Histidine, Lysine and Methionine: From Metabolism to Balanced Dairy Rations

Histidine, Lysine and Methionine: From Metabolism to Balanced Dairy Rations Histidine, lysine et méthionine: Du métabolisme à des rations laitères équilibrées H. Lapierre 1, D.R. Ouellet 1, L. Doepel 2, G. Holtrop 3 and G.E. Lobley 4 1 Dairy and Swine Research and Development Centre, Agriculture and Agri-Food Canada, STN Lennoxville, Sherbrooke, QC, J1M 1Z3; 2 Dept. of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, AB, T6G 2P5; 3 BioSS & 4 Obesity and Metabolic Health Division, Rowett Research Institute, Aberdeen, AB21 9SB, UK. Abstract: Expression of requirements of essential of amino acids (EAA) Two approaches are used to express recommendations and supply of EAA in dairy cows: the proportion and the factorial approach. The proportion approach has the merit of simplicity and was crucial in aiding the implementation of balancing dairy rations for AA. As our basic knowledge on AA utilization by the cow increases, however, we will be able to update the factorial approach that, in the long term, should be more accurate. Definition of requirements of essential amino acids The EAA are divided in two groups. Group 1-AA are mainly extracted by the liver with a postliver supply approximately equal to both mammary uptake and milk output. In contrast, Group 2- AA have low hepatic extraction with post-liver supply greater than mammary uptake itself exceeding milk output. The excess mammary uptake of Lys (Group 2) relative to milk protein output is used for the synthesis of non-eaa, a function maintained to a limited extent at very low supply of Lys. The contribution of Lys to the synthesis of non-eaa is therefore a role important enough to limit milk protein production. The metabolism of His (Group 1) suggests that His should have similar requirement to Met. The large variation observed in the recommendations of this AA (from 2.4 to 3.2 % of metabolizable protein supply) may be related to the potential contribution of carnosine, a dipeptide containing His. Supply rumen protected amino acids Finally, the examination of AA metabolism can be used to determine the effectiveness of supplements of AA. Practically, commercial supplements exist currently only for Met. The D- isomer form of Met, which contributes half of the commercial rumen protected Met products, is transformed into the L-isomer with at least 75% efficiency. Between 51-88% of an infused dose of the hydroxy analog of Met is also transformed into L-Met. 14

A better knowledge of the metabolism of AA by the dairy cows is a necessary tool to better determine AA requirements as well as quantify the effectiveness of products used to balance dairy rations. Résumé: Expression des besoins en acides aminés essentiels (AAE) Deux approches sont utilisées pour exprimer les recommandations en AAE chez la vache laitière : proportionnelle et factorielle. L approche proportionnelle a l avantage d être simple et a été déterminante pour faciliter l implantation des rations équilibrées pour les AAE. Cependant, comme nos connaissances sur l utilisation des AAE par la vache augmentent, nous pourrons améliorer l approche factorielle, qui devrait être, à long terme, plus précise. Définition des besoins en acides aminés essentiels Les AAE se divisent en 2 groupes. Les AAE du groupe 1 sont principalement extraits par le foie, avec un apport post-hépatique est à peu près égal à la fois au prélèvement mammaire ainsi qu à la sécrétion en protéines du lait. À l opposé, les AA du groupe 2 sont très peu prélevés par le foie avec un apport post-hépatique supérieur au prélèvement mammaire lui-même excédant la protéine du lait. L excès de prélèvement mammaire de la Lys (groupe 2), relativement à la production de protéine du lait, est utilisé pour la synthèse d AA non-e, un rôle qui est diminué mais maintenu à des approvisionnements limités en Lys. La contribution de la Lys à la synthèse est donc un rôle assez important pour limiter la production de protéines du lait Le métabolisme de l His nous suggère que les besoins en His devraient être très semblables à ceux de la Met. La grande variation observée dans les recommandations de cet AA (2.4 à 3.2% des protéines métabolisables) provient probablement de la contribution possible de la carnosine, un dipeptide contenant de l His. Apports Finalement, l examen du métabolisme des AA peut s avérer très utile afin de déterminer l efficacité de suppléments d AA. Présentement, des suppléments commerciaux n existent que pour la Met. L isomère-d de la Met, qui constitue la moitié des suppléments commerciaux de Met protégés de la dégradation ruminale, est transformé en isomère-l, avec une efficacité minimale de 75%. Entre 51 et 88% de la dose d un analogue de Met infusé est aussi transformé en L-Met. Une meilleure connaissance du métabolisme des AA par la vache laitière est un outil critique dans la détermination des besoins en AA ainsi que dans la quantification de l efficacité de suppléments utilisés pour équilibrer les rations laitières pour les AA. INTRODUCTION Already 25-30 years ago it was recognized that the essential amino acids (EAA) behaved differently and could be broadly classified into two groups (Clark, 1975; Mepham, 1982). This division was based on the relationship between the mammary uptake of these AA and their output in milk protein. The Group1-AA, including His, Met, Phe + Tyr and Trp, had a mammary uptake to output ratio in milk protein approximately equivalent to unity, indicating a 15

