/kg), an estimate of the energy content of offered feed, was calculated using NASEM 28 equation 3-2 (2021). Consumed feed energy (FeedE) was calculated as: FeedE= [Diet Energy x DMI]. IOFC was calculated as the difference between milk revenue and feed costs using average 2021 milk component prices (USDA, 2022a, 2022b). Statistical Analysis All data were analyzed using the mixed model of SAS (version 9.4; SAS Institute, Cary, NC) according to the following model: Yijkl = Cov + μ + Ci + Pj + CiPj + Bk(CiPj) + Tl + TlCi + TlPj + TlCiPj + eijkl where Yijkl = dependent variable, Cov = fixed effect of covariate, μ = overall mean, Ci = fixed effect of cohort (i = 1 to 2), Pj = fixed effect of parity (j = 1 to 2), CiPj = interaction of cohort and parity, Bk(CiPj) = random effect of block within cohort and parity, Tl = fixed effect of treatment (l = 1 to 3), TlCi = interaction of treatment and cohort, TlPj = interaction of treatment and parity, TlCiPj = interaction of treatment, cohort, and parity, and eijkl = residual error. BW change/d and BCS change/28 d were analyzed using the same model but excluding the covariate variable. Normality of results was tested using box plots, normal probability, and homogeneity of variances. SCC displayed non-normality so was log transformed (including covariate SCC data) for analysis, and back-transformed for presentation. Main effects were declared significant at P ≤ 0.05 and tendencies at 0.05 < P ≤ 0.15. All data were expressed as LSM and SEM, unless otherwise specified. Preplanned contrasts were control vs DFM and G2 vs G2P. 29 CHAPTER 4 RESULTS AND DISCUSSION Our study is the first to evaluate this combination of microorganisms supplemented to dairy cattle and the effect on milk production, digestibility, and blood metabolites. In our study, treatment did not alter yield of milk, protein, fat, lactose, MUN, or ECM (all P > 0.4, table 4.1). One of the proposed benefits of feeding DFM to cattle is improved milk yield and composition but responses have been variable. DFM containing S. cerevisiae and E. faciem improved milk yield (+1.20 kg/d) and protein concentration (+0.21%) in a study by Nocek et al. (2003) but not in a similar study by AlZahal et al. (2014). Similarly, a combination of L. acidophilus and P. fruendenrechii increased milk yield (2.4 kg/d, Boyd et al., 2011) and ECM (1.95 kg/d, West and Bernard, 2011) in some studies but not all (Raeth-Knight et al., 2007; Ferraretto and Shaver, 2015). However, Goetz et al. (2021) detected no milk production or composition differences when feeding a DFM containing C. beijerinckii and P kudriavzevii, two of the species utilized in our study. When conducting a retrospective analysis, Goetz et al. (2021) reported cows who were low milk producers in the baseline period responded to treatment differently than high producing cows. When fed DFM, low producing cows (<53kg/d ECM) significantly increased ECM yield whereas high producing cows (>53kg/d ECM) tended to decrease milk yield. This indicates that responses to DFM supplementation, specifically C. beijerinckii and P kudriavzevii, may be affected by cow’s milk production level. In our study, cows produced 44 kg of milk/d with 3.9% fat and 3.2% protein on average, and we observed no difference in treatment response for primiparous vs multiparous cows or by pretreatment milk yield. More experiments investigating differing responses to DFM supplementation between high and low producing cows may be warranted. 