Effects of rumen‐protected methionine on milk production in early lactation dairy cattle fed with a diet containing 14.5% crude protein

Abstract We evaluated the influence on milk production of feeding early lactation cows a diet that included 14.5% crude protein (CP) and that did not meet methionine (Met) requirements or that met them by supplying rumen‐protected Met (RPMet). Thirty‐nine multiparous Holstein cows were allocated into two groups. For 15 weeks after calving, each group was fed one of the two total mixed rations, Control (n = 20) or Treatment (n = 19). The Treatment group received added RPMet at 0.034% (8 g/day) of the Control diet on dry matter basis. The adequacies of Met for the Control and Treatment groups were 96% and 106%, respectively, and for other amino acids, >110%. The CP level (14.5%) was 1 percentage point lower than that recommended by the Japanese Feeding Standard (2006). No between‐group differences were found in milk yield (40 kg/day), milk composition, plasma profile, rumen fermentation, nitrogen balance, or cow health. Met intake and the amount of rumen‐undegradable feed Met were higher in the Treatment group (p < 0.05). Microbial Met and total metabolizable Met did not differ between groups. Supplying RPMet in a 14.5% CP diet during early lactation did not dramatically affect milk production, because the amount of total metabolizable Met was unchanged.

The dietary level of crude protein (CP) is one of the most important factors in milk production. The amount of nitrogen (N) in dietary CP that is excreted in manure is about two to three times the amount excreted in milk (Broderick, 2003). Overfeeding of CP results in an energy cost to the animal; this cost is associated with the conversion of excess protein to urea (Dinn, Shelford, & Fisher, 1998). Despite a decrease in dietary CP content, milk production and composition were unchanged in one study (Bahrami-Yekdangi, Ghorbani, Khorvash, Khan, & Ghaffari, 2016). Reducing the intake of CP and rumen-degradable protein can decrease blood plasma TA B L E 1 Ingredient and nutrient compositions of diets urea levels (Gordon & McMurray, 1979;Westwood, Lean, Garvin, & Wynn, 2000) and is favorably associated with conception rate (Ferguson, Galligan, Blanchard, & Reeves, 1993). According to the Japanese Feeding Standard (JFS), 15.5% dietary CP is an optimal level during early lactation (NARO, 2006). Nevertheless, we have found previously that diets with 17.5%-14.5% CP and 8.0%-5.0% rumen-undegradable protein (RUP) do not affect milk production (unpublished). The amounts of N in milk and feces were not altered, whereas the amount of urinary N was increased, when cows were fed diets containing more than 14.5% CP and 5.0% RUP. We concluded that 14.5% CP, 5.0% RUP, and 9.5% rumen-degradable protein gave an optimum dietary N level. Similarly, Bahrami-Yekdangi et al. (2016) reported that approximately 9.5% rumen-degradable protein in diets provided sufficient protein to optimize milk production. However, there is growing concern about the need to validate the Met balance to maintain milk production on a 14.5% CP diet: reducing dietary CP is likely to require carefully tailoring of the dietary amino acid content-particularly the content of Met, as the first-limiting essential amino acid.
Many studies of RPMet have been performed on diets (e.g., CP 17%, Batistel et al., 2017) that include CP at levels higher than that recommended by the JFS (NARO, 2006). Furthermore, some studies have reported that feeding RPMet to cows in early lactation does not improve production (Misciatteilli et al., 2003;Ohgi et al., 2002).
However, to our knowledge, no experiments have yet investigated the Met balance when diets containing 14.5% CP or lower are fed. Our object here was to quantify the influence on milk production when cows in early lactation were fed a diet that included 14.5% CP-a level lower than that recommended by the JFS-and that met, or did not meet, the Met requirement by supplying RPMet.
N. balance and cow health were also investigated.

| Experimental design, animals, and diets
There were two kinds of experiment: (a) production trials and ( Cows had free access to the ration. Cows were milked twice daily.
Metabolism trials were conducted for 3 days consecutively during a 14-to 17-week period after parturition. The cows used in the metabolism trial were the same as those used in the production trial; they were fed the same TMR as the one to which they had been assigned for the production trial. During the 3 days of sampling for the metabolism trial, the cows used were not released for exercise but were housed in individual tie-stalls.

| Sample collection and data recording
In the production trial, TMR intake and milk yield were recorded daily. Milk samples were collected weekly at two consecutive milkings until 15 weeks. Body weight was measured each week until 15 weeks. Blood samples were collected via the jugular vein approximately 4 h after feeding at 1, 3, 5, 9, and 13 weeks. The samples were placed on ice immediately, and the plasma was obtained by centrifugation. Ruminal juice samples taken from the ventral sac at the time of blood sampling were strained through two layers of cheesecloth. Ruminal fluid pH was measured immediately with a pH meter (F-22; Horiba Ltd., Kyoto, Japan). Blood plasma and ruminal samples were frozen at −20°C until analyzed.
For the metabolism trial, total feces were collected, TMR intake and total fecal weight were recorded, and TMR and fecal samples were collected daily. Milk yields were recorded, and samples were taken at each milking. Urine samples were taken in one of two ways, namely as total urine samples or as spot urine samples. Total urine was collected with a vulva urine cup (Sanshin Industrial Co. Ltd., Yokohama, Japan) into a container containing 700 ml of 20% (v/v) H 2 SO 4 . Total urine weight was recorded and the sample obtained every day. Spot urine samples were collected from spontaneous urination by manual stimulation of the groin every day. Five milliliters of 20% (v/v) H 2 SO 4 was added per 100 ml of spot urine. Urine samples were frozen at −20°C until analyzed. Samples (other than the wet samples of feces used for analysis of total N) were prepared as dried samples: wet feces and TMR were dried for 48 h at 55°C, ground, and then passed through a 1-mm screen.
Health disorder data were recorded from after parturition until the end of the metabolism trial.

