NEW APPROACHES TO ASSESS AND IMPROVE PROTEIN EFFICIENCY IN LACTATING DAIRY COWS By Enhong Liu A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Animal Science- Doctor of Philosophy 2020 ABSTRACT NEW APPROACHES TO ASSESS AND IMPROVE PROTEIN EFFICIENCY IN LACTATING DAIRY COWS By Enhong Liu The long-term goal of the work is to improve protein efficiency in lactating dairy cows. To achieve this goal, four specific objectives were proposed: 1) determine the relationship of residual feed intake (RFI) to protein efficiency in lactating Holstein cows fed high or low protein diets, 2) determine whether low protein resilience (LPR) is an indicator of protein efficiency in individual dairy cows, 3) examine the association of digestibility with RFI and LPR in lactating dairy cows, and 4) quantify the importance of including body weight (BW) change in the cow response to decreased dietary protein content and develop models for predicting BW change when dietary protein is altered. Lactating Holstein cows (n= 166; 92 primiparous, 77 multiparous) with initial milk yield (MY) of 41 ± 9.8 kg/d were fed high (HP) and low (LP) protein diets in crossover experiments of two 28-35 d periods. Experiments were repeated in 69 of the 166 cows (42 primiparous, 27 multiparous) in late lactation. Low protein diets were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate rumen-degraded protein to maintain rumen function. Expeller soybean meal was added to formulate the HP diet, which contained 18% CP in peak lactation and 16% CP in late lactation. Cows were milked twice daily; DMI and MY were recorded once daily. Milk composition was measured over 4 consecutive milkings weekly, and BW was measured 3 times weekly. Samples of feed ingredients, orts and feces were collected in the last 5 days of each period and analyzed to determine digestibilities of DM, NDF, and CP for each cow on each diet. Fixed effects of diet, parity, treatment sequence nested in experiment, treatment period nested in experiment, interaction of parity and diet, and random effects of experiment and cow nested within experiment were included in models to compare production of cows fed different levels of CP. Protein efficiency was calculated for each cow on each diet in both peak lactation and late lactation. Residual feed intake was estimated for each cow on each treatment based on the actual intake, milk energy output, metabolic BW, and body energy change (estimated from BW change and BCS). Low protein resilience was estimated for each cow in peak lactation and also late lactation, based on protein captured in milk and body tissue when fed the LP vs HP diet. A negative correlation was observed between RFI and protein efficiency in cows fed the HP and LP diets in peak lactation and cows fed the HP diet in late lactation. Cows with higher LPR values had similar protein efficiency on the HP diet but significantly higher protein efficiency on the LP diet. Neither RFI nor LPR was correlated with digestibility regardless of diets or lactation stages. When dietary protein content was reduced, 40-50 % of the total energy loss, 10-20 % of total protein loss, and 15-25% of total income loss were due to BW loss, indicating that considering only changes in milk production underestimates the impact of dietary protein changes. In conclusion, 1) cows with lower RFI values utilized protein more efficiently, and protein efficiency will be improved in the process of selecting dairy cattle for low RFI, 2) cows with higher LPR values are better able to maintain production and have higher protein efficiency to adapt to low-protein feeding conditions, 3) variation in digestibility cannot explain the variations of RFI or LPR among lactating dairy cows, and we suggest that post-absorptive metabolism explains most of the variation in RFI and LPR when lactating cows are fed diets with minimal NDF in peak lactation and 40% NDF in late lactation, and 4) body reserve mobilization should not be neglected when assessing the cow response to changes in dietary protein ACKNOWLEDGEMENTS I would like to first thank Dr. Michael VandeHaar for mentoring me through my Ph.D. training. Dr. VandeHaar taught me how to be an educated consumer of science and how to be a good scientist. He is more than an academic advisor to me. He is dedicated to supporting my personal development and raising me to be a man of confidence, persistence, and reliability. I’m really grateful for the opportunity to work with him and learn from him both academically and personally. I also want to thank Drs. Robert Tempelman, Mark Hanigan, Adam Lock, and Christopher Wolf for serving on my dissertation committee. I wish to thank Dr. Tempelman, my committee chair, for guiding me through each milestone and providing invaluable advice on statistical analyses. I wish to thank Dr. Hanigan for the insightful discussions that profoundly influence my understanding of protein metabolism and animal modeling. I would like to thank Dr. Lock for challenging me to think thoroughly about my research and guiding me to be a critical thinker and rigorous researcher. I also would like to thank Dr. Wolf for willing to serve as the outfield member and helping expand the scope of my dissertation. I also would like to thank Drs. Steven Bursian and Catherine Ernst for being so generous with their time and willing to talk to me whenever I needed professional advice for my career. I would like to thank Jim Liesman for his help in research team recruitment, sample collection, and SAS programming. I would like to thank Dave Main for his coaching in lab analyses/farm work and immersing me into American culture. I also want to express my thanks to staff at the Michigan State University Dairy Cattle Teaching and Research Center, especially Rob West, for their assistance with the experiments. I would like to thank all undergraduate students in Dr. VandeHaar’s lab for their assistance with sample collection and analyses. Without Danielle Andreen, Jared Sanderson, Andrea Luttman, Maddy Meyer, Kristina Bowen, Hannah Barnard, iv Alaina Ableidinger, Laura Livingston, and many others, I would not have completed 11 experiments in 3 years. I would also like to thank my graduate student fellows, my dear friends, who have always been there for me, bouncing research ideas with me and inspiring me - Rodrigo Araujo de Souza, Martin J. Mangual, Katie Kennedy, Yan Sun, Laura Gualdron-Duarte, Jonas de Souza, Gabriela Maldini, Rodrigo Albornoz, Julie Opgenorth, and Brandon Van Soest. Finally, I want to thank my family, who have been there for me through the ups and downs that come with getting a Ph.D. Thank you to my parents who have consistently shown me unconditional love and support. They have supported my decision in coming to the U.S. for a joint training program for two years and spending another seven years away from home pursuing a Ph.D. degree in the U.S. I also want to give special thanks to my dear wife, Yunying Le, who was my girlfriend when I started the training. We survived the four-year long-distance relationship, and together, made unforgettable memories. Thank you for always having my back, loving me and supporting me. Thank you for always believing in me even on days when I doubted myself. v TABLE OF CONTENTS LIST OF TABLES ..................................................................................................................... viii LIST OF FIGURES ....................................................................................................................... x KEY TO ABBREVIATIONS .................................................................................................... xii CHAPTER 1 ................................................................................................................................... 1 LITERATURE REVIEW ......................................................................................................... 1 FEED EFFICIENCY AND RESIDUAL FEED INTAKE ...................................................... 1 PROTEIN EFFICIENCY ......................................................................................................... 7 LOW PROTEIN RESILIENCE ............................................................................................. 17 MODELING COW RESPONSE TO DIETARY PROTEIN ................................................ 21 REFERENCES ....................................................................................................................... 24 CHAPTER 2 ................................................................................................................................. 35 RELATIONSHIP OF RESIDUAL FEED INTAKE TO PROTEIN EFFICIENCY IN LACTATING COWS FED HIGH OR LOW PROTEIN DIET ......................................... 35 ABSTRACT ........................................................................................................................... 35 INTRODUCTION .................................................................................................................. 37 MATERIALS AND METHODS ........................................................................................... 38 RESULTS ............................................................................................................................... 47 DISCUSSION ........................................................................................................................ 64 CONCLUSIONS .................................................................................................................... 68 ACKNOWLEDGEMENTS ................................................................................................... 68 APPENDIX ............................................................................................................................ 69 REFERENCES ....................................................................................................................... 71 CHAPTER 3 ................................................................................................................................. 76 LOW PROTEIN RESILIENCE IS AN INDICATOR OF RELATIVE PROTEIN EFFICIENCY OF INDIVIDUAL DAIRY COWS ............................................................... 76 ABSTRACT ........................................................................................................................... 76 INTRODUCTION .................................................................................................................. 78 MATERIALS AND METHODS ........................................................................................... 79 RESULTS ............................................................................................................................... 87 DISCUSSION ...................................................................................................................... 107 CONCLUSIONS .................................................................................................................. 110 ACKNOWLEDGEMENTS ................................................................................................. 111 APPENDIX .......................................................................................................................... 112 REFERENCES ..................................................................................................................... 114 CHAPTER 4 ............................................................................................................................... 121 ASSOCIATION AMONG DIGESTIBILITY, RESIDUAL FEED INTAKE AND LOW PROTEIN RESILIENCE IN LACTATING DAIRY COWS FED HIGH AND LOW PROTEIN DIETS .................................................................................................................. 121 vi ABSTRACT ......................................................................................................................... 121 INTRODUCTION ................................................................................................................ 123 MATERIALS AND METHODS ......................................................................................... 124 RESULTS ............................................................................................................................. 130 DISCUSSION ...................................................................................................................... 144 CONCLUSIONS .................................................................................................................. 149 ANOWLEDGEMENTS ....................................................................................................... 149 REFERENCES ..................................................................................................................... 150 CHAPTER 5 ............................................................................................................................... 155 IMPORTANCE OF CONSIDERING BODY WEIGHT CHANGE IN RESPONSE TO DIETARY PROTEIN REDUCTION IN LACTATING DAIRY COWS ........................ 155 ABSTRACT ......................................................................................................................... 155 INTRODUCTION ................................................................................................................ 157 MATERIALS AND METHODS ......................................................................................... 158 RESULTS ............................................................................................................................. 164 DISCUSSION ...................................................................................................................... 180 CONCLUSIONS .................................................................................................................. 190 ACKNOWLEDGEMENTS ................................................................................................. 190 APPENDIX .......................................................................................................................... 191 REFERENCES ..................................................................................................................... 196 CHAPTER 6 ............................................................................................................................... 200 GENERAL DISCUSSION AND CONCLUSIONS ............................................................ 200 REFERENCES ..................................................................................................................... 205 vii Table 2.2 Dry matter intake, milk production, milk components and feed efficiency for cows fed Table 2.3 Body weight, body condition score and calculated energy values for cows fed Table 2.4 Dry matter intake, milk production, milk components and feed efficiency for cows fed Table 2.5 Body weight, body condition score and calculated energy values for cows fed LIST OF TABLES Table 2.7 Repeatability of RFI across protein contents within lactation stage and across lactation Table 2.1 Feed Ingredients and Nutrient Composition of Experimental Diets 1,2 ......................... 40 treatment diets in peak lactation1,2 ......................................................................................... 49 experimental diets in peak lactation ....................................................................................... 50 treatment diets in late lactation1,2 ........................................................................................... 51 experimental diets in late lactation ......................................................................................... 52 Table 2.6 Partial regression coefficients of the RFI models in peak- and late- lactation cows ..... 54 stages (n = 69) ........................................................................................................................ 59 protein diets across lactation stage ......................................................................................... 63 peak and late lactation12 ......................................................................................................... 70 in peak and late lactation1,2 .................................................................................................... 88 and late lactation1,2 ............................................................................................................... 100 and low- LPR cows in peak and late lactation1,2 .................................................................. 102 protein efficiency terms ........................................................................................................ 106 lactation1,2 ............................................................................................................................. 131 Table 4.1 Energy output, protein efficiency and digestibility for cows fed treatment diets in peak Supplementary Table 2.1 Dry matter intake, milk production, milk components, feed efficiency, body weight, body condition score and calculated energy for cows fed treatment diets in Table 3.3 Comparisons of protein output and protein efficiency parameters of high-, medium- Table 3.4 Pearson Correlation coefficients between LPR (low protein resilience) and various Table 3.1 Dry matter intake, milk production and protein efficiency for cows fed treatment diets Table 3.2 Comparisons of production parameters of high-, medium- and low- LPR cows in peak Table 2.8 Protein efficiency and MUN of high-, medium- and low-RFI cows fed high- and low- viii Table 4.2 Energy output, protein efficiency and digestibility for cows fed treatment diets in late Table 4.3. Correlation coefficients of RFI, LPR with digestibilities of DM, CP, and NDF in peak- Table 4.4. Correlation coefficients of RFI, LPR with digestibilities of DM, CP, and NDF in late- Table 4.5 Nutrient digestibility for high-, medium- and low-RFI cows fed high- and low- protein Table 4.6 Nutrient digestibility for high-, medium- and low-LPR cows fed high- and low- protein lactation1,2 ............................................................................................................................. 133 lactation cows (n=166) ......................................................................................................... 136 lactation cows (n= 69) .......................................................................................................... 138 diets in peak and late lactation1,2,3,4 ...................................................................................... 140 diets in peak and late lactation,1,2,3,4 ..................................................................................... 142 diets in peak lactation1,2 ....................................................................................................... 165 diets in late lactation1,2 ......................................................................................................... 166 and late lactation cows across diets1,2 .................................................................................. 168 diets1,2 ................................................................................................................................... 172 different assumptions of protein gain per kg body weight change1,2 ................................... 184 Table 5.3 Mean, standard deviation, minimal and maximal values for body tissue change in peak Table 5.1. Dry matter intake, milk production, and body reserve change for cows fed treatment Table 5.2 Dry matter intake, milk production, and body reserve change for cows fed treatment Table 5.4 Income and IOFC in peak and late lactation cows when fed high and low protein Table 5.5 Protein captured in body tissue gain for HP and LP diets across lactation stages with ix LIST OF FIGURES Figure 1.1 Contributions of biological mechanisms to variation in residual feed intake as Figure 1.3 Change of total protein capture from 14CP to 18CP as a function of ECM per kg Figure 2.1 Repeatability of residual feed intake (RFI) across dietary protein contents in peak Figure 2.2 Repeatability of residual feed intake (RFI) across dietary protein contents in late Figure 1.2 Change of milk protein yield from 14CP to 18CP as a function of ECM per kg MBW. Figure 2.4 Association between residual feed intake and milk protein efficiency in peak lactation determined from experiments on divergently selected cattle (Richardson and Herd, 2004) ... 5 ................................................................................................................................................ 19 MBW ...................................................................................................................................... 19 lactation cows ......................................................................................................................... 56 lactation cows ......................................................................................................................... 57 Figure 2.3 Repeatability of residual feed intake (RFI) across lactation stages in lactating cows . 58 cows ........................................................................................................................................ 61 cows ........................................................................................................................................ 62 resilience) in peak- lactation cows ......................................................................................... 83 lactation cows ......................................................................................................................... 90 contents (HP vs. LP) in late lactation cows ............................................................................ 91 contents (HP vs. LP) based on average values across lactation stages .................................. 92 based on average values across diets ..................................................................................... 93 LP) in peak lactation cows ..................................................................................................... 95 Figure 3.2 Repeatability of protein efficiency across dietary protein contents (HP vs. LP) in peak Figure 2.5 Association between residual feed intake and milk protein efficiency in late lactation Figure 3.1 Relationship between milk protein yield and cows response (and LPR, low protein Figure 3.3 Repeatability of protein efficiency (MPE, GPE, and MUN) across dietary protein Figure 3.4 Repeatability of protein efficiency (MPE, GPE, and MUN) for dietary protein Figure 3.5 Repeatability of protein efficiency (MPE, GPE, and MUN) across lactation stage Figure 3.6 Relationship between MUN and MPE/GPE across dietary protein contents (HP vs. x Figure 3.7 Relationship between MUN and MPE/GPE across dietary protein contents (HP vs. Supplementary Figure 3.1 The relationship between DMI, predicted BW, predicted EBW and Figure 5.2 Energy capture, protein capture, and income in milk and body tissue in late lactation Figure 5.1 Energy capture, protein capture, and income in milk and body tissue in peak lactation Figure 3.8 Relationship between MUN and MPE across dietary protein contents (HP vs. LP) and LP) in late lactation cows ....................................................................................................... 96 lactation stage ......................................................................................................................... 97 day in experiment in peak lactation cows ............................................................................ 113 cows ...................................................................................................................................... 170 cows ...................................................................................................................................... 174 Figure 5.3 Sensitivity analysis for peak-lactation cows on HP and LP diets .............................. 177 Figure 5.4 Sensitivity analysis for late-lactation cows on HP and LP diets ................................ 178 body weight) to dietary protein reduction in peak lactation ................................................ 192 body weight) to dietary protein reduction in late lactation .................................................. 194 Supplementary Figure 5.1 Time series of cow response (dry matter intake, milk production, and Supplementary Figure 5.2 Time series of cow response (dry matter intake, milk production, and xi KEY TO ABBREVIATIONS AA = amino acids AMP = adenosine diphosphate mRNA = messenger RNA, or message Ribonucleic acid ATP = adenosine triphosphate BodyE = energy expended for body tissue gain BodyP = protein captured for body tissue gain BUN = body urea nitrogen BW = body weight CapE = total energy capture CapP = total protein capture CBW = calf birth weight CP = crude protein CW = conceptus weight d = day dBW = mean daily BW change dEBW = mean daily change of empty body weight DHIA = dairy herd improvement association DietNEL = apparent diet energy content DIM = days in milk DM = dry matter DMI = dry matter intake xii EAA = essential amino acids EBW = empty body weight ECM = energy corrected milk BCS = body condition score ECM: feed = energy-corrected milk per unit of feed Exp = experiment GE = gross energy GPE = gross protein efficiency h = hours HH = hip height HLPR = high LPR group HP = high protein HRFI = high RFI group IOFC = income over feed cost LifeNitroEff = lifetime nitrogen efficiency LLPR = low LPR group. LP = low protein LPR = low Protein Resilience LRFI = low RFI group MBW = metabolic body weight ME = metabolizable energy MCP = ruminal microbial protein Milk: feed = milk to feed ratio xiii MilkE = milk energy output MilkP: FeedP = milk protein efficiency MLPR = medium LPR group MNE = milk nitrogen efficiency MP = metabolizable protein MPE = milk protein efficiency MRFI = medium RFI group mTOR = mammalian target of rapamycin MUN = milk urea nitrogen MY = milk yield N = nitrogen NDF = neutral detergent fiber NE = net energy Par = parity PDV = portal drained viscera PregE = energy expended for pregnancy PregP = protein captured for pregnancy RDP = rumen degradable protein RE = retained energy REI = residual energy intake RFI = residual feed intake RUP = rumen undegradable protein VIF = variance inflation factors xiv CHAPTER 1 LITERATURE REVIEW FEED EFFICIENCY AND RESIDUAL FEED INTAKE Feed Efficiency in Dairy Industry: From Milk: Feed to Residual Feed Intake Many terms have been used to define feed efficiency in the dairy industry. The earliest and simplest definition is milk to feed ratio (milk: feed), or the amount of milk output per unit of feed (Hooven et al., 1972). Although milk: feed is straightforward and easy to understand, it does not account for milk components in the milk output. Milk: feed does not differentiate cows with different yields of milk protein and fat if milk yield is similar. To address this issue, Grieve et al. (1976) and Custodio et al. (1983) revised the definition of feed efficiency to the amount of energy-corrected milk per unit of feed (ECM: feed) by adjusting milk production based on yields of milk protein and fat. ECM accounts for the difference in milk energy, thus is more appropriate to measure feed efficiency in the modern dairy industry. However, ECM: feed is still limited as it does not account for body reserve mobilization. In lactating dairy cows, body reserve is generally mobilized to support milk production, especially in the early lactation. Excessive body reserve mobilization can lead to metabolic disorders and impair future production performance (NRC, 2001). Thus, cows that produce more milk at the expense of excessive body reserve mobilization should not be considered efficient. Gross efficiency is defined as the energy captured in both milk and body tissue divided by feed energy intake (VandeHaar and St-Pierre, 2006); gross efficiency can distinguish cows that convert more feed to product (milk and body reserve), from ones that mobilize more body reserve to milk. However, gross efficiency also has limitations. In the dairy industry, milk production is more valuable than body tissue gain. Moreover, for cows with BCS over 3.5, body tissue gain is not desirable (NRC, 2001). Thus, cows that convert more 1 feed to milk rather than body reserve gain are preferable. Also, BW change is difficult to measure and usually not known; even under the condition that BW change is known, it is still challenging to compute GPE correctly since there is likely genetic variability between cows in how cows convert feed to BW gain.Therefore, compared to milk: feed and ECM: feed, gross efficiency is not commonly used to define efficiency in the current dairy industry. Essentially, efficiency consists of two aspects:1) more energy directed towards milk (and body tissue) than to maintenance; 2) better ability to convert gross energy (GE) to net energy (NE). As modern dairy cows consume at intake levels of four to five times maintenance, marginal benefits from further increasing production are decreasing (VandeHaar, 1998). Future improvements in feed efficiency can be attainable by selecting cows that can convert GE to NE more efficiently. Residual feed intake (RFI), calculated as the difference between the actual feed intake and the predicted feed intake, is an alternative definition for feed efficiency in livestock. Cattle consuming less than expected will be assigned negative RFI values and considered more efficient in converting GE to NE. RFI was first proposed by Koch et al. (1963) as a direct measurement of energy efficiency in beef heifers and bulls. Over the years, RFI has been established, evaluated, and validated in the beef industry (Herd et al., 2004; Durunna et al., 2011); RFI was proposed for the dairy industry in the 1990s (Veerkamp et al., 1995), and has received more attention in the past 10 years (Connor, 2015). Different equations were used to estimate RFI in dairy heifers (Rius et al., 2012) and lactating dairy cows (Tempelman et al., 2015; Potts et al., 2017). Although the equations were different, there are commonly accepted ground rules when estimating RFI. That is, all production variables, including energy partitioned towards milk, maintenance, and BW change, should be taken into account when estimating RFI (Connor, 2015). This avoids bias regarding body size and milk production level when comparing 2 cows in the same cohort. As milk production and maintenance are already adjusted in the calculation of RFI, RFI is independent of predicted maintenance energy based on BW0.75 and milk energy output (Pryce et al., 2012; Connor, 2015). Genetic Selection for Residual Feed Intake The heritability of RFI in beef cattle has been repeatedly shown to be between 0.38 and 0.62 (Archer et al., 1997; Schenkel et al., 2004). However, in dairy cattle, there are limited data available and the estimated heritability of RFI varied widely from 0.01 to 0.38 (Veerkamp et al., 1995; Vallimont et al., 2011; Pryce et al., 2012). Recently, a large data set was pooled through collaboration across institutes, making it possible to more accurately estimate RFI heritability in dairy cows. Based on the data from 5000 Holstein cows across countries, the estimated RFI heritability for dairy cows was about 0.17 (Tempelman et al., 2015). With this level of heritability, genetic selection for reduced feed intake with no loss in milk should be fruitful (Tempelman et al., 2015). Given all the benefits that come with using RFI when selecting efficient dairy cows, the breeding programs in New Zealand and Netherlands have already incorporated RFI into the selection index. In the next 3-5 years, RFI will be included in the breeding programs in the U.S. as well. Repeatability of Residual Feed Intake across Diets and Physiological State By definition, repeatability is a measure of the strength of relationship between repeataed records for a trait in a population (Boake, 1989). Repeatability across diets for a selection index is essential for the genetic selection to be effective. High repeatability across diets can assure that the values obtained from the test population will apply to cows fed with various diets. In lactating dairy cattle, RFI is repeatable across diets that vary in starch and NDF contents (Potts et al., 2015; Mangual et al., 2016). However, no studies have examined RFI repeatability across 3 dietary protein content in dairy cows. Most diets in previous RFI studies contained adequate or even excessive protein (Mangual et al., 2016; Potts et al., 2015; Tempelman et al., 2015). However, as ~42% of the total feed cost for a lactating cow is associated with protein (St-Pierre, 2012), feeding diets with less protein is becoming more and more common. For example, from 2004 to 2010, there was a 1.1%-unit reduction of CP content in dairy rations in Wisconsin (Shaver, 2010). Thus, it is important to understand whether RFI remains the same across diets with high- versus low- protein diets. Using RFI to determine protein efficiency will be misleading if it is not repeatable across protein contents. Moreover, a useful breeding index should also be repeatable across lactation stages. However, the repeatability of RFI across physiological stages is still inconclusive in cattle. RFI in beef cattle appears to be repeatable over different stages of production cycles. In Hurley et al. (2017), the estimated repeatability of residual energy intake (REI) across lactation stages ranges from 0.19 to 0.23 in a sample of 1290 Holstein-Friesian cows. Phenotypic correlation of RFI was low either when compared across weaned beef heifers later tested as lactating cows (Archer et al., 2002), or when estimated in growing dairy heifers that were later tested during lactation (Nieuwhof et al., 1992; Williams et al., 2011; Waghorn et al., 2012). Thus, more effort is needed for us to better understand the reason for lack of repeatability of RFI across lactation stages before using it for genetic selection; otherwise, RFI information obtained from cows in a certain period of lactation can be misleading. Sources of Variation for Residual Feed Intake Richardson and Herd (2004) listed various sources of variation in RFI in beef cattle which included feeding behavior, body composition, protein turnover and tissue metabolism, heat increment of feeding, digestive efficiency, and physical activity. Among these sources, 12% 4 of the RFI variation was from feeding behavior and physical activity, 10% was from digestibility, and 37% was from protein turnover and tissue metabolism (Figure 1. 1; Richardson and Herd, 2004). Considering that 27% of RFI variation in Richardson and Herd (2004) was unknown and might be measurement errors, the contribution of the sources mentioned above (feeding behavior and physical activity, digestibility, and protein turnover and tissue metabolism) is even greater. Thus, the discussion of possible sources of RFI variation among dairy cows will focus on these sources. Figure 1.1 Contributions of biological mechanisms to variation in residual feed intake as determined from experiments on divergently selected cattle (Richardson and Herd, 2004) One source of RFI variation among dairy cows can be feeding behavior and general activity. For example, studies have shown that, compared to high-RFI dairy heifers, low-RFI dairy heifers eat less frequently and spend less time eating per meal (Durunna et al., 2011; Green et al., 2013; Fitzsimons et al., 2014). Later McGee et al. (2014) found that heifers fed with different levels of concentrate had varying feeding behaviors which were associated with varying 5 RFI across the diets. More studies are needed before any conclusion can be drawn regarding the association between RFI and feeding behavior. In addition to the feeding behavior, difference in general activity may also contribute to the RFI variation. Gomes et al. (2013) demonstrated that low- RFI steers tend to spend more time lying and less time standing. As a result, the energy expense in feeding behavior and activity for low-RFI cattle is less than those with high-RFI values. Another factor that has been investigated in explaining the variation in RFI is the variation in digestion. Prior work has examined the link between RFI variation and digestive variability among cattle; however, findings are inconclusive. Specifically, RFI was associated with digestibility in some studies (Richardson and Herd, 2004; Nkrumah et al., 2006; McDonald et al., 2010; Rius et al., 2012), but not others (Cruz et al., 2010; Lawrence et al., 2011). One possible explanation of the inconsistent findings might be due to the difference in diets across the studies (Rius et al., 2012). Potts et al. (2017) examined the association of RFI and nutrient digestibility across high- and low-starch diets, and found that RFI was only associated with digestibility in cows fed low-starch diets but not high-starch diets. Findings in Potts et al. (2017) supported the notion that the association between RFI and nutrient digestibility depended on the diet. However, it is still unclear what it is in the diet that influences the association between RFI and digestibility. The next set of important factors that can impact RFI values in cattle are protein turnover and tissue metabolism. Protein turnover is energetically expensive in cattle; higher rates of protein turnover were shown to be related to higher maintenance expense (Castro Bulle et al., 2007), and thus leading to cows having higher RFI values. Richardson and Herd (2004) found that protein turnover and tissue metabolism contributed up to 37% of the RFI variation in beef 6 cattle. However, the link between RFI and protein turnover in dairy cattle is less clear. In one of the few studies exploring the relationship between protein turnover and RFI, Lawrence et al. (2012) did not find any association between them two. No other work has validated the results in other physiological states of dairy cows. If protein turnover is a significant contributor to RFI variation among beef cattle, it will likely explain at least part of the variation in RFI in dairy cows. Linking Residual Feed Intake to Protein Efficiency Protein turnover rates accounted for some of the variation in RFI (Richardson and Herd, 2004) and were also negatively associated with protein utilization efficiency in dairy cows (Herd et al., 2004; Castro Bulle et al., 2007). RFI is usually calculated on an energy basis, and, although we expect that RFI and protein efficiency are associated with each other to some extent, the direct link between RFI and protein efficiency has not been demonstrated. In lactating dairy cows, Xi et al. (2016) and Mangual et al. (2016) hypothesized an association between the two. There has been some existing work done in dairy heifers (Rius et al., 2012; Thornhill et al., 2014; Marett et al., 2017); however, no evidence supported such a link. No work has directly examined the link in the lactating dairy cows. As protein requirement and protein metabolism in lactating dairy cows were different from those in heifers, research work on lactating cows is in need before any conclusion can be drawn. Terms to Define Protein Efficiency PROTEIN EFFICIENCY There are a number of different ways to define protein efficiency, and each one has its own advantages and limitations. Milk protein efficiency, milk protein yield per unit of feed 7 protein consumption (MilkP: FeedP), sometimes referred to as milk nitrogen efficiency (MNE), is the most commonly used term to describe protein efficiency in the dairy industry. According to a meta-analysis done by Huhtanen and Hristov (2009), using data from North America and Europe from 1979 to 2005, the average MNE was 25%, with a range from 10% to 40%. Although easier to calculate and commonly used, MNE is limited due to failing to consider body protein reserve mobilization. In addition, by including body protein capture in addition to milk protein output, the protein-efficiency measure can be applied to cows in other physiological states (e.g., dry cows, heifers, and 1st lactation animals that are not at mature BW). Dairy cows spend about half of their lives as replacement heifers and dry cows; thus, protein efficiency in those stages is definitely critical to measure. In addition, by considering BW change, the protein- efficiency term that includes body protein can also enable a comparison of protein efficiency among different production enterprises in the livestock industry, such as cattle, swine, and poultry. However, relative to a large body of literature on feed efficiency, few publications discussed this complete measure of protein efficiency that includes both milk production and body tissue gain. A term accounting for lifetime nitrogen efficiency (LifeNitroEff) has been proposed to measure the protein efficiency throughout the entire life cycle (Foskolos and Moorby, 2018). LifeNitroEff considers not only the milk protein production, but also the growth and body composition change, protein expenditure for reproduction, and immune responses. Due to the complexity of measurement, calculation, and modeling, LifeNitroEff has not been adopted for use by the