stoichiometric transfer of the mammary uptake into milk. In contrast, Group 2-AA, that comprised Lys and the branched-chain AA, Ile, Leu and Val, had a mammary uptake greater than milk output. Based on these observations, it was suggested, up to the early 90 s that the AA limiting milk protein output would be those from Group 1, which would exclude Lys. With increasing research on AA nutrition, Lys and Met appeared to be the two first limiting AA in typical diets from North America and Europe and similar recommendations for these two AA were proposed in the mid-90 s from the laboratories of Rulquin et al. (1993) and Schwab (1996). The last NRC (7 th edition, 2001), using a similar methodological approach to Rulquin et al. (1993) but with an updated database, evaluated that Lys and Met should represent respectively 7.2 and 2.4% of metabolizable protein (MP) supply, close to the recommendations of 7.3 and 2.5% of protein digestible in the intestine (PDI) estimated by Rulquin et al. (1993). More recently, His was suggested to be the first limiting AA in diets with a high proportion of grass (Vanhatalo et al., 1999; Kim et al., 2000; Korhonen et al., 2000). For this AA, however, recommendations vary widely from 2.4% of MP (Doepel et al., 2004) to 3.2% of PDI (Rulquin et al., 2001), with an intermediate value of 2.7% of MP recommended by CPM-Dairy (Chalupa and Sniffen, 2006). From these observations, several questions immediately arise: 1) Although estimation of AA requirement expressed as a proportion of total supply (MP or PDI) was selected initially as the best way to indicate AA recommendation due to limited knowledge on AA nutrition precluding a factorial approach (NRC, 2001), with recent increased knowledge on AA metabolism, is it still the most appropriate expression of requirement? 2) Why is Lys so often limiting if the mammary uptake to milk output ratio usually exceeds unity? 3) Why is the estimation of His requirement so variable and what should be the requirement? and 4) Actually, commercial supplements of rumen protected AA are only available for Met: are these supplements really supplying metabolically active Met? This presentation does not pretend to be an exhaustive literature review on these issues but rather presents recent work from our laboratories that addresses elements of these specific questions. EXPRESSION OF REQUIREMENTS OF ESSENTIAL AMINO ACIDS Although the AA requirements of dairy cattle are not known with much certainty (NRC, 2001), this must not preclude attempts to balance rations for individual AA rather than just protein. The need to balance for individual AA is not just that conceptual as empirical evidences have shown that in addition to the MP balance, a good balance in individual AA was also beneficial on the animal performance. There are two approaches used to define AA requirements in dairy cows. The factorial approach cumulates the requirements for individual functions (maintenance, growth, pregnancy and lactation) with a definite AA composition and a defined efficiency of transfer of the digested AA for each function. A good example is the Cornell Net Carbohydrate and Protein System (Fox et al., 2004), probably the most widely used model using this approach in North America. The other option is to define requirements with a dose-response relationship between AA proportion in MP supply for maximal use of MP for milk protein synthesis. This is 16

the approach adopted by NRC (2001), as well as by INRA (2007), with recommendations proposed for Lys and Met. Defining the experimental conditions for the projects described below has raised the question: How should the supply be best expressed to relate with changes in milk protein yield?. The expression of the requirements as a proportion of MP has the advantage of simplicity and has certainly helped to convey the importance of balancing dairy rations for AA. Some points, however, question the relevance of this approach if we want to further refine the estimation of EAA. These issues will be discussed using Lys, but identical conclusions would be derived for all EAA. First, the proportion of the sum of EAA in MP supply is not fixed. For example, from the database used by Doepel et al. (2004), where all the digestive flows of EAA had been estimated with NRC (2001), the proportion of total EAA relative to MP supply in control treatments varied from 42 to 48% (mean 45.4%). Lysine supply averaged 104.9 g/d in these studies. Assuming a fixed supply of EAA, including this fixed amount of Lys supply, the proportion of Lys present in the MP would vary between 5.87 and 6.71% with the EAA proportion increasing 42 to 48% of MP (Table 1). This is an important consideration as a decreased proportion of EAA indicates, obviously, an increment in non-eaa and these do not influence milk protein yield (Whyte et al., 2006). Extreme situations provoking these discrepancies between the ratio of EAA to MP are experimentations where AA were infused. In such studies, the proportions achieved with the infusion do not need to respect what is usually obtained with feeding. One extreme example of this concept is demonstrated in one study with abomasal infusion of water, EAA, non-eaa and total AA: it is clear that the parameter that predicts milk protein yield is the absolute amount rather than the proportion of the AA relative to MP (Table 2). Addition of 380 g of non-eaa did not improved milk protein yield while maximum milk yield was observed when 360 g of EAA was added but no further increase when total AA was highest: across treatments, the proportion of Lys relative to MP varied greatly and was not related to milk protein yield. Table 1. Effect of the proportion of essential amino acids (EAA) relative to metabolizable protein (MP) on the expression of a constant lysine (Lys) supply relative to MP. Lys supply, g/d % EAA / MP Total MP % Lys / MP 104.9 43.0 1746 6.01 104.9 45.4 1655 6.34 104.8 48.0 1564 6.71 We have to face, however, our limited knowledge on AA utilization if we want to use the factorial approach to determine AA requirements. This approach requires a good assessment of needs for the individual functions (maintenance, growth, gestation, lactation). Nonetheless, even models based on the proportion approach determine the total requirements for MP based on the factorial approach, before the assessment of the AA balance. This similarity gives the opportunity to compare, from a similar basis, requirements of individual AA based on the proportion approach (NRC, 2001) with the factorial approach using the estimation of MP from single functions obtained from NRC (2001), with an appropriate AA composition selected for each function. 17