30 In our study, treatment significantly decreased SCC. Average SCC of CON was 20,700 cells/ml. Treatment decreased SCC (P=0.05; 7,200 cells/ml for G2 and 4,200 cells/ml for G2P), but G2 and G2P were not different (P>0.2). The cause for this improvement is unknown but may be connected to the potential role of B. fibrisolvens in immune modulation. B. fibrisolvens produces the enzyme β-mannanase, which hydrolyzes β-mannan present in hemicellulose (Nakai et al.,1993). This reaction releases mannan oligosaccharides (MOS), a molecule found to play an important role in moderating expression of cytokines including TNF-α, IL-1β, and IL-17 in gastrointestinal tissue (Franco et al., 2004; Ibuki et al., 2014). The expression of these proteins has been positively related to SCC, indicating MOS may be able to influence SCC through mediation of these cytokines (Oviedo-Boyso et al., 2007). Tewoldebrhan et al. (2017) reported low doses of supplemented β-mannanase (0.1% of DM) decreased SCC in lactating dairy cows with no effect on milk or component yield. Thus, I suggest B. fibrisolvens included in the treatment increased production of MOS through β-mannan hydrolysis, improved immune function, and decreased SCC. In our study, multiple species were supplemented and therefore, the SCC effect cannot be conclusively attributed to one organism. Additional research is required to understand the underlying relationship and mechanisms of DFM species and their effect on SCC. 31 Table 4.1. Effects of DFM supplementation on milk yield, milk composition, BW, BCS, and efficiency. Treatment1 P-Value SEM Variable CON G2 G2P CON vs Trt G2 vs G2P DMI (kg) 29.6 29.3 28.8 0.275 0.08 0.14 Milk (kg/d) 43.8 44.4 44.2 0.523 0.46 0.75 ECM2 (kg/d) 461 45.9 45.7 0.512 0.60 0.73 Fat (%) 3.90 3.85 3.87 0.053 0.37 0.77 Fat (kg/d) 1.69 1.68 1.68 0.023 0.47 0.89 Protein (%) 3.21 3.19 3.19 0.017 0.23 0.88 Protein (kg/d) 1.41 1.40 1.39 0.018 0.58 0.52 Lactose (%) 4.92 4.93 4.92 0.007 0.37 0.47 Lactose (kg/d) 2.27 2.17 2.20 0.954 0.94 0.52 MUN (mg/ml) 13.0 13.0 12.9 0.019 0.89 0.57 SCC3 (x1,000 cells/ml) 20.7 13.5 16.5 2.18 0.05 0.28 Feed Efficiency 1.54 1.57 1.58 0.016 0.06 0.43 (ECM/DMI) Milk Energy (Mcal/kg) 32.6 32.5 32.3 0.36 0.60 0.72 Milk Energy/Feed 0.244 0.245 0.249 0.0022 0.17 0.30 Energy Captured Energy/Feed 0.267 0.266 0.269 0.0022 0.74 0.21 Energy BW (kg) 694 687 686 2.47 0.02 0.61 BW Change (kg/d) 0.508 0.375 0.380 0.029 0.001 0.89 BCS 3.20 3.14 3.16 0.021 0.05 0.48 BCS Change (unit/28d) 0.070 0.052 0.048 0.008 0.04 0.74 1 Treatment were 1) control (CON); 2) 5g of Galaxis 2.0 (G2; Native Microbials, Inc.); 3) 5g of Galaxis 2.0 Plus (G2P; Native Microbials, Inc). 2 Energy-corrected milk; ECM = [(0.324 × kg milk) + (12.95 × kg milk fat) + (7.20 × kg milk protein)]. 3 Due to non-normality, SCC was log transformed for analysis. Means and SEM were back transformed for reporting. In dairy cattle, milk yield is primarily determined by energy intake. Energy intake is determined by the diet’s net energy content and the animal’s intake of the diet, making DMI important for supporting milk production. In our study, average DMI of CON cows was 29.6 kg/d. Compared with CON, DMI tended to decrease with DFM treatments (P=0.08; G2= 29.3 32 kg/d, G2P=28.8 kg/d, SEM=0.275) with a trend for the difference between G2 and G2P (P=0.14). This response differs from that of Goetz et al. (2021), who found no effect on DMI from supplementing C. beijerinckii and P. kudriavzevii, two of the species used in our study. Ferraretto and Shaver (2015) reported a tendency for decreased DMI when cows were supplemented P. freudenreichii and L. acidophilus, although others did not detect DMI effects with the same species (Raeth-Knight et al., 2017; Boyd et al., 2011; West and Bernard, 2011). These species may increase rumen propionate concentrations (Raeth-Knight et al., 2017) and increased propionate concentrations have led to decreased meal size and subsequently, reduced DMI (Allen, 2000). None of the species supplemented in our study produce propionate but may have encouraged satiety through acetate and butyrate production. R. bovis and C. beijerinkii both produce acetate through the fermentation of starch. Past research has shown infusions of acetic acid can have hypophagic effects (Gauldrón-Duarte and Allen, 2018; Sheperd and Combs, 1998). Additionally, B. fibrisolvens and C. beijericnkii produce butyrate through fermentation. Forms of butyrate have decreased DMI in some studies (Urrutia et al., 2019; Simkins et al., 1965), but not all (Izumi et al., 2019). We speculate that DFM supplementation may have increased rumen concentrations of acetate and butyrate, consequently