Nitrogen metabolism and mammary gland amino acid utilization in lactating dairy cows with different residual feed intake

Objective This study was conducted to enhance our understanding of nitrogen (N) metabolism and mammary amino acid (AA) utilization in lactating cows with divergent phenotypes of residual feed intake (RFI). Methods Fifty-three multiparous mid-lactation Holstein dairy cows were selected for RFI measurements over a 50-d experimental period. The 26 cows with the most extreme RFI values were classified into the high RFI (n = 13) and low RFI (n = 13) groups, respectively, for analysis of N metabolism and AA utilization. Results Compared with the high RFI cows, the low RFI animals had lower dry matter intake (p<0.01) with no difference observed in milk yield between the two groups (p> 0.10). However, higher ratios of milk yield to dry matter intake (p<0.01) were found in the low RFI cows than in the high RFI cows. The low RFI cows had significant greater ratios of milk protein to metabolizable protein (p = 0.02) and milk protein to crude protein intake than the high RFI cows (p = 0.01). The arterial concentration and mammary uptake of essential AA (p<0.10), branched-chain AA (p<0.10), and total AA (p<0.10) tended to be lower in the low RFI cows. Additionally, the low RFI cows tended to have a lower ratio of AA uptake to milk output for essential AA (p = 0.08), branched-chain AA (p = 0.07) and total AA (p = 0.09) than the high RFI cows. Conclusion In summary, both utilization of metabolizable protein for milk protein and mammary AA utilization are more efficient in cows with lower RFI than in the high RFI cows. Our results provide new insight into the protein metabolic processes (related to N and AA) involved in feed efficiency.


INTRODUCTION
Improving protein efficiency is an established goal in the dairy industry, as it is expected to increase profits and environmental benefits. Residual feed intake (RFI) has been a popular indicator of feed efficiency in recent studies [1,2] and is calculated as the difference between the actual and predicted dry matter intake (DMI) [3]. Accumulating evidence indicates that cows with lower RFI have higher efficiency of utilization of protein or nitrogen (N) [4,5]. However, the underlying mechanism is not clear.
Metabolism of N or amino acid (AA) is an important biological process. It is strongly associated with N efficiency and may directly affect RFI. Animals with low RFI (LRFI) consume less dry matter (DM) than high RFI (HRFI) animals [6]. Studies show that micro bial protein (MCP) synthesis increases with increasing levels of feed intake [7]. However, increased feed consumption may decrease nutrient digestion in the rumen. Both MCP and metabolizable protein (MP) are vital to the lactation of dairy cattle [8]. In our previous study, we found no significant difference in rumen fermentation parameters or MCP pro duction between LRFI and HRFI cows [9], while cows with LRFI values had higher ratio of milk protein to crude pro tein (CP) intake. Thus, the steps of MCP utilization for milk production, including the efficiencies of conversion of MCP to MP and of MP to milk, may potentially contribute to the greater N efficiency in LRFI cows. It is hypothesized that variation exists between LRFI and HRFI cows in MCP and MP utilization.
The N efficiency of dairy cows may be improved by in creased utilization of AA in mammary gland [10]. However, information regarding AA utilization in mammary gland by lactating cows of differing RFI is lacking. Measurements of AA concentrations and mammary plasma flow (MPF) may provide an indication of AA utilization in mammary gland. Therefore, the objective of the current study was to investi gate the relationships between RFI and MCP synthesis or mammary AA utilization in midlactation dairy cows. Un derstanding the utilization of N and AA in lactating cows with divergent RFI may provide a new insight into metabolic differences among cows differing in RFI and is essential to effectively utilize RFI in production systems.

Animals and management
The animal care and experimental procedures were approved by the Animal Use and Care Committee of Zhejiang Univer sity (Hangzhou, China, No. 12410). Fiftythree multiparous midlactation Holstein cows, with body weight (BW) of 634 (±85 kg, standard deviation) and days in milk of 153 (±20, standard deviation) were selected for the experiment. The experiment lasted for 57 d, including a 7d period for adaptation. During this period, the DMI of each cow was determined daily by an automatic feed system (Zhenghong Co., Shanghai, China). All cows were housed in a freestall barn with access to a total mixed ration, and were fed three times a day (06: 30, 14:30, and 21:30). Animals were weighed immediately after the morning milking every week.  [11]. The chem ical composition of the diets is listed in Table 1.

