Optimizing Protein Nutrition for Milk Production: AA Metabolism and Milk Protein Synthesis, Papers of Animal Biology

Download Optimizing Protein Nutrition for Milk Production: AA Metabolism and Milk Protein Synthesis and more Papers Animal Biology in PDF only on Docsity! 1998 J Dairy Sci 81:2540–2559 2540 Received August 20, 1997. Accepted March 23, 1998. Current Concepts of Amino Acid and Protein Metabolism in the Mammary Gland of the Lactating Ruminant B. J. BEQUETTE,* F.R.C. BACKWELL,* and L. A. CROMPTON† *Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen, Scotland, AB21 9SB †Department of Agriculture, University of Reading, Reading, England, RG6 2AT ABSTRACT Milk protein responses to protein nutrition are typically poor and, in part, may be due to the low efficiency ( ∼25 to 30%) of converting dietary N into milk. Posthepatic availability of amino acids (AA) is not limited, yet only ∼30% is converted into milk. The poor capture of AA by the mammary gland may relate to the imbalanced and uncoordinated timing of nutrient delivery to the gland. The infusion of essen- tial AA improves the efficiency of utilization (0.31); however, further catabolism of AA within the mam- mary gland suggests that AA transport is not a major limitation. These losses may serve ancillary or func- tional roles, but mammary oxidation of some AA oc- curs only when AA extraction exceeds the stoichio- metric requirements for milk protein synthesis. Intracellular substrate supply may be more limiting than is the appartus for protein synthesis. Studies utilizing isotope labeling and conducted in vitro and in vivo now suggest that circulating peptides and proteins can serve as sources of perhaps all AA for casein synthesis, but the source of these remains elusive. Constitutive protein and casein turnover con- tribute significantly (42 to 72%) to mammary protein synthesis. All AA are extensively channeled through an intermediary protein pool or pools that have rapid turnover rates. The AA are then incorporated into casein, which appears to be fixed in association with protein turnover. The mammary gland is a major controller of its metabolism, and the mechanisms of AA extraction and conversion into milk protein are linked to secretion events. Blood flow may be a key point of regulation whereby mechanisms sense and respond to nutrient supply and balance to the gland via alterations in hemodynamics. ( Key words: amino acid, mammary gland, casein, lactation) Abbreviation key: BCAA = branched-chain AA, EAA = essential AA, EPV = external pudic vein, GIT = gastrointestinal tract, GSH = glutathione, MG = mammary gland, NEAA = nonessential AA. INTRODUCTION Dairy producers and researchers have the goals of increasing the conversion efficiency of dietary nutrients (energy and protein) into milk and improv- ing its healthful properties (reduced fat and higher protein) and processing characteristics (e.g., >3% milk protein, higher k-CN and lower b-LG contents, and reduced lactose). An ability to achieve these goals, preferably jointly and through approaches that are acceptable to consumers, should translate into greater profit margins at the marketplace. The efficiency of converting dietary N into milk protein output is poor at 25 to 30%. Milk protein content and yield can be increased by dietary protein supplementation or by gastrointestinal infusion of protein or AA; however, the responses attained are often unpredictable and are considerably less than would be predicted by current feeding schemes for dairy cows (1, 53, 82). This result is largely due to the incomplete understanding and the inadequate mathematical representation of the metabolic trans- formations of AA during absorptive and postabsorp- tive processes (2) . Currently, a fixed factor (0.64 to 0.80 on an incremental basis) is used to convert absorbed AA into milk protein despite evidence that the observed response to supplemental protein AA is curvilinear [i.e., diminishing partial efficiencies (47, 80, 110)]. For example, Guinard et al. (47) observed that the conversion of absorbed essential AA ( EAA) into milk protein decreases from 0.44 to 0.34 as levels of duodenal infusion of casein increase up to 762 g/d. A large proportion of this loss occurred across the mammary gland ( MG) such that the udder conver- sion of EAA decreased from 0.88 to 0.49 with infusion. These responses probably would not have been predicted by the current protein feeding schemes. Thus, in the future, the dynamics of postabsorptive metabolic events, especially those governing the availability of AA to the MG and the utilization of AA by the MG for milk protein synthesis, will need to be Journal of Dairy Science Vol. 81, No. 9, 1998 SYMPOSIUM: OPTIMIZING PROTEIN NUTRITION FOR REPRODUCTION AND LACTATION 2541 represented so that responses in milk protein output from changes in dietary nutrient availability can be predicted more accurately. As a consequence, dairy producers might be better able to formulate diets that are economical and biologically efficient to attain desired production goals. Our two research groups have been working jointly on a project investigating the organ tissue-specific metabolism of nutrients, primarily AA, in the lactat- ing dairy cow and goat to elucidate the mechanisms and quantify the important nutrient interactions oc- curring in the gastrointestinal tract ( GIT) , liver, and MG. The overall aims of this project are to develop the framework of a mechanistic model of dairy cow metabolism for on-farm use in managing expensive feed resources and to identify ways of altering the protein content of milk (71). Aspects of studies con- ducted on the MG in vivo form part of this review, which briefly considers mammary AA metabolism in the context of the whole body and attempts to identify limits to AA conversion into milk protein. Many of our studies have coupled stable isotope kinetics to the conventional arteriovenous difference technique across the MG to investigate the dynamics of AA and protein metabolism and the possible role of blood- derived peptides and proteins as sources of AA for milk protein synthesis. TISSUE PROTEIN METABOLISM BY THE WHOLE BODY Milk protein secretion immediately postpartum oc- curs at such high rates that, except for the equally dramatic increases (by ∼300%) in feed intake and the modulations (both up- and down-regulation) that oc- cur in nonmammary tissue AA and protein metabolism, the cow could quickly reach a metabolic crisis by depleting critical protein reserves. The estab- lishment and coordination of this higher plane of metabolism forms the basis of the homeorhesis con- cept proposed by Bauman and Currie (11) whereby acute and chronic interactions among nutrients, hor- mones, and tissues orchestrate the metabolism and physiology of the whole body in a way that allows for high rates of milk production while