dairy industry to date. The next measure of protein efficiency is called human edible protein efficiency. As noted by Broderick (2017), ruminants convert human-inedible protein to human-edible protein, while monogastric animals competed with humans for feedstuff that is potential human food. 8 Thus, Broderick (2017) argued that it was not appropriate to directly compare protein efficiency between ruminants and monogastric animals. To address this, Broderick (2017) suggested using a term, human edible protein efficiency. Similar to MilkP: FeedP, human edible protein efficiency was calculated as the protein in the product per unit of human-edible dietary protein, which adjusts protein based on the feed source. After the adjustment, ruminants are much more efficient than swine and poultry (2.08 for milk, 1.19 for beef, 0.29 for pork, and 0.62 for poultry). In addition to all the ratio terms mentioned above, a biological indicator, milk urea nitrogen (MUN), is also commonly used to indicate protein efficiency in the dairy industry. MUN was first proposed as an indicator of protein efficiency by Oltner and Wiktorsson (1983). When protein (especially rumen degradable protein; RDP) exceeds microbial needs, there will be a large amount of ammonia produced in the rumen. Ammonia is then converted to urea in the liver (Colmenero and Broderick, 2006), and urea equilibrates between body fluids (DePeters and Ferguson, 1992). A high concentration of MUN indicates high body urea nitrogen (BUN) and inefficient utilization of protein (Broderick and Clayton, 1997; Roseler et al., 1993; Gustafsson and Palmquist, 1993). Improving Protein Efficiency by Nutritional Means Many nutritional means have been explored to improve protein efficiency (Sinclair et al., 2014; Broderick et al., 2015; Gidlund et al., 2015). After years of examination, several critical areas of protein utilization in dairy cows have been identified to improve protein efficiency. These critical areas include protein degradation and synthesis in the rumen, digestion and absorption in the small intestine, absorbed amino acids (AA) passing through portal drained 9 viscera (PDV) and liver, and AA extraction and utilization in mammary glands. The discussion below will be focused on how nutrition practice can potentially impact these four areas. Protein degradation and synthesis in the rumen. Over 30% of the total AA pool comes from dietary AA (Apelo et al., 2014). Thus, the efficiencies of protein digestion and AA absorption play important roles in determining overall protein efficiency. In rumen, the efficiency of protein degradation and ruminal microbial protein (MCP) synthesis are critical. It was commonly assumed that maximal microbial growth is equal to maximal nitrogen efficiency in the rumen. However, more nitrogen outflow is achieved per unit of of N consumed in low RDP diets. Clearly, maximal microbial growth cannot be achieved under this condition. Thus, to clarify, this section mainly reviews factors that can maximize MCP synthesis. Extensive work has been done to examine the impacts of diet and feed management factors on MCP synthesis. These factors include energy content and source, RDP content, matching content and source of carbohydrate and RDP content, forage source and length, and RDP source. Dietary energy level was initially considered as one of the most influential factors on MCP synthesis, because MCP synthesis improved significantly when increasing dietary starch level (Febel and Fekete,1996). However, in Russell and Wallace (1997), higher starch contents decreased rumen pH and fiber digestion, reducing the synthesis of de novo amino acids, and therefore depressed MCP synthesis. In Hoover and Stokes (1991), dietary RDP content was found to increase MCP synthesis, however, the optimal level of RDP also depended on the content and source of starch. Therefore, RDP content should be matched with energy level, otherwise protein will be wasted in the form of ammonia, or energy will be wasted in the form of heat (Kolver et al., 1998; Moharrery, 2004). In order to achieve maximal utilization efficiency of both RDP and starch, the ideal ratio of RDP to starch is 1: 4 (Yang et al., 2010), but the optimal ratio varies along with 10 other dietary factors (e.g., energy source and RDP source; Yang et al., 2010). Besides matching contents of energy and RDP, sources of energy and RDP should also be matched. Cone et al. (1989) and Chamberlain et al. (1993) showed that different energy sources have different effects on the MCP synthesis. The work done by Cone et al. (1989) showed that oat and barley fermented faster than corn, and negatively impacted the MCP synthesis. Chamberlain et al. (1993) found that when RDP was mostly from urea, supplementing soluble sugars (saccharose, lactose, and fructose) increased MCP synthesis, compared to supplementing cereals (high in starch). In addition to energy and RDP, forage can also affect MCP synthesis. The reason why forage source affects MCP synthesis is similar to that of starch source. Forage with slower digestion rates and longer particle sizes would maintain a more consistent rumen pH, a more functional rumen, and thus greater MCP synthesis. Lastly, RDP source can also impact MCP synthesis. Supplementing diets with RDP from non-urea sources, versus urea sources, increased ruminal microbial protein synthesis (Kertz, 2010). However, not all microbial species can efficiently use a sole source of RDP; to optimize growth of rumen microbes, a mix of urea, AA, and peptide is preferred. Additionally, supplementing certain AA (e.g., Phe, Leu, Ile) can inhibit microbial growth. Previous work has shown that, when supplementing Phe, Leu, and Ile, ruminal de novo AA synthesis fell up to 80% (Atasoglu et al., 1999). Besides MCP synthesis, nitrogen recycling can also impact rumen nitrogen utilization efficiency. According to Wallace and McPherson (1987), microbial nitrogen recycling significantly affect ruminal protein utilization and it is mainly associated with protozoal predation on ruminal bacteria, where protozoal predation mainly depends on the energy availability (Dijkstra et al., 1992). Also, urea in saliva circulating back to rumen contributes to the RDP pool and thus influences rumen nitrogen utilization efficiency. Based on the model in Dijkstra et al. (1992), saliva circulation is related to 11 DMI and dietary NDF content. Thus, the factors that can impact DMI (which will be discussed in the section “Other means to improve protein efficiency”) can all potentially impact nitrogen recycling and overall rumen nitrogen efficiency. All the nutritional optimizations above are under the assumption that other nutrients (such as minerals and vitamins) are not limiting. In daily practice, minerals and vitamins should also be supplemented sufficiently. Digestion and absorption in the small intestine. More absorbable rumen undegradable protein (RUP) and protected AA can enhance intestinal AA profile, increase milk protein yield, and increase protein efficiency in dairy cows (NRC, 2001). For example, supplementing methionine for soy-based diets, lysine for corn-based diets, and histidine for grass-based diets can improve milk yield and protein efficiency (Schwab and Broderick, 2017). However, as reviewed by Santos et al. (1997), the effect of protected AA supplements sometimes might be lower than expected. The discrepancy could be due to the following reasons: 1) supplementing RUP decreased MCP synthesis (Schwab, 1994), 2) the RUP source did not balance out the AA shortage in the base diets (e.g., supplementing corn gluten meal to the corn-based diets; Chandler, 1991), 3) the RUP sources (e.g., feather meal, meat and bone meal, and blood meal) had low intestinal digestibility (Schingoethe, 1991), and 4) RUP level in the base diets might already have been high enough. Absorbed AA passing through portal drained viscera (PDV) and liver. 3-10% of the absorbed AA are catabolized when they first pass through PDV, and up to 50% are catabolized on a daily basis (Apelo et al., 2014). Based on Hanigan et al. (1998; 2004), AA not utilized by mammary glands would be catabolized in PDV, and the remaining AA that flow out from PDV would circulate back to mammary glands, becoming available for milk protein synthesis again. The catabolic rates of different AA are different in PDV. MacRae et al. (1997) found that 13- 12 25% of Leu, Ile, Val, Lys, Thr, and His were catabolized in PDV, while up to 54% of Phe was catabolized. This work suggested that the requirement of Phe was higher than other AA in PDV. Similar to the metabolism of PDV, AA were also catabolized and utilized when passing through liver (Lobley et al., 2000). The catabolized AA were utilized in several critical metabolic steps, including: 1) converting carbon skeleton of deaminated AA to glucose or lipids, and 2) synthesizing critical protein (e.g., albumin) from non-essential AA (Lobley et al., 2000). Similar to the AA metabolism in PDV, the AA catabolism in liver entirely depends on the AA availability in blood flow when passing through liver. The work done by Reynolds (2005; 2006) showed that, when protein synthesis in mammary glands decreased, plasma AA concentration increased, and more AA were removed by liver. Therefore, to minimize the AA catabolism in liver and PDV, maximizing protein utilization in mammary glands is required to minimize the AA concentration in blood flow. AA utilization in mammary glands. AA uptake and milk protein synthesis are the two critical steps in regulating protein utilization efficiency in mammary glands. Several factors alter AA uptake and milk protein synthesis; these included plasma AA concentration, blood flow rate, plasma energy/AA status, and factors influencing mammalian target of rapamycin (mTOR) pathways (e.g., certain EAA, insulin, prolactin, cortisol, AMP/ATP). Plasma AA concentrations can influence AA uptake by mammary glands. The AA that are taken up into mammary glands exhibit Michaelis-Menten kinetics, in other words, the amount of AA uptake is maximized at infinite concentrations of AA (Neal and Thornley, 1983). However, AA uptake efficiency, if calculated as moles removed per mole delivered to the tissue, would be maximized as supply approaches zero. An optimal range of AA should be the amount of each AA that results in maximal protein synthesis. Thus, an optimal range of AA is usually expressed as per unit of milk 13 protein output; AA concentration exceeding that optimal range will decrease efficiency while AA concentration below the optimum will not decrease efficiency. Another factor that influences protein synthesis and protein efficiency is the blood flow rate. In Rius et al. (2010), faster blood flow within mammary glands induced by arginine supplementation increased milk yield. This is consistent with previous studies showing a positive link between blood flow and milk yield/ protein efficiency (Cant and McBride, 1995; Hanigan et al., 2002). Lastly, nutrients and hormones influencing signaling pathways can also increase protein efficiency via activating translation of milk protein mRNA. mTOR is the most well-studied pathway among all the pathways that control the activity of milk protein synthesis. Activation of mTOR pathway enhances protein synthesis, and thus increases protein efficiency (Rius et al., 2010). There are several factors (e.g., EAA such as Leu/ Ile/ Thr, insulin, cortisol, and cellular energetic status such as AMP/ATP ratio) that can be manipulated to directly activate mTOR pathways (Apelo et al., 2014). Other means to improve protein efficiency. Protein efficiency is defined as the ratio of milk protein yield (the numerator) to dietary protein intake (the denominator). Thus, theoretically, all the factors that can impact protein intake and milk protein output have the potential to impact protein efficiency. If this is the case, in addition to all the nutritional means mentioned above, management practices, such as grouping precise feeding diet reformulation frequency, forage harvest and ensiling process, feeding consistency, stocking density, barn temperature and humidity, improving cow comfort, and managing photoperiod, can also impact protein efficiency on herd level (Jonker et al., 2002). 14 Improving Protein Efficiency by Genetic Selection Considerable effort has been made to explore the feasibility of using genetic selection to improve protein efficiency in dairy cattle. Although a medium level of heritability in protein efficiency was detected in lactating dairy cows (0.10- 0.31; Li et al.,1998; Zamani et al., 2011), a number of issues have been raised regarding selecting dairy cows based on the traditional protein efficiency term (milk protein per unit of dietary protein intake; Zetouni et al., 2017). Most of the doubt was due to the fact of protein efficiency being a ratio trait. The drawback of ratio traits is that they are usually not normally distributed. As a result, it is difficult to expect the selection response due to the disproportionate selection pressure on the component traits (Zetouni et al., 2017). In other words, using ratio traits (e.g., protein efficiency) in genetic selection induces large error variance and unexpected results. Another issue for using MilkP: FeedP as a selection index is the reliability of the heritability calculated from the small data set. Protein efficiency data in individual cows is limited, as individual intake data is usually not available in commercial production settings. Accordingly, the data set size used to calculate the heritability of protein efficiency is commonly less than 600 cows (Zamani et al., 2011), which is generally much lower than the number required for an accurate heritability estimate (Misztal, 1997). Due to the two concerns mentioned above, MUN was proposed as an alternative selection index to indirectly improve protein efficiency (Wood et al., 2003). The variation in MUN within cow and herd has been widely recognized (Huhtanen et al., 2015). The heritability for MUN concentration ranges from 0.13 to 0.22 (Mitchell et al., 2005; Stoop et al., 2007; Bastin et al., 2009). The studies using MUN data suggest that there is a genetic component in the variation of protein efficiency. However, using MUN to represent protein efficiency is questionable. Nousiainen et al. (2004) observed a quadratic relationship between protein efficiency and MUN concentration in a meta- analysis. The work done by Nousiainen et al. (2004) suggested that MUN may not be a good 15 representation of the true protein efficiency when cows are fed excessive protein. As mentioned earlier, dairy cows are typically fed excessive protein on commercial farms to maximize production. As most of the MUN data used for calculating the genetic variation is from commercial farms (Michell et al., 2005; Stoop et al., 2007; Bastin et al., 2009), precaution should be taken when interpreting the results calculated from MUN data. Additionally, MUN concentration can be affected by many other factors: dehydration (Burgos et al., 2001; Weeth and Lesperance, 1965), the season of the year, time of sampling (Depeters and Cant, 1992; Kauffman and St-Pierre, 2001; Broderick and Clayton, 1997) and variable transport activities in kidney and rumen wall (Aguilar et al., 2012; Stewart and Smith, 2005). For example, MUN concentration can be elevated merely due to less blood urea transported to urine. According to Kohn et al. (2004), MUN values can be different between regions and milk analysis laboratories using different analysis methods. To sum, MUN can be used to monitor protein feeding in daily practice; however, it is not suitable to help define true protein efficiency in dairy cattle. Ranking cows for their protein efficiency based on MUN concentration can be misleading. The Most Effective Nutritional Method: Lowering Dietary Protein Content? Due to the drawbacks mentioned above in genetic selection for improving protein efficiency, nutritional manipulation might be a more effective way to improve protein efficiency. Among all nutritional means, the most effective and economical way is to lower dietary protein content (Huhtanen and Hristov, 2009). However, it is also well recognized that the reduction of protein intake could lead to reduced DMI, and consequently MY (Cantalapiedra-Hijar et al., 2014). The emerging challenge is to figure out ways to lower dietary protein intake while mitigating against the effects of low protein on DMI and milk production (Sinclair et al., 2014; Huhtanen et al., 2008; Ingvartsen and Andersen, 2000). If the protein shortage decreases milk 16 production, then the savings from feeding less protein is outweighed by the lost milk revenue. In this case, we need to identify cows with better ability to maintain their milk production and body reserve when fed low- protein diets. LOW PROTEIN RESILIENCE Resilience in Animal Science The idea of resilience originated from developmental psychology (de Terte and Stephens, 2014). By definition, resilience is “the ability to mentally or emotionally cope with a crisis or to return to pre-crisis status quickly”. In the animal production system, resilience can be defined as the ability of the animal to maintain its normal state after exposure to environmental disturbances, or the ability to quickly return to a normal state (Colditz and Hine, 2016). Several definitions of resilience and resilience- associated concepts (robustness, tolerance, resistance, plasticity, environmental sensitivity, canalization, and stability) have been discussed in the literature (Knap, 2005; Mulder et al., 2013; Colditz and Hine, 2016). “General” resilience is considered as a composite trait, consisting of different resilience to various environmental disturbances (Colditz and Hine, 2016; Elgersma et al., 2018), where disturbances are categorized in two groups: macro-environmental factors and micro- environmental factors (Falconer and Mackay, 1996; Mulder et al., 2013). Macro-environmental factors are environmental factors that impact the majority, if not all of the whole population (e.g., disease pressure, ambient temperature); while micro-environmental factors are the factors that only impact a minority of the whole population within that macro-environment (e.g., diseases, social interactions). 17 Low Protein Resilience in Dairy Cows To mitigate the impact of low-protein diets on DMI and MY, identifying and selecting cows with better ability to maintain their protein output when fed low-protein diets are needed. To help the identification and selection process, we are proposing a term: Low Protein Resilience (LPR). We define LPR as the difference between individual cow decreases in protein output and the average decrease in protein output when cows are switched from high-protein diets to low- protein diets after adjusting for levels of MY, parity, DIM, etc. In pilot studies, we included cows in the crossover studies with 2 periods and 2 diets (14% and 18% CP) to examine the individual cow response to the low-protein diet. We found, on average, milk production and body weight were significantly decreased in cows fed 14% CP when switched from 18% CP. However, the response to the same 4% CP decrease varied a lot among cows. In the figure below, the differences in cows’ milk protein yield as diets changing from 18% CP to 14% CP were plotted (Figure 1.2). For cows not differing in productive ability (ECM per kg MBW), the differences in the milk protein yield in response to the 4% dietary CP decrease was considerable. The findings suggest that there are individual differences in the ability of maintaining protein output when fed low protein diets among cows. The model can be further improved by using total protein captured in both milk and body tissues to avoid misleading information (Figure 1.3). Production level, BW, BCS, DIM, and parity need to be considered when modeling LPR, as all of these factors can impact the extent of protein mobilization and deposition. To our knowledge, no prior study examined the variation among cows in terms of their resilience to low-protein diets, and none explore the possible mechanisms to explain the resilience variation. 18 0.5 0.2 0.18 -0.1 Milk Protein decrease from 4% CP change Parity 1 Parity 2+ 0.23 0.28 0.33 0.38 0.43 0.48 CP18 ECM_MBW Figure 1.2 Change of milk protein yield from 14CP to 18CP as a function of ECM per kg MBW. CP18 ECM_MBW is the energy corrected milk per kg metabolic body weight when cows on diets with 18% CP. Total Protein Capture decrease from 4% CP change Parity 1 Parity 2+ 0.7 0.4 0.1 0.18 0.23 0.28 0.33 0.38 0.43 0.48 -0.2 CP18 ECM_MBW Figure 1.3 Change of total protein capture from 14CP to 18CP as a function of ECM per kg MBW Total protein capture = milk protein + body protein gain, where body protein deposition was estimated as 0.07 ´ body weight change for multiparous cows (Parity 2+) and 0.12 ´ body weight change for primiparous cows (Parity 1). CP18 ECM_MBW is the energy corrected milk per kg metabolic body weight when cows on diets with 18% CP. 19 By definition, LPR should be independent of milk production, which provides the advantage of combining LPR with production traits to select cows with both high productivity and better resilience. In the genetic selection program, production traits and robustness traits (the traits that describe the resistance, tolerance, and resilience to various environmental stressors and challenges) are usually contraindicative to each other (Knap, 2005). That is, animals with higher production are commonly more vulnerable to environmental stressors (i.e., less resilient). In the classic dairy breeding program, milk yield and milk protein yield are the major focuses, while health traits are often neglected (Egger-Danner et al., 2015). As a result, modern dairy cattle having higher productivity, but are more susceptible to physiological and immunological imbalances (mastitis, heat stress, etc.; Rauw and Gomez-Raya, 2015). As animal welfare, production longevity and sustainability are given more attention, including robustness traits into genetic selection programs is being discussed (Calus et al., 2013). At first, heat stress indicators were the only traits included in the genetic selection programs aiming to improve dairy cow robustness; later on, other health traits were also included into the index, in order to select cows with better resistance to infectious and non-infectious diseases (König and May, 2019). However, as discussed before, selecting robust cows might outweigh efforts in improving milk production. Given that there might be some common genetic factors in animals’ resilience to various disturbances (disease, temperature, social stress; Mulder et al., 2013), the new definition of resilience, LPR, provides a potential solution to incorporate both production traits and resilience traits into one genetic selection program, and select more resilient cows among high- producing ones. 20 MODELING COW RESPONSE TO DIETARY PROTEIN In a typical U.S. farm, 42% of the total feed cost is spent on protein (St-Pierre, 2012). Therefore, optimizing protein feeding could significantly improve farm profitability. One way to improve profitability is to decrease feed cost by lowering dietary protein content. However, protein reduction could decrease milk production and in turn milk revenue, and the savings from feeding less protein may be outweighed by the lost milk revenue. Researchers have been examining the trade-off between dietary protein contents and milk production for a long time. One of the best studies examining this trade-off was conducted 20 years ago by Wu and Satter (2000) at Wisconsin. They measured the response of lactating dairy cows to different amounts of dietary protein and concluded that feeding cows 17-19% CP before week 30 and 16% CP after week 30 optimized milk production. Since then, many researchers have continued to study the effects of dietary protein content on milk production (Broderick, 2003; Lammers and Heinrichs, 2000), and documented a negative link between dietary protein reduction and milk production. Following that, many researchers have attempted to work out the optimal protein content by modeling cows’ response to dietary protein. For example, Hristov et al. (2005) added initial BW in the response model to account for available body reserves for milk production, and significantly improved the overall model fit. Brun-Lafleur et al. (2010) found that parity explained significant variation in the model, and that the response curve for primiparous cows was much different from that for multiparous cows. They speculated that the differential MY response for primiparous cows versus multiparous cows might be because it was easier for multiparous cows to mobilize body reserves to support milk production. Furthermore, Moraes et al. (2018) acknowledged that body reserves would significantly impact the milk protein yield, especially body protein mobilization in early lactation and deposition in late lactation. However, it could not be captured in their analyses given the design and scope of the studies included in the 21 paper. Thus, although the effect of body reserves mobilization on milk production has been widely accepted by researchers, body reserves change has not been properly incorporated in prior models for various reasons (Hristov et al., 2005; Brun-Lafleur et al., 2010; Moraes et al., 2018). Ignoring body reserves change significantly impeded the accuracy of the prediction of cow response to dietary protein. If milk loss is consistently accompanied by body weight loss when fed low-protein diets, then the loss in protein capture and the loss from feeding less protein is underestimated. Additionally, researchers have been working on profit response to dietary protein for years. Yet, findings on the effect of dietary protein reduction on farm profitability are inconclusive. For example, Stewart et al. (2012) did not find any financial penalty or benefit when reducing dietary CP from 18 to 16.5% whereas Phuong et al. (2013) noted in their study that the loss of milk income as a result of decreased dietary protein content (from 19% to 15%) greatly exceeded the extra cost of feeding excessive protein. More recently, contrary to findings in Phuong et al. (2013), Fadul-Pacheco et al. (2017) found in a sample of Eastern Canadian dairy herds in 2011 that improving protein efficiency by reducing dietary protein contents from 16.5% to 15% significantly increased income over feed cost (IOFC) from Can$14.3 to Can$18.2/ cow per day. It is likely that 16.5% CP in Stewart et al. (2012) has met or exceeded the genetic capacity of cows to generate protein output; thus, the first 1- 2% units reduction of CP might be simply removing the excess above requirement. In addition, the inconsistent findings in previous studies could also be partly due to the differential economic conditions over the years and not accounting for body reserve mobilization and response variation among cows. In Potts et al. (2015), IOFC was found to be higher for low-RFI cows than high-RFI cows only when cows were fed high-starch diets but not low-starch diets. Taken together, these findings suggested an 22 interaction between animal and nutrient content, meaning that the effect of nutrient content may vary across cows. 23 REFERENCES 24 REFERENCES Aguilar, M., M. D. Hanigan, H. A. Tucker, B. L. Jones, S. K. Garbade, M. L. McGilliard, and R. E. James. 2012. Cow and herd variation in milk urea nitrogen concentrations in lactating dairy cattle1. J. Dairy Sci. 95: 7261-7268. Åhs, F., T. Furmark, Å. Michelgård, B. Långström, L. Appel, O. T. Wolf, and M. Fredrikson. 2006. Hypothalamic blood flow correlates positively with stress-induced cortisol levels in subjects with social anxiety disorder. Psychosom. Med. 68: 859-862. Apelo, S. A., J. R. Knapp, and M. D. Hanigan. 2014. Invited review: Current representation and future trends of predicting amino acid utilization in the lactating dairy cow. J. Dairy Sci. 97: 4000-4017. Archer, J. A., A. Reverter, R. M. Herd, D. J. Johnston, and P. F. Arthur. 2002. Genetic variation in feed intake and efficiency of mature beef cows and relationships with postweaning measurements. 7th World Congr. Genet. Appl. Livest. Prod. 31:221–224. Archer, J. A., P. F. Arthur, R. M. Herd, P. F. Parnell, and W. S. Pitchford. 1997. Optimum postweaning test for measurement of growth rate, feed intake, and feed efficiency in British breed cattle. J. Anim. Sci. 75:2024–2032. 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. Bastin, C., L. Laloux, A. Gillon, F. Miglior, H. Soyeurt, H. Hammami, and N. Gengler. 2009. Modeling milk urea of Walloon dairy cows in management perspectives. J. Dairy Sci. 92: 3529-3540. Bequette, B.J., M.D. Hanigan, A.G. Calder, C.K. Reynolds, G.E. Lobley, and J.C. MacRae. 2000. Amino acid exchange by the mammary gland of lactating goats when histidine limits milk production. J. Dairy Sci. 83: 765-775. Berry, D. P., and J. J. Crowley. 2013. Cell Biology Symposium: Genetics of feed efficiency in dairy and beef cattle. J. Anim. Sci. 91: 1594-1613. Binsiya, T. K., V. Sejian, M. Bagath, G. Krishnan, I. Hyder, A. Manimaran, R. Bhatta. 2017. Significance of hypothalamic-pituitary-adrenal axis to adapt to climate change in livestock. Int. Res. J. Agri. Food Sci. 2: 1-20. Boake, Christine RB. 1989. Repeatability: its role in evolutionary studies of mating behavior." Evolutionary Ecology 3: 173-182. 25 Broderick, G. A., A. P. Faciola, and L. E. Armentano. 2015. Replacing dietary soybean meal with canola meal improves production and efficiency of lactating dairy cows1 J. Dairy Sci. 98: 5672-5687. Broderick, G. A., and M. K. Clayton. 1997. A statistical evaluation of animal and nutritional factors influencing concentrations of milk urea nitrogen1. J. Dairy Sci. 80: 2964-2971. Broderick, G.A.. 2018. Optimizing ruminant conversion of feed protein to human food protein. Animal. 12(8):1722-1734. Brun-Lafleur, L., L. Delaby, F. Husson, and P. Faverdin. 2010. Predicting energy× protein interaction on milk yield and milk composition in dairy cows. J. Dairy Sci. 93: 4128-4143. Burgos, M. S., M. Senn, F. Sutter, M. Kreuzer, and W. Langhans. 2001. Effect of water restriction on feeding and metabolism in dairy cows. Am. J. Physiol. Regul. Integr. Comp. Physiol. 280: 418-427. Calus, M. P. L, D. P. Berry, G. Banos, Y. de Haas, and R. F. Veerkamp. 2013. Genomic selection: the option for new robustness traits? Advances in Animal Biosciences. 4: 618–625. Cant, J. P., and B. W. McBride. 1995. Mathematical analysis of the relationship between blood flow and uptake of nutrients in the mammary glands of a lactating cow. J. Dairy Res. 62:405– 422. Cant, J.P., J.J. Kim, S.R. Cieslar, and J. Doelman. 2018. Symposium review: Amino acid uptake by the mammary glands: Where does the control lie?. J. Dairy Sci. 101: 5655-5666. Cantalapiedra-Hijar, G., J. L. Peyraud, S. Lemosquet, E. Molina-Alcaide, H. Boudra, P. Noziere, and I. Ortigues-Marty. 2014. Dietary carbohydrate composition modifies the milk N efficiency in late lactation cows fed low crude protein diets. Animal. 8: 275-285. Castro Bulle, F. C. P., P. V. Paulino, A. C. Sanches, and R. D. Sainz. 2007. Growth, carcass quality, and protein and energy metabolism in beef cattle with different growth potentials and residual feed intakes. J. Anim. Sci. 85:928–936. Chamberlain, D.G., S. Robertson, and J.J. Choung. 1993. Sugars vs. starch as supplements to grass silage: effects on ruminal fermentation and the supply of microbial protein to the small intestine, estimated from the urinary excretion of purine derivatives, in sheep. J. Sci. Food Agric. 63:189 Chandler, P. T. 1991. Quantitative and qualitative characteristics of protein sources and interrelationships with energy. Virginia Dairyman 12:10, 12. Colditz, I.G. and B.C. Hine. 2016. Resilience in farm animals: biology, management, breeding and implications for animal welfare. Anim. Prod., 56: 1961-1983. Colmenero, J.O. and G.A. Broderick. 2006. Effect of dietary crude protein concentration on milk production and nitrogen utilization in lactating dairy cows. J. Dairy Sci. 89(5): 1704-1712. 26 Cone, J.W., W. Cline-Theil, A. Malestein, and A. van ́t Klooster. 1989. Degradation of starch by incubation with rumen fluid. A comparison of different starch sources. J. Sci. Food Agric. 49:173 Connor, E. E. 2015. Invited review: Improving feed efficiency in dairy production: Challenges and possibilities. Animal. 9: 395-408. Cruz, G. D., J. A. Rodríguez-Sánchez, J. W. Oltjen, and R. D. Sainz. 2010. Performance, residual feed intake, digestibility, carcass traits, and profitability of Angus-Hereford steers housed in individual or group pens. J. Anim. Sci. 88:324-329. Custodio, A. A., R. W. Blake, P. F. Dahm, T. C. Cartwright, G. T. Schelling, and C. E. Coppock. 1983. Relationships between measures of feed efficiency and transmitting ability for milk of Holstein cows. J. Dairy Sci. 66:1937. De Terte, I., and C. Stephens. 2014. Psychological resilience of workers in high‐risk occupations. Stress and Health. 30: 353-355. DePeters, E. J., and J. P. Cant. 1992. Nutritional Factors Influencing the Nitrogen Composition of Bovine Milk: A Review1. J. Dairy Sci. 75: 2043-2070. DePeters, E.J. and J.D. Ferguson. 1992. Nonprotein nitrogen and protein distribution in the milk of cows. J. Dairy Sci. 75: 3192-3209. Doeschl-Wilson, A. B., B. Villanueva, and I. Kyriazakis. 2012. The first step toward genetic selection for host tolerance to infectious pathogens: obtaining the tolerance phenotype through group estimates. Front. Genet. 3:265. Durunna, O. N., F. D. N. Mujibi, L. Goonewardene, E. K. Okine, J. A. Basarab, Z. Wang, and S. S. Moore. 2011. Feed efficiency differences and reranking in beef steers fed grower and finisher diets. J. Anim. Sci. 89:158–167. Durunna, O. N., F. D. N. Mujibi, L. Goonewardene, E. K. Okine, J. A. Basarab, Z. Wang, and S. S. Moore. 2011. Feed efficiency differences and reranking in beef steers fed grower and finisher diets. J. Anim. Sci. 89:158–167. Egger-Danner C, J. B. Cole, J. E. Pryce, N. Gengler, B. Heringstad, A. Bradley, and K. F. Stock. 2015. Invited review: overview of new traits and phenotypin strategies in dairy cattle with a focus on functional traits. Animal 9: 191–207. Elgersma, G.G, G. de Jong, R. Van der Linde, and H. A. Mulder. 2018. Fluctuations in milk yield are heritable and ca be used as a resilience indicator to breed healthy cows. J. Dairy Sci. 101: 1240-1250. Fadul-Pacheco, L., D. Pellerin, P. Y. Chouinard, M. A. Wattiaux, M. Duplessis, and E. Charbonneau. 2017. Nitrogen efficiency of eastern Canadian dairy herds: Effect on production performance and farm profitability. J. Dairy Sci. 100: 6592-6601. 27 Falconer, D. S., and T. F. C. Mackay. 1996. Introduction to quantitative genetics (fourth edition). Harlow: Longman Group Ltd. Fébel, H. and S. Fekete. 1996. Factors influencing microbial growth and the efficiency of microbial protein synthesis: a review. Acta Veterinaria Hungarica, 44: 39-56. Fitzsimons, C., D. A. Kenny, and M. McGee. 2014. Visceral organ weights, digestion and carcass characteristics of beef bulls differing in residual feed intake offered a high concentrate diet. Animal, 8: 949-959. Foskolos, A., and J. M. Moorby. 2018. Evaluating lifetime nitrogen use efficiency of dairy cattle: A modelling approach. PloS one. 13.8: e0201638. Gidlund, H., M. Hetta, S. J. Krizsan, S. Lemosquet, and P. Huhtanen. 2015. Effects of soybean meal or canola meal on milk production and methane emissions in lactating dairy cows fed grass silage-based diets. J. Dairy Sci. 98: 8093-8106. Gomes, C., R. D. Sainz, and P. R. Leme. 2013. Protein metabolism, feed energy partitioning, behavior patterns and plasma cortisol in Nellore steers with high and low residual feed intake. Revista Brasileira de Zootecnia. 42:44–50. Green, T. C., J. G. Jago, K. A. Macdonald, and G. C. Waghorn. 2013. Relationships between residual feed intake, average daily gain, and feeding behavior in growing dairy heifers. J. Dairy Sci. 96:3098–3107. Grieve, D. G., G. K. Macleod, T. R. Batra, E. B. Burnside, and J. B. Stone. 1976. Relationship of feed intake and ration digestibility to estimated transmitting ability, body weight, and efficiency in first lactation. J. Dairy Sci. 59:1312–1318. Gustafsson, A.H. and D.L. Palmquist. 1993. Diurnal variation of rumen ammonia, serum urea, and milk urea in dairy cows at high and low yields. J. Dairy Sci. 76: 475-484. Hanigan, M. D., C. K. Reynolds, D. J. Humphries, B. Lupoli, and J. D. Sutton. 2004. A model of net amino acid absorption and utilization by the portal-drained viscera of the lactating dairy cow. J. Dairy Sci. 87:4247–4268. Hanigan, M. D., J. P. Cant, D. C. Weakley, and Beckett, J. L. (1998). An evaluation of post- absorptive protein and amino acid metabolism in the lactating dairy cow. J. Dairy Sci. 81: 3385-3401. Hanigan, M. D., L. A. Crompton, B. J. Bequette, J. A. Mills, and J. France. 2002. Modelling mammary metabolism in the dairy cow to predict milk constituent yield, with emphasis on amino acid metabolism and milk protein production: Model evaluation. J. Theor. Biol. 217:311–330. Herd, R. M., V. H. Oddy, and E. C. Richardson. 2004. Biological basis for variation in residual feed intake in beef cattle. 1. Review of potential mechanisms. Aust. J. Exp. Agric. 44:423– 430. 28 Hooven, N. W., R. H. Miller, and J. W. Smith. 1972. Relationships among whole-and part- lactation gross feed efficiency, feed consumption, and milk yield. J. Dairy Sci. 55:1113–1122. Hoover, W.H. and S.R. Stokes. 1991. Balancing carbohydrates and proteins for optimum rumen microbial yield. J. Dairy Sci. 74: 3630-3644. Hristov, A.N., W. J. Price, and B. Shafii. 2005. A meta-analysis on the relationship between intake of nutrients and body weight with milk volume and milk protein yield in dairy cows. J. Dairy Sci. 88: 2860-2869. Huhtanen, P., and A. N. Hristov. 2009. A meta-analysis of the effects of dietary protein concentration and degradability on milk protein yield and milk N efficiency in dairy cows. J. Dairy Sci. 92: 3222-3232. Huhtanen, P., E. H. Cabezas-Garcia, S. J. Krizsan, and K. J. Shingfield. 2015. Evaluation of between-cow variation in milk urea and rumen ammonia nitrogen concentrations and the association with nitrogen utilization and diet digestibility in lactating cows. J. Dairy Sci. 98: 3182-3196. Huhtanen, P., J. I. Nousiainen, M. Rinne, K. Kytölä, and H. Khalili. 2008. Utilization and partition of dietary nitrogen in dairy cows fed grass silage-based diets. J. Dairy Sci. 91: 3589- 3599. Hurley, A. M., N. López-Villalobos, S. McParland, E. Lewis, E. Kennedy, M. O'Donovan, and D. P. Berry. 2017. Genetics of alternative definitions of feed efficiency in grazing lactating dairy cows. J. Dairy Sci. 100: 5501-5514. Ingvartsen, K. L., and J. B. Andersen. 2000. Integration of metabolism and intake regulation: a review focusing on periparturient animals. J. Dairy Sci. 83: 1573-1597. Jonker, J. S., R. A. Kohn, and J. High. 2002. Dairy herd management practices that impact nitrogen utilization efficiency. J. Dairy Sci. 85: 1218-1226. Kauffman, A. J., and N. R. St-Pierre. 2001. The relationship of milk urea nitrogen to urine nitrogen excretion in Holstein and jersey cows1. J. Dairy Sci. 84: 2284-2294. Kertz, A.F. 2010. Urea feeding to dairy cattle: A historical perspective and review. The Professional Animal Scientist, 26: 257-272. Knap, P. W. 2005. Breeding robust pigs. Aust. J. Exp. Agric. 45: 763–773. Koch, R. M., L. A. Swiger, D. Chambers, and K. E. Gregory. 1963. Efficiency of feed use in beef cattle. J. Anim. Sci. 22:486–494. Kolver, E., L.D. Muller, G.A. Varga, and T.J. Cassidy. 1998. Synchronization of ruminal degradation of supplemental carbohydrate with pasture nitrogen in lactating dairy cows. J. Dairy Sci. 81: 2017-2028. 29 König, S., and K. May. 2019. Invited review: Phenotyping strategies and quantitative-genetic background of resistance, tolerance and resilience associated traits in dairy cattle. Animal 13: 897-908. Lammers, B. P., and A. J. Heinrichs. 2000. The Response of Altering the Ratio of Dietary Protein to Energy on Growth, Feed Efficiency, and Mammary Development in Rapidly Growing Pre-pubertal Heifers1. J. Dairy Sci. 83: 977-983. Lawrence, P., D. A. Kenny, B. Earley, and M. McGee. 2012. Grazed grass herbage intake and performance of beef heifers with predetermined phenotypic residual feed intake classification. Animal. 6:1648–1661. Lawrence, P., D. A. Kenny, B. Earley, D. H. Crews Jr., and M. McGee. 2011. Grass silage intake, rumen and blood variables, ultrasonic and body measurements, feeding behavior, and activity in pregnant beef heifers differing in phenotypic residual feed intake. J. Anim. Sci. 89:3248–3261. Leenhouwers, J.I., E.F. Knol, P.N. de Groot, H. Vos, and T. van der Lende. 2002. Fetal development in the pig in relation to genetic merit for piglet survival. J. Anim. Sci. 80: 1759– 1770. Li, J., D. Chen, and S. Xu. 1998. The analysis on genetic factors of feed energy and protein efficiency of Chinese Simmental. In Proceedings of 6th World Congress on Genetics Applied in Livestock Production (pp. 133-136). Lobley, G. E., G. D. Milano, and J. G. van der Walt. 2000. The liver: Integrator of nitrogen metabolism. Pages 149–168 in Ruminant Physiology: Digestion, Metabolism, Growth, and Reproduction. P. B. Cronjé, ed. CABI Publishing, New York, NY. Lu, Y., M. J. Vandehaar, D. M. Spurlock, K. A. Weigel, L. E. Armentano, E. E. Connor, M. Coffey, R. F. Veerkamp, Y. de Haas, C. R. Staples, Z. Wang, M. D. Hanigan, and R. J. Tempelman. 