Table 2. Effect of supply of amino acids on milk protein yield. Treatments 1 Water EAA non-eaa TAA Metabolizable protein (MP) supply 2, g/d 1221 1556 1600 1936 EAA supply 2, g/d 559 919 559 919 Lys supply 2, g/d 82 141 82 141 Lys, %MP 6.72 9.06 5.13 7.28 Milk protein yield, g/d 967 1104 966 1150 1 Abomasal infusion of essential amino acids (EAA), non-eaa (non-eaa) or total AA (TAA); from Whyte et al., 2006. 2 Supply is estimated from MP or digestive flow of AA plus the abomasal infusion. Before this can be attempted, however, we need to define exactly what each of these functions, and thus their requirements, entails. For example, the so-called maintenance requirement should indeed represent the requirement for the maintenance of a high producing animal and not the requirements of an animal at maintenance. In practice, the most important contributor to maintenance requirement is metabolic fecal protein (MFP), estimated based on dry matter intake (NRC, 2001). To transform the MP requirement for MFP into AA requirements, an AA composition is necessary. Currently, most models use empty body weight composition but, in reality, these losses are proteins secreted into the gut and then excreted in the feces. This corresponds exactly to the definition of endogenous protein losses and it would be more appropriate to use the AA composition of endogenous protein secretions. This exact composition is not yet well-defined in ruminants but, as a first step, the average of values obtained from the abomasum in ruminants and across the small intestine in pigs has been proposed (Lapierre et al., 2007a). On the same theme, it is necessary to also improve the accuracy of MFP estimation. Recent work has determined rates of endogenous protein secretions across the digestive tract, including fecal endogenous losses, in dairy cows offered different types of diets (Ouellet et al., 2007). More work is required to determine those factors affecting endogenous protein losses and to refine the model used but with the results obtained so far, a value of 21.5 g MP/kg DMI is proposed for lactating cows fed normal dairy rations. Scurf AA composition should be based on keratin (Doepel et al., 2004), but this is a minor contributor (< 1%) to MP requirement. Urinary endogenous-n losses represent less than 10% of MP requirement and AA requirements are estimated from empty whole body AA composition. This may alter based on a deeper analysis of the various urine-n fractions of urinary-n but the overall impact will probably be marginal. Bearing in mind the various limitations detailed above, we can compare the proportion and factorial approaches for Lys requirements of cows at different production levels. For this exercise, the average of cows on the medium MP supply in Raggio et al. (2004) is used plus two hypothetical cows, one with lower (20 kg) and one with higher (40 kg) milk production. Comparison of the approaches is given in Table 3 and this also includes estimates of MFP based on NRC (2001) or Ouellet et al. (2007). One major conclusion is that as milk production increases the proportion of required Lys in the MP with the factorial approach also increases, reflecting the higher content of this AA in milk compared with the other functions (urinary, scurf, and MFP). The requirement of the intermediate cow determined by the factorial approach 18

yielded a proportion of Lys in MP supply close to the NRC value (7.24 vs. 7.20; Table 3). Nonetheless, the factorial approach and logical consideration of the biology of the lactating animal both indicate that for high producing dairy cows a larger proportion would be necessary and conversely lower producers would require a smaller proportion of Lys in the MP supply. Table 3. Estimation of lysine requirements in lactating cows using the proportion 1 or the factorial approach 2. Level of production Level of production Low Med High Low Med High DMI, kg/d 22.0 24.2 25.3 MPY, g/d 3 600 853 1200 Requirements MP, g/d % Lys 4 Lys, g/d Urinary endogenous 105 105 105 7.0 7.4 7.4 7.4 Scurf 15 15 15 3.7 0.5 0.5 0.5 MFP 536 570 591 6.0 32.3 34.3 35.6 MFP-Ouellet 5 472 519 546 28.4 31.2 32.7 Duodenal endogenous 6 156 172 179 Milk 896 1273 1791 8.7 79.1 112.4 158.1 Total 1707 2134 2681 Total-Ouellet 1643 2083 2633 Estimations of Lys requirements Proportion Factoria l Lys, g/d 123 154 193 119 155 202 Lys, % MP 7.2 7.2 7.2 6.98 7.24 7.52 Using MFP from Ouellet Lys, g/d 119 150 190 115 152 199 Lys, % MP 7.2 7.2 7.2 6.76 7.10 7.41 1 The proportion approach uses MP requirement from NRC (2001) times the recommendations that Lys supply should represent 7.2% of MP supply. 2 The factorial approach uses the MP estimated from the NRC (2001) times a determined AA composition for each function (see text for definition). 3 MPY: milk protein yield. 4 Percentage of Lys in protein used to transform the MP requirement into Lys requirement. 5 Metabolic fecal protein requirement estimated from the data of Ouellet et al. (2007). 6 The duodenal endogenous flow is included in the NRC (2001) requirement of MP, but is not included in the factorial approach. One final consideration about the proportion approach is that when attempts were made to determine requirements for all the EAA, results cannot realistically be attained for all the EAA. For example, to attain the recommendations of Doepel et al. (2004), we have to supply a diet that contains 50% of MP as EAA. In this study, estimations of recommendations were first obtained as EAA relative to total EAA, but then transferred to MP using an average proportion of EAA on 19