also allowing the animal to partition nutrients for vital functions that are unrelated to the MG. The net result is that nutrients are delivered at adequate rates to the MG from exogenous (i.e., diet) or endogenous (e.g., fat depots and muscle tissue) sources to support synthe- sis of milk constituents while sufficient quantities of nutrients are held in reserve to protect the health and reproductive status of the cow during the current and subsequent lactations. For example, the tissues of the dairy goat and cow undergo dramatic changes in metabolism within the first few weeks of lactation, including substantial in- creases in the protein mass of the MG (+1000%) and of the splanchnic tissues (rumen, small and large intestines, and liver; +11 to 28% vs. dry animals); conversely, decreases (18 to 21% vs. dry animals) occur in carcass, head, and feet protein mass (8, 25, 43). These changes have been linked to alterations in the fractional and absolute rates of tissue protein synthesis, but rates of protein degradation have not yet been monitored at this early stage. During later lactation, however, there is some evidence that mus- cle tissue depletion and repletion are modulated mainly by the regulation of proteolytic activity. Insu- lin is thought to inhibit protein hydrolysis in the muscle of lactating goats by selectively acting on skeletal muscle tissues, resulting in a decrease in the mRNA levels of ubiquitin, a component of the ATP- ubiquitin-dependent proteolytic pathway (62). The inhibitory actions of insulin appear to increase as stage of lactation advances, but this activity appears to be augmented by AA supply (102). Thus, through the interdependent influences of tissue sensitivity and substrate supply, the depletion or repletion of muscle protein stores could be tightly regulated to maintain a balanced and continuous supply of precur- sors to the MG. Perhaps a clearer understanding of how this process is regulated through metabolite receptor interactions and postreceptor events could lead to the development of nutritional regimens to up- or down-regulate the proteolytic pathway and make better economic use of these potentially rate-limiting reserves of AA. In early lactation, energy and N intake are often less than adequate to support the high rates of milk protein output. In these circumstances, the depletion of carcass tissue protein is important to supplement dietary and microbial protein AA and to maintain an adequate supply of AA to the MG and carbon for gluconeogensis by the liver. In the dairy cow, mobili- zation of protein from tissue is capable of supplying 90 to 430 g of AA/d (43, 60); in the dairy goat, this contribution could range from 2 to 66 g/d (8, 9, 25, 102). Thus, for a cow producing 1 kg of milk protein/ d, the contribution of these reserves can be considera- ble in terms of the quantity supplied; equally impor- tant however, is the timeliness of this contribution with respect to the requirements of the MG in early lactation to synthesize parenchymal (secretory) tis- sue (57) as well as milk proteins. The contribution of tissue mobilization has been assessed in cows by employing an isotope dilution technique, which exploits the natural differences in Journal of Dairy Science Vol. 81, No. 9, 1998 BEQUETTE ET AL.2544 Figure 1. Comparison of the 24-h blood flow pattern of one-half the mammary gland of a goat milked ( M ) twice daily (0800 and 1600 h), fed molasses-treated hay every 2 h in equal portions (90 g), and fed the concentrate portion either in equal parts (A; 800 g) at the milking times (M + F), in equal parts (800 g) 4 h after (M, F) the milking times (B), or in equal parts (133 g) every 2 h along with the hay (C). Blood flow was averaged over 1-h periods during 24 h, and each 1-h time point represents the mean ( ±SE) of 3 consecutive d on the respective feeding regimens. corresponds to MG metabolism during the 2 to 5 min that the sample was being withdrawn and blood flow was being determined. The assumption when employ- ing this approach must be that blood flow and the arteriovenous difference remain constant during the sampling period; otherwise, net uptake would not be comparable with the hourly or daily milk output measurement. Posture and the regularity of feed consumption appear to have a major influence on MG blood flow and possibly nutrient uptake. The MG blood flow (monitored by flow probe) and arteriovenous blood differences of volatile fatty acids, glucose, and oxygen have been monitored in cows, both standing and lying (90). Those authors observed that, although blood flow increased by 18% when cows were lying down, net substrate uptake remained constant. A prelimi- nary study (B. J. Bequette, 1995, unpublished data) examined whether meal consumption or milking af- fect MG blood flow. Daily (24-h) patterns of MG blood flow were compared in goats under three milk- ing and feeding regimens. Goats were milked twice daily (0800 and 1600 h) and were fed equal portions (90 g) of molasses-treated hay every 2 h (12 times daily). In addition, the goats received the concentrate part of their diet in one of three regimens: as two equal portions (800 g) fed during milking (Figure 1A), as two equal portions (800 g) fed 4 h after milking (Figure 1B), or as equal portions (133 g) fed with the hay (Figure 1C). The most striking differ- ence was the 4- to 5-h shift of the two blood flow peaks when twice daily concentrate feeding was moved forward by 4 h (compare Figure 1A vs. 1B), suggesting a strong involvement of signals related to nutrient intake in regulating MG blood flow. No dis- cernible patterns or relationships were apparent when the ration was fed 12 times daily (Figure 1C). These temporal blood flow responses to feeding com- pare closely with the hormone and metabolite profiles characterized for dairy cows fed at different frequen- cies (2 vs. 6 times daily) (101). Identification of the mechanisms that signal nutrient intake and blood flow could be important to enhance the partition of nutrients to the MG. Although it is unclear whether postural and feed- ing effects on blood flow also alter the net utilization of substrates by the MG, these effects must be consid- ered to avoid misinterpretation of the exact mechan- isms (blood flow vs. nutrient extraction) involved. In our studies with goats, we have often attempted to minimize blood flow variability by ensuring that goats remain standing for the sampling periods and are fed frequently (12 times daily) (14, 16, 18). In addition, arteriovenous blood samples are continu- ously withdrawn over 1-h periods to provide a direct Journal of Dairy Science Vol. 81, No. 9, 1998 SYMPOSIUM: OPTIMIZING PROTEIN NUTRITION FOR REPRODUCTION AND LACTATION 2545 comparison with measurements of blood flow and milk output integrated during the sampling period. AA SUPPLY AND MG UPTAKE: ARE THESE