2018. Genome-wide association analyses based on a multiple-trait approach for modeling feed efficiency. J. Dairy Sci. 101: 3140-3154. MacRae, J. C., L. A. Bruce, D. S. Brown, D. A. Farningham, and M. Franklin. 1997. Absorption of amino acids from the intestine and their net flux across the mesenteric- and portal-drained viscera of lambs. J. Anim. Sci. 75:3307–3314. Mangual, M.J., E. Liu, and M. J. VandeHaar. 2016. Repeatability of residual feed intake across dietary forage concentration, J Animal Sci., 94 (suppl_5):348–349. Marett, L. C., S. R. O. Williams, B. J. Hayes, J. E. Pryce, and W. J. Wales. 2017) Partitioning of energy and nitrogen in lactating primiparous and multiparous Holstein–Friesian cows with divergent residual feed intake. Animal Prod. Sci. 57(7): 1499-1506. McDonald T. J., B. M. Nichols, M. M. Harbac, T. M. Norvell, and J. A. Paterson. 2010. Dry matter intake is repeatable over parities and residual feed intake is negatively correlated with dry matter digestibility in gestating cows. J. Anim. Sci. 88(E-Suppl. 2):12. 30 McGee, M., J. A. Ramirez, G. E. Carstens, W. J. Price, J. B. Hall, and R. A. Hill. 2014. Relationships of feeding behaviors with efficiency in RFI-divergent Japanese Black cattle. J. Anim. Sci. 92: 3580-3590. Misztal, I., 1997. Estimation of variance components with large-scale dominance models. J. Dairy Sci. 80: 965-974. Mitchell, R. G., G. W. Rogers, C. D. Dechow, J. E. Vallimont, J. B. Cooper, U. Sander-Nielsen, and J. S. Clay. 2005. Milk urea nitrogen concentration: heritability and genetic correlations with reproductive performance and disease. J. Dairy Sci. 88: 4434-4440. Moharrery, A., 2004. Investigation of different levels of RDP in the rations of lactating cows and their effects on MUN, BUN and urinary N excretion. Ital. J. Anim. Sci. 3: 157-165. Moraes, L.E., E. Kebreab, J. L. Firkins, R.R. White, R. Martineau, and H. Lapierre. 2018. Predicting milk protein responses and the requirement of metabolizable protein by lactating dairy cows. J. Dairy Sci. 101: 310-327. Morme`de, P., A. Foury, E. Terenina, and P.W. Knap. 2010. Breeding for robustness: the role of cortisol. Animal. 5: 651–657. Mulder, H. A., Rönnegård, L., Fikse, W. F., Veerkamp, R. F., and Strandberg, E. 2013. Estimation of genetic variance for macro- and micro-environmental sensitivity using double hierarchical generalized linear models. Genet. Sel. Evol. 45:23. Neal, H.D. S. C., and J. H. M. Thornley. 1983. The lactation curve in cattle: a mathematical model of the mammary gland. J. Agric. Sci. 101: 389-400. Nieuwhof, G. J., J. A. M. Van Arendonk, H. Vos, and S. Korver. 1992. Genetic relationships between feed intake, efficiency and production traits in growing bulls, growing heifers and lactating heifers. Livest. Prod. Sci. 32:189–202. Nkrumah, J. D., E. K. Okine, G. W. Mathison, K. Schmid, C. Li, J. A. Basarab, M. A. Price, Z. Wang, and S. S. Moore. 2006. Relationships of feedlot feed efficiency, performance, and feeding behavior with metabolic rate, methane production, and energy partitioning in beef cattle. J. Anim. Sci. 84:145–153. Nousiainen, J., K. J. Shingfield, and P. Huhtanen. 2004. Evaluation of milk urea nitrogen as a diagnostic of protein feeding. J. Dairy Sci. 87: 386-398. NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Oltner, R., and H. Wiktorsson. 1983. Urea concentrations in milk and blood as influenced by feeding varying amounts of protein and energy to dairy cow. Livest. Prod. Sci. 10: 457-467. Phuong, H. N., N. C. Friggens, I. J. M. De Boer, and P. Schmidely. 2013. Factors affecting energy and nitrogen efficiency of dairy cows: A meta-analysis. J. Dairy Sci. 96: 7245-7259. 31 Potts, S. B., J. P. Boerman, A. L. Lock, M. S. Allen, and M. J. VandeHaar. 2015. Residual feed intake is repeatable for lactating Holstein dairy cows fed high and low starch diets. J. Dairy Sci. 98: 4735-4747. Potts, S. B., J. P. Boerman, A. L. Lock, M. S. Allen, and M. J. VandeHaar. 2017. Relationship between residual feed intake and digestibility for lactating Holstein cows fed high and low starch diets. J. Dairy Sci. 100: 265-278. Pryce, J. E., J. Arias, P. J. Bowman, S. R. Davis, K. A. Macdonald, G. C. Waghorn, W. J. Whales, II, Y. J. Williams, R. J. Spelman, and B. J. Hayes. 2012. Accuracy of genomic predictions of residual feed intake and 250-day body weight in growing heifers using 625,000 single nucleotide polymorphism markers. J. Dairy Sci. 95:2108–2119. Rauw, W. M., and L. Gomez-Raja. 2015. Genotype by environment interaction and breeding for robustness in livestock. Front. Genet. 6: 310. Reynolds, C. K. 2005. Nitrogen metabolism by splanchnic tissues of ruminants. In Biology of metabolism of growing animals. pp. 197–220. Elsevier Science, Oxford, England. Reynolds, C. K. 2006. Splanchnic metabolism of amino acids in ruminants. In Ruminant physiology. Digestion, metabolism and impact of nutrition on gene expression, immunology and stress. pp. 225–248. Wageningen Academic Publishers, The Netherlands. Richardson, E. C. and R. M. Herd. 2004. Biological basis for variation in residual feed intake in beef cattle . 2 . Synthesis of results following divergent selection Cooperative Research Centre for Cattle and Beef Quality . Aust. J. Exp. Agric. 44:431–440. Rius, A. G., J. A. D. R. N. Appuhamy, J. Cyriac, D. Kirovski, O. Becvar, J. Escobar, M. L. McGilliard, B. J. Bequette, R. M. Akers, and M. D. Hanigan. 2010. Regulation of protein synthesis in mammary glands of lactating dairy cows by starch and amino acids. J. Dairy Sci. 93:3114–3127. Rius, A. G., S. Kittelmann, K. A. Macdonald, G. C. Waghorn, P. H. Janssen, and E. Sikkema. 2012. Nitrogen metabolism and rumen microbial enumeration in lactating cows with divergent residual feed intake fed high-digestibility pasture. J. Dairy Sci. 95:9. 5024–5034. Roseler, D.K., J. D. Ferguson, C.J. Sniffen, and J. Herrema, 1993. Dietary protein degradability effects on plasma and milk urea nitrogen and milk nonprotein nitrogen in Holstein cows. J. Dairy Sci. 76: 525-534. Russell, J. B., and R. J. Wallace. 1997. Energy-yielding and energy-consuming reactions. In The rumen microbial ecosystem. 246-282. Springer, Dordrecht. Santos, F.A.P., J.E.P. Santos, C.B. Theurer, and J.T. Huber. 1998. Effects of rumen- undegradable protein on dairy cow performance: A 12-year literature review. J. Dairy Sci. 81: 3182-3213. 32 Schenkel, F. S., S. P. Miller, and J. W. Wilton. 2004. Genetic parameters and breed differences for feed efficiency, growth, and body composition traits of young beef bulls. Can. J. Anim. Sci. 84:177–185. Schingoethe, D. J. 1991. Protein quality and amino acid supplementation in dairy cattle. Pages 101–106 in Proc. South- west Nutr. Manage. Conf., Tempe, AZ. Dep. Anim. Sci., Univ. Arizona, Tucson. Schwab, C. G. 1994. Optimizing amino acid nutrition for optimum yields of milk and milk protein. Pages 114–129 in Proc. Southwest Nutr. Manage. Conf., Phoenix, AZ. Dep. Anim. Sci., Univ. Arizona, Tucson. Schwab, C.G. and G.A. Broderick, 2017. A 100-Year Review: Protein and amino acid nutrition in dairy cows. J. Dairy Sci. 100: 10094-10112. Shaver, R. D. 2010. Diets fed in selected WI high-producing dairy herds. Retrieved on 15 January 2017 from http://shaverlab.dysci.wisc.edu/wp-content/uploads/ sites/87/2015/04/2010wihigh-producingherds.pdf. Sinclair, K. D., P. C. Garnsworthy, G. E. Mann, and L. A. Sinclair. 2014. Reducing dietary protein in dairy cow diets: implications for nitrogen utilization, milk production, welfare and fertility. Animal. 8: 262-274. St-Pierre, N. R. 2012. The costs of nutrients, comparison of feedstuffs prices and the current dairy situation. The Ohio State University Extension Buckeye News. Accessed Jul. 20, 2013. Stewart, B. A., R. E. James, M. D. Hanigan, and K. F. Knowlton. 2012. Cost of reducing protein and phosphorus content of dairy rations. The Professional Animal Scientist. 28: 115-119. Stewart, G. S., and C. P. Smith. 2005. Urea nitrogen salvage mechanisms and their relevance to ruminants, non-ruminants and man. Nutr. Res. Rev. 18: 49-62. Stoop, W. M., H. Bovenhuis, and J. A. M. Van Arendonk. 2007. Genetic parameters for milk urea nitrogen in relation to milk production traits. J. Dairy Sci. 90: 1981-1986. Tempelman, R. J., D. M. Spurlock, M. Coffey, R. F. Veerkamp, L. E. Armentano, K. A. Weigel, and M. J. VandeHaar. 2015. Heterogeneity in genetic and nongenetic variation and energy sink relationships for residual feed intake across research stations and countries. J. Dairy Sci. 98: 2013-2026. Thornhill, J. B., L. C. Marett, M. J. Auldist, J. S. Greenwood, J. E. Pryce, B. J. Hayes, and W. J. Wales. 2014. Whole-tract dry matter and nitrogen digestibility of lactating dairy cows selected for phenotypic divergence in residual feed intake. Animal Prod. Sci. 54(9): 1460- 1464. Vallimont, J. E., C. D. Dechow, J. M. Daubert, M. W. Dekleva, J. W. Blum, C. M. Barlieb, W. Liu, G. A. Varga, A. J. Heinrichs, and C. R. Baumrucker. 2011. Heritability of gross feed 33 efficiency and associations with yield, intake, residual intake, body weight, and body condition score in 11 commercial Pennsylvania tie stalls. J Dairy Sci. 94:2108–2113. VandeHaar, M. J. 1998. Efficiency of nutrient use and relationship to profitability on dairy farms. J. Dairy Sci. 81:272–282. VandeHaar, M. J., and N. St-Pierre. 2006. Major advances in nutrition: Relevance to the sustainability of the dairy industry. J. Dairy Sci. 89: 1280-1291. VandeHaar, M. J., L. E. Armentano, K. Weigel, D. M. Spurlock, R. J. Tempelman, and R. Veerkamp. 2016. Harnessing the genetics of the modern dairy cow to continue improvements in feed efficiency1. J. Dairy Sci. 99: 4941-4954. Veerkamp, R. F., and G. C. Emmans. 1995. Sources of genetic variation in energetic efficiency of dairy cows. Livest. Prod. Sci. 44:87–97. Waghorn, G. C., K. A. Macdonald, Y. Williams, S. R. Davis, and R. J. Spelman. 2012. Measuring residual feed intake in dairy heifers fed an alfalfa (Medicago sativa) cube diet. J. Dairy Sci. 95:1462–1471. Weeth, H. J., and A. L. Lesperance. 1965. Renal Function of Cattle under Various Water and Salt Loads 1, 2. J. Anim. Sci. 24: 441-447. Williams, Y. J., J. E. Pryce, C. Grainger, W. J. Wales, N. Linden, M. Porker, and B. J. Hayes. 2011. Variation in residual feed intake in Holstein-Friesian dairy heifers in southern Australia. J. Dairy Sci. 94:4715–4725. Wood, G. M., P. J. Boettcher, J. Jamrozik, G. B. Jansen, and D. F. Kelton. 2003. Estimation of genetic parameters for concentrations of milk urea nitrogen. J. Dairy Sci.: 2462-2469. Wu, Z., and L. D. Satter. 2000. Milk Production during the Complete Lactation of Dairy Cows Fed Diets Containing Different Amounts of Protein1. J. Dairy Sci. 83: 1042-1051. Xi, Y. M., F. Wu, D. Q. Zhao, Z. Yang, L. Li, Z. Y. Han, and G. L. Wang. 2016. Biological mechanisms related to differences in residual feed intake in dairy cows. Animal, 10(8): 1311- 1318. Yang, J. Y., J. Seo, H. J. Kim, S. Seo and J. K. Ha. 2010. Nutrient synchrony: is it a suitable strategy to improve nitrogen utilization and animal performance?. Asian-australas. J. Anim. Sci. 23: 972-979. Zamani, P., S. R. Miraei-Ashtiani, D. Alipour, H. Aliarabi, and A. A. Saki. 2011. Genetic parameters of protein efficiency and its relationships with yield traits in lactating dairy cows. Livest Sci. 138(1-3): 272-277. Zetouni, L., M. Henryon, M.Kargo and J. Lassen. 2017. Direct multitrait selection realizes the highest genetic response for ratio traits. J. Animal Sci. 95(5): 1921-1925. 34 CHAPTER 2 RELATIONSHIP OF RESIDUAL FEED INTAKE TO PROTEIN EFFICIENCY IN LACTATING COWS FED HIGH OR LOW PROTEIN DIET A version of this manuscript has been accepted by Journal of Dairy Science, DOI: https://doi.org/10.3168/jds.2019-17567 ABSTRACT Our objectives were to determine the repeatability of residual feed intake (RFI) across dietary protein contents and to determine the association between RFI and protein efficiency in lactating cows. Holstein cows (n=166; 92 primiparous, 74 multiparous) with initial milk yield (MY) 41.3(cid:1)9.8 kg/d were fed diets with high or low protein (HP or LP) in peak lactation. Experiments were conducted as crossovers with two treatment periods of 28-35 d. Production of 69 of the 166 cows (42 primiparous, 27 multiparous) was also measured in late lactation. Low- protein diets were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function. High-protein diets were 18% CP in peak lactation and 16% CP in late lactation and contained extra expeller soybean meal to increase absorbed protein. Cows were milked 2 times daily; DMI and MY were recorded daily. Milk composition was measured over 4 consecutive milkings weekly, and BW was measured 3 times weekly. Fixed effects of diet, parity, treatment sequence, and treatment period, interaction of parity and diet, interaction of diet and period, and random effects of experiment and cow nested within experiment were included in the model to compare intake and production performance 35 between cows fed different levels of CP. RFI value was calculated for each cow on each treatment based on the actual intake, milk energy output, metabolic BW, and body energy (calculated from BW change and BCS over the treatment period) change. Ranking of cows for RFI was moderately repeatable across dietary protein in peak lactation (r = 0.59) but less repeatable in late lactation (r = 0.41). A negative correlation was observed between RFI and protein efficiency values (dietary protein captured in milk) for cows in both peak lactation (r = - 0.42) and late lactation (r = -0.24), which suggested that cows with higher energy efficiency had greater protein efficiency. In conclusion, RFI was repeatable across dietary protein contents within lactation stage, and cows with lower RFI values utilized protein more efficiently. 36 INTRODUCTION Given that 40% of feed cost can be attributed to protein (St-Pierre, 2012), much effort has been made to improve protein efficiency in dairy cows. Many nutritional means have been explored to improve protein efficiency, such as altering protein sources, supplementing with non- protein nitrogen or specific amino acids, and lowering dietary protein contents (Sinclair et al., 2014; Broderick et al., 2015; Gidlund et al., 2015). However, the efficiency of converting feed protein to milk protein is still less than 30% in the modern dairy (Huhtanen and Hristov, 2009). We wondered if genetic means could be used to further improve protein efficiency. One possible way to enhance protein efficiency is to select cows based on residual feed intake (RFI). Calculated as the difference between actual feed intake and predicted feed intake, RFI is considered as a direct measurement of energy efficiency (Koch et al., 1963) and has drawn considerable attention in the genetic improvement of dairy cattle (Connor, 2015; VandeHaar et al., 2016). RFI takes into account all production variables, and avoids bias caused by body size or milk production level when comparing efficiency between cows (Pryce et al., 2012; Connor, 2015). RFI is usually calculated on an energy basis, and, although we expect that lower RFI would be associated with greater protein efficiency, this has not been demonstrated. We also wondered whether RFI ranking among cows is repeatable across diets with varying protein contents. In lactating dairy cattle, RFI is repeatable across diets with varying starch and NDF contents (Potts et al., 2015; Mangual et al., 2016). However, no studies have reported the RFI ranking of cows across dietary protein contents. Most diets in previous RFI 37 studies contained adequate or even excessive protein (Mangual et al., 2016; Potts et al., 2015; Tempelman et al., 2015). If RFI is not repeatable across protein contents, using the RFI information from cows fed excessive protein to determine protein efficiency might be misleading. Lack of repeatability might especially be a problem if cows are fed diets that limit protein to minimize N excretion. Thus, the objective of this study was to examine the repeatability of RFI across diets with high or marginally deficient protein contents and the relationship between RFI and protein efficiency. We hypothesized that 1) RFI was relatively repeatable across dietary protein contents, and 2) cows with lower RFI values would have higher protein efficiency. MATERIALS AND METHODS Cows, Experimental Design, and Diets Experimental procedures were approved by the Institutional Animal Care and Use Committee of Michigan State University. Data from 166 cows in 11 separate cross-over experiments, with 7 studies containing cows in peak lactation and 4 studies containing cows in late lactation, were used to determine the RFI and protein efficiency of individual cows across diets that were high (HP) and low (LP) in protein in different lactation stages. The LP diet were formulated to be marginally-deficient in protein so that milk production would likely drop. In the 7 experiments containing 166 peak lactation cows, days in milk (DIM) was 50 to 130 d for all cows at the beginning of each experiment. For all 7 experiments, the 2 experimental periods lasted 28 to 35 d per period. Within each experiment, cows were blocked based on their 38 parity and productivity (milk energy per unit of metabolic BW), and then randomly assigned to one of the two treatment sequences (HP-LP, or LP-HP). For cows in peak lactation, the LPpeak diet (LP diet for peak lactation cows) contained 31% NDF, 32% starch and 14% CP, and the HPpeak diet (the HP diet for peak lactation cows) contained 29% NDF, 30% starch and 18% CP. Both diets contained at least 9.8% RDP (DM basis) to maintain adequate rumen function (NRC, 2001). The extra protein of HPpeak was achieved by replacing soybean hulls and ground corn with expeller soybean meal (Table 2.1). 39 51:49 51:49 68:32 68:32 Late lactation HPlate 50.0 18.2 8.2 9.2 0.6 9.7 3.3 1.0 LPlate 50.0 18.2 11.4 15.8 0.4 ---- 3.3 1.0 LPpeak 35.2 15.7 29.7 12.8 2.5 ---- 3.2 0.9 Peak lactation HPpeak 35.2 15.7 25.6 5.7 0.3 13.4 3.2 0.9 Table 2.1 Feed Ingredients and Nutrient Composition of Experimental Diets 1,2 Ingredient, % DM Corn silage Alfalfa silage Corn grain, ground Soybean hulls Solvent extracted soybean meal Expeller soybean meal Vitamin and mineral mix3 Urea Forage: Concentrate Nutrient Composition, % DM DM NDF Forage NDF Starch CP RDP RUP Apparent NEl, Mcal/kg4 1HPpeak and LPpeak diet were high-protein diets and low-protein diets fed to peak lactation cows, and HPlate and LPlate diet were high-protein diets and low-protein diets fed to late-lactation cows. 2Experimental diets were fed to cows in crossover design with at least 28-d periods 3Vitamin and mineral mix contained 24.8% ground corn grain, 21.5% dehydrated cane molasses, 11.2% limestone, 9.6% blood meal, 9.0% sodium bicarbonate, 6.6% dicalcium phosphate, 4.2% ReaShure choline, 3.1% magnesium sulfate, 2.8% salt, 2.0% vegetable oil, 1.5% niacin, 1.3% trace mineral mix, 0.95% biotin, 0.7% YeastPlus, 0.54% vitamin ADE premix, 0.32% selenium yeast, and 0.09% Rumensin 90. 4Mean apparent net energy concentration of diets, based on average cow performance. For each diet, Diet NEL= the average of (MilkE + 0.08 ´ MBW + BodyE)/ DMI for all cows on the diet, where MilkE is net energy utilized for milk synthesis, MBW is metabolic body weight, and ΔBodyE is net energy captured in body tissue. 37.9 38.0 29.5 24.4 15.9 10.0 6.0 1.5 47.1 31.3 20.7 33.5 14.3 9.8 4.5 1.6 37.9 40.2 29.5 26.0 12.8 9.0 3.5 1.6 Treatments 47.1 29.4 20.7 31.5 18.0 10.3 7.7 1.7 40 In the 4 experiments containing 69 late-lactation cows, DIM was 190 to 250 d at the beginning of each experiment. Intake, BW, and milk production of the 69 cows were measured in both peak and late lactation. For all the 4 experiments, the 2 experimental periods lasted 28 to 35 d per period. Within each experiment, cows were blocked based on their parity and milk energy per unit of metabolic BW, and then randomly assigned to one of the two treatment sequences (HP-LP, or LP-HP). For cows in late lactation, the LPlate diet (LP diet for late lactation cows) contained 40% NDF, 26% starch and 13% CP, and HPlate diet (HP diet for late lactation cows) contained 38% NDF, 24% starch and 16% CP. Both diets contained at least 9.0% RDP (DM basis) for rumen function. The extra protein of HPlate was achieved by replacing soybean hulls and ground corn with expeller soybean meal (Table 2.1). All cows were housed in individual tie stalls and milked twice a day (0430 and 1530). Tie stalls were equipped with a double-cupped watering system to prevent contamination of feed with water and with side panels and a front gate to prevent other cows from stealing feed during cow movements. Water was available ad libitum. Cows were fed once a day (1200) at > 110% of expected intake based on intake of the previous day, and orts were removed (1000) and weighed prior to feeding. Milk yield was recorded electronically at each milking, and milk samples were obtained from 4 consecutive milkings each wk. Milk samples were analyzed for fat, protein, lactose, somatic cell count, and MUN with infrared spectroscopy (AOAC, 1990; method 972.160) by Michigan DHIA (Grand Ledge, MI). Body weight for each cow was recorded 3 times per week immediately after the afternoon milkings. At the beginning and end of each 41 period, BCS was determined by 3 trained investigators and averaged for each cow on a 5-point scale, where 1=thin and 5=fat (Wildman et al., 1982). Collection and analyses of diet ingredients were the same for all the experiments. During the last 5 d of experimental periods, samples of feed ingredients were obtained daily to determine the nutrient profile of the diets. All samples were frozen after collection until analysis. Samples were composited to obtain one sample per period and dried in a forced air oven (57°C for > 72 h) before grinding through a Wiley mill (5-mm and 1-mm screen; Arthur H. Thomas Co., Philadelphia, PA). Samples of feed were analyzed for CP, starch, NDF, and ether extract. Calculations Milk energy output (MilkE; Mcal/d) for individual cows was estimated by the following equation (NRC, 2001; Equation 2-15): MilkE = [9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg)], where each component was calculated as the average output of individual cows during the treatment period. The milk: feed ratio for a cow during a period was determined as the average daily energy-corrected milk yield (ECM; ECM = [0.327 ´ milk (kg) + 12.95 ´ fat (kg) + 7.20 ´ protein (kg)]; Tyrell and Reid, 1965) divided by the average daily dry matter intake (DMI) over the entire period. 42 For cows > 190 d pregnant, body weight (BW) was corrected for conceptus weight (CW) for use in the RFI equation and to calculate energy and protein change of body tissues. CW was calculated using the equation from NRC (2001), CW = [18 + (D - 190) x 0.665] ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Metabolic BW (MBW) of a cow was estimated as BW0.75, where BW was the mean measured BW for the cow during the treatment period. Mean daily BW change (dBW; kg/d) was calculated for each cow within the treatment period by linear regression after two rounds of removing outliers in the data; an outlier was any BW > 3.5 SD from the regression line. Energy expended for body tissue gain (BodyE; Mcal/d) was estimated by an equation derived from NRC (2001; Table 2-5): DBodyE = (2.88 + 1.036 ´ BCS) ´ dBW, where BCS was the average BCS for a cow during the treatment period. Energy expended for pregnancy (PregE; Mcal/d) was estimated using the equation from NRC (2001; Equation 2-19): PregE = [(0.00318 ´ D - 0.0352) ´ (CBW/45)] / 0.218, where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Apparent diet energy content (DietNEL; Mcal/kg) was calculated for each diet as the average NEL required by each cow divided by her average daily intake for the diet: 43 DietNEL = Average [(MilkE + 0.08 ´ MBW + DBodyE + PregE) / DMI], where DMI was the average DMI for a cow during the treatment period. Models and Statistics The RFI for each cow on each diet was calculated as the residual term in the prediction of DMI. DMI for an individual cow during each period was regressed as a function of major energy sinks using GLM Procedure in SAS (9.4). To define RFI for cows in the peak lactation, DMI was modeled as: DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ DBodyE + β4 ´ DIM+ β5 ´ DIM ´ DIM + β6 ´ BCS + Parity + Experiment + Cohort(Experiment)+ Diet(Cohort ´ Experiment)+ e, where DMI was the observed DMI, MilkE was the observed milk energy output, MBW was the average BW0.75, DBodyE was the predicted change in body energy based on measured BW and BCS, DIM was the average DIM during each treatment period, and BCS was the average BCS during each treatment period; parity (1 or 2+), experiment (1-7), cohort nested within experiment, and diet nested within cohort and experiment were fixed effects, where a cohort is a group of cows that ate the same diet at the same time. RFI was defined as the residual term (e) in the model. For cows in the late lactation, DMI was modeled as:: DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ DBodyE + β4 ´ PregE + β5 ´ DIM + β6 ´ BCS + Parity + Experiment + Cohort(Experiment) + Diet(Cohort ´ Experiment)+e, where PregE was the energy expended for pregnancy, and RFI was still the residual term (e) in the model. 44 To determine the number of animals that changed their efficiency classification when they were switched from one diet to the other, cows were grouped into high (HRFI), medium (MRFI), and low (LRFI) RFI groups. Cows > 0.5 SD of the mean RFI for a cohort were classified as HRFI, cows < -0.5 SD were classified as LRFI, and those ± 0.5 SD were classified as MRFI. Repeatability of RFI across dietary protein contents was calculated using Pearson correlation coefficients by CORR procedure of SAS (9.4). Two RFI values within each lactation stage for each cow (RFIHPpeak vs. RFILPpeak; RFIHPlate vs. RFILPlate) were included in the analyses. To examine the RFI repeatability across lactation stages, two RFI values were calculated: RFIpeak and RFIlate, where RFIpeak was the average RFI across diets in peak lactation, and RFIlate was the average RFI across diets in late lactation. Pearson correlation coefficient for RFIpeak and RFIlate was calculated. To further examine the RFI repeatability across dietary protein contents, two RFI values were calculated: RFIhigh and RFIlow, where RFIhigh was the average RFI for high-protein diets across lactation stage, and RFIlow was the average RFI for low-protein diets across diets lactation stage. Pearson correlation coefficient for RFIhigh and RFIlow was calculated. Correlation was considered as significant at P≤ 0.05 and trends at P ≤0.10. For each cow on each diet, protein efficiency was calculated as dietary protein captured in milk protein (milk protein efficiency, MPE), and dietary protein captured in milk protein and body tissues (gross protein efficiency, GPE), respectively. Protein captured for body tissue gain 45 (BodyP; kg/d) was calculated using the following equations, which were derived as averages for BCS of 3.0 based on NRC (2001): BodyP = 0.12 ´ dBW for primiparous cows, BodyP = 0.07 ´ dBW for multiparous cows. To quantify the association among RFI, MPE, and GPE, Pearson correlation coefficients were obtained using the CORR Procedure in SAS (9.4). Partial correlations accounting for effects of parity, cohort, and experiment were estimated using the PARTIAL option in the CORR Procedure. To further determine the differences in protein efficiency between the most and least efficient cows, cows with different RFI for each diet (HP or LP) in each lactation stage across all eleven experiments were compared. The effect of RFI was determined using the GLM Procedure of SAS according to the model Yi = μ + Ri + e, where μ was the overall mean, Ri was the fixed effect of RFI group, and e was the residual error. Production, efficiency, and energy partitioning responses to diets with each lactation stage were analyzed using the MIXED Procedure in SAS (9.4), with fixed effects of diet, parity, treatment sequence, period, interaction of parity and diet, interaction of diet and period, and the random effects of experiment and cow nested within experiment. Significance was considered at P ≤ 0.05 and tendency at P ≤ 0.10. Interactions were considered significant at P ≤ 0.10 and trends at P ≤ 0.15. 46 Animal Performance RESULTS Cows fed low protein in general ate less, produced less milk, and gained less BW than cows fed high protein, in both peak and late lactations. As shown in Table 2.2 and Table 2.3, LPpeak decreased DMI (P< 0.01), MY (P< 0.01), milk fat yield (P< 0.01), milk protein yield (P< 0.01), milk lactose yield (P< 0.01), milk protein percentage (P< 0.01), milk lactose percentage (P< 0.01), and MUN (P< 0.01). For these cows, LPpeak also decreased BW (P< 0.01), BW gain (P< 0.01), BCS (P= 0.04), and change in BCS (P= 0.06). In peak lactation, LPpeak also decreased ECM per kg DMI (P< 0.01), milk energy (P< 0.01), and estimated retained energy (P< 0.01). As shown in Table 2.4 and Table 2.5, LPlate decreased DMI (P< 0.01), milk yield (P< 0.01), 3.5% FCM (P< 0.01), milk fat yield (P< 0.01), milk protein yield (P< 0.01), milk lactose yield (P< 0.01), and MUN (P< 0.01). For these cows, LPlate also decreased BW (P< 0.01), non-pregnant BW (P< 0.01), BW gain (P< 0.01), non-pregnant BW gain (P< 0.01), and BCS (P= 0.04). LPlate also decreased ECM per kg DMI (P< 0.01), milk energy (P< 0.01), maintenance energy (P< 0.01), and estimated retained energy (P< 0.01). Primiparous cows in general ate less, and produced less milk, but with greater milk component concentration, in both peak and late lactations. As shown in Table 2.2 and Table 2.3, among peak lactation cows, compared to multiparous cows, primiparous cows had less DMI (P< 0.01), MY (P< 0.01), FCM (P< 0.01), milk fat yield (P< 0.01), milk protein yield (P< 0.01), MUN (P< 0.01), and milk lactose yield (P< 0.01), with higher milk protein percentage (P= 0.05), 47 and milk lactose percentage (P< 0.01). Primiparous cows also had less milk energy (P< 0.01), and maintenance energy (P< 0.01), compared to multiparous cows. As shown in Table 2.4 and Table 2.5, among all late lactation cows, compared to multiparous cows, primiparous cows had higher milk fat yield (P= 0.08), milk fat percentage (P= 0.02), and milk lactose percentage (P< 0.01). Primiparous cows also had lower BW (P< 0.01), non-pregnant BW (P< 0.01), BCS (P< 0.01), change in BW (P= 0.07), and maintenance energy (P< 0.01), compared to multiparous cows. 48 Table 2.2 Dry matter intake, milk production, milk components and feed efficiency for cows fed treatment diets in peak lactation1,2 P-value5 Parity Treatments3 SEM Parity4 TRT ´ Parity8 SEM TRT HPpeak n=166 24.3 LPpeak n=166 23.3 Primi. n=184 21.3 Multi. n=148 26.3 0.86 0.35 0.14 <0.01 <0.01 37.3 37.2 36.7 33.7 34.8 33.7 0.23 0.53 0.25 0.01 0.19 0.16 44.7 43.3 43.7 41.2 41.0 40.6 0.96 0.87 0.84 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.13 0.03 1.27 3.49 1.07 2.94 1.84 4.99 9.2 1.62 1.41 3.46 1.21 2.97 2.07 5.01 15.1 1.70 DMI, kg/d Milk Yield, kg/d Milk ECM6 3.5% FCM7 Milk Components Fat, kg/d Fat, % Protein, kg/d Protein, % Lactose, kg/d Lactose, % MUN, mg/dL ECM/DMI 1Average DIM was 125 for primiparous cows in HPpeak diet, 126 for primiparous cows in LPpeak diet, 122 for multiparous cows in HPpeak diet, and 121 for multiparous cows in LPpeak diet. 2Average parity for multiparous cows was 2.94 in peak lactation. 3 Treatments contained 18% and 14% crude protein on a DM basis for peak lactation cows. 4 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 5 P-value associated with treatment differences (HPpeak vs. LPpeak; TRT) and parity differences (Primi vs. Multi.; Parity) in peak lactation cows. 6 Energy-corrected milk; ECM = [(0.327 ´ kg milk) + (12.95 ´ kg milk fat) + (7.20 ´ kg milk protein)] (Tyrrell and Reid, 1965). 7 Fat-corrected milk; 3.5 % FCM = [(0.4324 ´ kg milk) + (16.216 ´ kg milk fat)]. 8 Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 <0.01 0.13 <0.01 <0.01 <0.01 0.01 <0.01 <0.01 <0.01 0.13 <0.01 0.05 <0.01 <0.01 <0.01 0.87 1.50 3.43 1.28 2.91 2.20 4.94 12.7 1.66 0.59 0.30 0.18 0.07 0.01 0.14 0.16 0.37 1.18 3.53 1.01 2.99 1.71 5.07 11.6 1.65 0.03 0.06 0.02 0.03 0.05 0.02 0.20 0.03 49 LPpeak n=166 653 3.19 0.20 0.02 HPpeak n=166 658 3.23 0.57 0.07 Primi. n=184 596 3.23 0.38 0.04 Multi. n=148 714 3.20 0.40 0.05 Table 2.3 Body weight, body condition score and calculated energy values for cows fed experimental diets in peak lactation P-value3 Parity TRT ´ Parity Treatments1 SEM Parity2 SEM TRT 8.86 0.05 0.10 0.03 1.74 42.1 28.1 3.64 10.4 <0.01 0.41 0.88 0.83 <0.01 0.04 <0.01 0.06 1.04 0.02 0.10 0.03 0.64 0.03 0.18 0.63 0.01 BW, kg BCS, unit Change in BW, kg/ d4 Change in BCS, unit/28 d Calculated energy values5 Apparent NEL of diet, Mcal/kg Apparent NEL, Mcal/d Milk, Mcal/d Body Tissue Gain, Mcal/d Maintenance, Mcal/d 1 Treatments contained 18% and 14% crude protein on a DM basis for peak lactation cows. 2 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 3 P-value associated with treatment differences (HPpeak vs. LPpeak; TRT) and parity differences (Primi vs. Multi.; Parity) in peak lactation cows. 4 Determined by linear regression using BW measurements throughout the period. 5 Milk (MilkE)=[ 9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg) ]. Body tissue gain (ΔBodyE) = [(2.88+1.036 ´ BCS) ´ ΔBW], Maintenance=0.08 ´ MBW, where MBW= BW0.75 <0.01 <0.01 <0.01 <0.01 <0.01 0.61 <0.01 <0.01 0.82 <0.01 1.67 35.4 23.4 2.38 9.6 1.65 43.4 29.9 2.52 11.0 1.58 36.7 25.1 1.26 10.3 0.85 0.03 0.57 0.66 0.10 0.16 0.54 0.23 0.10 0.39 0.51 0.15 0.30 0.10 50 Table 2.4 Dry matter intake, milk production, milk components and feed efficiency for cows fed treatment diets in late lactation1,2 Treatments3 Parity4 SEM P-value5 Parity Primi. n=84 18.0 Multi. n=54 20.2 DMI, kg/d Milk Yield, kg/d Milk ECM6 3.5% FCM7 Milk Components Fat, kg/d Fat, % Protein, kg/d Protein, % Lactose, kg/d Lactose, % MUN, mg/dL ECM/DMI HPlate n=69 19.8 25.1 27.8 26.8 0.98 3.92 0.80 3.23 1.21 4.79 12.1 1.41 LPlate n=69 18.4 22.2 24.4 23.7 0.86 4.05 0.58 3.21 1.05 4.79 8.1 1.32 SEM 0.20 0.42 0.45 0.43 0.02 0.03 0.01 0.01 0.02 0.03 0.16 0.02 24.2 27.2 26.3 0.98 4.13 0.77 3.23 1.21 4.99 9.9 1.51 23.1 25.1 24.2 0.87 3.85 0.72 3.21 1.06 4.58 10.2 1.22 0.48 1.37 1.53 1.42 0.06 0.12 0.04 0.04 0.14 0.05 0.27 0.06 TRT <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 0.34 <0.01 0.90 <0.01 <0.01 TRT ´ Parity8 <0.01 0.06 0.02 0.03 0.02 0.79 <0.01 0.10 0.03 0.72 0.25 0.06 0.32 0.42 0.18 0.15 0.08 0.02 0.25 0.79 0.05 <0.01 0.28 <0.01 1Average DIM was 258 for primiparous cows in HPlate diet, 257 for primiparous cows in LPlate diet, 263 for multiparous cows in HPlate diet, and 264 for multiparous cows in LPlate diet. 2Average parity for multiparous cows was 3.12 in late lactation. 3 Treatments contained 16% and 13% crude protein on a DM basis for late lactation cows. 4 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 5 P-value associated with treatment differences (HPlate vs. LPlate; TRT) and parity differences (Primi vs. Multi.; Parity) in late lactation cows. 6 Energy-corrected milk; ECM = [(0.327 ´ kg milk) + (12.95 ´ kg milk fat) + (7.20 ´ kg milk protein)] (Tyrrell and Reid, 1965). 7 Fat-corrected milk; 3.5 % FCM = [(0.4324 ´ kg milk) + (16.216 ´ kg milk fat)]. 8 Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 51 BW, kg Non-pregnant BW, kg BCS, unit Change in BW4, kg/ d Change in non-pregnant BW, kg/d Change in BCS, unit/28 d Calculated energy values5 Apparent NEL of diet, Mcal/kg Apparent NEL, Mcal/d Milk, Mcal/d Body Tissue Gain, Mcal/d Maintenance, Mcal/d Pregnancy, Mcal/d Table 2.5 Body weight, body condition score and calculated energy values for cows fed experimental diets in late lactation Parity2 SEM TRT P-value3 Parity Treatments1 LPlate n=69 693 679 3.56 0.09 -0.05 -0.45 HPlate n=69 702 694 3.62 0.67 0.43 0.07 1.64 32.3 18.4 2.82 10.8 0.34 1.46 26.6 16.0 -0.34 10.6 0.37 SEM 1.52 5.12 0.03 0.09 0.09 0.37 0.04 0.77 0.31 0.60 0.06 0.11 Primi. n=84 623 616 3.44 0.29 0.14 -0.01 1.62 29.2 18.1 0.89 9.9 0.27 Multi. n=54 772 757 3.74 0.47 0.23 -0.37 1.48 29.8 16.3 1.59 11.5 0.45 14.5 14.9 0.09 0.09 0.09 0.37 0.04 1.01 1.04 0.61 0.18 0.13 <0.01 <0.01 0.04 <0.01 <0.01 0.16 <0.01 <0.01 <0.01 <0.01 <0.01 0.77 <0.01 <0.01 <0.01 0.07 0.33 0.34 <0.01 0.55 0.09 0.26 <0.01 0.17 TRT ´ Parity 0.98 0.07 0.94 0.18 0.09 0.14 0.06 0.05 0.026 0.12 0.09 0.72 1 Treatments contained 16% and 13% crude protein on a DM basis for late lactation cows. 2 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 3 P-value associated with treatment differences (HPlate vs. LPlate; TRT) and parity differences (Primi vs. Multi.; Parity) in late lactation cows. 4 Determined by linear regression using BW measurements throughout the period. 5 Milk (MilkE)=[ 9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg) ]. Body tissue gain (ΔBodyE) = [(2.88+1.036 ´ BCS) ´ ΔBW]. Maintenance=0.08 ´ MBW, where MBW= BW0.75 6 Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 52 Repeatability of Residual Feed Intake across Protein Contents and Lactation Stages In the RFI model in peak-lactation cows, the coefficients for the major energy sinks were 0.44 (P< 0.01) for MilkE, 0.06 (P< 0.01) for MBW, and 0.03 (P= 0.03) for DBodyE. The model R2 and root mean square error were 0.87 and 1.50, respectively. In the RFI model in late- lactation cows, the coefficients for the major energy sinks were 0.37 (P< 0.01) for MilkE, 0.08 (P< 0.01) for MBW, 0.05 (P= 0.13) for DBodyE, and -0.04 (P= 0.83) for PregE. The model R2 and root mean square error were 0.80 and 1.31, respectively. Further details are shown in Table 2.6. 53 1.73 1.15 0.24 0.19 Table 2.6 Partial regression coefficients of the RFI models in peak- and late- lactation cows Peak lactation Late lactation SEM 2.48 0.03 0.01 0.02 0.31 0.28 0.03 0.0001 P-value 0.18 <0.01 <0.01 0.03 0.16 <0.01 0.02 0.04 Contribution to DMI Mean SD 10.8 8.99 0.07 2.18 0.85 0.18 Coefficient -3.22 0.44 0.06 0.03 0.45 -1.06 0.07 -0.0002 SEM Coefficient 2.46 2.26 Intercept 0.03 0.37 MilkE1 0.02 0.08 MBW2 0.04 0.05 DBodyE3 0.19 -0.04 PregE4 0.44 -0.19 BCS 0.41 -1.08 Parity 0.007 0.03 DIM DIM´DIM -1.09 to -0.49 Experiment 0.13 to 3.03 -0.92 to -0.22 Cohort -1.49 to 1.65 1Milk energy (Mcal/d)= [ 9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg) ]. 2 Metabolic BW (Kg) = BW0.75 3 Energy utilized in body tissue gain (Mcal/d) = [(2.88+1.036 ´ BCS) ´ ΔBW]. 4 Pregnancy energy (Mcal/d) = [(0.00318 ´ D - 0.0352) ´ (CBW/45)] / 0.218, where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Contribution to DMI Mean SD 6.51 10.7 0.07 0.08 P-value 0.36 <0.01 <0.01 0.13 0.83 0.68 <0.01 <0.01 54 Regarding the RFI repeatability, in general, cows with lower RFI values when fed high- protein diets still had low RFI when switched to low-protein diets. In peak lactation, RFI was moderately repeatable across high and low protein diets (r= 0.59, P< 0.01). Figure 2.1 illustrates the relationship between RFI in HPpeak and RFI in LPpeak. In late lactation, RFI was less repeatable across protein contents as it was in peak lactation (r= 0.41, P= 0.03). Figure 2.2 illustrates the relationship between RFI in HPlate and RFI in LPlate. A moderate level of correlation between RFIhigh and RFIlow was observed (r= 0.51, P< 0.01). The Pearson correlation coefficient across peak- and late- lactation was 0.52 (P< 0.01) for DMI, 0.05 (P= 0.67) for MilkE, 0.05 (P= 0.67) for DBodyE, and 0.91 (P< 0.01) for MBW. Based on the data from 69 cows examined in both peak and late lactations, the repeatability of RFI was low across lactation stage for the HP diet (r= 0.11, P= 0.39), but moderate across lactation stage for the LP diet (r= 0.28, P= 0.02). The correlation between RFIpeak and RFIlate for both diets combined was moderate (r= 0.25, P= 0.04). Further details are shown in Table 2.7 and Figure 2.3. 55 Figure 2.1 Repeatability of residual feed intake (RFI) across dietary protein contents in peak lactation cows d / g k , s t e i D n i e t o r p - w o L n o I F R -6 -4 -2 6 4 2 0 -2 -4 -6 0 2 4 6 Primiparous Cows Multiparous Cows RFI on High-protein Diets, kg/d Figure 2.1 Repeatability of residual feed intake (RFI) across dietary protein contents in peak lactation cows (n=166). Repeatability of RFI across dietary protein contents was r= 0.59. RFI on low-protein diets could be predicted using RFI on high-protein diets as Y= 0.628 (± 0.062; P< 0.01) ´ X – 0.000 (± 0.09). Each data point represents one cow’s RFI value for each diet (n=166). Open circles indicate primiparous cows (n=92), and filled triangles indicate multiparous cows (n=74). 56 Figure 2.2 Repeatability of residual feed intake (RFI) across dietary protein contents in late lactation cows d / g k , s t e i D n i e t o r p - w o L n o I F R -6 -4 -2 6 4 2 0 -2 -4 -6 0 2 4 6 Primiparous Cows Multiparous Cows RFI on High- protein Diets, kg/d Figure 2.2 Repeatability of residual feed intake (RFI) across dietary protein contents in late lactation cows (n=69). Repeatability of RFI across dietary protein contents was r= 0.41. RFI on low-protein diets could be predicted using RFI on high-protein diets as Y= 0.387 (± 0.101; P< 0.01) ´ X – 0.000 (± 0.14). Each data point represents one cow’s RFI value for each diet (n=69). Open circles indicate primiparous cows (n=42), and filled triangles indicate multiparous cows (n=27). 