MP of 48% (excluding Trp). With this approach, Lys and Met requirements were yielding similar recommendations to those of Rulquin et al. (1993) and NRC (2001; Table 4). However, achieving a supply of MP containing 50% of EAA is not realistic with practical diets. In contrast, if the latest estimations of Rulquin et al. (2007) are summed, then the sum of EAA is closer to reality, approximately 45%. Unfortunately, in this approach the estimated requirement for certain AA, namely the branched-chain AA and Arg, are lower than usually provided by the rations, meaning that such low proportions cannot be obtained naturally for these AA. With more realistic proportions of these AA included in the calculations, then the sum of EAA will also approximate to 50% of MP supply. In our studies presented below, we infused AA to reach recommendations of EAA relative to MP from Doepel et al. (2004) which is only achievable when the supply of MP contains 50% of EAA. Therefore, in the preliminary presentation of our results, the level of AA supply relative to MP was expressed as if EAA constituted 50% of MP because direct expression relative to total MP was yielding values really too low to be compared with other systems. Table 4. Comparison of estimation requirement and supply of essential amino acids (EAA) relative to digestible protein. AA %EAA/MP 1 %AA/PDI 2 %AA/MP supply 3 Arg 4.6 3.1 4.6 His 2.4 3.03 2.1 Ile 5.3 4.6 4.9 Leu 8.9 8.9 8.9 Lys 7.2 7.3 6.3 Met 2.5 2.5 1.9 Phe 5.5 4.6 5.0 Thr 5 4.0 4.9 Trp 1.7 1.7 1.2 Val 6.5 5.3 5.6 Total 49.6 44.9 45.4 1 MP: metabolizable protein, from Doepel et al., 2004. 2 PDI: protein digested in the intestine, from Rulquin et al., 2007. 3 AA digestive flow and MP supply estimated with NRC (2001) from all the control treatments used in Doepel et al. (2004). A crucial point that has not been debated is the utilization of a fixed factor of conversion of absorbed MP or AA towards protein anabolism. An average value of 0.67 has been used in all the comparisons presented here in order to simplify the discussion, but it is clear that supply under requirement is used at a higher efficiency whereas supply over requirement yields a lower efficiency (Doepel et al., 2004; Lapierre et al., 2007a). 20

In conclusion, although the perfect system is not yet available to determine requirements, there is compelling evidence that diets need to be balanced for AA. The proportion approach has the advantage of being simple to use and has certainly initiated implementation of AA balance in diets. Nonetheless, as our basic knowledge on AA utilization by the cow increases, we will be able to update the factorial approach and this will, in the long term, be a more soundly based and accurate scheme. DEFINITION OF REQUIREMENTS OF ESSENTIAL AMINO ACIDS Metabolism of essential amino acids Over the last decade, research on AA metabolism has expanded from the mammary gland to also include AA utilization between absorption and mammary gland output, including metabolism by the splanchnic tissues. This section will summarize and update the presentation made last year at this Conference (Lapierre et al., 2007a). First, the splanchnic tissues comprise the portal-drained viscera (gut, spleen, pancreas and associated mesenteric fat) plus the liver. In dairy cows, despite the fact that these tissues contribute less than 10% of body mass (Gibb et al., 1992), they account for approximately 50% of both whole body oxygen consumption (Huntington, 1990) and protein synthesis (Lapierre et al., 2002a). Interestingly, but not surprisingly, the behaviour of EAA across the splanchnic tissues complements mammary metabolism and, simplistically, the EAA can be classified in the same groups as those defined for mammary metabolism. The Group 1- AA (His, Met and Phe+Tyr) have most, if not all, of their catabolism occurring within the liver and consequently have a post-liver supply almost identical to both mammary uptake and milk output. In contrast, Group 2-AA have little, if any, removal across the liver and thus have a postliver supply greater than mammary uptake itself greater than milk output (Lapierre et al., 2005a; Figure 1). Lysine When we look at these observations the first question that comes to mind is why would Lys be limiting when more is extracted by the gland than is secreted in milk protein? What is its role? Is the excess mammary uptake relative to milk output obligatory? To examine these questions, two studies were conducted to determine the effect of the supply of Lys on its mammary uptake in relation with the transfer of the excess-n from Lys to non-eaa (Lapierre et al., 2003, 2005b). Keep in mind that although the EAA are taken up in equal or excess amount relative to their milk output, the inverse situation exists for a number of non-eaa, for which mammary uptake is less than milk output, indicating intra-mammary synthesis. Thirty years ago, studies in vitro showed that Leu carbons could be used for mammary tissue synthesis of Glu that was subsequently used for milk protein synthesis (Wohlt et al., 1977) while more recently it was demonstrated in vivo that the N of Leu was transferred to non-eaa in milk protein in goats (Rubert-Alemán et al., 1999). Our first study (Lapierre et al., 2003) showed that Lys-N was indeed transferred to non- EAA. What was unclear, however, was whether such a transfer was obligatory or simply reflected how excess supply might be utilized by the dairy cow. This would then influence the requirements for Lys. 21