LIMITED? There are numerous reports (18, 47, 76, 80) in the literature of small or often no increases in milk yield and protein output (content) when supplemental dietary protein or AA are given and when poor recov- ery (13 to 17%) of the supplemental AA in milk protein is the norm. Such observations serve to high- light that the knowledge is incomplete concerning how dietary nutrients influence or limit milk protein synthesis. We have conducted a series of intragastric and i.v. AA infusion studies to begin to delineate whether tissue partition (competition), pattern, amount supplied, stage of lactation, or MG uptake of AA limit milk protein synthesis (see Table 1). GIT and i.v. Supplementation of AA A logical first step to identify AA limitations to milk protein synthesis is one that considers the rela- tive importance of EAA versus nonessential AA ( NEAA) supplies. To test whether the NEAA supply is limiting to milk protein synthesis, midlactation cows were fed a low protein diet (140 g of CP/kg of DM) and given jugular infusions (5-d) of either a total AA (400 g/d) or an EAA (208 g/d) mixture, both being equivalent in composition to bovine milk pro- tein (78). Milk yield was not affected, but milk pro- tein content increased by 2.6 and 4.4 g/kg with total AA and EAA infusion translating into +87 and +143 g of milk protein/d, respectively (Table 1). Recovery of the infused AA in milk protein was greater for EAA (36%) than for total AA (22%) infusion, and both were higher than has been observed in studies of dietary supplementation (76, 80) and GIT infusion (44, 45, 46, 47). Thus, there appears to be no addi- tional advantage to increasing the posthepatic sup- plies of the NEAA. For midlactation cows that were fed the same low protein diet, mesenteric vein infu- sion (3 d) of total AA (600 g/d) increased (+2.8 kg/d) milk yield, but infusion of the NEAA (288 g/d) reduced milk yield slightly (93). Both total AA and EAA (312 g/d) infusions increased milk protein con- tent (3.8 and 4.5 g/kg); however, the recovery of the infused AA was greater for total AA (31%) than for EAA (21%) infusion. This result may suggest that some relationship exists between NEAA and EAA metabolism by the liver, which may be required to ensure that posthepatic availability of EAA is ade- quate for milk protein synthesis. Indeed, NEAA may be more important when EAA supplies are aug- mented prehepatically. When midlactation cows were fed a medium CP diet (170 g/kg of DM) and were given duodenal infusions of EAA at two levels (312 and 468 g/d), maximal responses in milk protein output were achieved at the lowest rate of infusion (+86 vs. +74 g/d), and the recovery of the infused EAA in milk protein was also highest at the lower rate (14% vs. 8%) (29). The responses to AA supply may also be dependent upon stage of lactation, perhaps as a consequence of changes in tissue depot depletion and repletion and hormone sensitivity, which may alter the pattern and availability of AA to the MG. For cows that were fed a high protein diet (200 g/kg of DM) [(28); L. A. Crompton, 1996, unpublished data], EAA infusion (312 g/d) during early lactation (wk 8) resulted in a greater increase (+2.0 g/kg) in milk protein content than when EAA were given to the same cows in midlactation (wk 16; +1.2 g/kg). The efficiency of converting the additional EAA into milk was higher in early lactation (14%) than in midlactation (8%). Responses to AA infusion are also dependent upon the CP content of the basal diet. In midlactation cows that were fed a medium protein diet (170 g of CP/kg of DM), larger increases in milk protein content (2.4 vs. 1.2 g/kg) and output (80 vs. 49 g/d) and greater efficiencies of recovery (13 vs. 8%) were achieved in response to a jugular infusion of 312 g of EAA/d than when cows were fed a higher protein diet (200 g of CP/kg of DM) (L. A. Crompton, 1996, unpublished data). Comparison across studies also seems to sug- gest that milk protein content is increased more when AA are infused via i.v. rather than the intragastric route (2.9 vs. 1.1 kg of milk/d, 0.7 vs. 0.2 g of protein/ kg of milk) but that milk protein output and recover- ies are identical or similar. These conclusions will require further confirmation by direct testing in the same animals by site infusions. AA Supply and MG Uptake Metcalf et al. (76) observed that only the arterial supply of Lys and the net uptakes of Lys, Leu, and Tyr were increased significantly in response to sup- plemental fish meal. In a subsequent study (80) in which three levels of CP were examined, Lys (and Val) uptake by the MG was also increased by sup- plementation, despite the fact that the concentrations of most EAA, except for Met, Phe and Thr, were increased. In both of these studies, however, Lys was Journal of Dairy Science Vol. 81, No. 9, 1998 BEQUETTE ET AL.2546 TABLE 1. Comparison of studies examining responses of dairy cows fed grass silage diets to infusion of AA mixtures.1 1Milk production values represent the control (saline infusion) mean and the response ( D) to the infused AA. 2JV = Jugular vein, MES = mesenteric vein, and DUO = duodenum. 3EAA = Essential AA; TAA = total AA (both equivalent to AA composition in milk protein). 4Calculated as grams AA infused divided by equivalent casein AA output. Milk Milk CP in Infusion2 Infusion Milk protein protein Study Lactation diet site rate3 yield content yield Recovery4 (wk) (g/kg of DM) (g/d) (kg/d) ( D) (g/kg) ( D) (g/d) ( D) (g/g) Crompton et al. (28) 7–9 200 JV EAA 312 31.4 0.5 31.0 2.0 986 83 0.14 Metcalf et al. (78) 11–28 140 JV EAA 208 22.4 1.1 32.5 4.4 726 143 0.36 TAA 400 23.8 0.6 32.4 2.6 765 87 0.22 Reynolds et al. (93) 13–18 140 MES EAA 312 22.0 0.7 32.4 4.5 709 126 0.21 TAA 600 21.8 2.8 32.4 3.5 706 183 0.31 L. A. Crompton et al. 13–18 170 JV EAA 312 22.9 0.7 36.7 2.4 841 80 0.13 (1996, unpublished) 200 JV EAA 312 24.1 0.7 36.8 1.2 887 49 0.08 Crompton et al. (29) 13–18 170 DUO EAA 312 23.7 0.9 36.0 1.1 833 86 0.14 EAA 468 24.4 0.6 35.2 1.9 857 74 0.08 always extracted in excess, which increased still fur- ther with supplementation. Conversely, Phe, Tyr, His, and Met uptakes were consistently less than milk protein outputs. The intragastric infusion studies of Guinard and Rulquin (44) and Guinard et al. (47) are among the few available studies that have allowed a comparison of the partition and utilization by the MG of con- trolled known rates of AA supply. Removal of EAA by the MG was increased in response to incremental duodenal infusions of casein. The net removal of EAA represented 0.81 of the casein EAA infused at the lowest rate but only 0.50 at the highest level of casein infusion. The efficiency of converting the extracted EAA into milk protein was also reduced considerably (from 0.88 to 0.49) with nearly all the EAA being extracted far in excess of net requirements for milk protein