promoting satiety and decreasing DMI. Rumen VFA concentrations were not recorded in this study and therefore, follow up research using this particular DFM should quantify its effects on rumen VFA concentration. Our study is the first to evaluate this combination of microorganisms supplemented to dairy cattle and the effect on digestibility. Digestibility plays an important role in DMI, energy partitioning, and efficiency. In our study, apparent total tract digestibility of organic matter (OM) tended (P=0.06) to decrease with treatment (CON=66.1%, G2=64.8%, G2P=65.6%, 33 SEM=0.004, table 4.2). A tendency for a difference between G2 and G2P was detected (P=0.15); this tendency for greater digestibility of OM in G2P than G2 could be related to the tendency for lower DMI in G2 than G2P. Decreased rumen digestibility of OM can increase rumen physical fill and retention, triggering satiety signals and decreasing the cow’s desire to eat (Allen, 2000). In our study, only apparent total tract digestibility was measured so conclusions cannot be drawn on difference in sites of digestion and their subsequent effect on satiety and DMI. Based on the fermentation characteristics of the species supplemented in our study, improved digestibility, rather than decreased, would have been expected, specifically of starch and fiber. Average apparent total tract digestibilities for starch, NDF, CP, and fat were 98%, 45%, 60%, and 64%, respectively, and were not altered by our DFM treatments (P>0.15). Intanoo et al. (2020) observed similar results when feeding P. kudriavzevii to cows given an aflatoxin contaminated diet. Several studies evaluating other DFM species combinations have reported no effects on apparent total tract digestibility of starch, NDF, OM, CP, or ADF (Yoon and Stern, 1996; Philippeau et al., 2017; Raeth-Knight et al., 2007; Ferraretto and Shaver, 2015; Boyd et al., 2011). Beauchemin et al. (2003) reported that feedlot cattle had decreased total tract digestibility of OM due to decreased ruminal digestion of ADF and intestinal digestion of fiber when receiving a DFM of E. faciem and S. cerevisiae. The reason for why endomicrobial supplementation decreased OM digestibility in our study is unclear and requires more investigation. 34 Table 4.2. Effect of DFM supplementation on nutrient digestibility. Treatment1 P-Value SEM Variable CON G2 G2P CON vs Trt G2 vs G2P DMI (kg) 29.6 29.3 28.8 0.28 0.08 0.14 OM digestibility (%) 66.1 64.8 65.6 0.44 0.06 0.15 NDF digestibility (%) 45.4 43.7 45.1 0.75 0.19 0.12 Starch digestibility (%) 98.5 98.5 98.4 0.18 0.88 0.66 CP digestibility (%) 60.8 59.8 60.6 0.75 0.49 0.48 Fat digestibility (%) 65.7 62.5 63.3 1.72 0.20 0.74 1 Treatment were 1) control (CON); 2) 5g of Galaxis 2.0 (G2; Native Microbials, Inc.); 3) 5g of Galaxis 2.0 Plus (G2P; Native Microbials, Inc). In our study, DFM treatments decreased BW change from 0.51 to 0.38 kg/d (P<0.0001, SEM=0.029, table 4.1) and BCS change from 0.7 to 0.05 BCS units/28d (P=0.04, SEM=0.008). No difference between G2 and G2P was detected (P>0.7). Our results differ from Goetz et al. (2021), who detected no difference in BW or BCS for cattle fed C. beijerinckii and P. kudriavzevii. Additionally, they noted a tendency for high producing cows (>53kg/d ECM) fed DFM to have increased BCS and suggested additional energy garnered from the supplement was partitioned to tissue. Many other studies evaluating DFM supplements report no effects of treatment on BW or BCS change (AlZahal et al., 2014; Beauchemin et al., 2003; Boyd et al., 2011; Ferreratto and Shaver, 2015; West and Bernard, 2011; Klieve et al., 2012). Body weight and body condition score are important indicators of animal energy status, milk production, and health (Roche et al., 2009). In our study, treated animals tended to consume less dry matter, have lower OM digestibility, and gain less weight and condition. Treated animals also maintained the same production of milk and ECM as control animals. This suggests treated animals may have partitioned more energy and nutrients away from tissue gain and toward milk production. 