Figure 1. Net flux of two essential AA representative of Group 1 (methionine) and Group 2 (lysine) in dairy cows, relative to portal absorption. 1 0.75 0.5 Portal absorption Liver removal Post-liver supply Mammary uptake Milk 0.25 0 Lys Met Average of 13 treatments (n=9, review Lapierre et al., 2005a; n=2, Doepel et al., 2006; n=2, Lapierre et al., unpublished). This aspect was examined in the second study, where multicatheterized cows were fed a protein deficient diet and infused abomasally with a mixture of AA, with the AA profile of casein including, or not, Lys (Lapierre et al., 2005b). For the Lys- and Lys+ treatments, Lys supply averaged 4.9 % and 7.3% of MP (cf discussion above for the expression of supply). All of the infused additional Lys was recovered as portal absorption (Figure 2), suggesting no oxidation by the portal-drained viscera of the extra Lys supplied. Furthermore, liver removal was not affected by treatment and so deletion of Lys from the infusate decreased quantitatively the post-liver supply of Lys. This was accompanied by a decreased mammary uptake and milk output (Figure 2). Nonetheless, the Lys- cows did exhibit some metabolic adjustment because peripheral tissue removal of Lys also decreased and thus the proportional reduction in milk output was smaller than the decline in post-liver supply. Furthermore, although the uptake to output ratio decreased from 1.37 to 1.12 between the Lys+ and Lys- treatments, the lower value was still greater than unity despite the severe deficiency. This indicates that while the excess uptake of Lys by the mammary gland relative to output can be decreased, this is only partial and thus a major Lys deletion has a detrimental effect on milk protein production. To determine where this shortage of Lys negatively impacted mammary metabolism, [2-15 N]Lys was infused for 7 h on the last day of each experimental period. The enrichment of 15 N-AA in both arterial supply and milk protein was then analyzed to determine the contribution of Lys to non-eaa synthesis (Figure 3). These observations indicate a significant contribution of Lys to the synthesis of non-eaa, especially Asx (sum of Asn + Asp), Glx (sum of Gln + Glu), and Ser, the AA with the greatest deficiency of mammary uptake relative to milk output. When Lys supply was reduced the contribution to these non-eaa was still present but decreased significantly, and appeared to have been deleterious to milk protein output. We did not observe, with the net flux measurement, an 22

increased uptake of another AA to replace the Lys contribution. This may be due to the fact that Lys is catabolized to Glu which, together with Gln, represents 18% of milk protein and is the major intermediate in transamination reactions. To summarize, although Lys is taken up in excess relative to milk output by the mammary gland, a fact that initially tempted people to assume that Lys could not be a limiting AA, it seems that Lys plays an important role in the udder for the synthesis of non-eaa. More detailed research is needed to determine if the mammary gland can replace part of the function of Lys (from another EAA or from Glu itself), but in normal feeding conditions, the limitation of Lys is not solely due to its direct contribution to milk protein synthesis, but also relates to a role in the synthesis of non-eaa. Figure 2. Net flux of lysine in dairy cows infused with a mixture of amino acids excluding (Lys-) or including lysine (Lys+) mmol/h 40 20 Portal absorption Liver flux Post-liver Mammary uptake Milk 0 Lys- Lys + Figure 3. Distribution of 15 N between AA (% of labelled AA-N) in milk collected after 7 hours of a [ 15 N]lysine infusion in dairy cows infused with a mixture of amino acids excluding (Lys-) or including lysine (Lys+) % of labelled AA-N 6 4 2 Ala Asx Glx Leu Ser 0 Lys- Lys+ 23

Figure 4. Net flux of two essential histidine and methionine in dairy cows. mmol/h 12 8 4 Portal absorption Liver removal Post-liver Mammary uptake Milk 0 His Met Average of 13 treatments (n=9, review Lapierre et al., 2005a; n=2, Doepel et al., 2006; n=2, Lapierre et al., unpublished). Histidine Histidine belongs to the Group 1-AA, as does Met, and these two EAA show similar behaviour across the both the liver and the mammary gland (Figure 4). In addition, milk composition of His is very similar to Met, 27 vs. 28 mg AA per g of crude protein (Jensen, 1995). It is therefore curious that estimation of requirements can be so different from those of Met, from 2.4% of MP (Doepel et al., 2004), 2.7% of MP (CPM-Dairy; Chalupa and Sniffen, 2006) and 3.2% (Rulquin et al., 2001). Histidine has a unique peculiarity compared with other EAA, having labile pools that provide a source of stored His during short periods of deficiency without the need to deplete body proteins. The main pools are the intramuscular carnosine pool and circulating hemoglobin. Carnosine is a dipeptide, formed of β-ala and His and acts as an intracellular buffer. The total body content in a dairy cow with an average muscle carnosine concentration of 10 mm would be approximately 420 g of His: this could certainly contribute substantially to supply, especially in studies conducted in a Latin Square with short experimental periods. This issue was addressed in a recent study with multi-catheterized cows and 4 levels of His supply, averaging 1.70, 2.03, 2.35 and 2.67 % of MP (cf discussion above for the expression of supply). Preliminary results are available on the relationship between His supply, milk protein yield and the potential impact of body carnosine (Figure 5).The first three levels of supply indicate that a plateau in milk protein yield was reached between 2.03 and 2.35% His, with decreasing supply resulting in lower carnosine concentration in muscle. Between 1.70 and 2.35% of His, this decrement in carnosine muscle concentrations would have increased the supply for milk protein synthesis by 3.3 g/d of His over an estimated digestible flow of 38 g/d (NRC, 2001), the equivalent of more than 100 g milk protein. For an unknown reason, however, carnosine muscle concentrations declined at the highest His supply with a slight increase in milk protein yield Although these data do not 24