Sampling and measurements
Milk: The milk yield of each cow was recorded at each milking (06: 00, 14:00, and 21:00). Each week, milk samples were obtained on 2 consecutive days at each milking and then pooled. One subsample was stored at 4°C for analysis of fat, lactose, protein, somatic cells, and urea N with infrared spectroscopy (Foss4000; Foss Electric A/S, Hillerød, Den mark). Another subsample was frozen at -20°C for analysis of AA.
Analysis of plasma and milk amino acid: Blood samples from coccygeal artery and the subcutaneous abdominal vein were collected from each cow for two consecutive days at approximately 07:00, 15:00, and 22:00. Blood was immedi ately put on ice until centrifugation (3,000×g) at 4°C for 15 min), and the plasma was stored at -20°C for later analysis. The pooled plasma was analyzed for AA by previously de scribed methods [14]. Briefly, an aliquot of 1 mL of plasma was deproteinized with 10% sulfosalicylic acid (1:1, plasma to 10% sulfosalicylic acid). Samples were then centrifuged at 10,000×g at 4°C for 15 min. The supernatant was filtered through 0.45 μm and 0.22 μm nylon syringe filter units (Fisher Scientific, Pittsburgh, PA, USA) and placed in microcentri fuge tubes (Fisher Scientific, USA). Before analysis, milk was hydrolyzed by adding 6 N HCl and incubating at 110°C for 24 h [11]. The AA concentrations of plasma and milk were analyzed using an automatic AA analyzer (Hitachi High Technologies Corporation, Tokyo, Japan).

Calculations
Residual feed intake: The RFI value of each animal was cal culated as described previously [9]. Briefly, the RFI was estimated as the difference between expected feed intake and actual feed intake. Expected feed intake was computed using a multiple linear regression model, regressing DMI on measures of energycorrected milk (ECM) yield, meta bolic BW (BW 0.75 ), and average daily gain (ADG) over the measurement period: where, Y i is the average DMI of the ith animal, α 0 is the in tercept, α 1 , α 2 , and α 3 are the partial regression coefficients for ECM, BW 0.75 , and ADG, respectively, and e i is the ran dom error associated with the ith animal. The ADG was calculated as the slope from the regression of BW in the ex perimental period. All animals were ranked by RFI. With a power of 99% for RFI, the 13 lowest RFI and 13 highest RFI cows were selected to form two RFI groups: LRFI (high effi ciency, n = 13) and HRFI (low efficiency, n = 13). Microbial protein and metabolizable protein: The MCP was estimated from PD excretion in urine [12]. Creatinine has been validated as a marker to estimate urine volume and is assumed to be excreted at a rate of 29 mg/kg of BW [15] for calculating the urine volume excretion rate. The MP was calculated as the sum of the intestinally absorbable dietary protein (IADP) and intestinally absorbable MCP (IAMCP). The IADP was calculated as follows: IADP = rumen unde graded protein (RUP) content × CP intake × IDP, where IDP is the intestinal digestibility of RUP, determined accord ing to the threestep in vitro procedure [16]. The IAMCP was calculated using the equation IAMCP = MCP×0.64 [8].
Amino acid utilization: The MPF was estimated using the Fick principle [17]. The equation was as follows: The indices reflecting mammary AA utilization were cal culated as follows: Mammary uptake (g/d) = MPF (L/d)×arteriovenous difference (AVD, g/L) U:O ratio = mammary AA uptake (g/d)/milk AA output (g/d)

Statistical analysis
The variation of the data in this study was analyzed using the general linear model procedure of SAS according to the model Y i = μ+β i +e, where Y i is the dependent variable, μ is the overall mean, β i is the fixed effect of RFI group, and e is the residual error. Significance was considered at p≤0.05, and a tendency was defined as 0.05<p≤0.10.

Feed intake and lactation performance
The DMI and productivity of the cows selected for high or LRFI are presented in Table 2. The cows with lower RFI values had lower DMI than the HRFI ones (p<0.01) but greater ra tios of milk to DMI (p<0.01) and ECM to DMI (p<0.01).
The cows with lower RFI values had lower milk urea N than the HRFI cows (p = 0.05). The percentages of milk fat (p = 0.04) and milk protein (p = 0.05) were lower in cows with lower RFI, but the yield of milk, milk fat and milk protein were similar between the two groups (p>0.10). protein

Production of microbial protein and metabolizable
The urinary PD data, estimated MP yield and MP utilization efficiency are shown in Table 3. No difference was found be tween the two groups in urinary PD (p = 0.21), MCP yield (p = 0.21) or MP supply (p = 0.23). The LRFI cows had lower IADP (p = 0.04) but higher proportion of MP to milk pro tein (p = 0.02) and greater ratio of milk protein to CP intake than the HRFI group (p = 0.01, Table 3). However, the parti tioning of RDP to MCP (p = 0.64) and dietary protein to MP (p = 0.73) was not different between the two groups.