57 Figure 2.3 Repeatability of residual feed intake (RFI) across lactation stages in lactating cows d / g k , s w o C n o i t a t c a L - e t a L n i I F R -6 -4 -2 6 4 2 0 -2 -4 0 2 4 6 Primiparous Cows Multiparous Cows RFI in Peak-Lactation Cows, kg/d -6 Figure 2.3 Repeatability of residual feed intake (RFI) across lactation stages in lactating cows (n=69). Repeatability of RFI across lactation stages was r= 0.25. RFI in late-lactation cows could be predicted using RFI in peak-lactation cows as Y= 0.198 (± 0.097; P= 0.02) ´ X – 0.009 (± 0.12). Each data point represents one cow’s RFI value in each stage (n=69). Open circles indicate primiparous cows (n=42), and filled triangles indicate multiparous cows (n=27). 58 Table 2.7 Repeatability of RFI across protein contents within lactation stage and across lactation stages (n = 69) RFIHPlate RFILPlate RFIlate5 0.13 0.31 0.28 0.02 0.25 0.04 0.70 < 0.01 0.85 < 0.01 0.82 < 0.01 RFIlow6 0.40 < 0.01 0.86 < 0.01 0.73 < 0.01 0.51 < 0.01 0.36 < 0.01 0.73 < 0.01 0.64 < 0.01 0.11 0.39 0.28 0.02 0.23 0.05 0.34 < 0.01 0.41 < 0.01 RFIHPpeak RFILPpeak RFIpeak3 RFIhigh4 RFIHPlate RFILPlate RFIlate5 RFILPpeak 0.591 < 0.012 RFIpeak3 0.84 < 0.01 0.87 < 0.01 RFIhigh4 0.77 < 0.01 0.46 < 0.01 0.84 < 0.01 0.11 0.39 0.20 0.10 0.23 0.06 0.76 < 0.01 1 The Pearson correlation coefficient of the linear relationship between 2 variables 2 The P value associated with the linear relationship between 2 variables 3 Averaged RFI across the diets fed to cows in peak lactation 4 Averaged RFI for the HP diet across lactation stages 5 Averaged RFI across the diets fed to cows in late lactation 6 Averaged RFI for the LP diet across lactation stages 59 Residual Feed Intake and Protein Efficiency Overall, cows with lower RFI values exhibited higher protein efficiency. For cows in peak lactation, the Pearson correlation coefficient between RFI and milk protein efficiency was - 0.59 (P< 0.01) in the HPpeak diet and -0.41 (P< 0.01) in LPpeak diet. For cows in late lactation, the Pearson correlation coefficient between RFI and milk protein efficiency was -0.36 (P= 0.02) in the HPlate diet and -0.13 (P= 0.34) in LPlate diet. The correlation coefficient between RFI and milk protein efficiency across diets was -0.42 (P< 0.01; Figure 2.4) in peak lactation cows and - 0.24 (P= 0.06; Figure 2.5) in late lactation cows. Similar associations between RFI and gross protein efficiency were also observed in the current study, as MPE and GPE were highly correlated in both peak lactation (r= 0.83, P< 0.01) and late lactation (r= 0.89, P< 0.01). As shown in Table 2.8, cows with lower RFI values in peak lactation had higher milk protein efficiency and gross protein efficiency regardless of protein content in the diet; however, cows with lower RFI values did not necessarily have lower MUN concentration. Similar trends were observed in cows fed the HP diet in late lactation. In contrast, when fed the LP diet, cows with lower RFI in late lactation did not exhibit greater protein efficiency (MPE, GPE) nor MUN. 60 Figure 2.4 Association between residual feed intake and milk protein efficiency in peak lactation cows - w o L d n a - h g i H n o % , s t e i D n i e t o r P y c n e i c i f f E n i e t o r P k l i M -6 50 40 30 20 Cows on high-protein diets Cows on low-protein diets 10 -4 RFI on High- and Low- Protein Diets, kg/d -2 2 0 4 6 Figure 2.4 Association between residual feed intake and milk protein efficiency in peak lactation cows (n=332). In peak-lactation cows, correlation between residual feed intake (RFI) and milk protein efficiency (MPE) across high and low protein diets was -0.42. The correlation between RFI and MPE in high-protein diets was -0.59 (The equation was MPE= -0.96 (± 0.14; P< 0.01) ´ RFI + 27.5 (± 0.20; P< 0.01)), and the correlation between RFI and MPE in low-protein diets was -0.41 (The equation was MPE= -1.34 (± 0.14; P< 0.01) ´ RFI + 32.3 (± 0.21; P< 0.01)). Each data point represents one cow’s RFI value for each diet. Open triangles indicate cows in high-protein diets (n=166), and filled squares indicate cows in low-protein diets (n=166). 61 Figure 2.5 Association between residual feed intake and milk protein efficiency in late lactation cows % , s t e i D n i e t o r P - w o L d n a - h g i H n o y c n e i c i f f E n i e t o r P k l i M -6 50 40 30 20 10 Cows on high-protein diets Cows on low-protein diets 2 4 6 -2 -4 RFI on High- and Low- Protein Diets, kg/d 0 Figure 2.5 Association between residual feed intake and milk protein efficiency in late lactation cows (n=138). In late-lactation cows, correlation between residual feed intake (RFI) and milk protein efficiency (MPE) across high and low protein diets was -0.24. The correlation between RFI and MPE in high-protein diets was -0.36 (The equation was MPE= -1.27 (± 0.42; P< 0.01) ´ RFI + 26.5 (± 0.58; P< 0.01)), and the correlation between RFI and MPE in low-protein diets was -0.13 (The equation was MPE= -0.71 (± 0.55; P= 0.20) ´ RFI + 29.4 (± 0.68; P< 0.01)). Each data point represents one cow’s RFI value for each diet. Open triangles indicate cows in high-protein diets (n=138), and filled squares indicate cows in low-protein diets (n=138). 62 Variable Peak lactation MPE3, % GPE4, % MUN, mg/dL Late lactation MPE, % GPE, % MUN6, mg/dL 25.9 26.9 14.7 24.8 25.8 12.1 27.5 28.7 15.1 26.0 27.1 12.3 29.1 30.9 15.3 29.0 30.5 12.1 P-value2 <0.01 <0.01 0.76 0.01 <0.01 0.65 30.5 31.4 9.1 27.9 27.9 8.2 32.1 32.6 9.1 30.2 29.7 8.1 34.6 34.9 9.5 29.6 30.8 8.2 P-value <0.01 <0.01 0.47 0.27 0.29 0.78 Table 2.8 Protein efficiency and MUN of high-, medium- and low-RFI cows fed high- and low- protein diets across lactation stage High-protein diets HRFI1 npeak=47 MRFI npeak= 80 LRFI npeak=39 Low Protein diets HRFI npeak=49 MRFI npeak=73 LRFI npeak=44 nlate= 19 nlate= 28 nlate= 22 nlate= 22 nlate= 26 nlate= 21 1 Cows were grouped into high (HRFI), medium (MRFI), and low (LRFI) RFI groups. Cows > 0.5 SD of the mean RFI for a cohort were classified as HRFI, cows < -0.5 SD were classified as LRFI, and those ± 0.5 SD were classified as MRFI. 2 P-value associated with group difference 3 MPE, milk protein efficiency, defined as the dietary protein captured in milk 4 GPE, gross protein efficiency, defined as dietary protein captured in milk and body tissue 63 DISCUSSION Animal Performance across Dietary Protein Contents and Lactation Stages Overall, cows had lower feed intake and milk production when fed low-protein diets regardless of lactation stage. Production differences between the HP and LP cows were most likely the result of additional RUP supplementation in the HP diet, and thus inadequate metabolizable protein in the LP diet. The low-protein diets also decreased gains in BW and BCS. Significantly less gains in BW than in body condition might be due to the slightly less intakes of LP cows or might indicate that cows tended to gain less body protein than fat when fed the LP diet. With the similar decrease in dietary protein content, the decrease of feed intake and milk production was similar between peak-lactation and late-lactation cows; however, late-lactation cows tended to lose more non-pregnant BW and BCS than peak-lactation cows (table 3 and table 5). We suggest that nutrients were prioritized to pregnancy and milk synthesis instead of body tissue gain when protein was limiting in late lactation. Indeed, Bauman and Currie (1980) described that the priority of nutrient partitioning in cattle was pregnancy, followed by milk production, and lastly body reserve gain; our data was consistent with this idea. Repeatability of Residual Feed Intake across Dietary Protein Contents and Lactation Stages Although production was significantly altered by the diets, RFI within cows was still repeatable across dietary protein contents within each lactation stage. The literature on RFI repeatability has predominantly been focusing on peak-lactation cows. The current study supports the previous studies on RFI repeatability and extends RFI repeatability across diets to late-lactation cows. Among peak-lactation cows, the moderate level of RFI repeatability found in the current study (0.59) was in line with the previous RFI repeatability studies, where RFI was 64 repeatable across starch contents (0.73; Potts et al., 2015) and forage NDF contents (0.54; Mangual et al., 2016). According to Richardson and Herd (2004), the major contributor to the variation of RFI in cattle is “tissue metabolism and protein turnover”. We expect that treatments altering these processes might alter RFI significantly, and therefore alter RFI repeatability. Wessels et al.(1997) showed that supplementing amino acids alters protein turnover. Thus, the lower level of RFI repeatability in the current study, compared to Potts et al. (2015), might be related to the expected changes in protein metabolism when altering dietary protein. Lower RFI repeatability across dietary protein contents in late lactation, compared to peak lactation, was expected due to the uncertainty of pregnancy weight gain, which will be further discussed below. RFI repeatability across physiological states, such as across lactation stages in the current study, has been reported previously. Phenotypic correlation of RFI was low either when compared across weaned beef heifers later tested as lactating cows (Archer et al., 2002), or when estimated in growing dairy heifers that were later tested during lactation (Nieuwhof et al., 1992; Williams et al., 2011; Waghorn et al., 2012). The work done by Liinamo et al., (2015) and Li et al. (2017) demonstrated that genetic RFI values estimated from various lactation stages were different, and the difference was extremely evident when comparing the RFI estimated from early lactation with that estimated from late lactation. Although the DMI was moderately repeatable (r=0.52) across lactation stages in the current study, given the low repeatability of the major energy sinks, especially MilkE, the low RFI repeatability was fully expected. The low RFI repeatability across lactation stage could be due to the following reasons: 1) mechanisms controlling energy efficiency (or partitioning) shifted as lactation proceeded, and 2) our estimates of body energy change were not accurate and were altered by lactation stage. Throughout lactation, dairy cows undergo physiological changes, including 1) body reserve mobilization in 65 early lactation, 2) body tissue replenishment in peak-lactation, and 3) extra body fat storage in late lactation. We used BW change and BCS to predict energy change; however, BW change also included change in gut content and pregnancy gain, and body composition could not be fully represented by BCS. Practically, RFI is an adjusted DMI after accounting for energy partitioning to milk, body tissue gain, maintenance, and pregnancy (in late lactation); thus, any errors in estimating the energy sinks mentioned above can introduce errors in calculating RFI. Errors in BW change could introduce significant bias in the RFI estimation (Potts et al., 2015). This becomes especially important when estimating RFI for late-lactation cows. BW change in late-lactation cows was calculated from adjusted BW after deducting conceptus weight from measured BW. Given the difficulty in getting a precise estimate of conceptus weight, BW change in late-lactation cows could not be quantified as accurately as it was in peak-lactation cows. Therefore, more errors could be introduced in estimating RFI among late-lactation cows. Additionally, the difficulty in assessing conceptus weight can also contribute to the errors in estimating energy utilized in pregnancy. As errors were introduced in the two primary energy sinks, we expected that the estimated value of RFI would be less accurate in late lactation cows. Indeed, due to the difficulty of estimating BW change, Prendiville et al. (2011) advised to estimate RFI based on data between DIM 150 and DIM 230 when tissue gain or loss was minimal, in order to generate the most accurate estimates of RFI. Residual Feed Intake and Protein Efficiency No prior study has directly examined the relationship between RFI and protein efficiency in lactating Holstein cows; however, the relationship between RFI and protein efficiency in growing heifers was examined. Rius et al. (2012) observed no difference in nitrogen efficiency between 2 groups of Holstein-Friesian heifers with divergent RFI values. Following that, the 66 work done by Thornhill et al. (2014) and Marett et al. (2017) further showed that cows selected for lower RFI when they were calves/heifers did not have higher nitrogen efficiency in the subsequent lactation. In contrast with the heifer studies, the results in the current study suggested that RFI is strongly associated with protein efficiency in peak lactation cows and also in late lactation cows when protein is not limiting. Xi et al. (2016) and Mangual et al. (2016) speculated that lactating cows with lower RFI values might have higher protein efficiency, as indicated by the lower MUN values in their low- RFI cows. Prior work showed that protein turnover rates could be negatively associated with protein utilization efficiency in dairy cows (Herd et al., 2004; Castro Bulle et al., 2007). There is also work showing that greater protein turnover rates were related to higher RFI values in cattle (Richardson et al., 2004). Based on the prior work, a negative association between RFI and protein efficiency was expected. Indeed, the current study directly proves that this negative relationship exists in most cases, unless protein is limiting for pregnant cows. This poor correlation could be due to the nutrient repartitioning to pregnancy when protein was limiting in pregnant cows. Therefore, when pregnancy does not take the priority over milk synthesis, cows with lower RFI should utilize protein more efficiently. The moderate correlation between RFI and protein efficiency provides a new means to genetically improve protein efficiency in dairy cattle. Although a medium level of heritability for protein efficiency existed in lactating dairy cows (0.10-0.31; Li et al.,1998; Zamani et al., 2011), directly selecting dairy cows based on traditional protein efficiency term was questioned. Most of the doubt was due to the drawbacks of protein efficiency being a ratio trait. Ratio traits are usually not normally distributed. As a result, it is difficult to expect the selection response due to the disproportionate selection pressure on the component traits (Zetouni et al., 2017). In other 67 words, using ratio traits (e.g., protein efficiency term) in genetic selection induces large error variance and unexpected results. In contrast, RFI, as a residual term, overcomes all the drawbacks in ratio traits and is favorable in cow selection. However, due to the complexity of collecting individual intake data in dairy cows, estimating RFI is still difficult in dairy cows. CONCLUSIONS Low-protein diets significantly decreased feed intake, milk production, BW, energy captured in milk and body tissue, and feed efficiency in both peak and late lactation cows. Within each lactation stage, RFI was moderately repeatable across dietary protein contents; similarly, average RFI in high- and low-protein diets across lactation stages was also moderately repeatable. Thus, we expect that cows with lower RFI when fed diets with adequate protein, as is typical for North America, will still have lower RFI when fed diets marginally deficient in protein. Lastly, cows with lower RFI values utilized protein more efficiently. We suggest that protein efficiency will be improved in the process of selecting dairy cattle based on RFI. ACKNOWLEDGEMENTS We would like to acknowledge J. S. Liesman and the staff of the Michigan State University Dairy Cattle Teaching and Research Center for their assistance in these experiments, and Landus Cooperative for donating Soyplus soybean meal. This project was supported by Agriculture and Food Research Initiative Competitive Grant no. 2011-68004-30340 from the USDA National Institute of Food and Agriculture and funds from the Michigan Alliance for Animal Agriculture and Michigan AgBioResearch. 68 APPENDIX 69 Late-lactation cows Primi. on LP 17.2 22.4 24.9 24.2 0.90 0.69 Multi. on HP 20.8 24.2 26.3 25.3 0.91 0.76 Multi. on LP 19.5 22.1 23.9 23.1 0.83 0.68 Primi. on HP 18.8 26.1 29.5 28.4 1.05 0.84 Supplementary Table 2.1 Dry matter intake, milk production, milk components, feed efficiency, body weight, body condition score and calculated energy for cows fed treatment diets in peak and late lactation12 Primi. on HP Peak-lactation cows Multi. Primi. on LP on HP Multi. on LP 47.0 42.5 35.4 32.1 2.92 2.33 2.96 1.61 3.01 1.81 2.90 2.07 DMI, kg/d Milk, kg/d ECM3, kg/d 3.5% FCM4, kg/d Milk fat, kg/d Milk protein, kg/d Milk protein, % Milk lactose, kg/d ECM/DMI Non-pregnant BW, kg Change in non-pregnant BW, kg/d Apparent NEL of diet, Mcal/kg Apparent NEL, Mcal/d Milk energy5, Mcal/d Maintenance energy6, Mcal/d 1 Treatments (HP vs. LP) contained 18% and 14% crude protein on a DM basis for peak lactation cows, and 16% and 13% crude protein on a DM basis for late lactation cows. 2 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 3 Energy-corrected milk; ECM = [(0.327 ´ kg milk) + (12.95 ´ kg milk fat) + (7.20 ´ kg milk protein)] (Tyrrell and Reid, 1965). 4 Fat-corrected milk; 3.5 % FCM = [(0.4324 ´ kg milk) + (16.216 ´ kg milk fat)]. 5 Milk (MilkE)=[ 9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg) ]. 6 Maintenance=0.08 ´ MBW, where MBW= BW0.75 1.09 1.45 614 -0.18 1.49 25.5 16.5 9.9 1.31 1.57 619 0.47 1.75 32.8 19.7 9.9 1.12 1.24 769 0.40 1.53 31.9 17.2 11.7 1.00 1.20 745 0.07 1.42 27.7 15.5 11.4 70 REFERENCES 71 REFERENCES AOAC. 1990. Official Methods of Analysis. 15th ed. Association of Official Analytical Chemists, Gaithersburg, MD. Archer, J. A., A. Reverter, R. M. Herd, D. J. Johnston, and P. F. Arthur. 2002. Genetic variation in feed intake and efficiency of mature beef cows and relationships with post weaning measurements. 7th World Congr. Genet. Appl. Livest. Prod. 31:221–224. Bauman, D. E., and W. B. Currie. 1980. Partitioning of nutrients during pregnancy and lactation: a review of mechanisms involving homeostasis and homeorhesis. J. Dairy Sci. 63(9):1514- 1529. Broderick, G. A., A. P. Faciola, and L. E. Armentano. 2015. Replacing dietary soybean meal with canola meal improves production and efficiency of lactating dairy cows1 J. Dairy Sci. 98: 5672-5687. Castro Bulle, F. C. P., P. V. Paulino, A. C. Sanches, and R. D. Sainz. 2007. Growth, carcass quality, and protein and energy metabolism in beef cattle with different growth potentials and residual feed intakes. J. Anim. Sci. 85:928–936. Connor, E. E. 2015. Invited review: Improving feed efficiency in dairy production: Challenges and possibilities. Animal. 9: 395-408. Gidlund, H., M. Hetta, S. J. Krizsan, S. Lemosquet, and P. Huhtanen. 2015. Effects of soybean meal or canola meal on milk production and methane emissions in lactating dairy cows fed grass silage-based diets. J. Dairy Sci. 98: 8093-8106. Herd, R. M., V. H. Oddy, and E. C. Richardson. 2004. Biological basis for variation in residual feed intake in beef cattle. 1. Review of potential mechanisms. Aust. J. Exp. Agric. 44:423– 430. Huhtanen, P., and A. N. Hristov. 2009. A meta-analysis of the effects of dietary protein concentration and degradability on milk protein yield and milk N efficiency in dairy cows. J. Dairy Sci. 92: 3222-3232. Huhtanen, P., E. H. Cabezas-Garcia, S. J. Krizsan, and K. J. Shingfield. 2015. Evaluation of between-cow variation in milk urea and rumen ammonia nitrogen concentrations and the 72 association with nitrogen utilization and diet digestibility in lactating cows. J. Dairy Sci. 98: 3182-3196. Koch, R. M., L. A. Swiger, D. Chambers, and K. E. Gregory. 1963. Efficiency of Feed Use in Beef Cattle 1. J. Anim. Sci. 22: 486-494. Li, J., D. Chen, and S. Xu. 1998. The analysis on genetic factors of feed energy and protein efficiency of Chinese Simmental. In Proceedings of 6th World Congress on Genetics Applied in Livestock Production (pp. 133-136). Li, B., B. Berglund, W. F. Fikse, J. Lassen, M. H. Lidauer, P. Mäntysaari, and P. Løvendahl. 2017. Neglect of lactation stage leads to naive assessment of residual feed intake in dairy cattle. J. Dairy Sci. 100: 9076-9084. Liinamo, A.-E., P. Mantysaari, M. H. Lidauer, and E. A. Mantysaa-ri. 2015. Genetic parameters for residual energy intake and energy conversion efficiency in Nordic Red dairy cattle. Acta Agric. Scand. A Anim. Sci. 65:63–72. Carrasquillo-Mangual, M.J., E. Liu, and M. J. VandeHaar. 2016. Repeatability of residual feed intake across dietary forage concentration, J Animal Sci., 94 (suppl_5):348–349. Marett, L. C., S. R. O. Williams, B. J. Hayes, J. E. Pryce, and W. J. Wales. 2017) Partitioning of energy and nitrogen in lactating primiparous and multiparous Holstein–Friesian cows with divergent residual feed intake. Animal Prod. Sci. 57(7): 1499-1506. Nieuwhof, G. J., J. A. M. Van Arendonk, H. Vos, and S. Korver. 1992. Genetic relationships between feed intake, efficiency and production traits in growing bulls, growing heifers and lactating heifers. Livest. Prod. Sci. 32(3):189-202. NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Potts, S. B., J. P. Boerman, A. L. Lock, M. S. Allen, and M. J. VandeHaar. 2015. Residual feed intake is repeatable for lactating Holstein dairy cows fed high and low starch diets. J. Dairy Sci. 98: 4735-4747. Prendiville, R., K. M. Pierce, L. Delaby, and F. Buckley. 2011. Animal performance and production efficiencies of Holstein-Friesian, Jersey and Jersey × Holstein-Friesian cows throughout lactation. Livest. Sci. 138:25–33. Pryce, J. E., J. Arias, P. J. Bowman, S. R. Davis, K. A. Macdonald, G. C. Waghorn, and B. J. Hayes. 2012. Accuracy of genomic predictions of residual feed intake and 250-day body 73 weight in growing heifers using 625,000 single nucleotide polymorphism markers. J. Dairy Sci. 95: 2108-2119. Richardson, E. C., and R. M. Herd. 2004. Biological basis for variation in residual feed intake in beef cattle. 2. Synthesis of results following divergent selection. Australian Journal of Experimental Agriculture. 44: 431-440. Richardson, E. C., R. M. Herd, J. A. Archer, and P. F. Arthur. 2004. Metabolic differences in Angus steers divergently selected for residual feed intake. Aust. J. Exp. Agric. 44:441–452. Rius, A. G., S. Kittelmann, K. A. Macdonald, G. C. Waghorn, P. H. Janssen, and E. Sikkema. 2012. Nitrogen metabolism and rumen microbial enumeration in lactating cows with divergent residual feed intake fed high-digestibility pasture. J. Dairy Sci. 95: 5024-5034. Sinclair, K. D., P. C. Garnsworthy, G. E. Mann, and L. A. Sinclair. 2014. Reducing dietary protein in dairy cow diets: implications for nitrogen utilization, milk production, welfare and fertility. Animal. 8: 262-274. St-Pierre, N. R. 2012. The costs of nutrients, comparison of feedstuffs prices and the current dairy situation. The Ohio State University Extension Buckeye News. Accessed Jul. 20, 2013. Tempelman, R. J., D. M. Spurlock, M. Coffey, R. F. Veerkamp, L. E. Armentano, K. A. Weigel, and M. J. VandeHaar. 2015. Heterogeneity in genetic and nongenetic variation and energy sink relationships for residual feed intake across research stations and countries. J. Dairy Sci. 98: 2013-2026. Tyrrell, H. F., and J. T. Reid. 1965. Prediction of the energy value of the milk. J. Dairy Sci. 48: 1215-1223. Thornhill, J. B., L. C. Marett, M. J. Auldist, J. S. Greenwood, J. E. Pryce, B. J. Hayes, and W. J. Wales. 2014. Whole-tract dry matter and nitrogen digestibility of lactating dairy cows selected for phenotypic divergence in residual feed intake. Animal Prod. Sci. 54(9): 1460- 1464. VandeHaar, M. J., L. E. Armentano, K. Weigel, D. M. Spurlock, R. J. Tempelman, and R. Veerkamp. 2016. Harnessing the genetics of the modern dairy cow to continue improvements in feed efficiency1. J. Dairy Sci. 99: 4941-4954. Waghorn, G. C., K. A. Macdonald, Y. Williams, S. R. Davis, and R. J. Spelman. 2012. Measuring residual feed intake in dairy heifers fed an alfalfa (Medicago sativa) cube diet. J. Dairy Sci. 95(3): 1462-1471. 74 Wessels, R. H., E. C. Titgemeyer, and G. St. Jean. 1997. Effect of amino acid supplementation on whole-body protein turnover in Holstein steers. J. Animal Sci. 75 (11): 3066-3073. Wildman, E. E., G. M. Jones, P. E. Wagner, R. L. Boman, H. F. Troutt, and T. N. Lesch. 1982. A dairy cow body condition scoring system and its relationship to selected production characteristics. J. Dairy Sci. 65:495–501. Williams, Y. J., J. E. Pryce, C. Grainger, W. J. Wales, N. Linden, M. Porker, and B. J. Hayes. 2011. Variation in residual feed intake in Holstein-Friesian dairy heifers in southern Australia. J. Dairy Sci. 94(9): 4715-4725. Xi, Y. M., F. Wu, D. Q. Zhao, Z. Yang, L. Li, Z. Y. Han, and G. L. Wang. 2016. Biological mechanisms related to differences in residual feed intake in dairy cows. Animal, 10(8): 1311- 1318. Zamani, P., S. R. Miraei-Ashtiani, D. Alipour, H. Aliarabi, and A. A. Saki. 2011. Genetic parameters of protein efficiency and its relationships with yield traits in lactating dairy cows. Livest Sci. 138(1-3): 272-277. Zetouni, L., M. Henryon, M.Kargo and J. Lassen. 2017. Direct multitrait selection realizes the highest genetic response for ratio traits. J. Animal Sci. 95(5): 1921-1925. 75 CHAPTER 3 LOW PROTEIN RESILIENCE IS AN INDICATOR OF RELATIVE PROTEIN EFFICIENCY OF INDIVIDUAL DAIRY COWS A version of this manuscript has been submitted to Journal of Dairy Science ABSTRACT Our objectives were to determine 1) the sources of variation in cow responses to dietary protein reduction, and 2) the association of low protein resilience (LPR) with protein efficiency. Lactating Holstein cows (n= 166; 92 primiparous, 77 multiparous) with initial milk yield (MY) 41.3(cid:1)9.8 kg/d were included in the crossover experiments with two treatments and two periods of 28-35 d each. Production of 69 of the 166 cows (42 primiparous, 27 multiparous) was also measured in late lactation. Low-protein diets (LP) were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function. High- protein diets (HP) were 18% CP in peak lactation and 16% CP in late lactation and contained extra expeller soybean meal to increase absorbed protein. Protein efficiency terms (MPE: dietary protein captured in milk; GPE: dietary protein captured in both milk and body tissues) were calculated for each cow on each diet in both peak lactation (n= 332) and late lactation (n= 138). Low protein resilience value was calculated for each cow in peak lactation (n= 166) and late lactation (n= 69). The ability to maintain total protein capture (CapP, milk protein + body protein gain) varied significantly among cows, and the variation was mostly explained by CapP per kg 76 metabolic body weight (MBW) on the HP diet , parity, treatment sequence (HP- LP, LP- HP), and experiment. Protein efficiency (MPE and GPE) was moderately repeatable across dietary protein contents regardless of lactation stages. Milk urea nitrogen (MUN) was not associated with MPE or GPE in individual cows after accounting for the diet effect. Compared to low- LPR cows, high- LPR cows had similar protein efficiency (GPE and MPE) on the HP diet, but significantly higher GPE on the LP diet. In conclusion, cows maintained their protein-efficiency rankings when switched from the HP to LP diet, or vice versa; however, using MUN to rank cows for their protein efficiency may be misleading. With similar milk production on the HP diet, high-LPR cows were better able to maintain production and utilize protein more efficiently to adapt to low- protein feeding conditions. 77 INTRODUCTION Dairy cattle convert protein in feeds (many of which have little direct value for human nutrition) into milk protein, and dairy products have provided high-quality protein for human consumption for centuries (Broderick, 2018). Improving the efficiency of protein use has been the goal of many studies in the past 40 years, and dairy cattle in North America are generally fed lower protein diets today than they were 30 years ago. Lower dietary protein with the same milk protein output increases protein efficiency and profitability. However, protein-deficient diets can reduce DMI and thus milk yield (MY), which ultimately defeats the initial purpose of feeding less protein. Thus, the emerging challenge is to identify ways to feed less protein while maintaining or enhancing milk production to meet the dietary protein needs of a growing human population (Ingvartsen and Andersen, 2000; Huhtanen et al., 2008; Sinclair et al., 2014). Nutritionists typically examine the average response to diet interventions, and variation in the response among cows to protein reduction or supplementation has not been extensively studied. Some cows need less protein to meet the requirement because their production level is low. For some cows, however, the lack of response to reduced protein may simply imply that they did not need as much protein to achieve their milk production potential because they were able to use protein more efficiently than the cohorts. We will define cows that can tolerate less protein to maintain protein output as being resilient to low protein. To our knowledge, no prior studies have quantified this resilience and its relationship with protein utilization efficiency. 78 It is widely recognized that protein utilization efficiency, represented by MUN, varies both within cow and within herd (Wattiaux et al., 2005; Stoop et al., 2007; Huhtanen et al., 2015). However, no prior work has directly examined the repeatability of protein efficiency for individual lactating cows across diets that are high or low in protein content. No existing literature examined whether cows that are more protein-efficient in general are also more resilient to low- protein diets. Thus, our objectives were to 1) determine the sources of variation in cow responses to dietary protein reduction, and 2) the association of LPR with protein efficiency. Data MATERIALS AND METHODS Experimental procedures were approved by the Institutional Animal Care and Use Committee of Michigan State University. Data from 166 lactating Holstein dairy cows were used in this study. Among the 166 cows, 69 were studied in both peak and late lactations. Data of MY, milk components (milk protein and MUN), BW, and hip height were collected in the experiments. These are the same animal as in Liu and VandeHaar (2020). In brief, lactating Holstein cows (n= 166; 92 primiparous, 77 multiparous) with initial MY 41.3(cid:1)9.8 kg/d were included in the crossover experiments with two treatments and two periods of 28-35 d. The two treatments were HP and LP. Production of 69 of the 166 cows (42 primiparous, 27 multiparous) also was measured in late lactation. For cows in peak lactation, the LP diet contained 31% NDF, 79 32% starch and 14% CP, and the HP diet contained 29% NDF, 30% starch and 18% CP. Both diets contained at least 9.8% RDP (DM basis) to maintain adequate rumen function (NRC, 2001). For cows in late lactation, the LP diet contained 40% NDF, 26% starch and 13% CP, and the HP diet contained 38% NDF, 24% starch and 16% CP. Both diets contained at least 9.0% RDP (DM basis) for rumen function. The extra protein of HP diet was achieved by replacing soybean hulls and ground corn with expeller soybean meal. Cows were milked 2 times daily; DMI and MY were recorded daily. Milk composition was measured over 4 consecutive milkings weekly, and BW was measured 3 times weekly. Calculations For cows > 190 d pregnant, BW was corrected for conceptus weight (CW) for use in the calculation of protein change of body tissues. Conceptus weight was calculated using the equation from NRC (2001), CW = [18 + (D - 190) ´ 0.665] ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Metabolic BW (MBW) of a cow was estimated as BW0.75, where BW was the mean measured BW for the cow during the treatment period. After plotting the BW data along with the experimental day in the peak lactation cows (Supplementary Figure 1), we suspected that some of, if not all, BW change in the current study might be attributed to gut fill. In order to measure BW change more accurately, empty BW 80 (EBW) was calculated for each cow to adjust BW for the gut fill (Gibbs et al., 1992; Andrew et al., 1994), EBW = BW – 5.2 ´ DMI - CW, where DMI was the daily DMI when BW was measured. Mean daily EBW change (dEBW; kg/d) was calculated for each cow within the treatment period by linear regression after two rounds of removing outliers in the data; an outlier was any BW > 3.5 SD from the regression line. For multiparous cows, EBW change was considered to be all body condition; thus, protein captured for body tissue gained or lost with changes in EBW (BodyP; kg/d) was calculated using the following equations, BodyP = (0.151− 0.0268 ´ BCS) ´ dEBW (derived from NRC 2001, Table 2-4) For primiparous cows, we assumed their mature BW would be 700 kg and that they had to gain 0.14 kg EBW/d of true growth across the first lactation to reach 92% of mature BW by their second calving (NRC, 2001). Based on NRC (2001) equations (11-1 and 11-2), 0.132 kg protein per kg dEBW was assigned to the 0.14 kg/d true growth. Any deviation in dEBW from 0.14 kg/d was considered to be body condition gain or loss, and the dEBW associated with body condition change was the same as for multiparous cows, (0.151- 0.0268 ´ BCS) kg protein per kg dEBW. BodyP was estimated as : 81 BodyP=6 (0.151− 0.0268 ´ BCS) ´ dEBW , 0.132 ´ 0.14+[ (0.151−0.0268 ´ BCS) ´ (dEBW−0.14)] , Parity>1 Parity=1 Protein captured for pregnancy (PregP; kg/d) was calculated using the equation from NRC (2001): PregP = 0.00069 ´ D – 0. 0692 ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Total protein capture (CapP, kg/d) was calculated for each cow in each treatment (HP and LP) as follows: Milk Protein+BodyP ,DIM<200 Milk Protein+BodyP+PregP ,DIM≥200 CapP=6 After plotting the dCapP (change of CapP from HP to LP) along with the milk protein yield on the HP diet (Figure 3. 1), we found that the cows that produced less milk on the HP diet were those that exhibited less of a drop in captured protein on the LP diet. This decreased response to the LP diet would not appropriately indicate that a cow is more resilient. 82 Figure 3.1 Relationship between milk protein yield and cows response (and LPR, low protein resilience) in peak- lactation cows Cow Response vs. Milk Protein Yield for HP 0 0.5 1 1.5 2 2.5 -0.6-0.4-0.200.20.4 . d g k , P L o t P H m o r f o r P p a C d d / g k , R P L 0.4 0.2 0 0 -0.2 -0.4 Milk Protein Yield for HP, kg/d Primiparous Cows Multiparous Cows LPR vs. Milk Protein Yield on HP 0.5 1 1.5 2 2.5 Milk Protein Yield for HP, kg/d Primiparous Cows Multiparous Cows Figure 3.1 Relationship between milk protein yield and cows response (and LPR, low protein resilience) in peak- lactation cows. The correlation coefficient between milk protein yield for the HP diet and dCapP (change of total protein capture from HP to LP) was -0.41 in primiparous cows, and -0.45 in multiparous cows. The correlation coefficient between milk protein yield for HP and LPR was 0.04 in primiparous cows, and 0.03 in multiparous cows. Each data point represents one value (n=166). Open circles indicate primiparous cows, and solid triangles indicate multiparous cows. 83 To account for the difference in milk production on the HP diet and other factors that can potentially influence body protein mobilization and milk protein production, low protein resilience (LPR) was essentially calculated as the difference between the actual change of CapP and the predicted change of CapP, where larger numbers indicated better resilience. To calculate LPR, the initial full model was as: dCapP = β0 + β1 ´ CapP_MBWHP + β2 ´ BCSHP + β3 ´ MBW_HHHP + β4 ´ dCP + β5 ´ CPHP +β6 ´ DIMHP +Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where dCapP was the change of CapP from HP to LP (dCapP= CapPLP – CapPHP); CapP_MBWHP was the CapP per kg metabolic BW when fed the HP diet; BCSHP was the BCS when fed the HP diet; MBW_HHHP was the metabolic body weight to height ratio when fed the HP diet; Par was parity (primiparous or multiparous); DIMHP was the starting days in milk when fed the HP diet; Seq was treatment sequence (HP-LP or LP-HP); CPHP was the actual CP% in the HP diet; dCP was the actual CP% change from HP to LP; Exp was experiment, and e was the residual term in the model. LPR was the residual term of the model. All covariates were jointly checked for multicollinearity through variance inflation factors (VIF) analysis (SAS, 9.4). No covariates had VIF greater than 10. Following that, the backward stepwise model selection was used to finalize the model (SAS. 9.4). The reduced equation was used to determine an LPR value for each animal at each stage of lactation. 84 Cows were then grouped into high (HLPR), medium (MLPR), and low LPR (LLPR) groups. Cows > 0.5 SD of the mean LPR were classified as high LPR, cows < -0.5 SD were classified as Low LPR, and those between ± 0.5 SD were classified as medium LPR. For each cow on each diet within each lactation stage, protein efficiency was calculated as dietary protein captured in milk protein (milk protein efficiency, MPE), and dietary protein captured in milk protein and body tissues (gross protein efficiency, GPE). Repeatability of MPE across dietary protein contents was calculated using GLM procedure of SAS (9.4) within each lactation stage, after accounting for effects of diet, parity, treatment sequence, and experiment. To further examine the MPE repeatability across dietary protein contents, two MPE values were calculated: MPEHP and MPELP, where MPEHP was the average MPE for the HP diet across lactation stages and MPELP was the average MPE for the LP diet across lactation stages. To examine the MPE repeatability across lactation stages, two MPE values were calculated: MPEpeak and MPElate, where MPEpeak was the average MPE across diets in peak lactation and MPElate was the average MPE across diets in late lactation. Pearson correlation coefficients between MPEHP and MPELP and between MPEpeak and MPElate were calculated by GLM procedure after accounting for effects of parity and experiment. Correlation was considered as significant at P ≤ 0.05 and trends at P ≤ 0.10. Similar calculations and analyses were performed for GPE and MUN, in order to determine repeatability of GPE and MUN across dietary protein contents and lactation stages, respectively. 85 dMUN was calculated as the change of MUN from the HP diet to the LP diet. dMPE and dGPE were calculated as the change of MPE and GPE, respectively, from the HP diet to the LP diet. To quantify the association of LPR with various protein efficiency terms (MPE, GPE, MUN, dMPE, dGPE, and dMUN), Pearson correlation coefficients were obtained using the GLM procedure in SAS (9.4) after accounting for effects of parity, treatment sequence, and experiment. To determine the differences between the most and least resilient cows, production performance and protein efficiency of the cows from different LPR groups within each lactation stage were compared. The effect of LPR was determined using the GLM procedure of SAS according to the model Yi = μ + LPRi + e, where μ was the overall mean, LPRi was the fixed effect of LPR group, and e was the residual error. Cow production performance and protein efficiency responses to diets within each lactation stage were analyzed using the HPMIXED procedure in SAS (9.4), with fixed effects of diet, parity, treatment sequence nested in experiment, period within experiment, interaction of parity and diet, and the random effects of experiment and cow nested within experiment. Significance was considered at P (cid:1) 0.05 and tendency at P (cid:1) 0.10. Interactions were considered significant at P (cid:1) 0.10 and trends at P (cid:1) 0.15. 86 Cow Performance RESULTS As shown in Table 3.1, during peak lactation, the LP diet decreased milk protein yield (P < 0.01), CapP (P < 0.01), and MUN (P < 0.01); the LP diet also increased MPE (P < 0.01) and GPE (P < 0.01), compared to the HP diet. During late lactation, the LP diet decreased milk protein (P < 0.01), CapP (P < 0.01), and MUN (P < 0.01); the LP diet also increased MPE (P < 0.01) and GPE (P = 0.01), compared to the HP diet. In peak lactation, compared to multiparous cows, primiparous cows had less DMI (P < 0.01), milk protein (P < 0.01), CapP (P < 0.01), and MUN (P < 0.01), with similar MPE (P = 0.11) and GPE (P = 0.39). In late lactation, compared to multiparous cows, primiparous cows had higher MPE (P < 0.01) and GPE (P < 0.01), with similar DMI (P = 0.32), milk protein (P = 0.25), CapP (P = 0.55), and MUN (P = 0.28). 