synthesis. Excess uptake may have been util- ized for the de novo synthesis of many of the NEAA for which uptake was found to be less than or equal to milk outputs at all levels of infusion. In a related study (45), Lys was infused into the duodenum of cows on the same diet as just described. The net uptake of Lys exceeded milk protein outputs at the level of Lys infusion producing the maximum increase in milk protein output. At this level, presumably, other AA or factors had become limiting. This limita- tion was probably not due to a limitation of Met supply because, in a companion study (46), there was no milk response to infusion of Met on this same diet despite the substantial increase in arterial concentra- tions and supplies of Met to the MG. In contrast to the response to Lys infusion, however, the uptake of Met by the MG was not enhanced by Met infusion, which may be one reason for the lack of a milk protein response. PRECURSOR AA FOR MILK PROTEIN SYNTHESIS Arteriovenous balance has been widely used to quantify the AA requirements of the MG for metabolism and milk protein synthesis; based upon this comparison, AA have been proposed that limit milk protein synthesis. The needs of the MG would appear to be greater for EAA such as Val, Leu, and Ile (branched-chain AA; BCAA) and Arg, Lys, and Thr because their extraction generally exceeded milk pro- tein outputs. Quantification of Met, His, Phe, and Trp, for which uptake is usually considerably less (10 to 70%) than milk protein outputs, has proved to be more troublesome. In part, this situation may be due to the relatively low concentrations of these AA in blood and their greater instability during derivatiza- tion procedures and AA analysis. Apparent deficits in the net uptake of these AA therefore has resulted in speculation that non-free AA sources, such as pep- tides and proteins, may contribute to the supply of AA for milk protein synthesis. Whether non-free AA con- tribute to mammary metabolism is of particular im- portance to our attempts to develop mechanistic models of organ and tissue metabolism for which the fluxes of all AA must be represented (35, 36, 71). Arteriovenous Measurement of Peptides The direct quantification of peptides in blood and plasma has proved difficult to assess because of the Journal of Dairy Science Vol. 81, No. 9, 1998 SYMPOSIUM: OPTIMIZING PROTEIN NUTRITION FOR REPRODUCTION AND LACTATION 2549 tions of circulating peptides that represented up to 80% of the total AA measured (peptide-bound plus free AA) were found, but there was no significant net portal appearance of peptides. Employing more rigorous deproteinization techniques, much lower peptide concentrations have been reported in studies with rats (38), lactating goats (4) , and growing sheep (5) . In the latter study, peptide fluxes were also assessed in plasma that had been prepared by a previous method (59) to establish whether differ- ences in techniques may account for the varying results reported. In this analysis, as expected, peptide concentrations were much higher, but there was still no evidence that peptide uptake occurred across the GIT of the sheep. Recently, we have confirmed this observation in two related studies where experimen- tal methodologies (21) and the influence of dietary components (20) on peptide and AA fluxes in the GIT were examined in sheep. It appears, therefore, that it may be necessary to look elsewhere for the origin of much of the circulat- ing peptide AA. One obvious other source is the liver export proteins. Albumin is the major contributor to the export protein pool. Based upon the mean frac- tional synthesis rate (0.05/d) reported for growing sheep [52 kg of BW (26)], plasma volume (0.05 of BW), and concentrations of albumin (29 to 38 g/L) in the normal lactating goat (65 kg of BW), the poten- tial supply of albumin for peripheral tissue use could be 5 to 6.7 g/d. This amount could provide 0.280 to 0.367 g of Phe daily, which, if transferred to the MG, could make up 50 to 100% of the deficit in Phe supply that is required by goats (16). It is probably unlikely that all of the albumin synthesized is transported to the MG, and so other sources of the bound AA supply need to be considered, which may include contribu- tions of peptides derived from the turnover of en- dogenous sources such as collagen and skeletal mus- cle proteins [(41), review; (85)]. PROTEIN METABOLISM OF THE MG Milk Protein Synthesis The MG tissues comprise secretory and nonsecre- tory cells. In the lactating goat, secretory cell mass is substantially increased between parturition and peak lactation. Most of this increase and the accompanying up-regulation of mammary metabolic activity is the result of a threefold increase in cellular differentia- tion (112). The reduction in milk yield during later phases of lactation appears to be due to the loss of secretory tissue through apoptosis, rather than by dedifferentiation of the epithelial cells (41). The primary function of the MG is to synthesize and assemble the constituents of milk and, relative to the neonate, to allow bulk transfer of water after birth. Milk proteins are synthesized at a phenomenal rate compared with the constitutive protein mass of the MG [ratio of milk protein secretion (grams per day) to mammary tissue protein mass (grams): cows, 0.45 (43); and goats, 0.50 (8, 25)]. The major milk proteins that are synthesized by ruminant species are the caseins ( a-CN, b-CN, k-CN, and g-CN) and whey proteins ( b-LG, a-LA, and proteose-peptones), representing approximately 92% of bovine milk pro- teins; the remainder is represented by BSA and the Ig, which are transported into milk via transcellular mechanisms. All epithelial cells of the MG are thought to contribute to this biosynthetic process, and each cell expressing all of the genes is thought to synthesize and secrete the casein and whey proteins. What remains unclear, however, is whether all of the cells of the gland synthesize milk proteins at a cons- tant rate or whether synthesis is phasic. Researchers in New Zealand (33, 81) have examined the distribu- tion of milk protein gene expression in the MG of lactating sheep, beef cattle, and dairy cows. Gene expression of as1-CN, a-LA, and lactoferrin and the immunohistochemical staining of a-LA were found to be distributed heterogeneously in core biopsies taken throughout the MG of these lactating animals. This group also observed two classes (fatty and nonfatty) of alveoli having contrasting levels of a-LA gene ex- pression, suggesting that there may be specialized cell types that are capable of synthesizing milk pro- teins versus fat or that some cells within an alveolus are actively lactating (nonfatty) while others are dormant (fatty). Both scenarios probably occur, and, thus, the challenge is to delineate the exquisite local mechanisms controlling these events so