Mammary utilization of amino acid
The arterial concentrations of leucine (p = 0.07), Val (p = 0.08), essential AA (EAA, p = 0.06), branchedchain AA (BCAA, p = 0.10) and total AA (p = 0.07) tended to be lower in the LRFI cows than in HRFI cows ( Table 4). The arterial supply of all AA was not affected by RFI (p>0.10). The LRFI cows tended to be lower in AVD of leucine (p = 0.07), lysine (p = 0.08), EAA (p = 0.07), BCAA (p = 0.10) and total AA (p = 0.10, Table 4), but no difference was observed in the other AA between two groups. The MPF did not differ between the two groups (p>0.10).

DISCUSSION
Investigation of the conversion efficiencies of dietary CP into MP and of MP to milk protein among cows with diver gent feed efficiency is instrumental for improving animal N efficiency. The lower IADP observed in the LRFI group in our study can be attributed to the lower consumption of feed in this group. The MCP, RUP and endogenous CP contribute to the passage of MP to the tissue, and MCP accounts for the majority of the MP flow [8]. However, we found no difference in MCP yield between the two groups, which may have con tributed to the similar MP supply between the groups. A higher ratio of milk protein to MP was found in the LRFI animals, which consumed less DM. Our results are consistent with Rius et al [18], who reported greater N utilization and efficiency of MP in cows with lower N intake. The higher ratio of milk protein to MP in the LRFI cows indicated more effi cient utilization of AA in mammary gland in these cows. Absorbed AA are provided by MCP, and RUP is vital as a precursor for protein synthesis. The AA utilization in mam mary gland is high but varies among individual animals [19], indicating the potential for improvements of AA utilization efficiency in mammary gland. However, few studies have reported the relationship between efficiency of mammary AA utilization and RFI. Many studies have been conducted to improve AA utilization efficiency through nutritional manipulation, such as infusing AA or hormones [20,21] and altering dietary levels of energy or protein [22]. In com parison with nutritional manipulation, genetic selection may provide cumulative, longerterm enhancements of traits [23]. Therefore, the relationship between RFI and AA utilization efficiency in mammary gland warrants investi gation. 2) Group 1, sum of methionine, phenylalanine, tyrosine, and histidine; Group 2, sum of arginine, valine, isoleucine, leucine, lysine, and threonine.
The AA availability, AA uptake and protein synthesis are factors regulating AA utilization in mammary gland [19]. Doepel and Lapierre [20] reported a negative relationship between MPF and EAA supply, suggesting a compensation mechanism of EAA and blood flow in the mammary gland. In the current study, the content of arterial plasma free AA tended to be lower in the LRFI cows, corresponding to the lower DMI in this group. Our results are consistent with Martineau et al [24], who reported that a low MP supply de creased the concentration of AA. However, the supply of arterial plasma AA to mammary gland was similar between the HRFI and LRFI animals, a finding possibly attributable to the similar MPF between the groups.
In mammary epithelial cells, BCAA is a potential source of energy for highly active metabolic processes and nitroge nous precursors for the NEAA synthesis [25]. Studies have shown that BCAA plays a vital role in cell signaling. Specifi cally, leucine and Isoleucine can stimulate protein synthesis with the phosphorylation of mammalian target of rapamycin [26]. Thus, a U:O ratio of BCAA higher than 1.00 indicates its potential use for NEAA synthesis and regulatory signals. Although we observed no difference in milk yield between the RFI groups, BCAA uptake was higher in the HRFI cows. Several studies have demonstrated a decreased efficiency of mammary utilization of AA and increased catabolism of AA when the supply of MP or AA is increased [27,28]. Thus, the higher uptake of BCAA in the HRFI cows and the higher DMI in this group could be attributable to higher BCAA catabolism in this group. However, the molecular mecha nisms underlying the difference in AA utilization in mammary gland between cows of differing RFI could not be deter mined from our study. Recent studies have reported that transcriptomics is a powerful tool to investigate gene ex pression related to feed efficiency [29,30]. Thus, further research using transcriptomics is needed to investigate the regulatory genes and molecular mechanisms underlying the difference in mammary AA utilization among cows of divergent RFI.
In conclusion, the results of the current study revealed variation in N utilization and AA utilization efficiency in mammary gland in lactating dairy cows with divergent RFI values. The cows with lower RFI values had lower DMI than, but similar milk yield to the higher RFI cows. Despite similar MCP and MP production between the groups, the ratio of milk protein to MP was higher in the lower RFI cows, which is consistent with the higher efficiency of AA utilization by mammary gland observed in this group. Further research is needed to reveal the underlying mechanisms.

CONFLICT OF INTEREST
We certify that there is no conflict of interest with any financial organization regarding the material discussed in the manu script.