87 Parity P-value4 Parity TRT ´ Parity SEM 0.14 0.01 0.01 0.13 0.18 0.32 0.20 0.01 0.02 0.16 0.43 0.56 Treatments3 LP 0.86 0.18 0.86 0.16 0.31 0.25 0.35 0.02 0.03 0.20 0.46 0.48 0.48 0.04 0.04 0.27 1.10 1.11 SEM TRT <0.01 <0.01 <0.01 <0.01 0.11 0.39 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 n=166 23.3 1.07 1.09 9.2 32.4 30.1 HP n=166 24.3 1.21 1.24 15.1 27.6 27.5 Primi. Multi. n=148 n=184 26.3 21.3 1.01 1.28 1.30 1.04 12.7 11.6 29.6 30.3 30.1 29.6 Table 3.1 Dry matter intake, milk production and protein efficiency for cows fed treatment diets in peak and late lactation1,2 Peak-lactation cows DMI, kg/d Milk protein, kg/d Protein capture5, kg/d MUN, mg/dL Milk protein efficiency6, % Gross protein efficiency7, % Late-lactation cows <0.01 DMI, kg/d <0.01 Milk protein, kg/d 0.54 Protein capture, kg/d 0.25 MUN, mg/dL 0.84 Milk protein efficiency, % Gross protein efficiency, % 0.08 1Average DIM was 125 for primiparous cows in HPpeak diet, 126 for primiparous cows in LPpeak diet, 122 for multiparous cows in HPpeak diet, and 121 for multiparous cows in LPpeak diet; average DIM was 258 for primiparous cows in HPlate diet, 257 for primiparous cows in LPlate diet, 263 for multiparous cows in HPlate diet, and 264 for multiparous cows in LPlate diet. 2Average parity for multiparous cows was 2.94 in peak lactation, and 3.12 in late lactation 3 Treatments contained 18% and 14% crude protein on a DM basis for peak lactation cows, and 16% and 13% crude protein on a DM basis for late lactation cows 4 P-value associated with treatment differences (HP vs. LP; TRT) and parity differences (Primi vs. Multi.; Parity). Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 5 Total protein capture = milk protein + body protein, where body protein was estimated as: BodyP = (0.151 – 0.0268 ´ BCS) ´ dEBW, when Parity >1; BodyP = 0.132 ´ 0.14 + ((0.151- 0.0268 ´ BCS) ´ (dEBW – 0.14)), when Parity = 1 6 Milk protein efficiency, milk protein per unit of dietary protein intake. 7 Gross protein efficiency, milk protein and body protein capture per unit of dietary protein intake. n=69 18.4 0.68 0.70 8.1 28.8 27.8 n=84 18.0 0.77 0.80 9.9 29.8 29.0 n=54 20.2 0.72 0.75 10.2 25.0 25.0 n=69 19.8 0.80 0.84 12.1 26.0 26.2 <0.01 <0.01 <0.01 <0.01 <0.01 0.01 0.32 0.25 <0.01 0.28 <0.01 <0.01 88 Repeatability of Protein Efficiency across Dietary Protein Contents and Lactation Stage As illustrated in Figure 3.2, protein efficiency was moderately repeatable across dietary protein contents in peak-lactation cows. The repeatability for MPE, GPE, and MUN across HP and LP diets were 0.72, 0.59, and 0.58, respectively. As illustrated in Figure 3.3, protein efficiency was moderately repeatable across dietary protein contents in late lactation cows. The repeatability for MPE, GPE, and MUN across HP and LP diets were 0.70, 0.69, and 0.57, respectively. As illustrated in Figure 3.4 and Figure 3.5, based on the average value across dietary protein contents and lactation stages, MPE was repeatable across dietary protein contents (r= 0.68, P < 0.05), but not across lactation stages (r= 0.19, P = 0.12). Similar trends were observed for GPE, with GPE repeatability being 0.58 (P < 0.05) across dietary protein contents, and 0.15 (P = 0.21) across lactation stages. In contrast, MUN was repeatable across both dietary protein contents (r= 0.68, P < 0.05) and lactation stages (r= 0.53, P < 0.05). 89 Figure 3.2 Repeatability of protein efficiency across dietary protein contents (HP vs. LP) in peak lactation cows % , t e i d P L r o f E P M 50 40 30 20 10 10 20 60 50 40 30 20 10 % , t e i d P L e o f E P G Primiparous Cows Multiparous Cows 30 40 MPE for HP diet,% 50 Primiparous Cows Multiparous Cows 40 50 60 30 GPE for HP diet,% 10 20 Primiparous Cows Multiparous Cows t e i d P L r o f N U M 20 15 10 5 5 15 10 MUN for HP diet, mg/dL 20 25 Figure 3.2 Repeatability of protein efficiency across dietary protein contents (HP vs. LP) in peak lactation cows (n=166). Repeatability of protein efficiency (MPE, GPE and MUN) across dietary protein contents were 0.72 for MPE (Y = 0.844 (± 0.063; P < 0.01) ´ X + 9.05 (± 1.73; P < 0.01)), 0.59 for GPE (Y = 0.680 (± 0.079; P < 0.01) ´ X + 13.55 (± 2.24; P < 0.01)), and 0.58 for MUN (Y = 0.409 (± 0.053; P < 0.01) ´ X + 3.01 (± 0.81; P < 0.01)). Each data point represents one cow’s protein efficiency value for each diet (n=166). Open circles indicate primiparous cows (n=92), and solid triangles indicate multiparous cows (n=74). Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 90 Figure 3.3 Repeatability of protein efficiency (MPE, GPE, and MUN) across dietary protein contents (HP vs. LP) in late lactation cows % , t e i d P L r o f E P M 45 35 25 15 5 Primiparous Cows Multiparous Cows 5 25 15 35 MPE for HP diet, % 45 % , t e i d P L r o f E P G 45 35 25 15 5 5 Primiparous Cows Multiparous Cows 25 15 35 GPE for HP diet, % 45 Primiparous Cows Multiparous Cows L d / g m , t e i d P L r o f N U M 20 15 10 5 5 10 15 MUN for HP diet, mg/dL 20 Figure 3.3 Repeatability of protein efficiency (MPE, GPE and MUN) across dietary protein contents (HP vs. LP) in late lactation cows (n=69). Repeatability of protein efficiency across dietary protein contents were 0.70 for MPE (Y = 0.784 (± 0.093; P < 0.01) ´ X + 8.56 (± 2.61; P < 0.01)), 0.69 for GPE (Y = 0.633 (± 0.081; P < 0.01) ´ X + 12.67 (± 2.28; P < 0.01)), and 0.57 for MUN (Y = 0.489 (± 0.086; P < 0.01) ´ X + 2.16 (± 1.06; P = 0.04)). Each data point represents one cow’s protein efficiency value for each diet (n=69). Open cycles indicate primiparous cows (n=42), and solid triangles indicate multiparous cows (n=27). Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 91 Figure 3.4 Repeatability of protein efficiency (MPE, GPE, and MUN) for dietary protein contents (HP vs. LP) based on average values across lactation stages % , t e i d P L r o f E P M e v a 35 25 15 15 Primiparous Cows Multiparous Cows 25 35 aveMPE across HP diet, % % , t e i d P L r o f E P G e v a 45 35 25 15 5 5 Primiparous Cows Multiparous Cows 15 35 aveGPE for HP diet, % 25 45 Primiparous Cows Multiparous Cows L d / g m , t e i d P L r o f N U M e v a 15 10 5 5 10 15 aveMUN for HP diet, mg/dL 20 Figure 3.4 Repeatability of protein efficiency (MPE, GPE and MUN) for dietary protein contents (HP vs. LP) based on average values across lactation stage (n=69). Repeatability of protein efficiency across dietary protein contents were 0.68 for MPE (Y = 0.683 (± 0.080; P < 0.01) ´ X + 13.05 (± 2.73; P < 0.01)), 0.58 for GPE (Y = 0.454 (± 0.063; P < 0.01) ´ X + 2.55 (± 1.03; P < 0.01)), and 0.68 for MUN (Y = 0.644 (± 0.073; P < 0.01) ´ X + 12.05 (± 2.37; P < 0.01)). Each data point represents one cow’s protein efficiency value for each diet (n=69). Open circles indicate primiparous cows (n=42), and solid triangles indicate multiparous cows (n=27). aveMPE is the average MPE for the HP diet across lactation stages. aveGPE is the average GPE for the HP diet across lactation stages. aveMUN is the average MUN for the HP diet across lactation stages. Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 92 Figure 3.5 Repeatability of protein efficiency (MPE, GPE, and MUN) across lactation stage based on average values across diets % , n o i t a t c a L e t a L r o f E P M 40 30 20 10 10 Primiparous Cows Multiparous Cows 30 20 40 MPE in Peak Lactation, % % , n o i t a t c a L e t a L r o f E P G L d / g m , n o i t a t c a L e t a L r o f N U M 15 10 5 5 40 30 20 10 10 Primiparous Cows Multiparous Cows 30 20 40 GPE for Peak Lactation, % Primiparous Cows Multiparous Cows 10 20 MUN for Peak Lactation, mg/dL 15 Figure 3.5 Repeatability of protein efficiency (MPE, GPE and MUN) across lactation stage based on average values across diets (n=69). Repeatability of protein efficiency across lactation stage were 0.19 for MPE (Y = 0.307 (± 0.184; P = 0.10) ´ X + 18.82 (± 5.66; P < 0.01)), 0.15 for GPE (Y = 0.224 (± 0.149; P = 0.14) ´ X + 21.56 (± 4.73; P < 0.01)), and 0.53 for MUN (Y = 0.467 (± 0.095; P < 0.01) ´ X + 4.63 (± 1.14; P < 0.01)). Each data point represents one cow’s protein efficiency value for each diet (n=69). Open circles indicate primiparous cows (n=42), and solid triangles indicate multiparous cows (n=27). Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 93 MUN and Protein Efficiency In peak-lactation cows, GPE was highly correlated with MPE in both HP (r = 0.86, P < 0.05) and LP diets (r = 0.91, P < 0.05); MUN was not correlated with MPE or GPE in neither HP nor LP diet (Figure 3.6). In late-lactation cows, GPE was highly correlated with MPE in both HP (r = 0.86, P < 0.05) and LP diets (r = 0.91, P < 0.05); MUN was not correlated with MPE or GPE in neither HP nor LP diet (Figure 3.7). Based on the average value across dietary protein contents and lactation stages, there was no correlation between MUN and MPE regardless of dietary protein contents and lactation stages. Further details are shown in Figure 3.8. Similar trends were found between MUN and GPE. 94 Figure 3.6 Relationship between MUN and MPE/GPE across dietary protein contents (HP vs. LP) in peak lactation cows 50 50 % , t e i d P H r o f E P M % , t e i d P L r o f E P M 40 30 20 10 50 40 30 20 10 % , t e i d P H r o f E P G Primiparous Cows Multiparous Cows 10 25 15 MUN for HP diet, mg/dL 20 % , t e i d P L r o f E P G Primiparous Cows Multiparous Cows 10 15 MUN for LP diet, mg/dL 5 40 30 20 10 50 40 30 20 10 Primiparous Cows Multiparous Cows 10 25 15 MUN for HP diet, mg/dL 20 Primiparous Cows Multiparous Cows 5 10 15 MUN for LP diet, mg/dL Figure 3.6 Relationship between MUN and MPE/GPE across dietary protein contents (HP vs. LP) in peak lactation cows (n=166). The correlation coefficient between MUN and MPE was - 0.05 in HPpeak (Y = 0.204 (± 0.119; P = 0.10) ´ X + 24.39 (± 1.80; P < 0.01)), -0.12 in LPpeak (Y = 0.055 (± 0.063; P = 0.75) ´ X + 31.77 (± 1.63; P < 0.01)). The correlation coefficients between MUN and GPE was -0.15 in HPpeak (Y = 0.105 (± 0.123; P = 0.39) ´ X + 26.58 (± 1.88; P < 0.01)), -0.13 in LPpeak (Y = -0.092 (± 0.193; P = 0.63) ´ X + 33.55 (± 1.77; P < 0.01)). Each data point represents one value (n=166). Open circles indicate primiparous cows (n= 92), and solid triangles indicate multiparous cows (n= 74). Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 95 Figure 3.7 Relationship between MUN and MPE/GPE across dietary protein contents (HP vs. LP) in late lactation cows Primiparous Cows Multiparous Cows 10 15 20 MUN for HP diet, mg/dL Primiparous Cows Multiparous Cows % , t e i d P H r o f E P M % , t e i d P L r o f E P M 45 35 25 15 5 45 35 25 15 5 5 0 Primiparous Cows Multiparous Cows % , t e i d P H r o f E P G 20 10 15 MUN for HP diet, mg/dL Primiparous Cows Multiparous Cows % , t e i d P L r o f E P G 15 5 10 MUN for LP diet, mg/dL 45 35 25 15 5 45 35 25 15 5 5 0 5 10 15 MUN for LP diet, mg/dL Figure 3.7 Relationship between MUN and MPE/GPE across dietary protein contents (HP vs. LP) in late lactation cows (n=69). The correlation coefficient between MUN and MPE was -0.06 in HPlate (Y = - 0.194 (± 0.419; P = 0.64) ´ X + 28.90 (± 5.16; P < 0.01)), -0.13 in LPlate (Y = - 0.581 (± 0.538; P = 0.29) ´ X + 34.05 (± 4.44; P < 0.01)). The correlation coefficients between MUN and GPE was -0.14 in HPlate (Y = -0.581 (± 0.462; P = 0.21) ´ X + 34.63 (± 5.68; P < 0.01)), -0.13 in LPlate (Y = -0.521 (± 0.493; P = 0.29) ´ X + 34.34 (± 4.06; P < 0.01)). Each data point represents one value (n=69). Open circles indicate primiparous cows (n=42), and solid triangles indicate multiparous cows (n=27). Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 96 Figure 3.8 Relationship between MUN and MPE across dietary protein contents (HP vs. LP) and lactation stage Primiparous Cows Multiparous Cows Primiparous Cows Multiparous Cows 20 15 L d / g m , t e i d P H r o f N U M 10 5 L d / g m , t e i d P L r o f N U M L d / g m , n o i t a t c a L e t a L n i N U M 15 10 5 5 15 10 5 5 15 25 35 MPE for HP diet, % L d / g m , n o i t a t c a L k a e P r o f N U M Primiparous Cows Multiparous Cows 20 15 10 5 15 25 35 45 MPE for Peak Lactation, % 15 25 35 MPE for LP diet, % 45 Primiparous Cows Multiparous Cows 15 25 35 45 MPE in Late Lactation, % Figure 3.8 Relationship between MUN and MPE across dietary protein contents (HP vs. LP) and lactation stage (n=69). The correlation coefficient between MUN and MPE was 0.10 for the HP diet (Y = 0.046 (± 0.055; P = 0.41) ´ X + 12.29 (± 1.52; P < 0.01)), 0.06 for the LP diet (Y = 0.018 (± 0.041; P = 0.66) ´ X + 9.09 (± 1.29; P < 0.01)), 0.12 in peak lactation (Y = 0.075 (± 0.056; P = 0.19) ´ X + 9.57 (± 1.75; P < 0.01)), and -0.11 in late lactation (Y = -0.044 (± 0.033; P = 0.19) ´ X + 11.45 (± 0.97; P < 0.01)). Each data point represents one value (n= 69). Open circles indicate primiparous cows (n= 42), and solid triangles indicate multiparous cows (n= 27). Milk protein efficiency (MPE), milk protein per unit of dietary protein intake. Gross protein efficiency (GPE), milk protein and body protein capture per unit of dietary protein intake. 97 Variation of Low Protein Resilience in Lactating Dairy Cows The average dCapP in peak-lactation cows when switched from the HP to LP diet was - 0.15 kg/d, with a standard deviation being - 0.13 kg/d. The final model for LPR in peak-lactation cows was: dCapP = β0 + β1 ´ CapP_MBWHP + Par (Seq ´ Exp) + Seq (Exp) + Exp + e. Among the variables, 31% of the dCapP variation was explained by CapP_MBWHP, with 9%, 7%, and 14% explained by parity, treatment sequence, and experiment, respectively; the remaining 39% was defined as LPR. The average dCapP in late-lactation cows when switched from the HP to LP diet was - 0.15 kg/d, with a standard deviation being - 0.14 kg/d. The final model for LPR in late-lactation cows was: dCapP = β0 + β1 ´ CapP_MBWHP + β2 ´ MBW_HHHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e. Among the variables, 24% of the dCapP variation was explained by CapP_MBWHP, with 7%, 15%, 2%, and 11% explained by MBW_HHHP, parity, treatment sequence, and experiment, respectively; the remaining 40% was defined as LPR. As shown in Tables 3.2 and 3.3, among peak-lactation cows, when fed the HP diet, high- LPR cows had similar DMI, MY, ECM, MBW, BCS, protein intake, milk protein yield, CapP, MPE, and MUN, with lower BodyP and GPE, compared to low-LPR cows. When switched to LP diet, high-LPR cows had higher DMI, MY, ECM, protein intake, milk protein yield, BodyP, CapP, and GPE, compared to low-LPR cows. Among late-lactation cows, when fed the HP diet, high-LPR cows had similar DMI, MY, ECM, BCS, protein intake, milk protein yield, CapP, MPE, GPE, and MUN, with lower dEBW and BodyP, compared to low-LPR cows. When 98 switched to the LP diet, high-LPR cows had higher DMI, MY, ECM, dEBW, protein intake, milk protein yield, BodyP, CapP, MPE, and GPE. 99 Item3 LPR value, kg/d DMI_HP, kg/d DMI_LP, kg/d dDMI, kg/d MY_HP, kg/d MY_LP, kg/d dMY, kg/d ECM4_HP, kg/d ECM_LP, kg/d dECM, kg/d MBW5_HP, kg MBW_LP, kg dMBW, kg BCS_HP BCS_LP dBCS dEBW6 _HP, kg/d dEBW _LP, kg/d ddEBW7, kg/d 0.096a 0.006 24.8a 0.59 24.5a 0.58 -0.48a 0.25 42.3a 1.18 39.8a 1.32 -2.47a 0.50 41.7a 1.17 38.6a 1.03 -2.89a 0.45 129a 1.83 129a 1.79 -0.48a 0.56 3.12a 0.05 3.10a 0.05 -0.02a 0.03 0.16a 0.11 0.49a 0.12 0.33a 0.16 LLPR -0.096c 0.009 19.4a 0.67 17.8ab 0.63 -1.60ab 0.45 26.3a 1.48 18.4b 1.34 -7.88b 0.98 26.2a 1.66 20.7b 1.41 -5.46b 0.94 136a 3.71 136a 3.67 -0.54a 0.63 3.73a 0.10 3.77ab 0.10 0.04b 0.05 0.80ab 0.17 0.09b 0.15 -0.72c 0.24 0.001b 0.006 19.8a 0.43 18.0ab 0.41 -1.80ab 0.29 26.0a 0.96 23.5a 0.88 -2.38a 0.65 27.2a 1.07 24.6a 0.95 -2.38a 0.61 133a 2.39 132a 2.39 -0.98a 0.41 3.62a 0.08 3.46a 0.08 -0.16a 0.04 0.54a 0.12 0.11b 0.10 -0.43b 0.16 Table 3.2 Comparisons of production parameters of high-, medium- and low- LPR cows in peak and late lactation1,2 Peak-lactation Cows N= 166 HLPR MLPR Late-lactation Cows N= 69 HLPR MLPR 0.004b 0.004 23.8a 0.64 22.7b 0.62 -1.21b 0.26 40.9a 0.87 35.3b 0.98 -5.67b 0.37 40.6a 0.87 35.8b 0.76 -3.87b 0.34 128a 1.35 128a 1.32 -0.25a 0.42 3.29a 0.04 3.27a 0.04 -0.02a 0.02 0.43ab 0.09 0.18b 0.09 -0.27b 0.12 LLPR -0.103c 0.006 24.1a 0.44 22.4b 0.43 -2.19c 0.18 41.6a 1.26 35.7b 1.42 -5.44b 0.54 40.9a 1.26 34.9b 1.11 -5.18c 0.49 130a 1.97 129a 1.92 -0.91a 0.61 3.26a 0.05 3.23a 0.05 -0.02a 0.03 0.61b 0.11 -0.32c 0.12 -0.92c 0.12 100 0.082a 0.008 19.9a 0.61 19.2a 0.57 -0.72a 0.41 26.1a 1.35 23.1a 1.23 -2.99a 0.91 26.9a 1.51 24.4a 1.34 -2.53a 0.86 138a 3.37 137a 3.35 -0.74a 0.74 3.78a 0.10 3.62a 0.10 -0.17a 0.05 0.41a 0.16 0.73a 0.13 0.32a 0.22 Table 3.2 (cont’d) Peak-lactation Cows Late-lactation Cows Item3 CapPMBW8_HP CapPMBW_LP dCapPMBW 0.0063a 0.0004 0.0058a 0.0002 -0.0005a 0.0002 LLPR 0.0095a 0.0002 0.0082c 0.0002 -0.0019c 0.0001 N= 166 HLPR MLPR 0.0095a 0.0095a 0.0002 0.0002 0.0077b 0.0092a 0.0002 0.0002 -0.0011b -0.0002a 0.0001 0.0001 N= 69 HLPR MLPR 0.0064a 0.0003 0.0055b 0.0002 -0.0010b 0.0001 LLPR 0.0063a 0.0004 0.0046b 0.0003 -0.0017b 0.0002 1 Cows were grouped into high-, medium-, and low- LPR (low protein resilience) groups. Cows > 0.5 SD of the mean LPR were classified as high LPR, cows < -0.5 SD were classified as low LPR, and those ± 0.5 SD were classified as medium LPR. In peak lactation, LPR was defined as the residual term in the model: dCapP = β0 + β1 ´ CapP_MBWHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where CapP_MBWHP was the CapP per kg metabolic BW when fed the HP; Par was parity (primiparous or multiparous); Seq was treatment sequence (HP-LP or LP-HP); Exp was experiment, and e was the residual term in the model. In late lactation, the LPR model was: dCapP = β0 + β1 ´ CapP_MBWHP + β2 ´ MBW_HHHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where MBW_HHHP was the metabolic body weight to height ratio when fed the HP. 2 Different superscripts indicate significant differences in means (P < 0.05). Superscripts intend to compare variables within lactation stage, not between lactation stages. 3 Upper values within a row = least squares mean (LSM) of the group. Lower values within a row = standard error of the LSM. 4 Energy-corrected milk; ECM = [(0.327 ´ kg milk) + (12.95 ´ kg milk fat) + (7.20 ´ kg milk protein)] (Tyrrell and Reid, 1965). 5 Metabolic BW = BW0.75. 6 Change of EBW(empty BW) per day through the period, determined by linear regression using EBW measurements throughout the period, where EBW was the adjusted BW after accounting for the gut fill effect 7 ddEBW = dEBWLP- dEBWHP 8 Captured protein per unit of MBW, where captured protein= milk protein + protein captured in body tissues 101 Table 3.3 Comparisons of protein output and protein efficiency parameters of high-, medium- and low- LPR cows in peak and late lactation1,2 Peak-lactation Cows N= 166 HLPR MLPR Late-lactation Cows N= 69 HLPR MLPR Item3 LPR value, kg/d CPp4_HP, % CPp_LP, % dCPp, % CPIntake_HP, kg/d CPIntake_LP, kg/d dCPIntake, kg/d MilkP5_HP, kg/d MilkP_LP, kg/d dMilkP, kg/d BodyP6_HP, kg/d BodyP_LP, kg/d dBodyP, kg/d PregP7_HP, kg/d PregP_LP, kg/d dPregP, kg/d CapP8_HP, kg/d CapP_LP, kg/d dCapP, kg/d 0.096a 0.006 17.9a 0.03 14.2a 0.05 -3.71a 0.06 4.45a 0.11 3.48a 0.08 -0.97a 0.047 1.23a 0.035 1.14a 0.031 -0.083a 0.013 0.014a 0.007 0.035a 0.008 0.021a 0.010 0.004b 0.004 17.9a 0.03 14.2a 0.04 -3.74a 0.05 4.28a 0.08 3.23b 0.07 -1.05a 0.035 1.19a 0.026 1.03b 0.023 -0.130b 0.011 0.031b 0.005 0.016b 0.006 -0.014b 0.008 LLPR -0.103c 0.006 17.9a 0.02 14.2a 0.05 -3.73a 0.06 4.38a 0.12 3.18b 0.08 -1.20b 0.051 1.23a 0.038 1.02b 0.033 -0.209c 0.015 0.040bc 0.007 -0.017c 0.008 -0.058c 0.010 1.24a 0.035 1.18a 0.037 -0.062a 0.016 1.20a 0.026 1.05b 0.023 -0.145b 0.012 1.27a 0.038 1.00c 0.033 -0.266c 0.017 102 LLPR -0.096c 0.009 15.8a 0.004 12.9a 0.001 -3.00a 0.001 3.06a 0.10 2.30ab 0.09 -0.77ab 0.073 0.832a 0.049 0.606b 0.043 -0.237b 0.029 0.061b 0.013 0.007b 0.013 -0.055b 0.012 0.013a 0.005 0.005a 0.005 -0.008a 0.007 0.844a 0.053 0.613c 0.039 -0.230b 0.029 0.082a 0.008 15.8a 0.003 12.9a 0.001 -2.93a 0.001 3.15a 0.09 2.50a 0.08 -0.65a 0.067 0.829a 0.045 0.731a 0.039 -0.098a 0.026 0.027a 0.012 0.059a 0.012 0.032a 0.018 0.014a 0.005 0.011a 0.005 -0.002a 0.007 0.823a 0.048 0.790a 0.035 -0.033a 0.027 0.001b 0.006 15.8a 0.003 12.9a 0.001 -2.93a 0.001 3.14a 0.06 2.32a 0.06 -0.82a 0.047 0.849a 0.032 0.712a 0.028 -0.136a 0.019 0.037a 0.008 -0.001b 0.008 -0.038b 0.019 0.009a 0.004 0.009a 0.005 -0.001a 0.005 0.886a 0.034 0.712ab 0.025 -0.174b 0.019 Table 3.3 (cont’d) Peak-lactation Cows Late-lactation Cows Item3 MUN_HP MUN_LP dMUN dGPE, % dMPE, % GPE_LP, % MPE_LP, % MPE9_HP, % GPE10_HP, % 26.4a 1.22 29.4a 1.29 2.98a 0.98 12.52a 0.35 8.80a 0.29 -3.72a 0.31 26.1a 1.35 32.0a 1.16 5.87a 0.86 LLPR 27.9a 0.46 32.1a 0.54 4.05b 0.36 14.9a 0.30 8.97a 0.24 -6.00a 0.27 29.0ab 0.47 31.5c 0.57 2.52c 0.47 N= 166 HLPR MLPR 27.2a 27.6a 0.34 0.43 32.0a 32.8a 0.50 0.37 4.88a 5.23a 0.26 0.35 15.2a 14.9a 0.28 0.21 9.31a 9.15a 0.17 0.22 -5.84a -5.91a 0.20 0.25 27.9a 27.9a 0.32 0.44 34.0a 32.6b 0.39 0.52 4.65ab 6.11a 0.43 0.32 N= 69 HLPR MLPR 26.9a 0.87 30.7a 0.92 3.76a 0.68 11.92a 0.25 8.67a 0.20 -3.27a 0.22 28.2a 0.96 30.7a 0.92 2.47b 0.61 LLPR 26.6a 1.34 26.1b 1.31 -0.49b 1.05 12.41a 0.38 8.44a 0.30 -3.97a 0.33 27.6a 1.48 26.5c 1.26 -1.14c 0.93 1 Cows were grouped into high, medium, and low LPR (low protein resilience) groups. Cows > 0.5 SD of the mean LPR were classified as high LPR, cows < -0.5 SD were classified as low LPR, and those ± 0.5 SD were classified as medium LPR. In peak lactation, LPR was defined as the residual term in the model: dCapPr o= β0 + β1 ´ CapP_MBWHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where CapP_MBWHP was the CapP per kg metabolic BW when fed the HP diet; Par was parity (primiparous or multiparous); Seq was treatment sequence (HP-LP or LP-HP); Exp was experiment, and e was the residual term in the model. In late lactation, the LPR model was: dCapP = β0 + β1 ´ CapP_MBWHP + β2 ´ MBW_HHHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where MBW_HHHP was the metabolic body weight to height ratio when fed HP. 2 Different superscripts indicate significant differences in means (P<0.05). Superscripts intend to compare variables within lactation stage, not between lactation stages. 3 Upper values within a row = least squares mean (LSM) of the group. Lower values within a row = standard error of the LSM. 4 CP% of the diet, kg CP per kg DM 5 Milk protein, kg/d 6 Protein captured in body tissue gain: BodyP = (0.151 – 0.0268 ´ BCS) ´ dEBW, when Parity >1; BodyP = 0.132 ´ 0.14 + ((0.151- 0.0268 ´ BCS) ´(dEBW – 0.14)), when Parity= 1; dEBW= Change of EBW(empty BW) per day through the period, determined by linear regression using EBW measurements throughout the period, where EBW was the adjusted BW after accounting for the gut fill effect. 103 Table 3.3 (cont’d) 7 Protein captured for pregnancy: PregP= 0.00069 ´ D – 0. 0692 ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight (NRC 2001). 8 Total protein captured: Cows in the peak lactation: CapP= Milk Protein+ BodyP; Cows in the late lactation: CapP = Milk Protein+ BodyP+ PregP. 9 Milk protein efficiency, milk protein per unit of dietary protein intake. 10 Gross protein efficiency, milk protein and body protein capture per unit of dietary protein intake. 104 Low Protein Resilience and Protein Efficiency As shown in Table 3. 4, LPR in peak- lactation cows was correlated with MPELP (r = 0.23, P = 0.09), dMPE (r = 0.13, P = 0.09), GPELP (r = 0.37, P < 0.05), and dGPE (r = 0.51, P < 0.05); LPR was not correlated with MPEHP, GPEHP, MUNHP, MUNLP, or dMUN. As shown in Table 3. 4, LPR in late- lactation cows was correlated with MPELP (r = 0.27, P < 0.05), dMPE (r = 0.35, P < 0.05), GPELP (r = 0.48, P < 0.05), and dGPE (r = 0.65, P < 0.05); LPR was not correlated with MPEHP, MUNHP, GPEHP, MUNLP, or dMUN. 105 MPEHP2 MPELP dMPE GPEHP3 GPELP dGPE MUNHP4 MUNLP dMUN 0.09 0.37 0.51 0.04 0.03 -0.01 0.24 < 0.05 < 0.05 0.57 0.66 0.81 0.04 0.48 0.65 0.02 0.01 -0.01 0.70 < 0.05 < 0.05 0.88 0.91 0.95 Table 3.4 Pearson Correlation coefficients between LPR (low protein resilience) and various protein efficiency terms Item1 P Value P Value LPR_Mid r 0.00 0.23 0.13 0.97 0.09 0.09 LPR_Late r 0.01 0.27 0.35 0.91 < 0.05 < 0.05 1 r stands for the Pearson correlation coefficient of the linear relationship between two variables (MPE vs. LPR, GPE vs. LPR, MUN vs. LPR); correlation was considered significant when P < 0.05. 2 MPE (milk protein efficiency) is milk protein per unit of dietary protein intake. MPEHP is the MPE for the HP diet; MPELP is the MPE for the LP diet; dMPE is the difference between MPEHP and MPELP. 3 GPE (gross protein efficiency) is milk protein and body protein captured per unit of dietary protein intake. GPEHP is the gross protein efficiency (GPE) for the HP diet; GPELP is the GPE for the LP diet; dGPE is the difference between GPEHP and GPELP. 4 MUN, milk urea nitrogen. MUNHP is the MUN for the HP diet; MUNLP is the MUN for the LP diet; dMUN is the difference between MUNHP and MUNLP. 106 DISCUSSION In general, the cows that produced less milk and had less total protein capture when fed the HP diet were those that exhibited less of a drop in captured protein when fed the LP diet (Figure 3.1). This was expected, and this decreased response to the LP diet would not in any way indicate that a cow uses protein more efficiently—these cows needed less protein because they produced less. Our hypothesis was that some cows can tolerate lower protein to produce the same amount of milk because they are more efficient metabolically. To identify these more efficient cows, we tried to account for all factors that might explain differences in the protein response (such as CapP per kg MBW); these factors are shown in Table 3. 2 and Table 3. 3. After accounting for all factors not related to protein efficiency per se, LPR accounted for 40% of the variation in cow responses when switched from the HP to LP diet. In the current study, when quantifying the total protein output in lactating dairy cows, we also considered protein captured in body tissues, as indicated by EBW change, in addition to milk protein output. Although EBW change might not be an accurate measure of body protein deposition/mobilization, it can still help provide a more complete measurement of protein capture and thus cows’ resilience to the LP diet. When dietary protein is limited, cows would mobilize body protein to support milk protein production, especially in early lactation (Chilliard and Robelin, 1983). Thus, cows maintaining milk production under the LP diet through excessive body protein mobilization should not be considered as resilient cows. Accordingly, our LPR model included all factors (production level, BW, BCS, DIM, parity, etc.) that may impact the level of protein mobilization and deposition in body tissues, as described by Bauman and Currie (1980) and Komaragiri and Erdman (1997). Furthermore, for late-lactation dairy cows, protein utilized for pregnancy was also included in our model. 107 Significant variation existed among cows in their ability to maintain protein capture in milk and body when fed the LP diet. Having explored an exhaustive list of phenotypic factors that may influence cows’ response to the LP diet, our model explained ~60 % of the variation in dCapP among cows. We hypothesize that some part of LPR may have a genetic basis. Indeed, genetic variation in resilience to various stressors has been well documented. For example, the resilience to mastitis is heritable, and Scandinavian breeding program has been using it since the 1960s (Martin et al., 2018). Al-Kanaan (2016) showed that resilience to heat stress is a heritable trait for cattle. Furthermore, Mulder et al. (2013) suggested that genetic factors explain animals’ resilience to various disturbances (e.g., disease, temperature, social stress). Because heritability studies require large numbers of animals, future collaborations should be encouraged to investigate the heritability of LPR, in order to facilitate potential genetic selection. Low Protein Resilience, Protein Efficiency, and Other Production Traits In this study, LPR was correlated with protein efficiency for the LP diet and also the difference of protein efficiency between HP and LP diets. In other words, cows with higher LPR might not be the most protein-efficient cows when fed the HP diet, but they are the more efficient cows when fed the LP diet. Interestingly, we also found that the high-LPR cows, compared to low-LPR cows, captured less body protein when fed the HP diet, but more when fed the LP diet. These results imply that cows with less body protein deposition in diets containing adequate protein were likely more resilient to diets deficient in protein. Given that milk yield was similar between high- and low- LPR cows (Table 3.2), LPR might be useful to be included in the selection index. In the classic dairy breeding program, milk yield and milk protein yield are the major focuses (Egger-Danner et al., 2015). As a result, modern dairy cows have higher productivity than the cows hundreds of years ago (Rauw and 108 Gomez-Raya, 2015). The tradeoff between maintaining production and improving resilience is being discussed. LPR provides a potential solution to incorporate both production traits and resilience traits into one genetic selection program, and select more resilient ones among high- producing cows. Repeatability of Protein Efficiency across Dietary Protein Contents and Lactation Stage Overall, cows in the current study maintained their protein efficiency rankings across dietary protein contents but did not maintain their protein efficiency rankings across lactation stages. This is consistent with prior work (Zamani et al., 2011) demonstrating low repeatability for protein efficiency across lactation using monthly records on 500 dairy cows (r= 0.12). The low repeatability of protein efficiency across lactation stages could be due to shifts in nutrient partitioning between production and reproduction. In the last weeks of lactation, as pregnancy progressively takes priority over milk production, more protein is utilized to support fetal growth instead of milk protein synthesis (Veerkamp, 1998; Dillon et., 2003; Friggens et al., 2013). Thus, cows producing more milk protein in peak lactation might not be the ones in late lactation. MUN and Protein Efficiency We did not find any correlation between protein efficiency (both MPE and GPE) and MUN of individual cows within diets and lactation stages. That is, for cows fed the same dietary protein in the same lactation stage, those with lower MUN values do not necessarily utilize protein more efficiently. Although Nousiainen et al. (2004) demonstrated negative links between MUN concentration and protein efficiency, they did not remove the influence of dietary protein levels in the model. As dietary protein content has significant influence on MUN and protein efficiency simultaneously (Huhtanen and Hristov, 2009), it is not surprising that MUN could be negatively linked with protein efficiency if diet effect was not removed. Indeed, in the current 109 study, we also found that 1) cows on the HP diet had higher MUN and lower protein efficiency (shown in Table 3.1), and 2) MUN was negatively associated with MPE (rpea k= -0.49, P < 0.05; rlate = -0.27, P < 0.05) and GPE (rpeak = -0.48, P < 0.05; rlate = -0.27, P < 0.05), if the effect of diets was not removed. Thus, the negative link between protein efficiency and MUN in Nousiainen et al. (2004) may be an artifact as a result of the manipulation of dietary protein contents. When accounting for the diet effect, Huhtanen et al. (2015) found a much lower change of protein efficiency per unit change of MUN, compared to Nousiainen et al. (2004). Even when cows were fed the same protein content, MUN concentration can also be affected by many other factors: dehydration (Weeth and Lesperance, 1965; Steiger Burgos et al., 2001), the season of the year, time of sampling (Depeters and Cant, 1992; Broderick and Clayton, 1997; Kauffman and St-Pierre, 2001) and variable transport activities in kidney and rumen wall (Stewart and Smith, 2005; Aguilar et al., 2012). For example, one cow can have a higher MUN concentration merely due to less urea being excreted in urine, thus more urea circulating in the body and excreting in the milk. In addition, according to Kohn et al. (2004), MUN values can be different between regions and milk analysis laboratories if different analysis methods are used. Thus, we suggest that MUN can be a good indicator for protein feeding in daily practice; however, ranking cows for protein efficiency based on MUN may be misleading. CONCLUSIONS In general, when cows were fed the LP diet compared to the HP diet, production decreased and efficiency of protein use increased, and cows maintained their protein efficiency ranking across HP and LP diets within lactation stage. Based on the results that protein efficiency was poorly associated with MUN, using MUN to rank cows for protein efficiency is misleading. Significant variation in response to protein reduction existed among cows with some cows 110 experiencing little drop when fed the LP diet. We could predict the change in captured protein (in milk and body tissue) of individual cows when changing dietary protein content with 60% accuracy based on their performance when fed the HP diet. The remaining 40% could be considered low protein resilience, or LPR. High-LPR cows had similar protein efficiency, as low-LPR cows, when fed the HP diet but higher protein efficiency when fed the LP diet. Given the existing variation among cows, LPR can potentially have a genetic basis. However, more work is needed to examine whether LPR is repeatable across other types of diet changes (for example, other types of base diets and other protein or amino acid supplements) and different lengths of time period (1 wk. vs. 4 wk. vs. 10 wks., etc.). If LPR is repeatable across diets and time, and it is indeed an individual cow trait, further work on a potential genetic basis for LPR would be warranted and would require collaboration among research institutes to collect adequate data. ACKNOWLEDGEMENTS We would like to acknowledge J. S. Liesman and the staff of the Michigan State University Dairy Cattle Teaching and Research Center for their assistance in these experiments, and Landus Cooperative for donating Soyplus soybean meal. This project was supported by Agriculture and Food Research Initiative Competitive Grant no. 2011-68004-30340 from the USDA National Institute of Food and Agriculture and funds from the Michigan Alliance for Animal Agriculture and Michigan AgBioResearch. 111 APPENDIX 112 30 28 26 24 22 20 18 d / g k , I M D y l i a D g k , W B E d e t c i d e r P 600 550 500 450 0 7 14 21 28 35 42 49 56 Experimental Day Cohort 1 Primi. Cohort 2 Primi Cohort 1 Multi. Cohort 2 Multi Predicted EBW vs. Day 0 7 28 21 42 14 Experimental Day 35 49 56 Cohort 1 Primi. Cohort 1 Multi. Cohort 2 Primi. Cohort 2 Multi. Supplementary Figure 3.1 The relationship between DMI, predicted BW, predicted EBW and day in experiment in peak lactation cows DMI vs. Day g k , W B d e t c i d e r P 750 700 650 600 550 Predicted BW vs. Day 0 7 14 21 28 35 42 49 56 Experimental Day Cohort 1 Primi. Cohort 2 Primi Cohort 1 Multi. Cohort 2 Multi Supplementary Figure 3.1 The relationship between DMI, predicted BW, predicted EBW and day in experiment in peak lactation cows (n=166). Solid circles indicate primiparous cows in cohort 1, open circles indicate multiparous cows in cohort 1, solid triangles indicate primiparous cows in cohort 2, and open triangles indicate multiparous cows in cohort 2, where cows in cohort 1 were assigned the treatment sequence from HP to LP while cows in cohort 2 were assigned the treatment sequence from LP to HP. Upton plotting the data of DMI and predicted BW, we could not determine whether the BW change was due to the change of DMI; thus we corrected BW for EBW (empty BW; EBW = BW- 5.2 ´ DMI). In multiparous cows, the EBW change (kg/d) was different from BW change (kg/d). 113 REFERENCES 114 REFERENCES AFRC (Agriculture and Food Research Council). 1992. Technical Committee on Responses to Nutrients. Report No. 9. Nutritive Requirements of Ruminant Animals: Energy. Nutr. Abstr. Rev. B 62:787–835. Aguilar, M., M. D. Hanigan, H. A. Tucker, B. L. Jones, S. K. Garbade, M. L. McGilliard, and R. E. James. 2012. Cow and herd variation in milk urea nitrogen concentrations in lactating dairy cattle1. J. Dairy Sci. 95: 7261-7268. Åhs, F., , T. Furmark, Å. Michelgård, B. Långström, L. Appel, O. T. Wolf, and M. Fredrikson. 2006. Hypothalamic blood flow correlates positively with stress-induced cortisol levels in subjects with social anxiety disorder. Psychosom. Med. 68: 859-862. Al-Kanaan, A. 2016. Heat stress response for physiological traits in dairy and dual purpose cattle populations on phenotypic and genetic scales. PhD thesis, Faculty of Organic Agriculture, University of Kassel, Kassel, Germany. Andrew, S. M., D. R. Waldo, and R. A. Erdman. 1994. Direct analysis of body composition of dairy cows at three physiological stages. J. Dairy Sci. 77:3022–3033. Apelo, S. A., J. R. Knapp, and M. D. Hanigan. 2014. Invited review: Current representation and future trends of predicting amino acid utilization in the lactating dairy cow. J. Dairy Sci. 97: 4000-4017. Barros, T., M. A. Quaassdorff, M. J. Aguerre, J. O. Colmenero, S. J. Bertics, P. M. Crump, and M. A. Wattiaux. 2017. Effects of dietary crude protein concentration on late-lactation dairy cow performance and indicators of nitrogen utilization. J. Dairy Sci.100: 5434-5448. Bauman, D. E., and W. B. Currie. 1980. Partitioning of nutrients during pregnancy and lactation: a review of mechanisms involving homeostasis and homeorhesis. J. Dairy Sci. 63:1514-1529. Broderick, G. A. 2018. Optimizing ruminant conversion of feed protein to human food protein. Animal 12.8: 1722-1734. Broderick, G. A., and M. K. Clayton. 1997. A statistical evaluation of animal and nutritional factors influencing concentrations of milk urea nitrogen1. J. Dairy Sci. 80: 2964-2971. Calus, M. P. L, D. P. Berry, G. Banos, Y. de Haas, and R. F. Veerkamp. 2013. Genomic selection: the option for new robustness traits? Advances in Animal Biosciences. 4: 618–625. 115 Cantalapiedra-Hijar, G., J. L. Peyraud, S. Lemosquet, E. Molina-Alcaide, H. Boudra, P. Noziere, and I. Ortigues-Marty. 2014. Dietary carbohydrate composition modifies the milk N efficiency in late lactation cows fed low crude protein diets. Animal. 8: 275-285. Cheung, K. L., and R. A. Lafayette. 2013. Renal physiology of pregnancy. Adv Chronic Kidney Dis. 20: 209-214. Chilliard, Y., and J. Robelin. 1983. Mobilization of body proteins byearly lactating cows measured by slaughter and D2O techniques. Pages 195–198 in IVth Int. Symp. Protein metabolism and nutrition. EAAP Publication no. 31, vol. 2. EAAP, Rome Italy. Colditz, I. G., and B. C. Hine. 2016. Resilience in farm animals: biology, management, breeding and implications for animal welfare. Anim. Prod. Sci. 56: 1961–1983. DePeters, E. J., and J. D. Ferguson. 1992. Nonprotein nitrogen and protein distribution in the milk of cows. J. Dairy Sci. 75: 3192-3209. Dijkstra, J., C. K. Reynolds, E. Kebreab, A. Bannink, J. L. Ellis, J. France, and A. M. van Vuuren. 2013. Challenges in ruminant nutrition: Towards minimal nitrogen losses in cattle. Pages 47–58 in Energy and Protein Metabolism and Nutrition in Sustainable Animal Production. W. J. Oltjen, E. Kebreab, and H. Lapierre, ed. Wageningen Academic Publishers, Wageningen, the Netherlands. Dillon, P., S. Snijders, F. Buckley, B. Harris, P. O’connor, and J. F. Mee. 2003. A comparison of different dairy cow breeds on a seasonal grass-based system of milk production: 2. Reproduction and survival. Livest Prod Sci. 83: 35-42. Doeschl-Wilson, A. B., B. Villanueva, and I. Kyriazakis. 2012. The first step toward genetic selection for host tolerance to infectious pathogens: obtaining the tolerance phenotype through group estimates. Front. Genet. 3:265. Dunstan, R. H., M. M. Macdonald, G. R. Murphy, B. Thorn, and T. K. Roberts. 2019. Modelling of protein turnover provides insight for metabolic demands on those specific amino acids utilized at disproportionately faster rates than other amino acids. Amino Acids. 1-15. Egger-Danner C, J. B. Cole, J. E. Pryce, N. Gengler, B. Heringstad, A. Bradley, and K. F. Stock. 2015. Invited review: overview of new traits and phenotypin strategies in dairy cattle with a focus on functional traits. Animal 9: 191–207. Elgersma, G. G., G. De Jong, R. Van der Linde, and H. A. Mulder. 2018. Fluctuations in milk yield are heritable and can be used as a resilience indicator to breed healthy cows. J. Dairy. Sci. 101: 1240–1250. 116 Friggens, N. C., L. Brun-Lafleur, P. Faverdin, D. Sauvant, and O. Martin. 2013. Advances in predicting nutrient partitioning in the dairy cow: recognizing the central role of genotype and its expression through time. Animal, 7: 89-101. Gibbs, M. J., W. E. Irvings, M. S. Dhanoa, and J. D. Sutton. 1992. Changes in body components of autumn-calving Holstein Friesian cows over the first 29 weeks of lactation. Anim. Prod. 55:339–360. Gustafsson, A. H., and D. L. Palmquist. 1993. Diurnal variation of rumen ammonia, serum urea, and milk urea in dairy cows at high and low yields1. J. Dairy Sci. 76:475-484. Huhtanen, P., E. H. Cabezas-Garcia, S. J. Krizsan, and K. J. Shingfield. 2015. Evaluation of between-cow variation in milk urea and rumen ammonia nitrogen concentrations and the association with nitrogen utilization and diet digestibility in lactating cows. J. Dairy Sci. 98: 3182-3196. Huhtanen, P., J. I. Nousiainen, M. Rinne, K. Kytölä, and H. Khalili. 2008. Utilization and partition of dietary nitrogen in dairy cows fed grass silage-based diets. J. Dairy Sci. 91: 3589- 3599. Ingvartsen, K. L., and J. B. Andersen. 2000. Integration of metabolism and intake regulation: a review focusing on periparturient animals. J. Dairy Sci. 83: 1573-1597. Kauffman, A. J., and N. R. St-Pierre. 2001. The relationship of milk urea nitrogen to urine nitrogen excretion in Holstein and jersey cows1. J. Dairy Sci. 84: 2284-2294. Knap, P. W. 2005. Breeding robust pigs. Aust. J. Exp. Agric. 