35 Although treatment tended to decrease DMI in our study, ECM was not affected. This resulted in a tendency for increased efficiency (ECM/DMI) in treated animals. Treatments tended to increase ECM/DMI from 1.54 to 1.57 (G2) or 1.58 (G2P; P=0.06, SEM=0.016, table 4.1) but with no difference between G2 and G2P. A similar study (Goetz et al., 2021) observed that supplementing C. beijerinckii and P. kudriavzevii increased ECM/DMI over time due to increased ECM and similar DMI. Similarly, West and Bernard (2011) saw cows supplemented with L. acidophilus and P. fruendenreichii had increased ECM/DMI due to increased ECM and no change in DMI. Increased ECM/DMI may indicate greater utilization and partitioning of feed energy through improved rumen conditions, function, or microbial yield. Efficiency responses are likely affected by factors such as dosage, diet, and animal’s physiological state as other studies feeding the same species as West and Bernard (2011) reported no differences in efficiency measures (ECM/DMI or FCM/DMI) (Ferraretto and Shaver, 2015; Boyd et al., 2011; Raeth-Knight et al., 2007). In our study, DFM did not significantly alter MilkE/FeedE or CapE/FeedE (P>0.2, table 4.1). Few other DFM studies have evaluated energy metrics to assess efficiency. Weiss et al. (2008) reported feeding Propionibacterium strain P169 to transition cows increased net energy of lactation use per unit of DMI despite no treatment effect on DMI or ECM/DMI. Supplementation of S. cerevisiae and E. faciem did not improve net energy balance in transition cows (AlZahal et al., 2014). Considering our results and others, it appears species of microorganisms or animal physiological state may affect energy efficiency. Direct-fed microbials have been proposed to improve energy metabolism for dairy cattle. Blood markers including glucose, insulin, and NEFAs are indicators of energy metabolism and 36 were analyzed in our study. Our study is the first known research to investigate supplementing these microorganisms and their effect on blood metabolites. Glucose is the main fuel source for dairy cattle and is critical to maintain good health and milk production. Propionate produced by the rumen microbiome are converted to glucose in the liver and thus, serve as the main glucose source for dairy cattle. Therefore, DFMs may increase glucose concentrations by improving the rumen microbiome’s production of propionate. In our study, treatment did not significantly alter plasma glucose concentrations (P=0.3, table 4.3), similar to the results of many other DFM trials (Beauchemin et al., 2003; AlZahal et al., 2014; West and Bernard, 2011; Boyd et al., 2011; Francisco et al., 2002; Ghorbani et al., 2002). Two DFM studies in transition cows feeding S. cerevisiae and multiple strains of Enterococcus bacteria reported increased plasma glucose concentrations following parturition (Nocek et al., 2003; Nocek and Kautz, 2006). Insulin, the essential regulator of blood glucose levels, is an equally important indicator of energy metabolism. In our study, plasma insulin concentrations tended to decrease by 11% with treatment (P=0.06; CON=0.95 μg/L, G2=0.81 μg/L, G2P=0.85 μg/L, SEM=0.057), but no difference between G2 and G2P was detected (P=0.6). This differs from the results of two transition cow studies by Francisco et al. (2002) and Nocek et al. (2003). Francisco et al. (2002) reported supplementing Propionibacteria to transition cows did not alter plasma glucose or insulin concentrations and suggested dosage or frequency of the DFM may have been inadequate. In comparison, when Nocek et al. (2003) fed S. cerevisiae and 2 Enterococcus strains pre- and post-partum, treated animals had increased plasma glucose and insulin concentrations. Nocek et al. (2003) suggests treated cows derived more energy from dietary carbohydrates to improve energy status. 