that fat- producing cells can be converted into cells producing milk protein, or the proportion of these cells that are actively lactating can be increased. Intracellular Channeling of AA The intracellular compartmentalization of AA pools and the protein synthetic and catabolic apparatus have been demonstrated in many tissue types (heart, skeletal muscle, and liver) from different species [humans, sheep, mice, swine, and rabbits (26, 34, 99, 117)] under a variety of nutritional conditions (starved vs. fed; protein intake level). Moreover, com- partmentalization of AA pools for the synthesis of different cellular proteins [acid-soluble vs. muscle protein (117); albumin vs. ferritin (34)] has also Journal of Dairy Science Vol. 81, No. 9, 1998 BEQUETTE ET AL.2550 Figure 2. Comparison of the temporal labeling patterns of [D5]Phe and [13C]Phe in secreted milk casein. The 12-h constant intravenous infusion of [D5]Phe was terminated at the same time as the 7-h infusion of [13C]Phe was initiated. Beginning at h 10 of [D5]Phe infusion and continuing until the end of the [13C]Phe infusion, the goat was milked out completely by hand, and, at the first 3 milking times (h 10, 11, and 12), an intravenous dose of oxytocin (1 IU) accompanied milking. The rate constants ( k ) for the decay and incorporation curves were derived by fitting the data to the equation form: y = A(m), where m denotes (1 – e–kt) and (e–kt) for the two curves, respectively. been observed. In studies of the MG and other tissues, the selection of the appropriate precursor pool for protein synthesis and catabolism is paramount to assess the qualitative and quantitative significance of these pathways. In order to determine the precursor pool(s) for synthesis of the caseins, we compared the rates of labeling of the free AA in blood and plasma with that of secreted milk casein. In studies with lactating cows (19, 35) and goats (early and late lactation) (15) that were given a constant jugular infusion of 13C-labeled AA (Leu, Phe, Val, and Met) for 12 to 16 h, the free AA pool in blood and plasma rapidly (1 to 3 h) reached an isotopic plateau. The rise to plateau (rate of incorporation) of the [13C]AA in milk casein and whey proteins lagged substantially behind, reach- ing an asymptote with the blood and plasma only after 10 to 13 h of infusion. At 1-h intervals, the glands were milked out completely (using oxytocin injections in the cow studies). Thus, the rapid (<1 h) appearance of isotope in secreted milk proteins is indicative that the intervals between synthesis and secretion are short and that storage or residual pools are small. Similarly, slow rates of incorporation into milk casein have now been observed to occur for 13 AA (14). In fact, this may be the case for all AA because the number of AA evaluated in that study was limited by the capabilities of the mass spectro- metric analyses. The overall slow rate of appearance of [13C]AA in the secreted milk proteins suggests that there is channeling of intracellular degradation products arising from the turnover of milk proteins, constitutive proteins, or both, into the immediate precursor pool for milk protein synthesis. In a study of goats (17), the casein proteins were separated by FPLC (Pharmacia Biotech, Uppsala, Sweden) into three crude fractions representing predominantly k-CN, a-CN, and b-CN. The incorpora- tion rates were slightly different between casein frac- tions, and k-CN had the slowest rate (i.e., lower enriched). Similarly, others (63) have observed in two lactating goats that the [35S]Met specific activity of whole casein was 19 to 33% lower than in b-LG after 7 to 9 h (nonplateau conditions) of isotope infusion. These observations suggested that the casein and whey proteins may be synthesized by specialized cells receiving a source of AA that is of lower specific activity (diluted), that these proteins are synthesized by the same cells but with different sources of AA (intracellular, extracellular free vs. peptide-bound AA) contributing to their precursor pools, or that these proteins are synthesized or processed more slowly than the other proteins. One approach to differentiate further between these mechanisms would be to examine the labeling kinetic relationships under plateau conditions. Nonetheless, substantiation of the observations of Lee et al. (63) would raise the possibility that AA could be specifi- cally targeted to certain milk proteins. The channeling hypothesis has been challenged (113) on the basis that residual milk had not been completely removed in the latter studies because animals were not given repeated injections of oxyto- cin. Rather, those authors suggest that the more likely explanation is that unlabeled residual milk at the start of isotope infusion had become progressively diluted with highly labeled, newly synthesized caseins. Some milk resides in the secretory cells and in the alveolar lumen and smaller ducts even after apparent complete milk removal; however, several pieces of evidence continue to support the process of precursor channeling. First, even when oxytocin is administered, the time lag and rate of labeling of secreted milk casein are the same as in other studies. Illustrated in the modeling paper of France et al. (35) is the temporal labeling pattern of casein from a dairy cow that had been given oxytocin (5 to 10 IU) every time that she had Journal of Dairy Science Vol. 81, No. 9, 1998 SYMPOSIUM: OPTIMIZING PROTEIN NUTRITION FOR REPRODUCTION AND LACTATION 2551 been milked (machine-milking and hand-stripping at 1-h intervals) during a 20-h infusion of [13C]Leu. The rate constant for casein labeling for this cow (0.31/h) is similar to those results reported in studies (15; mean: 0.25/h) in which the goats were not given oxytocin at each hourly milking time. Second, in a preliminary experiment, we (B. J. Bequette and F.R.C. Backwell, 1994, unpublished ob- servations) found that the rate constants for label disappearance from (–0.148/h) casein and incorpora- tion into (0.137/h) casein were similar in a goat that had been milked out three times with oxytocin during the final 2 h prior to ending a 12-h constant infusion of [D5]Phe and the start of a 7-h constant infusion of [13C]Phe (Figure 2). The similar kinetic behaviors of the isotopes suggested that they labeled the same newly synthesized casein pool, which would suggest that the residual pool was small or absent. Third, the size of the residual casein pool would have to be extraordinarily large to account totally for the kinetics observed. For instance, if isotope dilution of secreted casein is assumed to have occurred only as the result of a residual pool in the gland (cellular, luminal, and ductal) and that the average rate cons- tant for casein labeling was 0.31/h for a cow produc- ing 1 kg of milk protein, then the turnover of milk protein on a daily