definitively indicate that requirements are reached at a His supply of 2.35% of MP supply, they do confirm that milk protein yield (or percentage) is not the sole criteria that needs to be used when studying His supply as a substantial portion of supply can arise from, or contribute to, the muscle carnosine pool. It is unclear currently what factors regulate partition of His towards carnosine or milk protein synthesis. Figure 5. Effect of increasing supply of histidine (as % of metabolizable protein: MP) on muscle carnosine concentrations and milk protein yield. 10 Muscle carnosine Milk protein yield 1200 Muscle carnosine (mm) 9 8 7 1100 1000 900 Milk protein yield (g/d) 6 1.7 2.03 2.35 2.67 His (% MP) 800 In conclusion, the biology indicates that even if the mammary uptake of Lys exceeds output in milk protein, the contribution to the synthesis of non-eaa within the mammary gland is critical to maintain milk protein output. Utilization of AA across the splanchnic and mammary bed, as well as concentrations of AA in milk protein would suggest similar requirements for His compared with Met. Variations observed between studies to determine His requirement might be due to the potential endogenous contribution of carnosine to His supply. SUPPLY - RUMEN PROTECTED AMINO ACIDS Using the proportion approach, only through the use of rumen protected AA can we attain the level recommended for certain AA, especially Met, Lys and His. Using the factorial approach, recommendations of g/d of each AA can be reached by increasing supplementation of rumen undegradable protein, but this will lead to excess supply of other AA. Therefore, even with this approach it is probably more economical, and certainly more environmentally friendly, to use rumen protected AA to increase those that are deficient. In practice, only Met is commercially 25

offered currently in a rumen protected form that has been scientifically tested and reported. Lysine was offered for a short period a few years ago but then withdrawn from the market. It is probable, however, that the large demand for supplemental Lys created by the availability of distillers grains from corn residue as a by-product of the ethanol industry will stimulate the reintroduction of rumen protected Lys products to the market. This section will not compare the bioavailability of rumen protected Met (RPM) products from different companies but will focus on two specific aspects concerning the metabolic use of RPM. First, all the commercial RPM are an equimolar mixture of the D and L-isomers of Met, but the D-isomer cannot be directly used in the synthesis of protein by mammals! What happens to D- Met in the dairy cow? Second, an analog of Met is offered on the market; once absorbed, is it efficiently used by the dairy cow? Utilization of the D-isomer of Met The RPM products are derived from chemical synthesis and contain both the D and L-isomers in approximately equal proportions. The D-isomer is a mirror image of the L-isomer but cannot be used for the direct synthesis of protein by mammals. The AA contained in plant (or animal) feedstuff proteins are only present as the L-isomers. Preliminary observations revealed that the mammary gland of dairy cows cannot extract D-Met from plasma (Lapierre and Lobley, unpublished). Although support of the activity of the D-form was suggested when a similar N retention was obtained with supplementation of the L- vs. a mixture of the D and L-isomers in growing cattle (Campbell et al., 1996), but less than half the dose was needed to support the increased N retention. Therefore, we conducted a study to directly determine whether the D- isomer was transformed into the biologically active L-isomer. Cows fed a protein-deficient diet were infused abomasally with a mixture of AA devoid of Met. During that period, for 7 h, the cows were infused with Met labelled on the methyl group (L-[ 2 H 3 ]Met), and received a bolus injection of D-Met labelled on the carboxyl carbon ([D-[1-13 C]Met), thus allowing the direct determination of the proportion of the D-isomer converted to L-Met. The results showed that at least 75% of the D-Met was transformed into L-Met. This was a relatively slow process and thus the biological half-life of the D-form was much longer than the L-isomer (Lapierre et al., 2008). The mammary gland still was unable to use D-Met directly, but did use the L-isomer formed in other tissues. Utilization of the Met analog Another type of Met supplement is also available on the market. This is the Met hydroxy analog referred to earlier as MHA, but now known as HMB or HMTBA (according to its chemical nomenclature: 2-hydroxy-4-[methylthio]-butanoic acid). This analog is different from Met as it contains a hydroxy group instead of the amino group. The HMTBA is not offered in protected form, but its chemical nature offers some resistance to ruminal degradation (Belasco, 1972; Patterson and Kung, 1988). As for other products affected by rumen degradation, the rate of passage is a major factor affecting HMTBA availability, and estimation of rumen escape varied from 5% (Noftsger et al., 2005) to 44% (Koenig et al., 1999), with intermediate values of 18% measured in our lab (Lapierre et al., 2007b). Recently an isopropyl ester of HMTBA was also offered on the market that is absorbed rapidly across the rumen wall (Graulet et al., 2004). The 26