37 Insulin and glucose are interdependent. Therefore, we would expect if plasma insulin concentrations increased, it is in response to increased glucose concentrations. In our study, no change was detected in plasma glucose concentrations although insulin concentrations decreased. This may indicate treatment resulted in altered tissue insulin response such as increased insulin sensitivity. We did not conduct tests to measure insulin sensitivity and therefore, cannot conclude how this DFM affected insulin sensitivity or responsiveness in treated animals. The tendency for decreased plasma insulin concentrations in treated animals likely did not substantially alter glucose uptake as 84% of glucose uptake is insulin independent (De Koster and Opsomer, 2013). Future research is required to understand more fully how this DFM affects insulin and glucose uptake in dairy cattle. Non-esterified fatty acids are commonly released into the blood when dairy cattle mobilize adipose, usually to meet energetic requirements. Elevated plasma NEFA concentrations can indicate that cows are in negative energy balance, therefore, effects of DFM supplementation on NEFA concentrations are expected to be more pronounced during the transition period than in mid- and late-lactation cows (McArt et al., 2013). At the start of our study, all cows were greater than 55 DIM. Compared to CON, treatments did not significantly affect plasma NEFA concentration (P=0.4) but a difference was detected between G2 and G2P (P=0.03; Con=90 μM, G2=81 μM, G2P=92 μM, SEM=3.489). This is similar to results by Alzahal et al. (2014) and Nocek and Kautz (2006) who both reported no effect of DFM treatment in transition cows. In contrast, two studies in postpartum dairy cows reported decreased plasma NEFA concentrations in treated cows over time (Nocek et al., 2003; Francisco et al., 2002). Compared to control, treatment did not affect plasma NEFA concentration, but there was a significant difference between G2 and G2P. When compared to G2P, G2 treated cows had 38 12% lower plasma NEFA concentrations (G2=81 μM, G2P=92 μM). G2 contained a lower concentration of C. beijerinckii and P. kudriavzevii, indicating the dosage of these species may affect fat mobilization, but the underlying mechanisms are unclear. Plasma insulin and NEFAs generally share an antagonistic relationship. Increased levels of plasma NEFA lowers insulin secretion and glucose uptake (Hue and Taegtmeyer, 2009). For this reason, it was unexpected that treatment tended to have lower plasma insulin concentrations when compared to control as well as treatment G2 having significantly lower plasma NEFA concentrations compared to G2P. The explanation for decreased plasma NEFA and insulin concentrations in G2 animals in unknown. Overall, plasma metabolite concentrations in our study indicate that treatment may have affected energy metabolism but requires more focused research to determine how and through what mechanisms. Table 4.3. Effect of DFM supplementation on plasma metabolite concentrations. Treatment1 P-Value SEM Variable CON G2 G2P CON vs Trt G2 vs G2P Glucose (md/dL) 55.5 54.6 54.5 0.790 0.34 0.91 NEFA (μM) 90.4 81.2 92.4 3.49 0.41 0.031 Insulin (μg/L) 0.947 0.812 0.851 0.057 0.057 0.570 1 Treatment were 1) control (CON); 2) 5g of Galaxis 2.0 (G2; Native Microbials, Inc.); 3) 5g of Galaxis 2.0 Plus (G2P; Native Microbials, Inc). In summary, I found that supplementing post-peak dairy cows with a DFM containing Clostridium beijerinckii, Pichia kudriavzevii, Ruminococcus bovis, and Butyrivibrios fibrisolvens did not significantly alter milk yield or composition but significantly decreased SCC. Treated animals had significantly lower change in BW (0.13 kg/d less) and BCS (0.02 units/28d less) as well as a tendency for decreased DMI, leading to a tendency for improved efficiency. Additionally, treatment cows tended to have decreased plasma insulin concentrations with maintained glucose and NEFA concentrations, indicating treatment may have altered tissue 39 insulin response but further investigation using an insulin sensitivity test is required. Overall, DFM supplementation with these microorganism species did not notably improve performance or efficiency of lactating dairy cows. Different dosage of C. beijerinckii and P. kudriavzevii also did not significantly affect the performance outcomes of treated animals. 