basis would be nearly equivalent to the daily milk protein output (i. e., a protein pool in the gland that would be three times the size of the hourly rate of milk protein output (130-g pool vs. 42 g/h). This turnover would have to represent the mix- ing of old caseins with newly synthesized caseins. A more probable explanation, but one that would also be insufficient to account for the labeling kinetics observed, would involve the constant replacement of casein by degradative mechanisms rather than by the mixing of casein molecules. Nascent and mature caseins have been shown to be degraded intracellu- larly. Proteolytic cleavage of the signal (docking) sequence [7 to 10% of the molecule (27)] from nas- cent proteins occurs as part of the recognition process for targeting proteins destined for export, and studies in vitro have shown that mature caseins are also degraded (92). Based on a model of Leu kinetics of the lactating goat MG in vivo (87), it has been esti- mated that one-third of the milk proteins synthesized are degraded. Indeed, a group from the Hannah Research Institute (111, 114) has identified and iso- lated a protein from the whey fraction of residual milk, termed the feedback inhibitor of lactation, which appears to be involved in the regulation of casein degradation through its inhibitory actions on casein secretion in vitro. Finally, the high turnover rate (42 to 130%/d) that has been reported for mixed mammary tissue proteins of the early lactating goat (8, 25) would appear to be substantial enough to result in the dynamic mixing events of the intracellular pool. These measurements probably also include a population of tissue proteins that have much higher rates than the average values reported. However, even when based upon the rates reported for mixed proteins, significant recycling of labeled AA into the intracellular pool will occur, lead- ing to the curvilinear pattern of casein labeling ob- served (15). In consequence, although it is likely that some cells may retain or fail to secrete milk proteins after initiation of milk letdown and that a significant portion (maximum of 40%) of the caseins may be degraded intracellularly, it is highly unlikely that the gland stores a pool of milk proteins three times the rate of output. Thus, channeling through intermedi- ary constitutive or casein pools must occur. The existence of an intracellular residual protein pool or pools has been implicated in previous studies with mammary tissue explants (111) and with the perfused MG of the goat (95, 96). Intracellular chan- neling phenomena are not restricted to observations in ruminants because similar rate constants of casein labeling (employing [13C]Leu) of secreted milk have been observed in the lactating sow [goats 0.21 to 0.27/ h (15) vs. sows 0.24/h (52)]. Channeling has also been observed in other tissues. For example, 60% of AA for protein synthesis in muscle is derived from products of protein degradation (99). The reasons for the extensive channeling in the MG remains to be elucidated. However, based upon the fact that all AA are involved (14, 15) and that these kinetics are maintained under a range of physiological conditions, this obligatory process may be necessary to guarantee an adequate rate of delivery and concentration of AA reaching the site of casein synthesis, even through periods of AA deficiency (13). In other words, this process may represent a natural buffering system. Constitutive protein channeling would serve two pur- poses. First, it would prioritize the maintenance of cell integrity and the synthetic machinery; otherwise, unregulated casein synthesis would deplete the cell of vital structural and regulatory proteins (e. g., en- zymes, transporters, receptors, and growth factors) that are required to guarantee that milk is produced every hour and every day for the suckling young. Second, channeling of metabolically related pathways is more efficient than are free ones. This process would allow local concentrations of substrates to be increased without destruction of unstable intermedi- ates or losses from competitive side reactions, thus Journal of Dairy Science Vol. 81, No. 9, 1998 BEQUETTE ET AL.2554 On a whole body basis, however, the contribution was much greater and increased from 0.10 to 0.18. The greater conversion rates observed peripheral to the MG may suggest a greater requirement for Tyr out- side the MG tissues or may reflect the role of the liver in eliminating excess AA supplies in the body. Lys Despite many reports demonstrating the rate- limiting status of Lys on most corn-based dairy ra- tions, there is little information regarding its metabolism by the gland. This AA presents an anomaly in that, despite its often limiting situation, net extraction of Lys is significantly greater than milk protein output (45). Net accretion of Lys may occur during the latter stages of lactation when the gland undergoes apoptosis and remodeling whereby Lys could be accreted in e- (g-glutamyl)-Lys crosslink- ages (83). Alternatively, Lys may be oxidized in a regulatory or passive manner to provide ketogenic intermediates. To determine whether the excess up- take represented oxidation, we (S. J. Mabjeesh and B. J. Bequette, 1997, unpublished data) monitored Lys metabolism by the MG of the goat in late lacta- tion by employing the arteriovenous kinetic technique and found that, in response to a 5-d i.v. infusion of Lys plus Met (9 and 3 g/d, respectively), Lys oxida- tion increased from 16 to 30%, representing a nearly twofold increase in the absolute rate of oxidation. Milk protein output was not affected by infusion of these two AA; thus, Lys oxidation also appears to respond to AA supply. However, a test that is similar to that conducted in the Leu study with goats (18) needs to be undertaken to determine the obligatory nature of this process. Sulfur AA Along with Lys, Met is often considered to be one of the limiting AA of corn-based rations, particularly when heated soybeans make up most of the protein source. In addition to incorporation into protein, Met is involved in a multitude of pathways leading to synthesis of specialized compounds, such as phos- pholipids, carnitine, creatine, and the polyamines. At the same time, Met provides methyl groups for a number of transmethylation reactions involved in the regulation of DNA activity and oncogene status and sulfur groups for the synthesis of Cys. In goats, 28% of the methyl group of Met contributes to the choline pool, and 10% of Met molecules are irreversibly lost through oxidation. One consequence of the latter process is the synthesis of Cys (32). The difference in these contributions reflects the remethylation of homocysteine to produce Met. The net demands of the transmethylation routes versus transsulfuration routes may place limits on each of these pathways while the sum of the