45: 763–773. Knot, S.A., L.J. Cummins, F.R. Dunshea, and B.J. Leury. 2008. Rams with poor feed efficiency are highly responsive to an exogenous adrenocorticotropin hormone (ACTH) challenge. Domestic Anim. Endocrinology. 34: 261–268. Komaragiri, M. V., and R. A. Erdman. 1997. Factors affecting body tissue mobilization in early lactation dairy cows. 1. Effect of dietary protein on mobilization of body fat and protein. J. Dairy Sci. 80: 929-937. Kohn, R. A., K. R. French, and E. Russek-Cohen. 2004. A comparison of instruments and laboratories used to measure milk urea nitrogen in bulk-tank milk samples. J. Dairy Sci. 87: 1848-1853. 117 König, S., and K. May. 2019. Invited review: Phenotyping strategies and quantitative-genetic background of resistance, tolerance and resilience associated traits in dairy cattle. Animal 13: 897-908. Leenhouwers, J.I., E.F. Knol, P.N. de Groot, H. Vos, and T. van der Lende. 2002. Fetal development in the pig in relation to genetic merit for piglet survival. J. Anim. Sci. 80: 1759– 1770. M. Steiger Burgos, M. Senn, F. Sutter, M. Kreuzer, and W. Langhans. 2001. Effect of water restriction on feeding and metabolism in dairy cows. Am. J. Physiol. Regul. Integr. Comp. Physiol. 280: 418-427 Martin P, Barkema H. W., Brito L. F., Narayana S. G., and Miglior F. 2018. Symposium review: Novel strategies to genetically improve mastitis resistance in dairy cattle. J. Dairy Sci. 101: 2724–2736. Morme`de, P., A. Foury, E. Terenina, and P.W. Knap. 2010. Breeding for robustness: the role of cortisol. Animal. 5: 651–657. Mulder, H. A., Rönnegård, L., Fikse, W. F., Veerkamp, R. F., and Strandberg, E. 2013. Estimation of genetic variance for macro- and micro-environmental sensitivity using double hierarchical generalized linear models. Genet. Sel. Evol. 45:23. NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Nousiainen, J., K. J. Shingfield, and P. Huhtanen. 2004. Evaluation of milk urea nitrogen as a diagnostic of protein feeding. J. Dairy Sci. 87: 386-398. Ogoh, S., K. Sato, K. Okazaki, , T. Miyamoto, A. Hirasawa, K. Morimoto, and M. Shibasaki. 2013. Blood flow distribution during heat stress: cerebral and systemic blood flow. J. Cerebr. Blood F. Met. 33: 1915-1920. Pryce, J. E. and Y. de Haas. 2017. Genetic selection for dairy cow welfare and resilience to climate change. In Achieving sustainable production of milk. Burleigh Dodds Science Publishing Limited. Rauw, W. M., and L. Gomez-Raja. 2015. Genotype by environment interaction and breeding for robustness in livestock. Front. Genet. 6: 310. 118 Roche, J. R., N. C. Friggens, J. K. Kay, M. W. Fisher, K. J. Stafford, and D. P. Berry. 2009. Invited review: Body condition score and its association with dairy cow productivity, health, and welfare. J. Dairy Sci. 92: 5769-5801. Roseler, D. K., J. D. Ferguson, C. J. Sniffen, and J. Herrema. 1993. Dietary protein degradability effects on plasma and milk urea nitrogen and milk nonprotein nitrogen in Holstein cows. J. Dairy Sci. 76: 525-534. SAS Institute Inc. 2011. SAS/Stat 9.4 User’s Guide. SAS Inst. Inc., Cary, NC. Shaver, R. D. 2010. Diets fed in selected WI high-producing dairy herds. Retrieved on 15 January 2017 from http://shaverlab.dysci.wisc.edu/wp-contents/uploads/ sites/87/2015/04/2010wihigh-producingherds.pdf. Shirali, M., P. F. Varley, and J. Jensen. 2018. Bayesian estimation of direct and correlated responses to selection on linear or ratio expressions of feed efficiency in pigs. Genet Select Evol. 50: 33. Sinclair, K. D., P. C. Garnsworthy, G. E. Mann, and L. A. Sinclair. 2014. Reducing dietary protein in dairy cow diets: implications for nitrogen utilization, milk production, welfare and fertility. Animal. 8: 262-274. Stewart, G. S., and C. P. Smith. 2005. Urea nitrogen salvage mechanisms and their relevance to ruminants, non-ruminants and man. Nutr. Res. Rev. 18: 49-62. Stoop, W. M., H. Bovenhuis, and J. A. M. Van Arendonk. 2007. Genetic parameters for milk urea nitrogen in relation to milk production traits. J. Dairy Sci. 90: 1981-1986. Vanrobays, M. L., H. Hammami, H. Soyeurt, J. Vandenplas, E. Froidmont, and N. Gengler. 2015. Assessing resilience of dairy cattle by studying impact of heat stress on predicted feed intake. In Proceedings of the Third Dairycare Conference 2015 (pp. 49-49). DairyCare COST Action FA1308. Veerkamp, R. F. 1998. Selection for economic efficiency of dairy cattle using information on live weight and feed intake: a review. J. Dairy Sci. 81:1109-1119. Wattiaux, M. A., E. V. Nordheim, and P. Crump. 2005. Statistical evaluation of factors and interactions affecting dairy herd improvement milk urea nitrogen in commercial Midwest dairy herds. J. Dairy Sci. 88: 3020-3035. Weeth, H. J., and A. L. Lesperance. 1965. Renal Function of Cattle under Various Water and Salt Loads 1, 2. J. Anim. Sci. 24: 441-447. 119 Zamani, P., S. R. Miraei Ashtiani, D. Alipour, M. Tabatabaei, H. Aliarabi, A. A. Saki, and A. Abdolmohammadi. 2011. Statistical analysis of some factors affecting crude protein balance in lactating dairy cows. J. Agr. Sci. Tech. 13: 1033-1043. Zetouni, L., M. Henryon, M.Kargo and J. Lassen. 2017. Direct multitrait selection realizes the highest genetic response for ratio traits. J. Animal Sci. 95(5): 1921-1925. 120 CHAPTER 4 ASSOCIATION AMONG DIGESTIBILITY, RESIDUAL FEED INTAKE AND LOW PROTEIN RESILIENCE IN LACTATING DAIRY COWS FED HIGH AND LOW PROTEIN DIETS ABSTRACT Our objective was to determine whether variation in total tract digestibility could account for the variation in residual feed intake (RFI) and low protein resilience (LPR) in lactating dairy cows. Lactating Holstein cows (n = 166; 92 primiparous, 77 multiparous) with initial milk yield (MY) 41.3 ± 9.8 kg/d were included in the crossover experiments with two treatments (high- protein diets, HP; low-protein diets, LP) and two periods of 28- 35 d each. Experiments were repeated in 69 of the 166 cows (42 primiparous, 27 multiparous) in late lactation. Low-protein diets were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function. Expeller soybean meal was added in place of corn and soyhulls to create high-protein diets, which were 18% CP in peak lactation and 16% CP in late lactation. Cows were milked 2 times daily; DMI and MY were recorded daily. Milk composition was measured over 4 consecutive milkings weekly, and BW was measured 3 times weekly. Samples of feed ingredients, orts and feces were collected in the last 5 days of each treatment period and analyzed to determine the digestibilities of DM, NDF, and CP for each cow on each diet. RFI was calculated for each cow on each diet based on the actual intake, milk energy output, metabolic BW, and retained body energy (calculated from BW change and BCS over the treatment period). LPR was calculated for each cow in each lactation stage based on the captured protein (milk protein + body protein gain) and feed intake. Neither RFI nor LPR was correlated with digestibilities of DM, NDF, or CP in either diet or lactation stage. The changes in 121 digestibilities of DM, NDF, and CP from the HP to LP diet did not account for LPR. In conclusion, variation in digestibility among cows could not explain the variations in RFI or LPR; we suggest that post-absorptive metabolism explains most of the variation in RFI and LPR when lactating cows are fed diets with minimal NDF in peak lactation and 40% NDF in late lactation. 122 INTRODUCTION Variation in residual feed intake (RFI) of lactating cows is well documented (Tempelman et al., 2015; VandeHaar et al., 2016). Variability in digestion is one factor that has been investigated to explain this variation. Richardson and Herd (2004) found that digestive variability accounted for 10% of the RFI variation in finishing beef steers. Consistently, Nkrumah et al. (2006) and Rius et al. (2012) found that growing heifers with low RFI values had better nutrient digestibility than those with high RFI values. In contrast with those results, no relationship was found between RFI and nutrient digestibility by Cruz et al. (2010) and Lawrence et al. (2011). Given different diets being fed across those studies, Rius et al. (2012) raised the possibility that the inconsistent findings among studies perhaps are due to differences in diets of these studies. Following this idea, Potts et al. (2017) examined the association of RFI and nutrient digestibility in lactating dairy cows across high- and low-starch diets. In their study, RFI was only correlated with DM digestibility in the low-starch diet but not in the high-starch diet. Whether differences in other nutrients (e.g., CP) also influence the possible correlation of digestibility with RFI is not clear. Some cows are more resilient to low-protein diets, in other words, they are able to continue to produce normal quantities of milk per unit of BW when fed diets that have insufficient protein for the average cow, even after adjusting for all factors that would be expected to alter protein requirements. The underlying mechanisms for low protein resilience (LPR) are not clear, but the resilience could be due to a better ability to efficiently digest foods in the face of a low-protein diet or to have more efficient post-absorptive metabolisms. Thus, the objective of the current study was to determine whether variation in total tract digestibility could account for the variation in RFI and LPR in lactating dairy cows. We hypothesized that 1) 123 digestive efficiency would account for some of the variation in RFI and this relationship would be altered by dietary protein content, and 2) cows with higher LPR value would have greater digestibility when fed low-protein diets. MATERIALS AND METHODS Cows, Experimental Design, and Diets Experimental procedures were approved by the Institutional Animal Care and Use Committee of Michigan State University. Data from 166 lactating Holstein dairy cows were used in this study. Data of MY, milk components, BW, and hip height were collected in the experiments. These are the same animal as in Liu and VandeHaar (2020). In brief, lactating Holstein cows (n= 166; 92 primiparous, 77 multiparous) with initial milk yield (MY) 41.3 ± 9.8 kg/d were included in the crossover experiments with two treatments (high-protein diets, HP; low-protein diets, LP) and two periods of 28-35 d each. Experiments were repeated in 69 of the 166 cows (42 primiparous, 27 multiparous) in late lactation. Low-protein diets were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function. Expeller soybean meal was added in place of corn and soyhulls to create high-protein diets, which were 18% CP in peak lactation and 16% CP in late lactation. Cows were milked 2 times daily; DMI and MY were recorded daily. Milk composition was measured over 4 consecutive milkings weekly, and BW was measured 3 times weekly. 124 Sample Collection and Analyses Collection and analyses of diet ingredients, orts, and fecal samples followed similar procedures in all the experiments. During the last 5 days of experimental periods, samples of feed ingredients (0.5 kg) and orts (12.5% of the amount) were obtained daily to determine the nutrient profile of the diets. Samples of feces were collected every 15 h in the last 5 d to obtain 8 samples per cow to represent every 3 h of a day to account for variations. All samples were frozen after collection until analysis. The reported nutrient and ingredient composition of diets were calculated by averaging across both periods for each experiment. Samples of feed ingredients, orts and feces were analyzed for CP, NDF, indigestible NDF, and ash. Crude protein was determined according to Hach et al. (1987). Neutral detergent fiber was determined according to Mertens (2002). Indigestible NDF, which was used as an internal marker to estimate fecal output and nutrient digestibility (Cochran et al., 1986), was estimated as NDF residue after 240 h of in vitro fermentation (Goering and VanSoest, 1970); flasks were re-inoculated at 120 h to ensure a viable microbial population. Rumen fluid for the in vitro incubations was collected from a cow fed only dry hay. Ash was determined after 5 h combustion at 500°C. Calculations and Statistical Analyses Milk energy output (MilkE; Mcal/d) for individual cows was estimated by the following equation (NRC, 2001; Equation 2-15): MilkE = [9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg)], where each component was calculated as the average output of individual cows during the treatment period. 125 For cows > 190 d pregnant, BW was corrected for conceptus weight (CW) for use in the RFI equation and to calculate energy and protein change of body tissues. Conceptus weight was calculated using the equation from NRC (2001), CW = [18 + (D - 190) x 0.665] ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Metabolic BW (MBW) of a cow was estimated as BW0.75, where BW was the mean measured BW for the cow during the treatment period. Empty BW (EBW) was calculated for each cow to adjust BW for the gut fill (Gibbs et al., 1992; Andrew et al., 1994), EBW = BW – 5.2 ´ DMI- CW, where DMI was the daily DMI when BW was measured. Mean daily EBW change (dEBW; kg/d) was calculated for each cow within the treatment period by linear regression after two rounds of removing outliers in the data; an outlier was any BW > 3.5 SD from the regression line. For multiparous cows, EBW change was considered to be all body condition, and the body energy gained or lost with changes in EBW (BodyE; Mcal/d) was estimated by the equation as: BodyE = RE (Mcal/kg) ´ dEBW, where RE= 3.52 + 1.27 ´ BCS (derived from NRC 2001, Table 2-4). For primiparous cows, we assumed their mature BW would be 700 kg and that they had to gain 0.14 kg EBW/d of true growth across the first lactation to reach 92% of mature BW by their second calving (NRC, 2001). Based on the NRC (2001) equations (11-1 and 11-2), the RE content of true growth is 4.4 Mcal/kg dEBW. Any deviation in dEBW from 0.14 kg/d was 126 considered to be body condition gain or loss, and the dEBW associated with body condition change was the same as for multiparous cows (3.52 + 1.27 ´ BCS) Mcal/kg dEBW. Thus, the equation to calculate BodyE was: BodyE=' (3.52 + 1.27 ´ BCS) ´ dEBW ,Parity>1 4.4 ´ 0.14+[(3.52 + 1.27 ´ BCS) ´ (dEBW−0.14)] ,Parity=1 Energy expended for pregnancy (PregE; Mcal/d) was estimated using the equation from NRC (2001; Equation 2-19): PregE = [(0.00318 ´ D - 0.0352) ´ (CBW/45)] / 0.218, where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Apparent diet net energy content (DietNEL; Mcal/kg) was calculated for each diet as the average NEL required by each cow for maintenance, milk, pregnancy, and body gain divided by her average daily intake for the diet: DietNEL = Average [(MilkE + 0.08 ´ MBW + BodyE + PregE) / DMI], where DMI was the average DMI for a cow during each treatment period. Under the similar assumptions and information used in NRC energy calculation (2001; Table 2-4, Equations 11-4 and 11-5), protein captured for body tissue gain (BodyP; kg/d) was calculated from dEBW and BCS as: BodyP (0.151− 0.0268 ´ BCS) ´ dEBW , =' 0.132 ´ 0.14+[ (0.151−0.0268 ´ BCS) ´ (dEBW−0.14)] , Parity>1 Parity=1 127 where (0.151- 0.0268 ´ BCS) kg protein per kg dEBW was assumed when dEBW was considered as body condition gain or loss, and 0.132 kg protein per kg dEBW was considered for the 0.14 kg/d growth. Protein captured for pregnancy (PregP; kg/d) was calculated using the equation from NRC (2001): PregP = 0.00069 ´ D – 0. 0692 ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Total protein capture (CapP, kg/d) was estimated for each cow in each treatment (HP and LP) as: CapP=' Milk Protein+BodyP ,DIM<200 Milk Protein+BodyP+PregP ,DIM≥200 Residual feed intake was calculated similar to the residual term in the prediction of DMI, as previously described in Liu and VandeHaar (2020). Intake for an individual cow during each period was regressed as a function of major energy sinks using GLM procedure in SAS (9.4): DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ BodyE + β4 ´ DIM+ β5 ´ DIM ´ DIM + Parity + Experiment + Cohort (Experiment)+ Diet (Cohort ´ Experiment)+ e, where DMI was the observed DMI, MilkE was the observed milk energy output, MBW was the average BW0.75, BodyE was the predicted change in body energy based on dEBW and BCS, DIM was the average DIM during each treatment period, and BCS was the average BCS during each treatment period; parity (1 or 2+), experiment (1-7), cohort nested within experiment, and diet nested within cohort and experiment were fixed effects, where a cohort is a group of cows that ate the same diet at the same time. RFI was defined as the residual term (e) in the model. For cows in the late lactation, a term for pregnancy energy was included. 128 Low protein resilience was calculated as described by Liu and VandeHaar (2019, submitted) as: dCapP = β0 + β1 ´ CapP_MBWHP + β2 ´ BCSHP + β3 ´ MBW_HHHP + β4 ´ dCP + β5 ´ CPHP + β6 ´ DIMHP +Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where dCapP was the change of captured protein (milk protein, body protein and pregnancy protein) when switched from the HP to LP diet, CapP_MBWHP was the CapP per kg metabolic BW when fed the HP diet; BCSHP was the BCS when fed the HP diet; MBW_HHHP was the metabolic body weight to height ratio when fed the HP diet; Par was parity (primiparous or multiparous); DIMHP was the starting days in milk when fed the HP diet; Seq was sequence (HP- LP or LP-HP); CPHP was the actual CP% in the HP diet; dCP was the actual CP% change from HP to LP; Exp was experiment, and e was the residual term in the model. LPR was defined as the residual term. Cows were grouped into high (HRFI), medium (MRFI), and low (LRFI) RFI groups. Cows > 0.5 SD of the mean RFI were classified as HRFI group, cows < -0.5 SD were classified as LRFI, and those ± 0.5 SD were classified as MRFI. Cows were also grouped into high (HLPR), medium (MLPR), and low (LLPR) LPR groups based on similar criteria. Comparison among different groups was performed. The effect of efficiency group (RFI or LPR) was determined using the GLM Procedure of SAS according to the model Yi = μ + Ri + e, where μ was the overall mean, Ri was the fixed effect of efficiency group, and e was the residual error. For each cow on each diet within each lactation stage, protein efficiency was calculated as dietary protein captured in milk protein (milk protein efficiency, MPE). 129 To quantify the association among RFI, LPR, MPE and nutrient (CP, NDF and DM) digestibility within each diet, Pearson correlations were obtained using the GLM procedure in SAS (9.4) after accounting for effects of parity, treatment sequence, and experiment. Production responses to diets and digestibility difference within each lactation stage were analyzed using the HPMIXED procedure in SAS (9.4), with fixed effects of diet, parity, treatment sequence nested in experiment, period within experiment, interaction of parity and diet, and the random effects of experiment and cow nested within experiment. Significance was considered at P (cid:1) 0.05 and tendency at P (cid:1) 0.10. Interactions were considered significant at P (cid:1) 0.10 and trends at P (cid:1) 0.15. Cow Performance for High- and Low- Protein Diets RESULTS As shown in Table 4.1, in peak-lactation cows, compared to the LP diet, the HP diet increased apparent NEL intake, BodyE, and MilkE; HP also increased milk protein yield and MUN concentration, and decreased MPE. Compared to the LP diet, the HP diet increased digestibilities of DM, NDF, and CP by 2.8, 2.8, and 6.2 percentage units (P < 0.05), respectively. As shown in Table 4.2, in late-lactation cows, compared to the LP diet, the HP diet increased apparent NEL intake, BodyE, and MilkE; HP also increased milk protein yield and MUN concentration, and decreased MPE. Compared to the LP diet, the HP diet increased digestibilities of DM, NDF, and CP by 2.0, 1.8, and 7.4 percentage units (P < 0.05), respectively. 130 Table 4.1 Energy output, protein efficiency and digestibility for cows fed treatment diets in peak lactation1,2 Treatments3 SEM Parity SEM TRT P-value4 HPpeak n=166 LPpeak n=166 Primi. n=184 Multi. n=148 Parity TRT ´ Parity 0.03 0.57 0.75 0.10 0.02 0.01 0.20 0.46 44.4 29.9 3.49 11.0 36.7 25.1 1.34 10.3 0.51 0.15 0.30 0.10 1.28 0.01 12.7 30.3 1.01 0.03 11.6 29.6 1.07 0.01 9.2 32.4 42.3 28.1 3.81 10.4 1.21 0.03 15.1 27.6 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 34.6 23.4 1.66 9.6 0.03 0.18 0.71 0.01 0.01 0.01 0.13 0.18 0.42 0.58 0.49 Calculated Energy Values Apparent NEL5, Mcal/d Milk, Mcal/d Body Tissue Gain6, Mcal/d Maintenance7, Mcal/d N Metabolism Milk Protein, kg/d Body Protein8, kg/d MUN, mg/dL MPE9, % Nutrient Digestibilities 0.23 DM, % 0.70 NDF, % 0.09 CP, % 1Average DIM was 125 for primiparous cows in HPpeak diet, 126 for primiparous cows in LPpeak diet, 122 for multiparous cows in HPpeak diet and 121 for multiparous cows in LPpeak diet. 2Average parity for multiparous cows was 2.94 in peak lactation 3 Treatments contained 18% and 14% crude protein on a DM basis for peak lactation cows 4 P-value associated with treatment differences (HPpeak vs. LPpeak; TRT) and parity differences (Primi vs. Multi.; Parity) in peak lactation cows. Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 5 NEL= MilkE + 0.08 ´ MBW + BodyE, where MilkE is net energy utilized for milk synthesis, MBW is metabolic body weight, and BodyE is estimated energy captured in body tissue. 6 BodyE = (3.52 + 1.27 ´ BCS) ´ dEBW, when Parity >1; BodyE = 4.4 ´ 0.14 + ((3.52 + 1.27 ´ BCS) ´ (dEBW – 0.14)), when Parity= 1. dEBW is the estimated change of EBW(empty BW) per day through the period, determined by linear regression using EBW measurements throughout the period, where EBW was the adjusted BW after accounting for the gut fill effect. 7 Maintenance energy = 0.08 ´ MBW, where MBW= BW0.75 <0.01 <0.01 <0.01 <0.01 <0.01 0.04 <0.01 0.11 0.18 0.36 0.16 0.31 <0.01 <0.01 <0.01 65.7 48.3 69.4 64.2 46.3 66.4 64.5 47.5 66.2 0.54 0.06 0.78 62.9 45.5 63.2 0.43 0.60 0.50 131 Table 4.1 (cont’d) 8 Protein captured in body tissue gain: BodyP = (0.151 – 0.0268 ´ BCS) ´ dEBW, when Parity >1; BodyP = 0.132 ´ 0.14 + ((0.151- 0.0268 ´ BCS) ´(dEBW – 0.14)), when Parity = 1. 9 MPE, milk protein efficiency, defined as the dietary protein captured in milk. 132 Table 4.2 Energy output, protein efficiency and digestibility for cows fed treatment diets in late lactation1,2 Treatments3 LPlate n=69 HPlate n=69 Parity Primi. n=84 Multi. n=54 SEM TRT P-value4 Parity TRT ´ Parity Calculated energy values Apparent NEL5, Mcal/d Milk, Mcal/d Body Tissue Gain6, Mcal/d Maintenance7, Mcal/d Pregnancy8, Mcal/d N Metabolism Milk Protein, kg/d Body Protein9, kg/d MUN, mg/dL MPE10, % Nutrient Digestibilities DM, % NDF, % CP, % 33.3 18.4 3.81 10.8 0.34 0.80 0.04 12.1 26.0 69.8 57.6 74.7 28.3 16.0 1.34 10.6 0.37 0.58 0.02 8.1 28.8 67.8 55.8 67.3 SEM 0.77 0.31 0.71 0.06 0.11 0.01 0.01 0.16 0.43 0.48 0.90 0.59 29.9 18.1 1.66 9.9 0.27 0.77 0.03 9.9 29.8 69.2 57.0 70.9 31.8 16.3 3.50 11.5 0.45 0.72 0.03 10.2 25.0 68.5 55.4 71.0 1.01 1.04 0.75 0.18 0.13 0.04 0.01 0.27 1.10 0.50 0.81 0.62 <0.01 <0.01 <0.01 <0.01 0.77 <0.01 0.03 <0.01 <0.01 <0.01 0.09 <0.01 <0.01 0.17 0.25 0.71 0.28 <0.01 <0.01 0.09 <0.01 0.12 0.06 0.88 0.12 0.02 0.32 0.09 0.72 <0.01 0.21 0.25 0.84 0.91 0.33 0.37 1Average DIM was 258 for primiparous cows in HPlate diet, 257 for primiparous cows in LPlate diet, 263 for multiparous cows in HPlate diet, and 264 for multiparous cows in LPlate diet. 2Average parity for multiparous cows was 3.12 in late lactation 3 Treatments contained 16% and 13% crude protein on a DM basis for late lactation cows 4 P-value associated with treatment differences (HPlate vs. LPlate; TRT) and parity differences (Primi vs. Multi.; Parity) in late lactation cows. Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 5 NEL= MilkE + 0.08 ´ MBW + BodyE, where MilkE is net energy utilized for milk synthesis, MBW is metabolic body weight, and BodyE is estimated energy captured in body tissue. 6 BodyE = (3.52 + 1.27 ´ BCS) ´ dEBW, when Parity >1; BodyE = 4.4 ´ 0.14 + ((3.52 + 1.27 ´ BCS) ´ (dEBW – 0.14)), when Parity= 1. dEBW is the estimated change of EBW(empty BW) per day through the period, determined by linear regression using EBW measurements throughout the period, where EBW was the adjusted BW after accounting for the gut fill effect. 7 Maintenance energy = 0.08 ´ MBW, where MBW = BW0.75 133 Table 4.2 (cont’d) 8 Energy expended for pregnancy = [(0.00318 ´ D - 0.0352) ´ (CBW/45)] / 0.218, where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. 9 Protein captured in body tissue gain: BodyP = (0.151 – 0.0268 ´ BCS) ´ dEBW, when Parity >1; BodyP = 0.132 ´ 0.14 + ((0.151- 0.0268 ´ BCS) ´(dEBW – 0.14)), when Parity = 1. 10 MPE, milk protein efficiency, defined as the dietary protein captured in milk 134 Residual Feed Intake and Digestibility The relationships between RFI and digestibilities of DM, NDF, and CP are illustrated in Table 4.3 and Table 4.4. For both peak- and late- lactation cows, RFI was not correlated with digestibilities of DM, NDF, or CP, regardless of diets. Based on the comparisons between LRFI cows and HRFI cows in Table 4.5, regardless of diets and lactation stages, cows with lower RFI values did not have greater digestibilities of DM, NDF, or CP, except that LRFI cows tended to have a higher DM digestibility in peak lactation. 135 Table 4.3. Correlation coefficients of RFI, LPR with digestibilities of DM, CP, and NDF in peak- lactation cows (n=166) Item1 DMdHP2 NDFdHP3 CPdHP4 DMdLP5 NDFdLP6 CPdLP7 dDMd8 dNDFd9 dCPd10 RFIHP11 RFILP12 LPR13 MPEHP14 MPELP15 -0.10 0.19 0.02 0.85 0.01 0.86 -0.07 0.36 -0.01 0.88 -0.08 0.30 -0.14 0.11 0.00 0.96 0.15 0.08 -0.06 0.46 0.14 0.11 -0.13 0.12 -0.07 0.38 0.06 0.42 -0.12 0.14 -0.05 0.56 0.13 0.12 -0.14 0.11 -0.08 0.30 -0.06 0.49 -0.08 0.30 1 Upper values within a row = Pearson correlation coefficient of the linear relationship between 2 variables. Lower values within a row = P-value associated with the linear relationship between 2 variables. LP, low- protein diets; HP: high-protein diets 2 DM digestibility for HP 3 NDF digestibility for HP 4 CP digestibility for HP 5 DM digestibility for LP 6 NDF digestibility for LP 7 CP digestibility for LP 8 Change of DM digestibility from HP to LP 9 Change of NDF digestibility from HP to LP 10 Change of CP digestibility from HP to LP 136 Table 4.3 (cont’d) 11 Residual feed intake (RFI) in cows fed the HP diet, where RFI was defined as the residual term in the model: DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ BodyE + β4 ´ PregE + β5 ´ DIM + β6 ´ BCS+ Parity + Experiment + Cohort(Experiment) + Diet(Cohort ´ Experiment)+e. DMI was the observed DMI, MilkE was the observed milk energy output, MBW was the average BW0.75, BodyE was the predicted change in body energy based on measured BW and BCS, DIM was the average DIM during each treatment period, and BCS was the average BCS during each treatment period; parity (1 or 2+), experiment (1-7), cohort nested within experiment, and diet nested within cohort and experiment were fixed effects, where a cohort is a group of cows that ate the same diet at the same time. 12 Residual feed intake in cows fed LP 13 Low protein resilience (LPR) was defined as the residual term in the model: dCapP = β0 + β1 ´ CapP_MBWHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where CapP_MBWHP was the CapP per kg metabolic BW when fed HP; Par was parity (primiparous or multiparous); Seq was treatment sequence (HP-LP or LP-HP); Exp was experiment, and e was the residual term in the model. 14 Milk protein efficiency in cows fed HP, defined as the dietary protein captured in milk 15 Milk protein efficiency in cows fed LP, defined as the dietary protein captured in milk 137 Table 4.4. Correlation coefficients of RFI, LPR with digestibilities of DM, CP, and NDF in late- lactation cows (n= 69) Item1 DMdHP2 NDFdHP3 CPdHP4 DMdLP5 NDFdLP6 CPdLP7 dDMd8 dNDFd9 dCPd10 RFIHP11 RFILP12 LPR13 MPEHP14 MPELP15 -0.11 0.38 0.01 0.91 -0.13 0.37 -0.17 0.18 0.08 0.57 0.09 0.50 -0.07 0.54 0.03 0.78 -0.14 0.35 0.05 0.66 0.08 0.56 -0.04 0.71 0.06 0.63 0.00 0.98 -0.02 0.88 0.15 0.25 0.11 0.41 -0.03 0.77 0.07 0.60 0.05 0.72 0.06 0.63 1 Upper values within a row = Pearson correlation coefficient of the linear relationship between 2 variables. Lower values within a row = P-value associated with the linear relationship between 2 variables. LP, low- protein diets; HP: high-protein diets. 2 DM digestibility for HP 3 NDF digestibility for HP 4 CP digestibility for HP 5 DM digestibility for LP 6 NDF digestibility for LP 7 CP digestibility for LP 8 Change of DM digestibility from HP to LP 9 Change of NDF digestibility from HP to LP 10 Change of CP digestibility from HP to LP 138 Table 4.4 (cont’d) 11 Residual feed intake (RFI) in cows fed HP, where RFI was defined as the residual term in the model: DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ BodyE + β4 ´ PregE + β5 ´ DIM + β6 ´ BCS+ Parity + Experiment + Cohort(Experiment) + Diet(Cohort ´ Experiment)+e. DMI was the observed DMI, MilkE was the observed milk energy output, MBW was the average BW0.75, BodyE was the predicted change in body energy based on measured BW and BCS, DIM was the average DIM during each treatment period, and BCS was the average BCS during each treatment period; parity (1 or 2+), experiment (1-7), cohort nested within experiment, and diet nested within cohort and experiment were fixed effects, where a cohort is a group of cows that ate the same diet at the same time. 12 Residual feed intake in cows fed LP 13 Low protein resilience (LPR) was defined as the residual term in the model: dCapP=β0 + β1 ´ CapP_MBWHP + β2 ´ MBW_HHHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where CapP_MBWHP was the CapP per kg metabolic BW when fed HP; MBW_HtHP was the metabolic body weight to height ratio when fed HP; Par was parity (primiparous or multiparous); Seq was treatment sequence (HP-LP or LP-HP); Exp was experiment, and e was the residual term in the model. 14 Milk protein efficiency in cows fed HP, defined as the dietary protein captured in milk 15 Milk protein efficiency in cows fed LP, defined as the dietary protein captured in milk 139 Table 4.5 Nutrient digestibility for high-, medium- and low-RFI cows fed high- and low- protein diets in peak and late lactation1,2,3,4 Low Protein diets HRFI npeak= 49 60.8a 61.5a 44.2a nlate= 22 69.5a 69.2a 57.8a MRFI npeak= 73 63.1ab 63.4a 46.9a nlate= 26 67.6ab 66.7ab 55.5a LRFI npeak= 44 63.0ab 62.9a 46.1a nlate= 21 68.2a 67.3a 56.3a Variable Peak lactation DM, % CP, % NDF, % Late lactation DM, % CP, % NDF, % High-protein diets HRFI npeak= 47 MRFI npeak= 80 LRFI npeak= 39 64.5a 67.8a 48.4a nlate= 19 70.1a 74.8a 56.6a 65.2a 68.8a 48.9a nlate= 28 71.1a 76.1ab 58.7a 65.2a 69.3a 48.5a nlate= 22 70.0a 74.5a 57.2a 1 In peak lactation cows, residual feed intake (RFI) was defined as the residual term in the model: DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ BodyE + β4 ´ DIM+ β5 ´ DIM ´ DIM + β6 ´ BCS + Parity + Experiment + Cohort(Experiment)+ Diet(Cohort ´ Experiment)+ e, where DMI was the observed DMI, MilkE was the observed milk energy output, MBW was the average BW0.75, BodyE was the predicted change in body energy based on measured BW and BCS, DIM was the average DIM during each treatment period, and BCS was the average BCS during each treatment period; parity (1 or 2+), experiment (1-7), cohort nested within experiment, and diet nested within cohort and experiment were fixed effects, where a cohort is a group of cows that ate the same diet at the same time. 2 In late lactation cows, RFI was defined as the residual term in the model: DMI = β0 + β1 ´ MilkE + β2 ´ MBW + β3 ´ BodyE + β4 ´ PregE + β5 ´ DIM + β6 ´ BCS+ Parity + Experiment + Cohort(Experiment) + Diet(Cohort ´ Experiment)+e. 3 Cows were grouped into high (HRFI), medium (MRFI), and low (LRFI) RFI groups. Cows > 0.5 SD of the mean RFI for a cohort were classified as HRFI, cows < -0.5 SD were classified as LRFI, and those ± 0.5 SD were classified as MRFI. 4 P-values associated with group difference were all >0.05. 140 Low Protein Resilience and Digestibility The relationships between LPR and digestibilities of DM, NDF, and CP are illustrated in Table 4.3 and Table 4.4. For both peak- and late- lactation cows, LPR was not correlated with digestibilities of DM, NDF, or CP. In addition, LPR was also not correlated with the change in digestibilities of DM, NDF, or CP from HP to LP diet. Based on the comparisons between LLPR cows and HLPR cows in Table 4.6, regardless of diets and lactation stages, cows with higher LPR values did not have greater digestibilities of DM, NDF, or CP. 141 Variable Peak DM, % CP, % NDF6, % Late DM, % CP, % NDF, % 3.16a 5.88a 3.01a 2.98a 6.81a 3.64a 1.65a 4.39a 1.21a Table 4.6 Nutrient digestibility for high-, medium- and low-LPR cows fed high- and low- protein diets in peak and late lactation,1,2,3,4 High-protein diets (HP) HLPR MLPR LLPR Low-Protein diets (HP) HLPR MLPR LLPR Change from HP to LP HLPR MLPR LLPR 65.2a 68.9a 48.8a 65.1a 69.0a 48.5a 64.6a 68.2a 48.6a 62.0a 62.9a 45.7a 62.2a 62.2a 44.9a 63.4a 63.7a 47.3a 68.0a 67.8a 56.4a 68.4a 66.8a 57.0a 70.2a 69.4a 58.1a 1.09a 6.03a 0.15a 70.0a 75.7a 57.1a 71.3a 75.2a 58.1a 70.8a 75.2a 59.0a 2.09a 7.39a 0.90a 1 In peak lactation cows, low protein resilience (LPR) was defined as the residual term in the model: dCapP = β0 + β1 ´ CapP_MBWHP + Par (Seq ´ Exp) + Seq (Exp) + Exp + e., where CapP_MBWHP was the CapP per kg metabolic BW when fed HP; Par was parity (primiparous or multiparous); Seq was treatment sequence (HP-LP or LP-HP); Exp was experiment, and e was the residual term in the model. 2 In late lactation cows, low protein resilience (LPR) was defined as the residual term in the model: dCapP = β0 + β1 ´ CapP_MBWHP + β2 ´ MBW_HHHP + Par (Seq ´ Exp) + Seq (Exp) +Exp + e, where CapP_MBWHP was the CapP per kg metabolic BW when fed HP; MBW_HtHP was the metabolic body weight to height ratio when fed HP; Par was parity (primiparous or multiparous); Seq was treatment sequence (HP-LP or LP-HP); Exp was experiment, and e was the residual term in the model. 3 Cows were grouped into high (HLPR), medium (MLPR), and low (LLPR) RFI groups. Cows > 0.5 SD of the mean LPR for a cohort were classified as HLPR, cows < -0.5 SD were classified as LLPR, and those ± 0.5 SD were classified as MLPR. 4 P-values associated with group difference were all > 0.05. 2.34a 8.64a 2.04a 142 Protein Efficiency and Digestibility The relationships between MPE and digestibilities of DM, NDF, and CP are illustrated in Table 4.3 and Table 4.4. For peak-lactation cows fed the HP diet, MPE was positively correlated with CP digestibility (r = 0.15; P =0.08), but not correlated with DM and NDF digestibility; when fed the LP diet, MPE was not correlated with digestibilities of DM, CP, or NDF digestibility. For late-lactation cows, MPE was not correlated with digestibilities of DM, NDF, or CP in either HP or LP diet. 143 DISCUSSION Cow Performance and Digestibilities of DM, NDF and CP in High- and Low- Protein Diets Along with increasing milk production, the HP diet increased digestibilities of DM, NDF and CP in both stages of lactation. The effect of protein on digestibility of NDF (and thus DM digestibility) has been shown previously by Broderick and Reynal (2009), with the possible mechanism being that supplementing nitrogen supported growth of rumen microbes (Russell et al., 1992; Allen, 2000). However, NDF digestibility was not expected to be different in the current study, as both diets (HP and LP) were calculated to contain at least 9.8% RDP for peak- lactation cows and 9.0% for late-lactation cows. According to NRC (2001), these are expected to be adequate to support maximal ruminal microbial function. Thus, our results suggest that supplementing extra RDP that exceeds NRC 2001 recommendation can still improve NDF digestibility. The greater CP digestibility for the HP diet is likely due to the higher digestibility of protein from expeller soybean meal, compared to the LP-diet protein where 44% of the protein was from forage. The major difference between our treatment diets was that the HP diet contained 4% expeller soybean meal in addition to the protein in the LP diet. Lee et al. (2012) showed that supplementing expeller soybean meal increased digestibilities of DM, NDF, and CP. Although the expeller soybean meal provides primarily RUP, it also contains RDP. This RDP along with the N recycled back to the rumen from extra RUP provides extra rumen available nitrogen to improve digestibilities of DM and NDF (NRC, 2001). Another potential mechanism to explain the difference in NDF digestibility between HP and LP diets is the extra 2% starch in the LP diet. When formulating the LP diet, we added starch and fiber in place of protein. Based on de Souza et al. (2018), increasing dietary starch by 2% would decrease NDF digestibility by 1.2%; this is 43% of the 2.8% drop in NDF digestibility of peak-lactation cows and 67% of the 1.8% drop in NDF digestibility of late-lactation cows, when switched from the HP diet to the LP 144 diet. When comparing the results between lactation stages, digestibilities of DM, NDF, and CP in peak-lactation cows were lower than those in late-lactation cows. The difference between the two lactation stages is likely due to 1) faster passage rates for peak-lactation cows with higher intake levels, and 2) higher starch content in the diets fed to peak-lactation cows. Peak-lactation cows in the current study consumed 25% more DM than late-lactation cows. Given that digestibilities of DM and fiber are negatively correlated to intake (de Souza et al., 2018), it was expected that digestibilities of DM and NDF were lower in peak-lactation cows. However, 25% more DMI might not be able to lead to the 4%-unit difference in NDF digestibility in the current study; thus, the significant difference in NDF digestibility between peak- and late- lactation cows was expected to be mostly due to the higher starch content in peak-lactation diets. Based on de Souza et al. (2018), increasing dietary starch by 4.5%, as it was from peak to late lactation in the current study, would decrease NDF digestibility by 2.7%, which was 68% of the drop of NDF digestibility in the current study when comparing peak- and late- lactations. Residual Feed Intake and Digestibility Based on our previous work, we realized that some, if not all, of the BW change in the current study could be attributed to change in gut fill, and BodyE in primiparous cows should be different from that in multiparous cows. In order to calculate BodyE more accurately, BW was adjusted based on DMI (Liu and VandeHaar, submitted 2019), and BodyE was calculated based on new equations. RFI values based on the new method in the current study were strongly correlated with the RFI values calculated from unadjusted BW (Liu and VandeHaar, 2020), with the correlation coefficients being 0.98 in peak-lactation cows on the HP diet, 0.99 in peak- lactation cows on the LP diet, 0.99 in late-lactation cows on the HP diet, and 0.98 in late- lactation cows on the LP diet. 145 In the current study, no association was observed between RFI and digestibilities of DM, CP, or NDF, regardless of dietary protein contents and lactation stages. Several previous studies have shown similar results; for example, no correlation existed between RFI and nutrient digestibility in studies with steers (Cruz et al., 2010), dairy heifers (Lawrence et al., 2011; 2013), beef cattle (Fitzsimons et al., 2014), and lactating cows (Thornhill et al., 2014; Olijhoek et al., 2018). However, some studies have demonstrated correlations between RFI and nutrient digestibility. Specifically, Nkrumah et al. (2006) found RFI to be negatively correlated with digestibilities of DM and CP in steers fed diets containing 18-21% NDF and 12-13% CP. Rius et al. (2012) showed that RFI and nitrogen digestibility were negatively correlated in lactating dairy cattle fed diets with 36% NDF and 23% CP. In McDonnell et al. (2016), low-RFI heifers had higher digestibilities of CP and DM than high-RFI heifers in a nutrient-limiting diet but not in nutrient-adequate diets. Potts et al. (2017) found that DM digestibility explained 9-31% of the variation in RFI in cows fed low-starch diets (less than 17% starch), but 0% in cows fed diets containing ~30% starch. Based on the results from prior work, we suggest that perhaps nutrient digestibility only accounts for the difference in RFI among individual animals when they are fed nutrient-deficient diets. In Nkrumah et al. (2006), steers were fed at 2.5 maintenance level, which is considered as restricted feeding (Olijhoek et al., 2018). Cows in Rius et al. (2012) were fed on pasture, where nutrient deficiency could also be a potential issue for certain animals. In McDonnell et al. (2016), the association of RFI with nutrient digestibility was only present in grass-silage fed animals but not in pasture or TMR-fed animals. However, the idea of nutrient availability was not able to explain the results in Potts et al. (2017). In Potts et al. (2017), cows in both treatment groups were fed with adequate nutrients. Although the dietary starch content was 12-16% in low-starch diets, dietary NE was still ~ 42 Mcal/d, which was similar to the high- 146 starch diets in their study. The relationship between RFI and nutrient digestibility in low-starch diets could be because low-starch (or high-fiber) diets can allow low-RFI cows to express their superior digestive ability, while high-starch diets are already highly digestible and thus barely allow low-RFI cows to express their superior digestibility. In one study of broiler chicken, Rougière et al. (2009) found that the digestive ability of chicken with lower efficiency was improved when fed coarse-particle diet, while the digestive efficiency in chicken with higher- efficiency could not. Rougière et al. (2009) concluded that chicken with lower efficiency would need the stimulation of coarse particles to achieve greater digestive efficiency, while chicken with higher efficiency that have already achieved superior digestibility did not respond to the stimuli. This result suggests that animals with different efficiency status would respond differently to diets. Taken together, both ideas, nutrient deficiency and high-fiber stimuli, can potentially explain the varying relationships between RFI and nutrient digestibility among diets. However, limitations still exist; further examination is in need. Interestingly, in the current study, there was no association observed between RFI and nutrient digestibility in cows fed the LP diet. The seemingly contradictory finding to the idea of nutrient availability indeed suggests that it is the nutrient availability specifically to ruminal microbes that influence the association between RFI and nutrient digestibility. In the current study, the rumen is expected to be fully functional in both HP and LP diets, as both diets contained adequate RDP and energy (starch), with the LP diet only deficient in RUP. Taken together, when RDP and starch are adequate to allow rumen to function to the fullest extent, RFI is not related to digestion regardless of species, growth stages, or physiological states, given that no association was detected in steers or beef (Cruz et al., 2010; Fitzsimons et al., 2014), heifers or cows (Lawrence et al., 2011; McDonnel et al., 2016; Potts et al., 2017), Holstein cows or 147 Jersey cows (Olijhoek et al., 2018), peak-lactation cows or late lactation cows (current study). Thus, when fed nutrient-sufficient diets with low fiber contents, cows’ variability in RFI should be largely attributed to post-digestive metabolisms. Low Protein Resilience, Protein Efficiency, and Digestibility In the current study, the results suggested that cows with better resilience do not necessarily have better digestibility in low-protein diets, nor better ability to maintain their digestibility when switching from the HP to LP diet. That is, post-absorptive mechanisms should contribute more to cows’ resilience to low protein. In the current study, no association was observed between protein efficiency and nutrient digestibility in both peak-lactation and late-lactation cows, except for a low-moderate association between CP digestibility and MPE in peak-lactation cows fed the HP diet. This is consistent with findings in Huhtanen and Hristov (2009), where protein efficiency was poorly associated with rumen protein degradation. Given that protein degradation significantly impacts nutrient digestibility (NRC, 2001), the findings in Huhtanen and Hristov (2009) indicated that protein efficiency might not be associated with digestibility in dairy cows. Indeed, according to Apelo et al. (2014), compared to the digestion ability, post-absorption metabolisms play more important roles in regulating protein efficiency. About 60% of the nitrogen lost occurs in amino acid (AA) catabolism after absorption, especially in portal-drained viscera (PDV) and liver (Hanigan et al., 2004). The maximal theoretical efficiency to convert an ideal absorbed essential amino acids (EAA) profile into milk protein in dairy cows is 75- 85% (AFRC, 1992). Baker (1996) demonstrated that when each EAA supply matched exactly with tissue needs, a similar maximal post-absorptive nitrogen efficiency can be achieved in pigs. If this maximal efficiency is achieved, the theoretical protein efficiency for dairy cows can be as high as 0.49- 0.60, assuming 148 that 65-70% of the dietary protein can be digested and absorbed (NRC, 2001). However, the protein efficiency in the modern dairy cows averages as 0.25 and ranges between 0.15 and 0.40 (Huhtanen and Hristov, 2009). Thus, there is a great potential for the protein efficiency in dairy cow to improve. In order to improve protein efficiency, future nutrition studies should focus more on identifying ways to minimize post-absorptive AA catabolism. CONCLUSIONS High-protein diets significantly increased energy output, MUN, and digestibilities of DM, NDF, and CP in both peak- and late- lactation cows. RFI was not associated with digestibilities of DM, CP, or NDF in neither HP nor LP diet. In other words, we expect that cows’ energy efficiency is more related to post-absorptive metabolisms when fed diets with adequate ruminal N and energy. Lastly, cows with higher LPR values did not have higher digestibilities of DM, NDF, and CP in the LP diet, nor better ability to maintain their digestibility. Work on post- absorptive metabolisms should be performed to further explore the mechanisms of this resilience ability. ANOWLEDGEMENTS We would like to acknowledge J. S. Liesman, D. G. Main, and the staff of the Michigan State University Dairy Cattle Field Laboratory for their assistance in these experiments. This project was supported by Agriculture and Food Research Initiative Competitive Grant no. 2011- 68004-30340 from the USDA National Institute of Food and Agriculture and the fund from Michigan Alliance for Animal Agriculture. 