40 CHAPTER 5 OVERALL CONCLUSIONS Direct-fed microbials have been used by the dairy industry to promote milk production and efficiency. Most organisms in commercial DFM are not native to the rumen, which may limit their ability to interact with the native microbiome (Weimer, 2015). Using species native to the dairy cow rumen may allow for improved interaction and manipulation of the rumen microbiome. Few studies have investigated the use of native rumen organisms in DFMs and their effects. The objective of this thesis was to evaluate DFM products containing organisms native to the dairy cow rumen and their effects on milk production and efficiency. This study examined a DFM containing Clostridium beijerinckii, Pichia kudriavzevii, Ruminococcus bovis, and Butyrivibrios fibrisolvens on milk production and efficiency of dairy cows. In this research, DFM did not significantly alter milk yield or composition but significantly decreased SCC. DFM significantly lowered change in BW and BCS as well as tended to decrease DMI and improve milk/feed. We suggest that the reduced gains in BW and BCS of cows fed native DFM indicate that these treatments caused cows to partition energy differently than control cows. Treated cows also tended to eat less and digest OM less efficiently but produce as much milk, which might explain the lower weight gain. Additionally, DFM did not alter plasma NEFA concentrations indicating fat mobilization was not altered to sustain production. The lack of effect on milk production and composition in this study is similar to a study by Goetz et al. (2021) who fed two of the same species as this study, C. beijerinckii and P. kudriavzevii, and also saw no change in production. This study is the first to evaluate nutrient digestibility with these supplemented species. The organisms R. bovis and C. beijerinckii were expected to enhance starch digestion (Gaffney et al., 2021; Leschine, 2005). Digestibilities of several nutrients (NDF, starch, CP, fat) were not 41 improved by our DFM treatments, similar to studies of other DFM products (Yoon and Stern, 1996; Raeth-Knight et al., 2007; Ferraretto and Shaver, 2015; Philippeau et al., 2017). Perhaps the reason we saw no improvement in digestion or production from native DFM in our study was because our diet was high in starch (29%) that was from highly digestible sources (98.5% diet starch digestibility). Thus, further improvements in starch digestion were unlikely. Results might have been different if our basal diet had a lower starch digestibility. In conclusion, these results indicate supplementation of this native rumen DFM does not significantly improve milk production or efficiency. This work provides foundational in vivo research for evaluation of native rumen species for use as a DFM. Further research is needed to identify how this DFM affects rumen microbiome dynamics, energy partitioning, with diets containing different amounts or sources of starch, and between high and low producing cows. 42 APPENDIX 43 APPENDIX Income Over Feed Cost Income over feed cost (IOFC) estimates the economic outcomes of changes in DMI, ration cost, and milk production and composition. In our study, treatment did not significantly alter IOFC (P=0.7) indicating that although there was a tendency for decreased DMI in treated animals, this did not translate to a significant economic benefit using our initial prices. The IOFC for CON at baseline daily milk income and feed costs was $8.04 (Table A.1). Therefore, DFM cows would need to achieve a baseline IOFC of $8.04 to break-even and $8.14 to achieve a 2:1 return on investment when compared to the control animals. The average performance of DFM supplemented cows in our study did not break-even or achieve a 2:1 return, earning −$0.01/d less than CON (G2P baseline IOFC=8.03) (Table A.2). I then conducted a sensitivity analysis to determine how shifts in milk prices and feed costs would affect the IOFC of