pathways alter the availability of Met for protein synthesis; thus, identification of the regulators balancing these opposing processes is im- portant. It has been examined whether provision of choline plus creatinine to sheep fed to maintenance requirements could alleviate the recycling of Met through the transmethylation pathway (66). Although less recycling was observed to occur with supplementation of the methyl donors, this decrease did not lead to a greater incorporation of [35S]Cys into wool. Recent in vitro (88, 107) and in vivo (16) evidence supports a role for peptides and proteins in the extra- cellular supply of Met for milk protein synthesis. Thus, the total uptake of Met probably balances with its output in milk protein. Conversely, the net uptake of Cys is often inadequate to meet the requirements for the synthesis of milk proteins and possibly for glutathione ( GSH) synthesis. Thus, additional in- puts of Cys may be required either from blood-derived peptides containing Cys or via MG synthesis from Met. These latter possibilities have been investigated by a group in New Zealand. During jugular infusion of [35S]Cys in goats, intracellular specific activity of GSH in MG was 30-fold greater than the blood GSH activity, implying that the MG synthesizes considera- ble GSH (58). Conversely, GSH uptake can be a source of mammary Cys. Glutathione is concentrated in the erythrocyte and appears to be transferred to tissues such as the MG via this pool (89). Glutathione may also be synthesized within the MG and transported extracellularly, and, thus, net uptake measurements may underestimate GSH influx. One approach to assess the uptake of GSH from the erythrocyte and the contribution toward Cys supply would be to infuse [35S]GSH-labeled red blood cells. The contribution of sulfur-Met to the MG synthesis of Cys via Cys synthetase has also been examined (63). In that study in goats, a-LA, which does not contain Met residues, had greater incorporation of [35S] when [35S]Met was infused close arterially (i. e., into the external pudic artery) than when the isotope was infused peripherally via the jugular vein. Also, the specific activity of the intracellular pool of free [35S]Cys was higher than that in plasma. It was estimated (63) that roughly 10% of the Met-sulfur is transferred to Cys in the whey proteins. Journal of Dairy Science Vol. 81, No. 9, 1998 SYMPOSIUM: OPTIMIZING PROTEIN NUTRITION FOR REPRODUCTION AND LACTATION 2555 Arg Arginine is extracted in the greatest quantities relative to milk protein outputs. Arginine has other metabolic functions in addition to protein synthesis. Recently, the role of Arg as a precursor of nitric oxide has received attention because of the potential effect of nitric oxide in regulating the nutrient perfusion of mammary tissue through dilation of the microvas- culature (61). The endothelial cells of the mammary vasculature and the epithelium lining alveoli and ducts exhibit nitric oxide synthase III activity, and secretory cells may therefore be capable of regulating their own local nutrient environment through altera- tion of the capillary blood supply. The MG possesses a partial urea cycle, which sug- gests an intermediary role for Arg and other inter- mediates of the cycle in MG function. In rat mam- mary tissue, the activity of arginase, which hydrolyzes Arg to form Orn and urea, increases 3-fold during lactation (55). There are two forms of the enzyme in the MG; the AII form is found in greatest quantities at midlactation. Arginase activity appears to be under the influence of other AA, and in studies (37) of rat MG (37) Pro, Orn, Lys, and certain BCAA inhibited the enzyme. The activity of this pathway may be important for the synthesis of Pro. In studies utilizing labeling of perfused sheep and goat udders, citrulline, Arg, and Orn contributed approximately 20% of casein-Pro (97, 104). This pathway may provide an alternative and perhaps crit- ical supply of Pro that is typically not extracted in adequate quantities for casein synthesis (80). The synthesis of Pro via Arg may be inherently limited, however, because, in bovine mammary tissue, the key enzyme in this pathway, ornithine-d-transferase (EC 2.6.1.13), has a high Michaelis constant (8.4 mM) , which would require that high pools of Orn be main- tained to achieve maximal rates of conversion through this pathway (10). Alternatively, the re- quirement for de novo synthesis of Pro may restrict the availability of Arg for other functions (e.g., poly- amine synthesis via Orn). Indeed, in vitro studies indicate that the contribution of Arg to Pro synthesis can be reduced when additional Pro is supplied (49). This possibility has been considered (24), and milk protein output was increased by 16% when two mid- lactation cows were given duodenal infusion of Pro (80 g/d). Although MG uptake of Arg was decreased with Pro infusion, this decrease could not be linked to the effects of Pro on mammary metabolism because changes in Pro uptake were not reported. Neverthe- less, the responses to Pro infusion are intriguing and warrant further investigation. His and MG Blood Flow The relationship between His uptake and milk out- put is quite variable, and, in part, this variation may be related to whether concentrations in whole blood or plasma are monitored for uptake comparisons (14). In the perfused guinea pig MG (30), His is not oxi- dized, and so utilization for protein synthesis or nonoxidative pathways (e.g., carnosine synthesis and deamination via histidase) predominate. However, the guinea pig MG appears to express the His cata- bolic pathways and enzymes differently from expres- sion in the cow (72). The two species vary primarily because of differences in the His degradative enzymes for the synthesis and inactivation of histamine, the latter having been shown to cause constriction of mammary pudendal arteries of goats (54). In con- trast to the laboratory species, bovine mammary tis- sue contains concentrations of histamine that parallel those of His, and levels of the enzyme responsible for this conversion, His decarboxylase (EC 4.1.1.22), also appear to parallel these changes (72). Those authors also found histamine to be localized in mast and nonmast cell tissues of the MG, but whether hista- mine functions here to initiate milk secretion and letdown, or in regulation of MG blood flow, or both, has not been determined. We have evidence that may link the His supply to the regulation of MG blood flow. Employing a goat model in which milk produc- tion response was limited by His (B. J. Bequette, 1995, unpublished data), we observed upon removal of His from the AA mixture that was abomasally infused (77 g/d for 7-d periods) that milk protein output decreased by 15 to 35%. In contrast, MG blood flow (monitored 24-h/d by flow probes) increased by 26 to 34% upon removal of His. Upon