149 REFERENCES 150 REFERENCES AFRC (Agriculture and Food Research Council). 1992. Technical Committee on Responses to Nutrients. Report No. 9. Nutritive Requirements of Ruminant Animals: Energy. Nutr. Abstr. Rev. B 62:787–835. Ahs, F., T. Furmark, A. Michelgard, B. Langström, L. Appel, O. T. Wolf, and M. Fredrikson. 2006. Hypothalamic blood flow correlates positively with stress-induced cortisol levels in subjects with social anxiety disorder. Psychosom. Med. 68: 859-862. Allen, M. S. 2000. Effects of diet on short-term regulation of feed intake by lactating dairy cattle. J. Dairy Sci. 83:1598–1624. Apelo, S. A., J. R. Knapp, and M. D. Hanigan. 2014. Invited review: Current representation and future trends of predicting amino acid utilization in the lactating dairy cow. J. Dairy Sci. 97: 4000-4017. Baker, D. H. 1996. Advances in amino acid nutrition and metabolism of swine and poultry. Pages 41–52 in Nutrient Management of Food Animals to Enhance and Protect the Environment. E. T. Kornegay, ed. CRC Lewis, Boca Raton, FL. Binsiya, T. K., V. Sejian, M. Bagath, G. Krishnan, I. Hyder, A. Manimaran, R. Bhatta. 2017. Significance of hypothalamic-pituitary-adrenal axis to adapt to climate change in livestock. Int. Res. J. Agri. Food Sci. 2: 1-20. Broderick, G. A., and S. Reynal. 2009. Effect of source of rumen-degraded protein on production and ruminal metabolism in lactating dairy cows. J. Dairy Sci. 92:2822–2834. Broderick, G. A., M. Stevenson, R. Patton, N. Lobos, and J. O. Colmenero. 2008. Effect of supplementing rumen-protected methionine on production and nitrogen excretion in lactating dairy cows. J. Dairy Sci. 91:1092–1102. Cochran, R. C., D. C. Adams, J. D. Wallace, and M. L. Galyean. 1986. Predicting digestibility of different diets with internal markers: Evaluation of four potential markers. J. Anim. Sci. 63:1476–1483. Colditz, I. G., and B. C. Hine. 2016. Resilience in farm animals: biology, management, breeding and implications for animal welfare. Anim. Prod. Sci. 56: 1961–1983. Cruz, G. D., J. A. Rodríguez-Sánchez, J. W. Oltjen, and R. D. Sainz. 2010. Performance, residual feed intake, digestibility, carcass traits, and profitability of Angus-Hereford steers housed in individual or group pens. J. Anim. Sci. 88:324-329. 151 de Souza, R. A., R. J. Tempelman, M. S. Allen, W. P. Weiss, J. K. Bernard, and M. J. VandeHaar. 2018. Predicting nutrient digestibility in high-producing dairy cows. J. Dairy Sci. 101: 1123-1135. Doeschl-Wilson, A. B., B. Villanueva, and I. Kyriazakis. 2012. The first step toward genetic selection for host tolerance to infectious pathogens: obtaining the tolerance phenotype through group estimates. Front. Genet. 3:265. Elgersma, G. G., G. De Jong, R. Van der Linde, and H. A. Mulder. 2018. Fluctuations in milk yield are heritable and can be used as a resilience indicator to breed healthy cows. J. Dairy. Sci. 101: 1240–1250. Fitzsimons, C., D. A. Kenny, and M. McGee. 2014. Visceral organ weights, digestion and carcass characteristics of beef bulls differing in residual feed intake offered a high concentrate diet. Animal, 8: 949-959. Goering, H. K., and P. J. Van Soest. 1970. Forage Fiber Analysis (Apparatus, Reagents, Procedures, and Some Applications). Agric. Handbook No. 379. US Department of Agriculture-Agricultural Research Service, Washington, DC. Hach, C. C., B. K. Bowden, A. B. Kopelove, and S. V. Brayton. 1987. More powerful peroxide Kjeldahl digestion method. J. Assoc. Off. Anal. Chem. 70:783–787. Hanigan, M. D., C. K. Reynolds, D. J. Humphries, B. Lupoli, and J. D. Sutton. 2004. A model of net amino acid absorption and utilization by the portal-drained viscera of the lactating dairy cow. J. Dairy Sci. 87:4247–4268. Huhtanen, P., and A. N. Hristov. 2009. A meta-analysis of the effects of dietary protein concentration and degradability on milk protein yield and milk N efficiency in dairy cows. J. Dairy Sci. 92: 3222-3232. Knot, S.A., L.J. Cummins, F.R. Dunshea, and B. J. Leury. 2008. Rams with poor feed efficiency are highly responsive to an exogenous adrenocorticotropin hormone (ACTH) challenge. Domestic Anim. Endocrinology 34: 261–268. Lawrence, P., D. A. Kenny, B. Earley, and M. McGee. 2013. Intake of conserved and grazed grass and performance traits in beef suckler cows differing in phenotypic residual feed intake. Livest. Sci. 152:154–166. Lawrence, P., D. A. Kenny, B. Earley, D. H. Crews Jr., and M. McGee. 2011. Grass silage intake, rumen and blood variables, ultrasonic and body measurements, feeding behavior, and activity in pregnant beef heifers differing in phenotypic residual feed intake. J. Anim. Sci. 89:3248–3261. Lee, C., A. N. Hristov, T. W. Cassidy, K. S. Heyler, H. Lapierre, G. A. Varga, M. J. de Veth, R. A. Patton, and C. Parys. 2012. Rumen- protected lysine, methionine, and histidine increase milk protein yield in dairy cows fed a metabolizable protein-deficient diet. J. Dairy Sci. 95:6042–6056. 152 Leenhouwers, J.I., E.F. Knol, P.N. de Groot, H. Vos, and T. van der Lende. 2002. Fetal development in the pig in relation to genetic merit for piglet survival. J. Anim. Sci. 80: 1759– 1770. Leenhouwers, J.I., E.F. Knol, P.N. de Groot, H. Vos, and T. van der Lende. 2002. Fetal development in the pig in relation to genetic merit for piglet survival. J Anim Sci 80: 1759– 1770. McDonnell, R. P., K. J. Hart, T. M. Boland, A. K. Kelly, M. McGee, and D. A. Kenny. 2016. Effect of divergence in phenotypic residual feed intake on methane emissions, ruminal fermentation, and apparent whole-tract digestibility of beef heifers across three contrasting diets. J. Dairy Sci. 94: 1179-1193. Mertens, D. R. 2002. Gravimetric determination of amylase-treated neutral detergent fiber in feeds with refluxing in beakers or crucibles: Collaborative study. J. AOAC Int. 85:1217– 1240. Morme`de, P., A. Foury, E. Terenina, and P.W. Knap. 2010. Breeding for robustness: the role of cortisol. Animal 5: 651–657. Nkrumah, J. D., E. K. Okine, G. W. Mathison, K. Schmid, C. Li, J. A. Basarab, M. A. Price, Z. Wang, and S. S. Moore. 2006. Relation- ships of feedlot feed efficiency, performance, and feeding behavior with metabolic rate, methane production, and energy partitioning in beef cattle. J. Anim. Sci. 84:145–153. NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Oldham, J. D. 1984. Protein energy relationships in dairy cows. J. Dairy Sci. 67:1090–1114. Olijhoek, D. W., P. Løvendahl, J. Lassen, A. L. F. Hellwing, J. K. Höglund, M. R. Weisbjerg, and P. Lund. 2018. Methane production, rumen fermentation, and diet digestibility of Holstein and Jersey dairy cows being divergent in residual feed intake and fed at 2 forage-to- concentrate ratios. J. Dairy Sci. 101: 9926-9940. Potts, S. B., J. P. Boerman, A. L. Lock, M. S. Allen, and M. J. VandeHaar. 2017. Relationship between residual feed intake and digestibility for lactating Holstein cows fed high and low starch diets. J. Dairy Sci. 100: 265-278. Richardson, E. C., and R. M. Herd. 2004. Biological basis for variation in residual feed intake in beef cattle. 2. Synthesis of results following divergent selection. Australian Journal of Experimental Agriculture. 44: 431-440. Rius, A. G., S. Kittelmann, K. A. Macdonald, G. C. Waghorn, P. H. Janssen, and E. Sikkema. 2012. Nitrogen metabolism and rumen microbial enumeration in lactating cows with divergent residual feed intake fed high-digestibility pasture. J. Dairy Sci. 95:5024– 5034. 153 Russell, J. B., J. D. O’Connor, D. G. Fox, P. J. Van Soest, and C. J. Sniffen. 1992. A net carbohydrate and protein system for evaluating cattle diets: I. Ruminal fermentation. J. Anim. Sci. 70:3551– 3561. Tempelman, R. J., D. M. Spurlock, M. Coffey, R. F. Veerkamp, L. E. Armentano, K. A. Weigel, and M. J. VandeHaar. 2015. Heterogeneity in genetic and nongenetic variation and energy sink relationships for residual feed intake across research stations and countries. J. Dairy Sci. 98: 2013-2026. Thornhill, J. B., L. C. Marett, M. J. Auldist, J. S. Greenwood, J. E. Pryce, B. J. Hayes, and W. J. Wales. 2014. Whole-tract dry matter and nitrogen digestibility of lactating dairy cows selected for phenotypic divergence in residual feed intake. Anim. Prod. Sci. 54:1460–1464. Van Soest, P. J. 1994. Nutritional Ecology of the Ruminant, 2nd ed. Cornell University Press, Ithaca, NY. VandeHaar, M. J., L. E. Armentano, K. Weigel, D. M. Spurlock, R. J. Tempelman, and R. Veerkamp. 2016. Harnessing the genetics of the modern dairy cow to continue improvements in feed efficiency1. J. Dairy Sci. 99: 4941-4954. 154 CHAPTER 5 IMPORTANCE OF CONSIDERING BODY WEIGHT CHANGE IN RESPONSE TO DIETARY PROTEIN REDUCTION IN LACTATING DAIRY COWS ABSTRACT Our objectives were to 1) determine the importance of including body weight (BW) change in the response of lactating dairy cows to a shortage of dietary protein, and 2) develop models to predict BW change when cows are fed low-protein diets. Cows in peak-lactation (n=166) were fed high and low protein diets in a cross-over design of two periods; the study was repeated with 69 of these cows in late-lactation. BW change was used to predict energy and protein changes. Feed intake, milk protein yield, BW change, total protein capture (CapP, calculated as the sum of milk protein and body protein gain), and total energy capture (CapE, calculated as the sum of milk energy and retained body energy) were calculated for each cow on each diet. Income over feed cost (IOFC) for each cow on each diet and the decrease of IOFC in response to protein reduction were calculated. Fixed effects of diet, parity, treatment sequence nested in experiment, treatment period nested in experiment, interaction of diet and parity, and the random effects of experiment and cow nested within experiment were included in the model to compare production performance and IOFC between diets. A sensitivity analysis was performed to examine the significance of different factors influencing IOFC. Prediction models of BW change from high-protein to low- protein diets included fixed (e.g., CP%, BW, and DIM when cows were on high-protein diets) and random (e.g., experiment) effects. In peak lactation, reducing protein from 18% to 14% saved $1.06 per cow in daily feed cost but resulted in estimated daily losses of: 1) 2.9 Mcal milk energy and 2.2 Mcal body energy, 2) 0.13 kg milk protein and 0.02 kg body protein, 3) $1.80 milk income and $0.36 body salvage value. 155 Therefore, BW loss accounted for 43% of the estimated energy loss, 11% of estimated protein loss, and 17% of total income loss. In late-lactation, body tissue loss resulting from feeding less CP (13% vs. 16%) accounted for 1) 51% of estimated energy loss, 2) 14% of estimated protein loss, and 3) 25% of total income loss. In the sensitivity analysis, when calculating IOFC, milk fat price was the most influential factor when cows were fed specific diets, regardless of lactation stages. When calculating the decrease of IOFC from high-protein to low-protein diets, feed cost was the most influential factor. In conclusion, body reserve change should be considered when assessing the cow response to changes in dietary protein. 156 INTRODUCTION Given that protein accounts for 40% of the total feed cost for lactating cows (St-Pierre, 2012), feeding diets with less protein would reduce feed cost and in turn improve farm profitability. However, studies examining the effect of reducing dietary protein on farm profitability were inconclusive. For example, Fadul-Pacheco et al. (2017) found that reducing dietary protein from 16.5 to 15.0% increased income over feed cost (IOFC); whereas Stewart et al. (2012) observed no change in IOFC when dietary CP was reduced from 18.0 to 16.5%. These inconsistencies could be due to different economic conditions, different base diets or protein sources, different animals, or different environments. It is also likely that 16.5% CP in Stewart et al. (2012) has already met or nearly met cows’ genetic capacity to synthesize milk protein. As demonstrated by the Law of dimishing returns, milk response to each unit of successive increase of protein becomes smaller as consumption of dietary protein increases (VandeHaar and St- Pierre; 2006). To maximize profit, we should minimize the protein feeding, while at the same time meeting or nearly meeting the metabolizable protein requirements for a cow’s genetic potential. A deficiency of protein would diminish milk income and excessive protein feeding would increase feed costs with no production benefit. Researchers have examined this tradeoff for many years. Wu and Satter (2000) measured the response of lactating dairy cows to different amounts of dietary protein content during a complete lactation and suggested that feeding cows 17-19% CP before week 30 and 16% CP after week 30 maximized milk production. Hundreds of studies in the last 50 years have examined effects of dietary protein content and source on milk production. Dietary protein is commonly assumed to be sufficient or deficient based on milk production responses. However, dietary protein should not be considered sufficient if body 157 protein is lost to make up for a shortage of dietary protein. This is an important aspect that is commonly overlooked. The importance of body reserve mobilization in supporting milk production in early lactation is widely accepted (NRC, 2001). According to Chilliard and Robelin (1983), when dietary protein is limited, cows might mobilize body protein to support milk production. Milk response to dietary protein reduction has been modeled by Moraes et al. (2018). According to Morae’s et al. (2018), body protein mobilization should be considered when assessing cow response; however, due to the limitation of data, they did not examine this idea. In nutrition trials of lactating cows, responses in BW gain or loss have not been extensively examined in the response to dietary protein reduction. If milk loss is consistently accompanied with BW loss when cows are fed low protein, then failing to account for BW change in individual cows can greatly underestimate the loss. Thus, the objectives of the current study were to 1) determine the importance of including BW change in cow response to dietary protein reduction, and 2) develop models to predict BW change when reducing dietary protein contents. Data MATERIALS AND METHODS Experimental procedures were approved by the Institutional Animal Care and Use Committee of Michigan State University. Data from 166 lactating Holstein dairy cows were used in this study. Among the 166 cows, 69 were studied in both peak and late lactations. Data of MY, milk components (milk protein and MUN), BW, and hip height were collected in the experiments. These are the same animals described by Liu and VandeHaar (2020). In brief, lactating Holstein cows (n= 166; 92 primiparous, 77 multiparous) with initial milk yield (MY) 158 41.3 ± 9.8 kg/d were included in the crossover experiments with two treatments (high-protein diets, HP; low-protein diets, LP) and two periods of 28-35 d each. Experiments were repeated in 69 of the 166 cows (42 primiparous, 27 multiparous) in late lactation. Low-protein diets were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function. Expeller soybean meal was added in place of corn and soyhulls to create high-protein diets, which were 18% CP in peak lactation and 16% CP in late lactation. Cows were milked 2 times daily; DMI and MY were recorded daily. Milk composition was measured over 4 consecutive milkings weekly, and BW was measured 3 times weekly. Calculations Data from peak lactation and late lactation was analyzed separately. All the data used in the analyses below were for cows in the peak lactation, and the methods were applied to late- lactation cows as well. Energy-corrected milk (ECM; kg/d) was calculated for each cow on each diet based on the equation in NRC 2001: ECM = 0.327 ´ milk yield (kg/d) + 12.95 ´ milk fat (kg/d) + 7.2 ´ milk protein (kg/d) For cows > 190 d pregnant, BW was corrected for conceptus weight (CW) for use in the calculation of energy and protein change of body tissues. CW was calculated using the equation from NRC (2001), CW = [18 + (D - 190) ´ 0.665] ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Metabolic BW (MBW) of a cow was estimated as BW0.75, where BW was the mean measured BW for the cow during the treatment period. 159 Empty BW (EBW) was calculated for each cow to adjust BW (after being corrected for CW) for gut fill (Gibbs et al., 1992; Andrew et al., 1994), EBW = BW – 5.2 ´ DMI - CW, where DMI was the daily DMI when BW was measured. For multiparous cows, EBW change was considered to be all body condition, and the body energy gained or lost with changes in EBW (BodyE; Mcal/d) was estimated by the equation as: BodyE = RE (Mcal/kg) ´ dEBW, Where RE = 3.52 + 1.27 ´ BCS (derived from NRC 2001, Table 2-4). For primiparous cows, we assumed their mature BW would be 700 kg and that they had to gain 0.14 kg EBW/d of true growth across the first lactation to reach 92% of mature BW by their second calving (NRC, 2001). Based on the NRC (2001) equations (11-1 and 11-2), the RE content of true growth is 4.4 Mcal/kg dEBW. Any deviation in dEBW from 0.14 kg/d was considered to be body condition gain or loss, and the dEBW associated with body condition change was the same as for multiparous cows (3.52 + 1.27 ´ BCS) Mcal/kg dEBW. Thus, the equation to calculate BodyE was: BodyE=' (3.52 + 1.27 ´ BCS) ´ dEBW ,Parity>1 4.4 ´ 0.14+[(3.52 + 1.27 ´ BCS) ´ (dEBW−0.14)] ,Parity=1 Milk energy output (MilkE; Mcal/d) was estimated using the following equation (NRC, 2001; from Equation 2-15): MilkE = 9.29 ´ fat (kg) + 5.63 ´ true protein (kg) + 3.95 ´ lactose (kg), where each component was based on the period average output of a cow. 160 Energy expended for pregnancy (PregE; Mcal/d) was estimated using the equation from NRC (2001; 2-19): PregE = (0.00318 ´ D - 0.0352) ´ (CBW / 45) / 0.218, where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Total energy capture (CapE, Mcal/d) was estimated as: CapE=' MilkE+BodyE , MilkE+BodyE+PregE , DIM<200 DIM≥200 Under the similar assumptions and information used in NRC energy calculation (2001; Table 2-4, Equations 11-4 and 11-5), protein captured for body tissue gain (BodyP; kg/d) was calculated from dEBW and BCS as, BodyP (0.151− 0.0268 ´ BCS) ´ dEBW , =' 0.132 ´ 0.14+[ (0.151−0.0268 ´ BCS) ´ (dEBW−0.14)] , Parity>1 Parity=1 where (0.151- 0.0268 ´ BCS) kg protein per kg dEBW was assumed when dEBW was considered as body condition gain or loss, and 0.132 kg protein per kg dEBW was considered for the 0.14 kg/d growth. Protein captured for pregnancy (PregP; kg/d) was calculated using the equation from NRC (2001): PregP = 0.00069 ´ D – 0. 0692 ´ (CBW/45), where D was the day of gestation between 190 and 279, and CBW was the calf birth weight. Total protein capture (CapP, kg/d) was estimated for each cow in each treatment (HP and LP) as: CapP=' Milk Protein+BodyP ,DIM<200 Milk Protein+BodyP+PregP ,DIM≥200 161 Feed cost was calculated as the average feed cost from 2016 to 2018. Historical feed prices in the Midwestern area of the U.S. from 2016 to 2018 were used to calculate the average price for each ingredient (Ishler, 2020). The prices ($/kg DM) used in the current study were: $0.13/kg corn silage, $0.19/kg legume silage, $0.18/kg soybean hulls, $0.17/kg ground corn, $0.40/kg solvent extracted soybean meal, $0.47/kg expeller soybean meal, $1.37/kg mix of urea, vitamins, and minerals. Milk income was determined based on individual production of fat ($5.48/kg), protein ($4.13/kg), and lactose ($1.44/kg), then adjusted for the premium including volume ($0.04/kg) and somatic cell count ($0.00080/kg; if SCC < 350,000). Milk component prices were determined based on the 2016- 2018 Class and Components Prices for Federal Milk Marketing Order 33 (Mideast Marketing Area). BW gain was assigned a value of $ 1.36/kg, calculated as the average value of a cull cow ($/kg) from 2016 to 2018. Income over feed cost was calculated as: IOFC = milk price ($/kg) ´ milk yield (kg/d) + gain value ($/kg) ´ BW gain (kg/d) – feed cost ($/kg) ´ feed intake (kg/d). The difference of IOFC (dIOFC) between HP and LP diets was calculated for each cow. A sensitivity analysis was performed on critical financial factors (e.g., feed cost, milk price, and cull cow value), with an assumption of ± 30% variation applied to each of the factors. The change in EBW gain associated with the diet change (dEBWg; kg/d) was calculated as: dEBWg = dEBWLP – dEBWHP. The following model was used to examine the extent to which dEBWg can be predicted by the factors listed on the right side of the equation. 162 dEBWg = β0 + β1 ´ dEBWHP+ β2 ´ ECMHP + β3 ´ MBWHP + β4 ´ BCSHP + β5 ´ HHHP + β6 ´ DIMHP + Par + Seq (Exp) + Exp + e, where dEBWHP was the dEBW when fed HP; ECMHP was the ECM when fed HP; MBWHP was the MBW when fed HP; HHHP was the hip height when fed HP; BCSHP was the BCS when fed HP; Par was parity (primiparous or multiparous); DIMHP was the starting days in milk when fed HP; Seq was sequence (HP-LP or LP-HP); Exp was experiment; and e was the residual term in the model. In the model, all covariates were jointly checked for multicollinearity through variance inflation factors (VIF) analysis (SAS, 9.4). When covariates had VIF greater than 10, the covariate with lesser interest was removed from the analysis. The final model was selected based on backward and stepwise selection criteria (SAS, 9.4). Production responses to diets with each lactation stage were analyzed analyzed using the HPMIXED procedure in SAS (9.4), with fixed effects of diet, parity, treatment sequence nested in experiment, treatment period within experiment, interaction of parity and diet, and the random effects of experiment and cow nested within experiment. Significance was considered at P (cid:1) 0.05 and tendency at P (cid:1) 0.10. Interactions were considered significant at P (cid:1) 0.10 and trends at P (cid:1) 0.15. As mentioned before, all the analyses above were for cows in the peak lactation. The same analyses were performed for the late- lactation cows as well. 163 Animal Performance RESULTS Cows fed the LP diet ate less, produced less milk, and gained less BW than cows fed the HP diet, in both peak and late lactations. Primiparous cows in general ate less, and produced less milk, but with greater milk component concentration, in both peak and late lactations. Further details are shown in Table 5.1 and Table 5.2. Further descriptive information regarding BW change is shown in Table 5.3. Time series of cow response (DMI, milk production, and body weight) to the dietary protein reduction are shown in Supplementary Figure 5.1 and Supplementary Figure 5.2. 164 Table 5.1 Dry matter intake, milk production, and body reserve change for cows fed treatment diets in peak lactation1,2 P-value5 Parity SEM Parity4 SEM TRT TRT ´ Parity Treatments3 LPpeak n=166 23.3 HPpeak n=166 24.3 Primi. n=184 21.3 33.7 Multi. n=148 26.3 44.7 0.01 41.2 37.3 0.35 0.96 0.86 0.14 0.23 <0.01 <0.01 <0.01 <0.01 1.18 3.53 1.01 2.99 1.71 5.07 11.6 1.50 3.43 1.28 2.91 2.20 4.94 12.7 0.01 0.02 0.01 0.01 0.01 0.01 0.13 1.41 3.46 1.21 2.97 2.07 5.01 15.1 0.03 0.06 0.02 0.03 0.05 0.02 0.20 1.27 3.49 1.07 2.94 1.84 4.99 9.2 <0.01 0.13 <0.01 0.05 <0.01 <0.01 <0.01 <0.01 0.13 <0.01 <0.01 <0.01 0.01 <0.01 DMI, kg/d Milk Yield, kg/d Milk Components Fat, kg/d Fat, % Protein, kg/d Protein, % Lactose, kg/d Lactose, % MUN, mg/dL BW, kg 658 BCS, unit 3.23 Change in BW, kg/ d6 0.57 Changes in EBW, kg/ d7 0.39 Change in BCS, unit/28 d 0.071 1Average DIM was 125 for primiparous cows in HPpeak diet, 126 for primiparous cows in LPpeak diet, 122 for multiparous cows in HPpeak diet, and 121 for multiparous cows in LPpeak diet. 2Average parity for multiparous cows was 2.94 in peak lactation. 3 Treatments contained 18% (HPpeak) and 14% (LPpeak) crude protein on a DM basis for peak lactation cows. 4 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 5 P-value associated with treatment differences (HPpeak vs. LPpeak; TRT) and parity differences (Primi vs. Multi.; Parity) in peak lactation cows. Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 6 Determined by linear regression using BW measurements throughout the period. 7 Determined by linear regression using EBW (empty BW= BW- 5.2´ DMI) throughout the period. 0.59 0.30 0.18 0.07 0.01 0.14 0.16 <0.01 0.04 <0.01 <0.01 0.06 <0.01 0.41 0.88 0.83 0.83 596 3.23 0.38 0.27 0.044 653 3.19 0.20 0.12 0.017 714 3.20 0.40 0.25 0.045 8.86 0.05 0.10 0.07 0.03 0.16 0.54 0.23 0.45 0.10 1.04 0.02 0.10 0.07 0.03 165 Table 5.2 Dry matter intake, milk production, and body reserve change for cows fed treatment diets in late lactation1,2 P-value5 Parity Treatments3 SEM Parity4 SEM TRT DMI, kg/d Milk Yield, kg/d Milk Components Fat, kg/d Fat, % Protein, kg/d Protein, % Lactose, kg/d Lactose, % MUN, mg/dL BW, kg Non-pregnant BW, kg BCS, unit Change in BW, kg/ d6 Change in non-pregnant BW, kg/d7 Change in EBW, kg/ d8 Change in BCS, unit/28 d HPlate n=69 19.8 25.1 0.98 3.92 0.80 3.23 1.21 4.79 12.1 702 694 3.62 0.67 0.43 LPlate n=69 18.4 22.2 0.86 4.05 0.68 3.21 1.05 4.79 8.1 693 679 3.56 0.09 -0.05 Primi. n=84 18.0 Multi. n=54 20.2 0.20 0.42 0.02 0.03 0.01 0.01 0.02 0.03 0.16 1.52 5.12 0.03 0.09 0.09 24.2 0.98 4.13 0.77 3.23 1.21 4.99 9.9 623 616 3.44 0.29 0.14 23.1 0.87 3.85 0.72 3.21 1.06 4.58 10.2 772 757 3.74 0.47 0.23 0.48 1.37 0.06 0.12 0.04 0.04 0.14 0.05 0.27 14.5 14.9 0.09 0.09 0.09 <0.01 <0.01 <0.01 <0.01 <0.01 0.34 <0.01 0.90 <0.01 <0.01 <0.01 0.04 <0.01 <0.01 0.32 0.42 0.08 0.02 0.25 0.79 0.05 <0.01 0.28 <0.01 <0.01 <0.01 0.07 0.33 TRT ´ Parity <0.01 0.06 0.02 0.79 <0.01 0.10 0.03 0.72 0.25 0.98 0.07 0.94 0.18 0.09 0.62 0.067 0.23 -0.005 0.14 0.37 0.30 -0.020 0.56 0.082 0.13 0.37 <0.01 0.16 0.01 0.34 0.34 0.14 1Average DIM was 258 for primiparous cows in HPlate diet, 257 for primiparous cows in LPlate diet, 263 for multiparous cows in HPlate diet, and 264 for multiparous cows in LPlate diet. LP diet were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function but deficient to support milk production. The HP diet were 18% CP in peak lactation and 16% CP in late lactation and contained extra expeller soybean meal to increase RUP. 2Average parity for multiparous cows was 3.12 in late lactation. 3 Treatments contained 16% (HPlate) and 13% (LPlate) crude protein on a DM basis for late lactation cows. 166 Table 5.2 (cont’d) 4 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 5 P-value associated with treatment differences (HPlate vs. LPlate; TRT) and parity differences (Primi vs. Multi.; Parity) in late lactation cows. Values within each TRT ´ Parity interaction are shown in Supplementary Table 2.1 6 Determined by linear regression using BW measurements throughout the period. 7 Determined by linear regression using adjusted BW measurements (subtracting conceptus weight) throughout the period. 8 Determined by linear regression using EBW (empty BW= BW- 5.2´ DMI) throughout the period. 167 Mean Peak Lactation 50 600 BW3 1.13 0.62 dBW4 0.69 0.44 dEBW5 0.29 3.26 BCS6 dBCS7 0.093 0.25 HP 718 0.51 0.36 3.22 0.050 65 1.16 0.82 0.44 0.29 471 791 -1.70 6.49 2.06 -1.15 2.58 4.08 -0.47 0.92 524 734 -0.21 1.64 -0.96 2.36 2.83 4.42 -2.07 0.35 915 3.57 1.87 4.50 0.83 930 1.33 2.60 4.75 1.52 Table 5.3 Mean, standard deviation, minimal and maximal values for body tissue change in peak and late lactation cows across diets1,2 Primi Multi Primi SD Min Max Mean SD Min Max Mean SD Min Max Mean LP Multi SD Min Max 714 0.29 0.17 3.19 0.038 64 0.65 0.72 0.44 0.28 597 -1.15 -1.62 2.42 -0.75 593 0.14 0.11 3.22 0.002 51 0.59 0.95 0.32 0.22 583 908 -4.55 4.71 -1.54 3.61 2.42 4.58 -1.00 0.75 651 904 -0.39 2.06 -0.41 1.87 3.25 4.83 -0.74 1.31 448 798 -1.69 2.66 7.25 -1.57 2.42 4.17 -0.50 0.67 527 713 -1.03 1.17 -1.14 1.10 2.92 4.08 -1.06 2.04 73 0.54 0.62 0.46 0.37 788 0.70 0.62 3.83 0.169 50 619 0.57 -0.06 0.58 0.04 3.42 0.34 -0.005 0.43 Late Lactation 50 628 BW 0.48 0.64 dBW 0.72 dEBW 0.53 3.48 BCS 0.37 dBCS 0.035 0.38 1 HP stands for high-protein diets and LP stands for low-protein diets. The LP diet were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function but deficient to support milk production. The HP diet were 18% CP in peak lactation and 16% CP in late lactation and contained extra expeller soybean meal to increase RUP. 2 Primi. stands for primiparous cows and Multi. stands for multiparous cows. 3 Average BW (kg) measured throughout the treatment period. 4 Change of BW (kg/d), determined by linear regression using BW measurements throughout the period. 5 Change of EBW (kg/d), determined by linear regression using EBW (empty BW= BW- 5.2´ DMI) throughout the period. 6 Average BCS (unit), determined by three investigators at the beginning and end of each period, and averaged for the period. 7 Change of BCS (unit/ 28 day). 74 779 0.67 0.23 0.76 0.33 3.78 0.45 -0.005 0.62 649 -1.48 -0.97 3.25 -2.04 168 Importance of Including BW Change in Cow Response As shown in Figure 5.1, reducing protein from the HP to LP diet in peak-lactation cows saved $1.06 per cow in daily feed cost but resulted in estimated daily losses of: 1) 2.9 Mcal MilkE and 2.2 Mcal BodyE, 2) 0.13 kg milk protein and 0.02 kg BodyP, 3) $1.80 milk income and $0.36 body salvage value. Therefore, body tissue loss resulting from the 4% CP reduction in peak lactation cows contributed to 1) 42% of estimated energy loss, 2) 11% of estimated protein loss, and 3) 17% of gross income loss. As shown in Table 5.4, when cows in peak lactation were underfed protein, the major saving of feed cost was from decreased DMI (24.3 kg/d vs. 23.3 kg/d), and the major loss of milk income was from decreased milk production (41.2 kg/d vs. 37.3 kg/d). The milk price ($/cwt) and feed price ($ /kg DM) were barely different between HP and LP diets. 169 Figure 5.1 Energy capture, protein capture, and income in milk and body tissue in peak lactation cows 35 30 25 20 15 1.5 1.0 0.5 0.0 Captured Energy, Mcal/d Distribution of Energy Loss 2.9 28.0 HP 0.7 25.1 LP Body Milk Captured Protein, kg/d 0.03 0.01 1.21 HP 1.08 LP Body Milk 2.2 Mcal/d 42% 2.9 Mcal/d 63% Milk Body Distribution of Protein Loss 0.02 kg/d 11% 0.13 kg/d 89% Milk Body 170 Figure 5.1 (cont’d) Income, $/d Distribution of Income Loss 17.0 15.0 13.0 11.0 0.53 17.4 HP 0.17 15.6 LP Body Milk 0.4 $/d 17% 1.8 $/d 83% Milk Body Figure 5.1 Energy capture, protein capture, and income in milk and body tissue in peak lactation cows (n=166). Reducing protein from 18P to 14P in peak-lactation cows resulted in estimated daily losses of: 1) 2.9 Mcal milk energy and 2.2 Mcal body tissue energy, 2) 0.13 kg milk protein and 0.02 kg body protein, 3) $1.80 milk income and $0.36 body salvage value. Body tissue loss resulting from the 4% units CP reduction in peak lactation cows contributed to 1) 42% of the estimated energy loss, 2) 11% of estimated protein loss, and 3) 17% of gross income loss. Milk energy was estimated based on production of milk fat, protein, and lactose. Body energy and body protein were estimated based on change of empty body weight and BCS. Milk income was determined based on individual production of fat ($5.48/kg), protein ($4.13/kg), and lactose ($1.44/kg), then adjusted for the premium including volume ($0.04/kg) and somatic cell count ($0.0008/kg). Milk components price was determined based on the 2016-2018 Class & Components Prices for Federal Milk Marketing Order 33 (Mideast Marketing Area). The profit gain of BW was assigned to $ 1.36/kg, calculated as the average value of a cull cow ($/kg) from 2016 to 2018. 171 Table 5.4 Income and IOFC in peak and late lactation cows when fed high and low protein diets1,2 Peak Lactation Late Lactation HPPeak LPPeak HPLate LPlate Delta 1.50 0.24 2.91 0.52 Delta3 1.80 0.17 3.89 0.37 5.00 ± 0.14 0.22 ± 0.01 23.3 ± 0.12 6.06 ± 0.15 0.25 ± 0.01 24.3 ± 0.14 11.45 ± 0.53 20.69 ± 0.55 25.1 ± 0.41 0.84 ± 0.13 9.96 ± 0.52 20.45 ± 0.55 22.2 ± 0.41 0.31 ± 0.13 17.35 ± 0.33 19.31 ± 0.16 41.2 ± 0.85 0.53 ± 0.11 15.56 ± 0.32 19.14 ± 0.15 37.3 ± 0.85 0.17 ± 0.10 Income_Milk4, $/d Milk Price, $/cwt Milk Production, kg/d Income_Body Tissue Gain5, $/d Feed Cost6, $/d Feed Cost, $/kg DM DMI, kg/d IOFC_Milk only7, $/d IOFC_Milk+Body8, $/d 1Average parity for multiparous cows was 2.94 in peak lactation, and 3.12 in late lactation 2 Treatments contained 18% and 14% crude protein on a DM basis for peak lactation cows, and 16% and 13% crude protein on a DM basis for late lactation cows 3 Difference between HP and LP (All P values associated with treatments (HP vs. LP) were less than 0.05) 4 Milk income was determined based on individual production of fat ($5.48/kg), protein ($4.13/kg), and lactose ($1.44/kg), then adjusted for the premium including volume ($0.04/kg) and somatic cell count ($0.0008/kg). Milk components price was determined based on the 2016-2018 Class & Components Prices for Federal Milk Marketing Order 33 (Mideast Marketing Area). 5 The profit gain of BW was assigned to $ 1.36/kg, calculated as the average value of a cull cow ($/kg) from 2016 to 2018. 6 The prices ($/kg DM) used were: $0.13/kg corn silage, $0.19/kg legume silage, $0.18/kg soybean hulls, $0.17/kg ground corn, $0.40/kg solvent extracted soybean meal, $0.47/kg expeller soybean meal, $1.37/kg mix of urea, vitamins, and minerals. 7 Income over feed cost (IOFC) was calculated as milk income ($/d) – feed cost ($/d) 8 Income over feed cost (IOFC) was calculated as milk income ($/d) + body tissue gain ($/d) – feed cost ($/d) 3.80 ± 0.11 0.21 ± 0.01 18.4 ± 0.20 4.67 ± 0.12 0.24 ± 0.01 19.8 ± 0.20 0.87 0.03 1.40 0.63 1.15 1.06 0.03 1.01 0.74 1.01 6.79 ± 0.17 7.61 ± 0.21 6.16 ± 0.17 6.46 ± 0.20 11.27 ± 0.19 11.80 ± 0.20 10.53 ± 0.18 10.80 ± 0.20 172 As shown in Figure 5.2, reducing protein from the HP to LP diet in late-lactation cows saved $0.87 per cow in daily feed cost but resulted in estimated daily losses of: 1) 2.4 Mcal MilkE and 2.5 Mcal BodyE, 2) 0.12 kg milk protein and 0.02 kg BodyP, 3) $1.50 milk income and $0.52 body salvage value. Therefore, body tissue loss resulting from the 3% units CP reduction in late lactation cows contributed to 1) 51% of estimated energy loss, 2) 14% of estimated protein loss, and 3) 25% of gross income loss. As shown in Table 5.4, when cows in late lactation were underfed protein, the major saving of feed cost was from decreased DMI (19.8 kg/d vs. 18.4 kg/d), and the major loss of milk income was from depressed milk production (25.1 kg/d vs. 22.2 kg/d). The milk price ($/cwt) and feed price ($ /kg DM) were not different between HP and LP diets. 173 Figure 5.2 Energy capture, protein capture, and income in milk and body tissue in late lactation cows 30 25 20 15 10 5 1.0 0.5 0.0 Captured Energy, Mcal/d Distribution of Energy Loss 2.5 Mcal/d 51% 2.4 Mcal/d 49% Milk Body Distribution of Protein Loss 0.02 kg/d 14% 0.12 kg/d 86% Milk Body 3.8 20.1 HP 1.3 17.6 LP Body Milk Captured Protein, kg/d 0.04 0.02 0.80 HP 0.68 LP Body Milk 174 Figure 5.2 (cont’d) Income, $/d Distribution of Income Loss 15.0 13.0 11.0 9.0 7.0 0.84 11.5 HP 0.32 10.0 LP Body Milk 0.5 $/d 25% 1.5 $/d 75% Milk Body Figure 5.2 Energy capture, protein capture, and income in milk and body tissue in late lactation cows (n=69). Reducing protein from 16P to 13P in late-lactation cows resulted in estimated daily losses of: 1) 2.4 Mcal milk energy and 2.5 Mcal body tissue energy, 2) 0.12 kg milk protein and 0.02 kg body protein, 3) $1.50 milk income and $0.52 body salvage value. Body tissue loss resulting from the 3% units CP reduction in late lactation cows contributed to 1) 51% of estimated energy loss, 2) 14% of estimated protein loss, and 3) 25% of gross income loss. Milk energy was estimated based on production of milk fat, protein, and lactose. Body energy and body protein were estimated based on change of empty body weight and BCS. Milk income was determined based on individual production of fat ($5.48/kg), protein ($4.13/kg), and lactose ($1.44/kg), then adjusted for the premium including volume ($0.04/kg) and somatic cell count ($0.0008/kg). Milk components price was determined based on the 2016-2018 Class & Components Prices for Federal Milk Marketing Order 33 (Mideast Marketing Area). The profit gain of BW was assigned to $ 1.36/kg, calculated as the average value of a cull cow ($/kg) from 2016 to 2018. 