objective of this section is not to discuss the rumen by-pass of HMTBA, but rather investigate the mode of action once absorbed. First, contrary to the initial hypothesis originating from work in poultry, a large fraction of absorbed HMTBA bypasses the liver in ruminants (Lapierre et al., 2002b; Wester at al., 2006) and is available to peripheral tissues that convert the analog to L- Met. The newly synthesized Met can be retained within the tissues for anabolic purposes (protein synthesis) or exported into the plasma to be used by other tissues. This explains why supplementation with HMTBA can result in increased availability of Met within tissues without a major increment in circulating concentrations of Met. In dairy cows infused into the jugular vein with 1.5 g HMTBA/h, approximately 20% of milk protein Met was derived from HMTBA, with the Met being synthesized either within the mammary gland directly or transported there from other tissues (Lapierre et al., 2002b). One feature of HMTBA is that, although supplied as a mixture of D and L-isomers (as for Met), the tissues convert both forms into L-Met, so no D-Met is present in the animals (Lobley et al., 2001). In conclusion dairy cows can effectively transform Met provided as the D-isomer or as HMTBA into L-Met for the synthesis of protein, including milk protein. CONCLUSION In conclusion, a better knowledge of the metabolism of proteins and AA will greatly help to define requirements of the dairy cow. Globally, protein- or AA-derived secretary losses from the body (as metabolic fecal protein or urinary-n excretion) plus milk protein secretion (plus growth and gestation, when present) represents MP requirements of the lactating animal. As their quantification improves, this will enhance the accuracy of the factorial approach and permit better AA balance of dairy rations. In addition, for individual AA, a greater understanding of their utilization at the tissue level will allow determination of their roles and how this might limit milk protein synthesis. Such information will help delineate with increased accuracy the requirement for each AA and allow prediction of potential metabolic flexibility in the face of any deficiency. Indeed, an improved factorial approach will be a vital intermediate step in the progression of prediction systems towards the ultimate goal of realistic and functional mechanistic models. Acknowledgements: The authors wish to thank the financial support of Agriculture and Agri- Food Canada, the Rural and Environment Research and Analysis Directorate of the Scottish Government, the Natural Sciences and Engineering Research Council of Canada and the Dairy Farmers of Canada. 27

REFERENCES Belasco, I.J. 1972. Stability of methionine hydroxy analog in rumen fluid and its conversion in vitro to methionine by calf liver and kidney. J. Dairy Sci. 55:353-357. Campbell, C.G., E.C. Titgemeyer, and G. St-Jean. 1996. Efficiency of D- vs L-methionine utilization by growing steers. J. Anim. Sci. 74:2482-2487. Chalupa, W. and C. Sniffen. 2006. Balancing Rations on the Basis of Amino Acids: The CPM- Dairy Approach. Pages 96-109. In 21 st Annual Southwest Nutrition & Management Conference, Tempe, AZ. Clark J.H. 1975. Lactational responses to postruminal administration of proteins and amino acids. J. Dairy Sci. 58:1178-1197. Doepel, L., D. Pacheco, J.J. Kennelly, M.D. Hanigan, I.F. López, and H. Lapierre. 2004. Milk protein synthesis as a function of amino acid supply. J. Dairy Sci. 87:279-1297. Fox, D.G., L.O. Tedeschi, T.P. Tylutki, J.B. Russell, M.E. van Amburg, L.E. Chase, A.N. Pell, and T.R. Overton. 2004. The Cornell Net Carbohydrate and Protein System model for evaluating herd nutrition and nutrient excretion. Anim. Feed Sci. Tech. 112:29-78. Gibb, M.J., W.E. Ivings, 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. Graulet, B., C. Richard, and J.C. Robert. 2004. The isopropyl ester of methionine hydroxyanalogue is absorbed through the rumen wall of the cow. J. Anim. Feed Sci. 13 (Suppl.1): 269-272. Huntington, G.B. 1990. Energy metabolism in the digestive tract and liver of cattle: influence of physiological state and nutrition. Repr. Nutr. Dev. 30:35-47. INRA. 2007. Alimentation des bovines, ovins et caprins. Besoins des animaux Valeurs des aliments. Ed. Quæ. 307p. Jensen, R.G. 1995. Handbook of milk composition. R.G. Jensen, ed. Academic Press. Toronto, ON. Koenig, K.M., L.M. Rode, C.D. Knight, and P.R. McCullough. 1999. Ruminal escape, gastrointestinal absorption, and response of serum methionine to supplementation of liquid methionine hydroxy analog in dairy cows. J. Dairy Sci. 82:355-361. Kim, C.H., J.J. Choung, and D.G. Chamberlain. 2000. Variability in the ranking of the three most-limiting amino acids for milk protein production in dairy cows consuming grass silage and a cereal-based supplement containing feather meal. J. Sci. Food. Agric. 80:1386 1392. 28