replenishment of His, milk protein output increased, and MG blood flow decreased, both returning to normal levels. Ar- terial plasma His was also reduced from 50 to 15 mM during the limitation period. One might speculate that blood flow decreased because of lower mammary tissue concentrations of histamine, which may have occurred in response to the lower concentrations of plasma His. Conversely, a reduced rate of histamine synthesis and, thus, milk protein secretion may have been the initial response to the limiting His supply. The changes in blood flow did influence MG function because, despite the lower plasma His, MG fractional extraction of His was considerably enhanced (15 to 20% vs. 85 to 95% extraction), resulting in extremely low (1 to 2 mM) plasma venous concentrations of His for some of the goats. Moreover, these findings may suggest that the MG has the ability to sense a sub- strate limitation and respond through mechanisms controlling blood flow. Journal of Dairy Science Vol. 81, No. 9, 1998 BEQUETTE ET AL.2556 IMPLICATIONS AND FUTURE PERSPECTIVES Are scientists getting any closer to identifying con- sumer acceptable ways to increase milk protein per- centage (>3%) and output consistently and to im- prove the efficiency of milk production above the current upper limit of 30%? The answer to this ques- tion is a resounding yes, but progress has been and will continue to be cautiously slow, at least to those in search of the “magic bullet”. The MG is the major controller of its own metabolic fate, but, as yet, we have not been able to identify the right combination of factors that the MG requires to take full advantage of its potential to synthesize and secrete milk. Although many have continued, some- times exhaustively, to search for limiting AA, others have taken a different approach and have begun to ask important questions such as what these limita- tions are and why and where these limitations occur. Finding the answers may require looking in places not previously contemplated and looking for metabolic roles not yet identified. The recent evidence support- ing a role for peptides or proteins in interorgan AA exchange and as transport vehicles of possibly every AA to the MG for milk protein synthesis may be one of these unforeseen places. The ability to extract AA from the blood supply, a function of transporter ac- tivity, may not be as much a limitation as the surface area that is exposed to AA in the capillary networks. Optimizing this situation may be a function of the relative distribution of blood flow to the MG on a whole body basis and that passing through the capil- lary microvasculature [i.e., the nutritive blood flow concept advanced by the New Zealand group (61)], which appears to be regulated locally within the mammary tissues, possibly via mechanisms that are capable of sensing substrate deficiencies. The cellular metabolism of the MG is quite active, with many synthetic and catabolic mechanisms to be considered, and yet we have probably only now begun to identify how these mechanisms interrelate and how impor- tant, necessary, or rate-limiting they can be. What is the potential of molecular biology to eliminate or enhance the throughput of some of these enzymatic pathways? The role of energy substrates was not con- sidered in this review, but may be equally as impor- tant as rate-limiting sources for ATP and NADPH production or for intermediary metabolic carbon sup- plies. The metabolism and functions of the MG are much more complex than we have considered. As more in- formation accumulates, greater emphasis will be placed on mathematical modeling to collate and sim- plify all of the interrelated mechanisms of metabolism and regulation (7) . The ability to represent MG bio- chemistry and physiology in mathematical terms is close to becoming a reality, but there is still the challenge of integrating these terms within the whole-animal system. Obviously knowledge and representation of the endocrine systems and their components will be required as well as ways in which these integrate or are integrated by incoming dietary nutrients (11, 12). The challenge will then be to test and represent these processes at the different organ, tissue, cellular, and subcellular levels so that ulti- mately it will be possible to evaluate quantitatively the influence of dietary nutrient intake on the pat- terns of substrate assimilation into milk components. REFERENCES 1 Agricultural and Food Research Council. 1993. Energy and Protein Requirements of Ruminants. An Advisory Manual Prepared by the AFRC Technical Committee of Responses to Nutrients. CAB Int., Wallingford, United Kingdom. 2 Armentano, L. E. 1994. Impact of metabolism by extragas- trointestinal tissues on secretory rate of milk proteins. J. Dairy Sci. 77:2809–2820. 3 Backwell, F.R.C., B. J. Bequette, D. Wilson, A. G. Calder, J. A. Metcalf, D. Wray-Cahen, J. C. MacRae, D. E. Beever, and G. E. Lobley. 1994. Utilization of dipeptides by the caprine mam- mary gland for milk protein synthesis. Am. J. Physiol. 267: R1–R6. 4 Backwell, F.R.C., B. J. Bequette, D. Wilson, J. A. Metcalf, M. F. Franklin, D. E. Beever, G. E. Lobley, and J. C. MacRae. 1996. Evidence for the utilization of peptides for milk protein synthesis in the lactating dairy goat in vivo. Am. J. Physiol. 271:R955–R960. 5 Backwell, F.R.C., M. Hipolito-Reis, D. Wilson, L. A. Bruce, V. Buchan, and J. C. MacRae. 1997. Quantification of circulat- ing peptides and assessment of peptide uptake across the gastrointestinal tract of sheep. J. Anim. Sci. 75:3315–3322. 6 Backwell, F.R.C., D. Wilson, and A. Schweizer. 1995. Evidence for a glycyl-proline transport system in ovine enterocyte brush-border membrane vesicles. Biochem. Biophys. Res. Com- mun. 215:561–565. 7 Baldwin, R. L., R. S. Emery, and J. P. McNamara. 1994. Metabolic relationships in the supply of nutrients for milk protein synthesis: integrative modelling. J. Dairy Sci. 77: 2821–2836. 8 Baracos, V. E., J. Brun-Bellut, and M. Marie. 1991. Tissue protein synthesis in lactating and dry goats. Br. J. Nutr. 66: 451–465. 9 Barnes, D. M., and D. L. Brown. 1990. Protein reserves in lactating dairy goats. Small Ruminant Res. 3:19–24. 10 Basch, J. J., E. D. Wickham, H. M. Farrell, Jr., and J. E. Keys. 1995. Ornithine-d-aminotransferase in lactating bovine mam- mary glands. J. Dairy Sci. 78:825–831. 11 Bauman, D. E., and W. B. Currie. 1980. Partitioning of nutrients during pregnancy and lactation: a review of mechan- isms involving homeostasis and homeorhesis. J. Dairy Sci. 63: 1514–1529. 12 Bauman, D. E., T. R. Mackle, M. A. McGuire, and J. M. Griinari. 1997. Amino acid supply and physiological controls of milk synthesis. J. Dairy Sci. 80(Suppl. 1):140.(Abstr.) 13 Bequette, B. J., and F.R.C. Backwell. 1997. Amino acid supply and metabolism by the ruminant mammary gland. Proc. Nutr. Soc. 56:593–605.