CON and G2P as well as the difference between them (Table A.3). In this sensitivity analysis, IOFC was calculated using individual treatment least-squares means for daily milk income and feed costs of CON and G2P. IOFC was calculated as the difference between daily milk income and feed costs. Daily milk income was calculated as: Daily Milk Income = [Milk Price * (Milk lbs/100)]. Milk Price was calculated as: Milk Price = [(Fat% * $/lb of Butterfat) + (Protein% * $/lb of Protein) + (5.73 * $/lb Other Solids) + PPD]. The milk price was calculated using average 2021 milk component prices (USDA, 2022a, 2022b). I assumed the DFM would cost $0.10/cow/d (Torres, 2021) and average feed costs were $0.15 per lb of dry matter (Liu et al., 2021). Therefore, daily feed cost for control animals was calculated as: Daily Feed Cost = [DMI lbs * $0.15]. For cows receiving the DFM treatments: Daily Feed Cost = [(DMI lbs * $0.15) + $0.10]. 44 Milk income and feed costs were decreased or increased by 15% or 30% from baseline and IOFC was recalculated. As expected, IOFC was greatest for both CON and G2P when milk income was increased 30% and feed costs decreased 30% (CON=$16.33, G2P=$16.18). Additionally, IOFC for both treatments were lowest when milk income was decreased 30% and feed costs increased 30% (CON= −$0.25, G2P= −$0.20). As milk income decreased and feed cost increased, G2P garnered greater IOFC compared to CON but still did not achieve a 2:1 return. G2P generated the most additional IOFC compared to CON when milk income decreased 30% and feed costs increased 30% (difference = $0.15). In contrast, as milk income increased and feed costs decreased, G2P generated even less IOFC compared to CON. When milk income increased 30% and feed costs decreased 30%, G2P generated the least IOFC compared to CON (difference = −$0.16). Overall, the DFM supplemented in this trial did not significantly alter IOFC compared to CON or achieve a 2:1 return on investment not only at baseline but also when milk income and feed costs increased or decreased. 45 Table A.1. Income over feed costs when milk income or feed costs change for control treatment. Change in Milk Income ($/d)1 −30% −15% 0% +15% +30% +30% −0.25 2.42 5.10 7.78 10.45 Change in +15% 1.22 3.89 6.57 9.25 11.92 0% 2.69 5.36 8.04 10.72 13.39 Feed Costs2 −15% 4.16 6.83 9.51 12.19 14.86 −30% 5.63 8.30 10.98 13.66 16.33 1 Baseline milk income for control was $17.84. 2 Baseline feed costs for control was $9.80. Table A.2. Income over feed costs when milk income or feed costs change for G2P treatment. Change in Milk Income ($/d)1 −30% −15% 0% +15% +30% +30% −0.20 2.45 5.10 7.74 10.39 Change in +15% 1.31 3.96 6.60 9.25 11.90 0% 2.74 5.38 8.03 10.68 13.32 Feed Costs2 −15% 4.16 6.81 9.46 12.10 14.75 −30% 5.59 8.24 10.88 13.53 16.18 1 Baseline milk income for G2P was $17.64. 2 Baseline feed costs for G2P was $9.61. Table A.3. Difference between G2P and CON income over feed costs when milk income or feed costs change1. Change in Milk Income ($/d)2 −30% −15% 0% +15% +30% +30% 0.15 0.12 0.09 0.06 0.03 Change in +15% 0.09 0.06 0.03 0.00 −0.03 0% 0.05 0.02 −0.01 −0.04 −0.07 Feed Costs3 −15% 0.01 −0.02 −0.05 −0.08 −0.11 −30% −0.04 −0.07 −0.10 −0.13 −0.16 1 Calculated as: Difference = [G2P IOFC – CON IOFC]. 2 Baseline milk income for treatments were Con = $17.84 and G2P = $17.64. 3 Baseline feed costs for treatments were Con = $9.80 and G2P = $9.61. 46 Differences Between Cohorts Due to barn space and animal availability, cows in this study were divided into two cohorts. Cohort 1 (39 cows) completed the experiment from 11/13/2020 to 3/19/2021 and Cohort 2 (51 cows) from 1/29/2021 to 6/4/2021. During statistical analysis, we discovered significant differences in average production and efficiency variables between the two cohorts (Table A.4). Also, a significant interaction of cohort by treatment was identified for BW change/d (P=0.02). There was tendency for the interaction of cohort by treatment for milk fat concentration, ECM/DMI, MilkE/FeedE, CapE/FeedE, BW, and plasma insulin concentration (all 0.05