Korhonen, M., A. Vanhatalo, and P. Huhtanen. 2000. Responses to graded postruminal doses of histidine in dairy cows fed grass silage diets. J. Dairy Sci. 83:2596 2608. Lapierre, H., J.P. Blouin, J.F. Bernier, C.K. Reynolds, P. Dubreuil and G.E. Lobley. 2002a. Effect of diet quality on whole body and splanchnic protein metabolism in lactating dairy cows. J. Dairy Sci. 85:2631-2641. Lapierre, H., J.J. Dibner, M. Vázquez-Añón, D. Parker, P. Dubreuil, G. Zuur and G.E. Lobley. 2002b. Use of 2-hydroxy-4-[methylthio]-butanoic acic (HMB) by lactating dairy cows. J. Anim. Sci. 80 / J. Dairy Sci. 85, Suppl. 1: 72. Lapierre H, E. Milne, J. Renaud and G.E. Lobley. 2003. Lysine utilization by the mammary gland. Pages 777-780 In Progress in research on energy and protein metabolism. EAAP publication No.109. Ed. W.B. Souffrant and C.C. Metges. Lapierre H., R. Berthiaume, G. Raggio, M.C. Thivierge, L. Doepel, D. Pacheco, P. Dubreuil, and G.E. Lobley. 2005. The route of absorbed nitrogen into milk protein. Anim. Sci. 80:11-22. Lapierre, H., L. Doepel, E. Milne, and G.E. Lobley. 2005b. Effect of lysine (Lys) supply on its utilization by the mammary gland. J. Anim. Sci. 83 / J. Dairy Sci. 88 Suppl.1:89. Lapierre, H., G. Raggio, D.R. Ouellet, R. Berthiaume, L. Doepel, and D. Pacheco. 2007a. Balancing dairy rations for amino acids: why and how. Eastern Nutrition Conference, Animal Nutrition Association of Canada, Québec City, Electronic version: 155-179. Lapierre, H., M. Vázquez-Añón, D. Parker, K.M. Koenig, P Dubreuil, and G.E. Lobley. 2007b. Absorption of 2-hydroxy-4-methylthiobutanoate (HMTBA) in Dairy Cows. J. Dairy Sci. 90: 2937-2940. Lapierre, H., G. Holtrop, A.G. Calder, J. Renaud, and G.E. Lobley. 2008. Is D-methionine (Met) used by the dairy cow? J. Dairy Sci. Abstract Lobley, G.E., H. Lapierre, and M. Vázquez-Añón. 2001. HMB metabolism in ruminants. Pages 15-24 In Southwest Nutrition & Management Conference proceedings. February 2001. Mepham, T.B. 1982. Amino acid utilization by lactating mammary gland. J. Dairy Sci. 65:287-298. National Research Council (NRC). 2001. Nutrient Requirements of Dairy Cattle. 7 th rev. ed. Natl. Acad. Sci., Washington, DC. Noftsger, S., N.R. St-Pierre, and J.T. Sylvester. 2005. Determination of rumen degradability and ruminal effects of three sources of methionine in lactating cows. J. Dairy Sci. 88:223-237. 29

Ouellet, D.R., D. Valkeners, G. Holtrop, G.E. Lobley, and H. Lapierre. 2007. Contribution of endogenous secretions and urea recycling to nitrogen metabolism. Proc. Cornell Nutrition Conference for feed manufacturers. Dpt. Anim. Science, Cornell University, NY: 1-24. Patterson, J. A., and L. Kung, Jr. 1988. Metabolism of DL-methionine and methionine-analogs by rumen microorganisms. J. Dairy Sci. 71:3292-3301. Raggio, G., D. Pacheco, R. Berthiaume, G.E. Lobley, D. Pellerin, G. Allard, P. Dubreuil, and H. Lapierre. 2004. Effect of metabolizable protein on splanchnic flux of amino acids in lactating dairy cows. J. Dairy Sci. 87:3461-3472. Rubert-Alemán J, G. Rychen, F. Casseron, F. Laurent, and G.J. Martin.1999. 15 N enrichment of casein amino acids in the milk from goats given a single intravenous dose of L-[ 15 N]leucine. J. Dairy Res. 66:283-288. Rulquin, H., P.M. Pisulewski, R. Vérité, and J. Guinard. 1993. Milk production and composition as a function of postruminal lysine and methionine supply : a nutrient-response approach. Livest. Prod. Sci. 37:69-90. Rulquin, H., R. Vérité, J. Guinard-Flament and P.M. Pisulewski. 2001. Acides amines digestibles dans l Intestin. Origines des variations chez les ruminants et répercussions sur les protéines du lait. INRA Prod. Anim. 14:201-210. Rulquin, H., G. Raggio, H. Lapierre, and S. Lemosquet. 2007. Relatiosnhip between intestinal supply of essential amino acids and their mammary metabolism in the lactating dairy cows. In Energy and protein metabolism and nutrition. EAAP publication No.124. Ed. I. Ortigues-Marty. pp. 587-588. Schwab, C.G. 1996. Rumen-protected amino acids for dairy cattle: Progress towards determining lysine and methionine requirements. Anim. Feed Sci. Technol. 59:87-101. Vanhatalo, A., P. Huhtanen, V. Toivonen, and T. Varvikko. 1999. Response of dairy cows fed grass silage diets to abomasal infusion of histidine alone or in combinations with methionine and lysine. J. Dairy Sci. 82:2674-2685. Wester, T.J., M. Vázquez-Añón, J.J. Dibner, D.S. Parker, A.G.Calder, and G.E. Lobley. 2006. Hepatic metabolism of 2-hydroxy-4-methylthiobutanoic acid in growing lambs. J. Dairy Sci.89:1062-1071. Weekes T.L., P.H. Luimes, and J.P. Cant. 2006. Responses to amino acid imbalances and deficiencies in lactating dairy cows. J. Dairy Sci. 89:2177-2187. Wohlt JE, J.H. Clark, R.G. Derrig, and C.L. Davis. 1977. Valine, leucine and isoleucine metabolism by lactating bovine mammary tissue. J. Dairy Sci. 60:1875-1882. 30

Whyte, T., A. Hayirli, H. Lapierre, and L. Doepel. 2006. Effect of post-ruminal supplementation of amino acids on production performance of lactating dairy cows. J. Anim. Sci 84 / J. Dairy Sci. 89 (suppl 1):78. 31