175 Sensitivity Analysis As the key parameters varied between the range of -30% to +30%, IOFC (calculated as milk income ($/d) + body tissue gain ($/d) – feed cost ($/d)) varied as shown in Figures 5.3 and 5.4. The most influential factor was milk fat price, causing ~20% variation of IOFC regardless of diets and lactation stages. Following that, feed price and milk protein price were the second and third most influential factors on IOFC. Body salvage value was not important in determining IOFC when cows were on specific diet (HP or LP); however, the impact of body salvage value became greater, when estimating change of IOFC from the HP to LP diet. 176 Figure 5.3 Sensitivity analysis for peak-lactation cows on HP and LP diets IOFC Sensitivity Analysis_HPpeak IOFC Sensitivity Analysis_LPpeak Milk fat price, $/kg Milk protein price, $/kg Milk lactose price, $/kg Body salvage, $/kg Feed cost, $/kg DM -20% HPpeak 30%- -19% 19% -13% -8% -1% 13% 8% 2% Milk fat price, $/kg Milk protein price, $/kg Milk lactose price, $/kg Body salvage, $/kg -20% 19% -13% 11% -8% -1% 6% 1% -16% 0% -10% HPpeak 30%+ 16% 10% 20% Feed cost, $/kg DM -15% 13% -20% LPpeak 30%- -10% 0% LPpeak 30%+ 10% 20% dIOFC Sensitivity Analysis_Peak Milk fat price, $/kg Milk protein price, $/kg Milk lactose price, $/kg Body salvage, $/kg -10% -10% 30% 30% 0% 0% 20% 10% Feed cost, $/kg DM -21% 50% -25% -15% -5% 5% 15% 25% 35% 45% 55% dIOFC_peak 30%- dIOFC_peak 30%+ Figure 5.3 Sensitivity analysis for peak-lactation cows on HP and LP diets. With 30% change of each factor (listed on the Y-axis), income over feed cost (IOFC) varied. The solid filled sections are the response of IOFC to +30% change of each factor, and pattern filled sections are the response of IOFC to -30% change of each factor. IOFC (income over feed cost) =milk income ($/d) + gain value of BW gain ($/d) – feed cost ($/d). dIOFC is the difference of IOFC between HP and LP diets. 177 Figure 5.4 Sensitivity analysis for late-lactation cows on HP and LP diets IOFC Sensitivity Analysis_HPlate IOFC Sensitivity Analysis_LPlate Milk fat price, $/kg Milk protein price, $/kg Milk lactose price, $/kg Body salvage, $/kg -21% 21% -13% 13% -7% 7% -3% 3% Milk fat price, $/kg Milk protein price, $/kg Milk lactose price, $/kg Body salvage, $/kg -22% 22% -13% 13% -7% 7% -1% 1% Feed cost, $/kg DM -18% 18% Feed cost, $/kg DM -18% 18% -25% -15% -5% 5% 15% 25% HPlate 30%- HPlate 30%+ -25% -15% -5% LPlate 30%- LPlate 30%+ dIOFC Sensitivity Analysis_Late 5% 15% 25% Milk fat price, $/kg Milk protein price, $/kg Milk lactose price, $/kg Body salvage, $/kg -17% -13% 17% 13% -6% 6% -14% 14% Feed cost, $/kg DM -22% -25% -15% dIOFC_late 30%- dIOFC_late 30%+ -5% 5% 23% 15% 25% Figure 5.4 Sensitivity analysis for late-lactation cows on HP and LP diets. With 30% change of each factor (listed on the Y-axis), income over feed cost (IOFC) varied. The solid filled sections are the response of IOFC to +30% change of each factor, and pattern filled sections are the response of IOFC to -30% change of each factor. IOFC (income over feed cost) =milk income ($/d) + gain value of BW gain ($/d) – feed cost ($/d). dIOFC is the difference of IOFC between HP and LP diets. 178 Prediction Model for BW Change Responding to Dietary Protein Reduction For peak-lactation cows, the dEBWg prediction model was: dEBWg (kg/d) = -1.33+ 1.09 ´ dEBWHP + 0.02 ´ ECMHP + Parity (R2 = 0.49), where dEBWHP was the dEBW when fed the HP diet and ECMHP was the ECM when fed the HP diet. For late-lactation cows, the dEBWg prediction model was: dEBWg (kg/d) = -1.44 + 1.27 ´ dEBWHP + 0.03 ´ ECMHP + Parity (R2 = 0.58), where dEBWHP was the dEBW when fed the HP diet and ECMHP was the ECM when fed the HP diet. 179 DISCUSSION Dietary protein supplements are expensive, but milk loss resulting from inadequate dietary protein can be even more expensive (VandeHaar and St-Pierre, 2006). To minimize the risk of losing milk production, producers usually feed cows with excessive protein (Colmenero and Broderick, 2006; Edouard et al., 2016). However, excessive protein feeding increases feed costs with no production benefit. To maximize profit, efforts have been made to find solutions to minimize protein feeding but still meet or nearly meet metabolizable protein (MP) requirements for lactating dairy cows. Studies examining cow responses to changes in dietary protein have focused on milk production, and generally have not included changes in body tissues (Ipharraguerre and Clark, 2005; Lean et al., 2018; Moraes et al., 2018). Thus, dietary protein was considered to be sufficient or deficient based on milk production alone. However, dietary protein should not be considered sufficient if body protein mass is lost or desired growth or condition gain is decreased to support milk production. This is an important aspect that is commonly overlooked. To our knowledge, the current study is the first to quantify changes in body mass in response to a reduction in dietary protein content. Our low protein diet was designed to be protein deficient, with a goal of meeting RDP requirements to maintain normal rumen function but feeding low RUP so that MP requirements to maintain milk production and body reserve were not met. With this diet, milk production was impaired, and thus our model enabled us to determine how much of the total response to a change in protein was body tissue and how much was milk production. 180 Importance of Considering BW Change when Evaluating Nutritional Responses. Energy efficiency in dairy cows is commonly defined as milk energy per unit of dietary energy intake. According to the current study, this simplified calculation can potentially underestimate true energy efficiency by 3-17% units in dairy cows. When assessing energy loss to dietary protein reduction, the proportion of body energy capture in total energy capture was even more significant. Specifically, the proportion of total energy loss that was due to body energy loss was 42% for peak-lactation cows and 51% for late-lactation cows, respectively. Therefore, body energy capture should not be neglected when calculating energy loss resulted from feeding less protein. The proportion of body protein capture in total protein gain was relatively small for cows fed either diet (HP: 3% and 5%; LP: 1% and 3%). However, when reducing dietary protein, the proportion of total protein loss that can be attributed to body protein loss was significant. Specifically, it was 11% for peak-lactation cows and 14% for late-lactation cows. In the current study, we recognize that we did not measure body protein mass or N balance directly; the body protein loss from dietary protein reduction was estimated from BW change. Accurate measurement of N loss would have precluded our ability to accurately measure milk and BW responses to 2 diets in >160 cows. However, we improved our estimation accuracy by correcting for changes in DMI that might have influenced gut fill and using changes in EBW to predict tissue energy and protein balance. To further examine whether the value assigned to each kg weight change influences the result, we compared the method used in the current study with the 181 one in our previous study (Liu and VandeHaar, 2020). Using the same cows, Liu and VandeHaar (2020) assumed that all the EBW change in multiparous cows was due to body condition change and all the EBW change in primiparous cows was due to growth; and thus, in Liu and VandeHaar (2020), we assigned 0.07 kg protein per kg dEBW for multiparous cows and 0.12 kg protein per kg dEBW for primiparous cows. We then performed a sensitivity analysis on the original coefficients assumed for protein gain/loss in body tissue change (0.07 kg protein per kg EBW change for multiparous cows, and 0.12 kg protein per kg EBW change for primiparous cows). Scenarios listed in Table 5.5 are different conditions. For example, in scenario 2, we assumed that all the EBW change was due to body condition regardless of parity and contained 7% protein; in scenario 4, we assumed that all the EBW change was from growth and contained 12% protein regardless of parity. These are extreme conditions that can help understand the range for the contribution of BodyP to total protein loss. As shown in Table 5.5, in peak-lactation cows, the contribution was no less than 11%, where larger coefficients led to larger contribution of BodyP to total protein loss. In late-lactation cows, the contribution was no less than 18%. These results clearly suggest that no matter what assumptions we made, body protein change is large enough that it should be considered when estimating cow responses to changes in dietary protein. To further determine the importance of including body reserve mobilization into cow response to dietary protein reduction, we estimated the NEL- allowable milk in peak- lactation cows. If no energy was utilized for body tissue gain, 4.4 kg/d milk would be lost when cows 182 switched from the HP to LP diet. Comparing to the actual loss of 3.9 kg/d milk in the current study, 0.5 kg/d milk loss was compensated by body reserve mobilization. In other words, 11% of milk loss was potentially compensated by body reserve mobilization. The assumptions that we made in the estimation of NEL- allowable milk were: 1) DEdensity (digestible energy density of diet, Mcal/kg) = % NDFdiet ´ NDFdigestibility ´ 4.2 + % NFCdiet ´ 0.90 ´ 4.2 + % Lipiddiet ´ 0.75 ´ 9.5 + % CPdiet ´ CPdigestibility ´ 5.65 2) DE (digestible energy, Mcal/d) = DEdensity (Mcal/kg) ´ DMI (kg/d) 3) ME (metabolizable energy, Mcal/d) = DE ´ 0.85 4) total NEL(net energy of the diet, Mcal/d) = ME ´ 0.66 5) NEmaintenance (maintenance energy, Mcal/d) = 0.08 ´ BW0.75 6) NEL available for milk (Mcal/d) = total NEL – NEmaintenance. 7) NEL - allowable milk (kg/d) ="# %&&'(%)&* +', -.&/ 012345 (7894/5;) 183 Primi Coefficients3,4,5 15.2 11.4 12.5 14.6 17.6 18.6 12.5 19.1 11.4 Contribution of BW to Total Protein Loss6, % 0.025 0.018 0.020 0.024 0.030 0.032 0.020 0.033 0.018 0.042 0.032 0.028 0.038 0.054 0.051 0.035 0.055 0.030 0.017 0.014 0.008 0.014 0.024 0.019 0.015 0.022 0.012 HP, kg/d LP, kg/d Delta, kg/d Table 5.5 Protein captured in body tissue gain for HP and LP diets across lactation stages with different assumptions of protein gain per kg body weight change1,2 Peak Lactation Multi Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Late Lactation Multi Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 1Average parity for multiparous cows was 2.94 in peak lactation, and 3.12 in late lactation 2 The LP diet were 14% CP in peak lactation and 13% CP in late lactation and were formulated to contain adequate RDP to maintain rumen function but deficient to support milk production. The HP diet were 18% CP in peak lactation and 16% CP in late lactation and contained extra expeller soybean meal to increase RUP. 0.021 0.017 0.025 0.019 0.029 0.026 0.018 0.029 0.015 0.036 0.027 0.042 0.036 0.046 0.048 0.031 0.051 0.028 0.057 0.044 0.067 0.055 0.075 0.074 0.049 0.080 0.043 23.1 18.4 25.9 23.1 27.7 28.6 20.5 29.8 18.9 7 7 7 ´1.3 12 7 7 12 7 12 12 12 7 ´ 0.7 12 ´ 1.3 12 ´ 0.7 7 ´ 1.3 12 ´ 1.3 7 ´ 0.7 12 ´ 0.7 7 ´ 1.3 7 ´ 0.7 12 ´ 1.3 12 ´ 0.7 7 ´ 1.3 12 ´ 1.3 7 ´ 0.7 12 ´ 0.7 12 7 12 12 12 12 7 7 Primi 7 7 184 Table 5.5 (cont’d) 3 Protein gain was assumed to be 0.07 kg per kg BW change for multiparous cows, and 0.12 kg per kg BW change for primiparous cows 4 An assumption of ± 30% variation was applied to the two coefficients (0.07 and 0.12) to examine the contribution of BodyP in different scenarios 5 Multi= multiparous cows; Primi= primiparous cows 6 Milk protein loss was 0.14 kg/d when switching from HP to LP in peak-lactation cows, and 0.12 kg/d when switching from HP to LP in late-lactation cows 185 Importance of Considering BW Change when Estimating Changes in IOFC Gain or loss of BW has rarely been taken into account when estimating profitability of diet changes. Based on the current study, the contribution of BW change to calculations of IOFC was small for cows on a specific diet (HP or LP) but was economically significant when calculating the response in IOFC to a reduction in dietary protein. Among peak-lactation cows, when reducing protein from 18% to 14%, the decrease in profit was 27% greater when considering BW loss. The difference was even larger in late-lactation cows. The economic analysis performed in the current study was based on the salvage value of cull cows and thus only the direct cost of BW loss was considered. Loss of BW might also affect health and fertility, but these indirect costs were not considered in the current study. In other words, the difference between IOFC when only considering milk and IOFC when considering milk and body responses would be even greater when including the indirect cost of losing BW. Therefore, at the very least, we suggest that changes in BW must be considered when considering the economic returns to changes in dietary protein; evaluating only milk responses underpredicts the total economic response. We recognize that there are some pitfalls in the current financial analyses. First, certain lost protein could be replenished later in lactation or during the dry period when feed costs per unit of energy and protein is lower. Given that the value of dietary protein and energy declines as dietary protein content declines, the cost of restoring weight loss from early lactation would be much cheaper in late lactation than in early lactation. Moreover, to maintain an optimal body 186 condition in late lactation, energy and protein consumption is commonly restricted for dairy cows. Therefore, maintaining or limiting body reserve gain in late lactation would have a positive economic value rather than a negative one. Such an outcome would have to be added to the calculation for late lactation cows, which would certainly reduce the cost of the BW loss. Second, limitations associated with the sensitivity analysis also existed. We found that milk fat price was the most influential factor for profitability when calculating IOFC in specific diet (HP or LP); this was not surprising because milk fat price was higher than the price for all other milk components during the period we sampled. If the milk protein price was higher than milk fat, as it was from August 2016 to December 2016 ($4.97 per kg milk fat vs. $5.71 per kg milk protein), milk protein price would be the most influential factor for profitability. Another limitation in the sensitivity analysis was that all the prices in the analysis were from the Midwestern U.S. and may not be relevant in other areas, although the relative rankings of dIOFC of individual cows would likely not change across regions. Prediction Model on EBW Change Response to Dietary Protein Deficiency The response of cows to changes in dietary protein content has been widely studied and modeled (Hristov et al., 2005; Lean et al., 2018; Moraes et al., 2018); however, previous studies mainly focused on milk production. We showed that BW change in response to decreased dietary protein should also be considered, and we attempted to predict the BW response based on information that could easily be measured. Based on the individual data from ~170 cows, factors that were significant in the prediction model included production level, BW, and parity. The 187 effect of experiment was also important in modeling BW change because it accounted for all environmental and dietary differences among experiments, such as forage quality, temperature, and humidity. The current study did not propose prediction models to estimate changes in body protein or energy in response to a shortage of dietary protein. However, the reader can use the equations used in the current study to estimate body protein and energy change based on dEBW. BodyE=' BodyP=' (3.52 + 1.27 ´ BCS) ´ dEBW , Parity>1 4.4 ´ 0.14+[(3.52 + 1.27 ´ BCS) ´ (dEBW−0.14)] , Parity=1 (0.151− 0.0268 ´ BCS) ´ dEBW , 0.132 ´ 0.14+[ (0.151−0.0268 ´ BCS) ´ (dEBW−0.14)] , Parity>1 Parity=1 Milk price, feed price, and body salvage value are region specific and vary over time as supply and demand fluctuate. Thus, a prediction model for profitability also was not proposed in the current study. In addition, because we had only 2 levels of protein, previous models based on many protein levels will be more accurate for predicting changes in milk (Hristov et al., 2005; Moraes et al., 2018). However, without additional data on BW change, we suggest that our predictions for change in EBW should be more accurate than assuming no change in EBW. These estimated responses in milk production and BW change could be combined with prices to estimate the economic return to changes in dietary protein for a group of cows. 188 We also tried to predict EBW change based on MY response. After exploring an exhaustive list of factors to account for (parity, ECM when cows on the HP diet, ECM per kg MBW when cows on the HP diet, DIM, treatment sequence, and etc.), we still could not find any relationship between MY response and EBW change. This result suggests that considerable variation exists in EBW change relative to the response in milk production of individual cows. More specifically, cows losing more milk when fed diets with less protein do not necessarily lose more body reserve. However, interestingly, BW change can be predicted based on the ECM change, after adjusting for several factors (ECM per kg MBW when cows on the HP diet, parity, treatment sequence, and experiment). Based on the model of dBW on dMY, for each kg decrease of ECM, BW would be expected to decrease by 1.89 kg. However, the prediction of BW change is not useful to farmers, because cull cows are generally sold based on carcass weight, not live weight. It drew our attention that no correlation existed between EBW change and ECM change while BW change was correlated with ECM change. As ECM change was highly associated with DMI change (P < 0.01) in the current study, it could be validated that part, if not all, of the BW change was due to DMI change. Future work on BW change in lactating dairy cows must adjust BW based on DMI; otherwise, the BW change information might be misleading. Based on all the information above, we suggest that BW change be routinely measured in studies evaluating responses of lactating cows to dietary changes in protein content, protein source, or amino acid supplements. 189 CONCLUSIONS Low-protein diets significantly decreased feed intake, milk production, BW, energy captured in milk and body tissue, and feed efficiency in both peak and late lactation cows. Within each lactation stage, BW change in dietary protein reduction significantly contributed to the total change of energy capture, protein capture, and income. When cows in peak lactation were underfed protein, the loss in net profit was estimated as 27% greater if BW change was included in the response; in late lactation cows, the loss in net profit could be 45% greater. Therefore, BW change should be monitored to fully assess cow response to dietary protein. ACKNOWLEDGEMENTS We would like to acknowledge J. S. Liesman and the staff of the Michigan State University Dairy Cattle Teaching and Research Center for their assistance in these experiments, and Landus Cooperative for donating Soyplus soybean meal. This project was supported by Agriculture and Food Research Initiative Competitive Grant no. 2011-68004-30340 from the USDA National Institute of Food and Agriculture and funds from the Michigan Alliance for Animal Agriculture and Michigan AgBioResearch. 190 APPENDIX 191 Supplementary Figure 5.1 Time series of cow response (dry matter intake, milk production, and body weight) to dietary protein reduction in peak lactation d / g k , e k a t n I r e t t a M y r D 31 29 27 25 23 21 19 17 0 d / g k , d l e i Y k l i M 55 50 45 40 35 30 25 d / g k , d l e i Y n i e t o r P k l i M 1.6 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 7 14 21 28 35 42 49 56 Day of Experiment, day Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 0 7 14 21 28 35 42 49 56 Day of Experiment, day Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 0 1 2 3 4 5 Week of Experiment, week Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 6 7 8 192 Supplementary Figure 5.1 (cont’d) g k , t h g i e W y d o B 750 700 650 600 550 g k , t h g i e W y d o B y t p m E 610 590 570 550 530 510 490 470 450 1 2 3 4 Week of Experiment, week 5 6 7 8 Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 1 2 3 4 5 6 7 8 Week of Experiment, week Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous Supplementary Figure 5.1. Dry matter intake and milk yield were averaged across experiments (1,…,7) by day. Milk protein yield, body weight, and empty body weight were averaged across experiments (1,…,7) by week. Empty body weight (kg)= body weight (kg) – 5.2 ´ DMI (dry matter intake, kg/d). Cohort 1 included the cows fed 18% CP in period 1 and 14% CP in period 2; LP included the cows fed 14% CP in period 1 and 18% CP in period 2. Primiparous= primiparous cows (parity=1); multiparous= multiparous cows (parity > 1). 193 Supplementary Figure 5.2 Time series of cow response (dry matter intake, milk production, and body weight) to dietary protein reduction in late lactation 25 d / g k , e k a t n I r e t t a M y r D 23 21 19 17 15 0 7 14 21 28 35 42 49 56 Day of Experiment, day Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous d / g k , d l e i Y k l i M 33 31 29 27 25 23 21 19 17 1.1 1 0.9 0.8 0.7 0.6 d / g k , d l e i Y n i e t o r P k l i M 0 7 14 21 28 35 42 49 56 Day of Experiment, day Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 0 1 2 3 4 5 6 7 8 Week of Experiment, week Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 194 Supplementary Figure 5.2 (cont’d) g k , t h g i e W y d o B 850 800 750 700 650 600 g k , t h g i e W y d o B y t p m E 750 700 650 600 550 500 1 2 3 4 5 6 7 8 Week of Experiment, week Cohort 1_Primiparous Cohort 2_ Primiparous Cohort 1_Multiparous Cohort 2_Multiparous 1 2 3 4 5 6 7 8 Week of Experiment, week Cohort 1_Primiparous Cohort 2_Primiparous Cohort 1_Multiparous Cohort 2_Multiparous Supplementary Figure 5.2. Dry matter intake and milk yield were averaged across experiments (1,…,4) by day. Milk protein yield, body weight, and empty body weight were averaged across experiments (1,…,4) by week. Empty body weight (kg)= body weight (kg) – 5.2 ´ DMI (dry matter intake, kg/d). Cohort 1 included the cows fed 16% CP in period 1 and 13% CP in period 2; LP included the cows fed 13% CP in period 1 and 16% CP in period 2. Primiparous= primiparous cows (parity=1); multiparous= multiparous cows (parity > 1). 195 REFERENCES 196 REFFERENCE Andrew, S. M., D. R. Waldo, and R. A. Erdman. 1994. Direct analysis of body composition of dairy cows at three physiological stages. J. Dairy Sci. 77:3022–3033. Apelo, S. A., J. R. Knapp, and M. D. Hanigan. 2014. Invited review: Current representation and future trends of predicting amino acid utilization in the lactating dairy cow. J. Dairy Sci. 97: 4000-4017. Bauman, D. E., and W. B. Currie. 1980. Partitioning of nutrients during pregnancy and lactation: a review of mechanisms involving homeostasis and homeorhesis. J. Dairy Sci. 63:1514-1529. Broderick, G. A. 2003. Effects of varying dietary protein and energy levels on the production of lactating dairy cows1. J. Dairy Sci. 86: 1370-1381. Broderick, G. A., M. J. Stevenson, R. A. Patton, N. E. Lobos, and J. J. Olmos Colmenero. 2008. Effect of supplementing rumen-protected methionine on production and nitrogen excretion in lactating dairy cows. J. Dairy Sci. 91:1092–1102. Brun-Lafleur, L., L. Delaby, F. Husson, and P. Faverdin. 2010. Predicting energy× protein interaction on milk yield and milk composition in dairy cows. J. Dairy Sci. 93: 4128-4143. Colmenero, J. O., and G. A. Broderick. 2006. Effect of Dietary Crude Protein Concentration on Milk Production and Nitrogen Utilization in Lactating Dairy Cows1. J. Dairy Sci. 89: 1704- 1712. Chilliard, Y., and J. Robelin. 1983. Mobilization of body proteins byearly lactating cows measured by slaughter and D2O techniques. Pages 195–198 in IVth Int. Symp. Protein metabolism and nutrition. EAAP Publication no. 31, vol. 2. EAAP, Rome Italy. Davidson, S., B. A. Hopkins, D. E. Diaz, S. M. Bolt, C. Brownie, V. Fellner, and L. W. Whitlow. 2003. Effects of amounts and degradability of dietary protein on lactation, nitrogen utilization, and excretion in early lactation Holstein cows. J. Dairy Sci. 86: 1681-1689. Edouard, N., M. Hassouna, P. Robin, and P. Faverdin. 2016. Low degradable protein supply to increase nitrogen efficiency in lactating dairy cows and reduce environmental impacts at barn level. Animal. 10: 212-220. Fadul-Pacheco, L., D. Pellerin, P. Y. Chouinard, M. A. Wattiaux, M. Duplessis, and E. Charbonneau. 2017. Nitrogen efficiency of eastern Canadian dairy herds: Effect on production performance and farm profitability. J. Dairy Sci. 100: 6592-6601. Farahani, T. A. H. A., N. E. Farsuni, and M. Kazemi-Bonchenari. 2019. Interactions of protein levels fed to Holstein cows pre-and postpartum on productive and metabolic responses. J. Dairy Sci. 102: 246-259. 197 Gibbs, M. J., W. E. Irvings, M. S. Dhanoa, and J. D. Sutton. 1992. Changes in body components of autumn-calving Holstein Friesian cows over the first 29 weeks of lactation. Anim. Prod. 55:339–360. Hanigan, M. D., C. K. Reynolds, D. J. Humphries, B. Lupoli, and J. D. Sutton. 2004b. A model of net amino acid absorption and utilization by the portal-drained viscera of the lactating dairy cow. J. Dairy Sci. 87:4247–4268. Hristov, A. N., W. J. Price, and B. Shafii. 2005. A meta-analysis on the relationship between intake of nutrients and body weight with milk volume and milk protein yield in dairy cows. J. Dairy Sci. 88: 2860-2869. Ipharraguerre, I. R., and J. H. Clark. 2005. Varying protein and starch in the diet of dairy cows. II. Effects on performance and nitrogen utilization for milk production. J. Dairy Sci. 88: 2556- 2570. Ishler, V. “Feed Price List”. 2020. Personal Communication. Accessed. Jones, A. P., R. D. Riley, P. R. Williamson, and A. Whitehead. 2009. Meta-analysis of individual patient data versus aggregate data from longitudinal clinical trials. Clin. Trials 6:16–27. Komaragiri, M. V., and R. A. Erdman. 1997. Factors affecting body tissue mobilization in early lactation dairy cows. 1. Effect of dietary protein on mobilization of body fat and protein. J. Dairy Sci. 80: 929-937. Lammers, B. P., and A. J. Heinrichs. 2000. The Response of Altering the Ratio of Dietary Protein to Energy on Growth, Feed Efficiency, and Mammary Development in Rapidly Growing Pre-pubertal Heifers1. J. Dairy Sci. 83: 977-983. Lean, I. J., M. B. De Ondarza, C. J. Sniffen, J. E. P. Santos, and K. E. Griswold. 2018. Meta- analysis to predict the effects of metabolizable amino acids on dairy cattle performance. J. Dairy Sci. 101: 340-364. Liu, E, and M. J. VandeHaar. 2019 Resilience to low protein as a potential trait for selecting dairy cattle toward greater protein efficiency. J. Dairy Sci. (submitted 12/31/19) Liu, E, and M. J. VandeHaar. 2020. Relationship of residual feed intake to protein efficiency in lactating cows fed high or low protein diets. J. Dairy Sci. (in press) Moraes, L. E., E. Kebreab, J. L. Firkins, R. R. White, R. Martineau, and H. Lapierre. 2018. Predicting milk protein responses and the requirement of metabolizable protein by lactating dairy cows. J. Dairy Sci. 101:310-327. NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Riley, R. D., P. C. Lambert, and G. Abo-Zaid. 2010. Meta-analysis of individual participant data: Rationale, conduct, and reporting. BMJ 340:c221–235. 198 Santos, J. E. P., E. J. DePeters, P. W. Jardon, and J. T. Huber. 2001. Effect of prepartum dietary protein level on performance of primigravid and multiparous Holstein dairy cows. J. Dairy Sci. 84: 213-224. Shaver, R. D. 2010. Diets fed in selected WI high-producing dairy herds. Retrieved on 15 January 2017 from http://shaverlab.dysci.wisc.edu/wp-content/uploads/ sites/87/2015/04/2010wihigh-producingherds.pdf. St-Pierre, N. R. 2012. The costs of nutrients, comparison of feedstuffs prices and the current dairy situation. The Ohio State University Extension Buckeye News. Accessed Jul. 20, 2013. Stewart, B. A., R. E. James, M. D. Hanigan, and K. F. Knowlton. 2012. Cost of reducing protein and phosphorus content of dairy rations. The Professional Animal Scientist. 28: 115-119. Stewart, L. A., and M. K. B. Parmar. 1993. Meta-analysis of the literature or of individual patient data: Is there a difference? Lancet 341:418–422. Vandehaar, M. J. 1998. Efficiency of nutrient use and relationship to profitability on dairy farms. J. Dairy Sci. 81: 272-282. VandeHaar, M. J., and N. St-Pierre. 2006. Major advances in nutrition: Relevance to the sustainability of the dairy industry. J. Dairy Sci. 89: 1280-1291. Wu, Z., and L. D. Satter. 2000. Milk Production during the Complete Lactation of Dairy Cows Fed Diets Containing Different Amounts of Protein1. J. Dairy Sci. 83: 1042-1051. 199 CHAPTER 6 GENERAL DISCUSSION AND CONCLUSIONS Residual feed intake (RFI) was repeatable across various dietary protein content within each lactation stage. Among peak-lactation cows, the moderate level of RFI repeatability was in line with previous studies examining repeatability of RFI across diets (Potts et al., 2015; Mangual et al., 2016). Lower RFI repeatability across dietary protein contents in late lactation, compared to peak lactation, might have been due to inaccuracies in measuring tissue gain of pregnant cows. The RFI repeatability across lactation stages was lower than expected, which could be due to the following reasons: 1) mechanisms controlling energy efficiency (or partitioning) shifted as lactation proceeded, and 2) our estimates of body energy change were not accurate and were altered by lactation stage. To generate the most accurate estimates of RFI, RFI estimation should be based on data between DIM 150 and DIM 230 when tissue gain or loss was minimal. Aligning with what Xi et al. (2016) and Mangual et al. (2016) speculated, RFI was associated with protein efficiency in peak-lactation cows and also in late-lactation cows when protein was not limiting. The poor correlation in late-lactation cows fed low-protein diets could be due to the nutrient partitioning to pregnancy; when pregnancy does not take the priority over milk synthesis, cows with lower RFI utilized protein more efficiently. To estimate RFI more accurately in late-lactation cows, better estimates of pregnancy weight gain and pregnancy energy gain are needed. Despite the high repeatability observed across dietary protein contents, cows did not maintain their protein efficiency rankings across lactation stages. This is consistent with the prior work (Zamani et al., 2011) demonstrating low repeatability for protein efficiency across lactation 200 using monthly records on 500 dairy cows (r= 0.12). The low repeatability of protein efficiency across lactation stage could be due to shifts in nutrient partitioning between production and reproduction. For cows fed the same dietary protein in the same lactation stage, those individual cows with lower MUN values do not necessarily utilize protein more efficiently. Thus, MUN of groups of cows can be a good indicator for protein feeding in daily practice; however, ranking cows for protein efficiency based on MUN may be misleading. To identify the cows that need less feed protein to produce the same amount of milk protein, low protein resilience, or LPR, was proposed. After accounting for all factors that can be measured, LPR accounted for 40% of the overall variation in cow responses when switched from high-protein to low-protein diets. Cows with higher LPR were the more efficient cows when fed low protein diets; thus, LPR could be a useful way to think about the relative protein efficiency of individual dairy cows in the future, especially if protein efficiency ever becomes a target trait for genetic selection. Interestingly, it was also observed that the cows with high LPR, compared to cows with low LPR, captured less body protein when fed high protein diets, but more when fed low protein diets. These results imply that cows with less body protein deposition in high-protein diets were likely more resilient to low protein diets. Given the existing variation among cows, LPR can potentially have a genetic basis. However, more work is needed to examine whether LPR is repeatable across other types of diet changes (for example, other types of base diets and other protein or amino acid supplements) and different lengths of the time period (1 wk. vs. 4 wk. vs. 10 wk., etc.). If LPR is repeatable across diets and time, and it is indeed an individual cow trait, further work on a potential genetic basis for LPR would be warranted and would require collaboration among research institutes to collect adequate data. 201 To investigate the underlying mechanisms of RFI and LPR, I examined the association of total tract digestibility to RFI and LPR. No association was observed between RFI and digestibilities of DM, CP, or NDF, regardless of dietary protein contents and lactation stages. The findings in the association of RFI to total tract digestibility were inconclusive in the literature, and results from previous studies suggest that the association of RFI and digestibility varies among diets. Potts et al. (2017) found that RFI and digestibility were correlated for low- starch, high-fiber diets but not high-starch diets. Perhaps the high fiber diets allowed low-RFI cows to express their superior digestive ability, while high-starch diets are already highly digestible and thus a more efficient digestive ability had no impact on overall efficiency. It is also likely that the nutrient availability to ruminal microbes influences the association between RFI and nutrient digestibility. Both ideas can potentially explain the findings, and further examination is in need. In any case, both the low and high protein diets in the present study were relatively high in starch and low in fiber, so perhaps differing digestive abilities of cows in this study had little impact on RFI. High-fiber diets containing different sources of NDF (forage vs. non-forage) can help examine the two ideas. Difference in total tract digestibility also did not contribute to difference in LPR. The results suggested that cows with better resilience do not necessarily have better digestibility in low-protein diets, nor better ability to maintain their digestibility when switching from high- protein to low-protein diets. Thus, post-absorptive mechanisms must contribute more to cows’ resilience to low protein. Previous studies investigating underlying physiological mechanisms in animal resilience to various stressors indicate that these mechanisms are very specific to the type of stressor (Doeschl-Wilson et al., 2012; Colditz and Hine, 2016; Elgersma et al., 2018). The only common factor that was considered across all studies on resilience traits is blood cortisol 202 level. For example, newborn piglet with larger adrenal glands and higher concentration of cortisol in blood were found to be more resilient to disease and had higher survival rates (Leenhouwers et al., 2002). Poultry with higher glucocorticoid levels adapted better to social stress when moved to a new group (Morme` de et al., 2010). In the same study, Morme` et al. also found that birds with a more intense hypothalamic–pituitary–adrenocortical axis stress response had greater immune responses and resistance to disease. Taken together, it seems that cortisol plays an important role in differentiating animals for stress resilience. Indeed, there are reasons to believe that blood cortisol could influence LPR. Cortisol has been shown to increase metabolic rate and catabolic processes such as protein degradation (Knot et al., 2008). Given that protein degradation contributes to protein turnover, a higher level of cortisol may lead to a greater protein turnover rate. Thus, it could be that cows with generally higher levels of blood cortisol already have higher protein turnover rates, and in turn, are less responsive to low-protein diets (as the stressor). Blood cortisol concentrations of the cows should be examined in the future LPR studies. When adjustments are made in the protein content or source of a diet for lactating dairy cows, the dietary protein is commonly assumed to be sufficient based on responses in milk or milk protein production. However, dietary protein should not be considered sufficient if body protein mass is reduced in the process. When reducing dietary protein, the proportion of total energy loss that was due to body energy loss was estimated as 42% for peak-lactation cows and 51% for late-lactation cows; the proportion of estimated protein loss that was attributed to estimated body protein loss was 11% for peak-lactation cows and 14% for late-lactation cows. In addition, if only milk responses are considered when reducing dietary protein, the loss in profit could be underestimated by 27% and 45%, in peak lactation cows and late lactation cows, 203 respectively, compared to considering both milk and body changes. The economic analysis performed in the current study was based on the salvage value of cull cows and thus only the direct cost of BW loss was considered. Loss of BW might also affect health and fertility, so the difference between IOFC when only considering milk and IOFC when considering both milk and body responses might be even greater when including these indirect costs of losing BW. Based on the individual data from ~170 cows in this study, factors that were significant in predicting BW change in response to a reduction in dietary protein included production level of cows when fed high-protein diets, change in BW when fed high-protein diets, and parity. Because there were only 2 levels of protein in the study, previous models based on many protein levels would be more accurate for predicting changes in milk (Hristov et al., 2005; Moraes et al., 2018). However, without additional data on BW change, the predictions for change in EBW should be more accurate than assuming no change in EBW. These estimated responses in milk production and BW change could be combined with prices of feed and milk to estimate the economic return to changes in dietary protein for a group of cows. To improve the BW prediction model, more data from cows fed diets with various protein contents is needed. This will require collaboration among research institutes to collect adequate data on production performance with different feeds specific to different geographic areas. 204 REFERENCES 205 REFERENCES Colditz, I. G., and B. C. Hine. 2016. Resilience in farm animals: biology, management, breeding and implications for animal welfare. Anim. Prod. Sci. 56: 1961–1983. Elgersma, G. G., G. De Jong, R. Van der Linde, and H. A. Mulder. 2018. Fluctuations in milk yield are heritable and can be used as a resilience indicator to breed healthy cows. J. Dairy. Sci. 101: 1240–1250. Hristov, A. N., W. J. Price, and B. Shafii. 2005. A meta-analysis on the relationship between intake of nutrients and body weight with milk volume and milk protein yield in dairy cows. J. Dairy Sci. 88: 2860-2869. Knot, S.A., L.J. Cummins, F.R. Dunshea, and B. J. Leury. 2008. Rams with poor feed efficiency are highly responsive to an exogenous adrenocorticotropin hormone (ACTH) challenge. Domestic Anim. Endocrinology 34: 261–268. Leenhouwers, J.I., E.F. Knol, P.N. de Groot, H. Vos, and T. van der Lende. 2002. Fetal development in the pig in relation to genetic merit for piglet survival. J. Anim. Sci. 80: 1759– 1770. Mangual, M. J., E. Liu, and M. J. VandeHaar. 2016. Repeatability of residual feed intake across dietary forage concentration. J. Animal Sci. 94(Suppl. 5):348–349. Moraes, L. E., E. Kebreab, J. L. Firkins, R. R. White, R. Martineau, and H. Lapierre. 2018. Predicting milk protein responses and the requirement of metabolizable protein by lactating dairy cows. J. Dairy Sci. 101:310-327. Morme`de, P., A. Foury, E. Terenina, and P.W. Knap. 2010. Breeding for robustness: the role of cortisol. Animal. 5: 651–657. Potts, S. B., J. P. Boerman, A. L. Lock, M. S. Allen, and M. J. VandeHaar. 2015. Residual feed intake is repeatable for lactating Holstein dairy cows fed high and low starch diets. J. Dairy Sci. 98: 4735-4747. Potts, S. B., J. P. Boerman, A. L. Lock, M. S. Allen, and M. J. VandeHaar. 2017. Relationship between residual feed intake and digestibility for lactating Holstein cows fed high and low starch diets. J. Dairy Sci. 100: 265-278. Xi, Y. M., F. Wu, D. Q. Zhao, Z. Yang, L. Li, Z. Y. Han, and G. L. Wang. 2016. Biological mechanisms related to differences in residual feed intake in dairy cows. Animal, 10(8): 1311- 1318. Zamani, P., S. R. Miraei-Ashtiani, D. Alipour, H. Aliarabi, and A. A. Saki. 2011. Genetic parameters of protein efficiency and its relationships with yield traits in lactating dairy cows. Livest Sci. 138(1-3): 272-277. 206