| Sample analysis and calculations
DMI was measured in the same way as reported in a study by Koga et al. (2001). Samples of TMR and feces were analyzed according to standard methods (AOAC, 1990). Samples of control diet, RPMet, and rumen- Hitachi High-Technologies Corp.). Total urine excretion from spot urine was estimated from creatinine and body weight by using the method of Tamura, Inoue, Shinohara, and Koga (2007). Urine samples were analyzed for allantoin, and rumen microbial N outflow into the intestine was determined by using the method of Chen (1989).
Amounts of metabolizable Lys and Met were estimated from the amount of rumen-incubated residual (as measured by using the nylon bag technique; Lykos & Varga, 1995) and the quantity of absorbable microbial protein in the intestine. The specific method of estimation was as follows: Control diet samples were ground to 2 mm, whereas RPMet samples were left as unground granules. Approximately 7.5 g of control sample or RPMet was placed into a nylon bag. Four bags per sample type (control diet and RPMet) were prepared. The bags were incubated in the rumens of two fistulated cows. The cows were multiparous and in midlactation, and were being fed the Control diet. In each cow, one set of control and RPMet bags was incubated from 06:00 to 18:00 hr and the other set from 18:00 to 06:00 hr. Ruminal disappearance of Lys and Met was calculated as the difference between Lys and Met contents of the initial samples and the residues after incubation in the rumen. We assumed that incubated residing amino acids flowed on to the intestine. The amounts of these amino acids (estimated absorbable rumen-undegradable feed Met or Lys) were calculated by using DMI and the rate of ruminal disappearance. It was assumed that 80% of the amino acid flow is absorbed (NRC, 1989) and 64% of the microbial CP outflow is absorbed as protein (NRC, 2001

| Statistical analysis
Data in production trials were analyzed by using the MIXED procedure in SAS (SAS Institute, Inc., Cary, NC, USA) according to the following model: where Y ijklm = the dependent variable, μ = the overall mean, D i = fixed effect of diet (i = control or treatment), W j = fixed effect of week,

| RE SULTS
Some cows exhibited health problems such as mastitis and displaced abomasum during the production and metabolism trials.
Those cows were excluded. At the end of the production trial, data from 20 Control group cows and 18 Treatment group cows were included in the analyses. In the metabolism trial, data sets from 19 Control group cows and 17 Treatment group cows were used.

| Production
Body weight, DMI, and production parameters are presented in Table 2. No item differed between the two groups (p > 0.05).
Daily intake of RPMet in the Treatment group was approximately 8 g by DMI (23.7 kg). Milk yield was over 40 kg/day with each diet.

| Blood plasma and Ruminal fluid
Blood plasma profiles are shown in Table 3. Total protein content differed between the two diets (p < 0.05). Ruminal fermentation characteristics are shown in Table 4. The caproate content as a percentage of VFAs differed between the two diets (p < 0.05).

| Digestibility
Digestibility and TDN are shown in Table 5. All digestibility and TDN did not differ significantly between the two diets (p > 0.10).

| Microbial N and metabolizable Lys and Met
The daily amounts of urinary allantoin and N did not differ between the results obtained from the total urine samples and the estima-  Table 6. Excretion of urinary allantoin and microbial N flows to the intestine did not differ between the two diets (p > 0.10). The amount of estimated total metabolizable Lys did not differ (p > 0.10). The intake of Met and the amount of estimated absorbable rumen-undegradable feed Met in the Treatment group were significantly higher than in the Control group (p < 0.05). Estimated absorbable microbial Met and estimated total metabolizable Met did not differ between the two groups (p > 0.10).

| Nitrogen balance
The N balance data are shown in Table 7. Input and output N did not differ between the two diet groups (p > 0.10). The

TA B L E 3 Blood plasma clinical chemistry profiles
N partition ratio as a percentage of N intake did not differ (p > 0.10).

| Health
The incidence of health disorders is shown in Table 8. The incidence of individual health disorders did not differ between dietary groups (p > 0.05). The percentage of animals with mastitis tended to be higher in the Control group (p < 0.10).
According to AminoCow, using the data from the production trial (DMI, milk yield, milk content of protein and fat), only the adequacies of Lys and Met were less than 110%. The adequacies in the Control and Treatment groups were, respectively, for Lys, 109% and 108%, and for Met, 96% and 104%. Only the Met requirement in the Control group was lacking (<100%) as designed, and supplementing with RPMet meant that the requirements for all essential amino acids were met. However, from our results we infer that feeding our 14.5% CP diet with RPMet does not affect milk production.
The results of the plasma analysis (Table 3) were within the normal reference ranges (Merck, 2012;Smith, 2009). The data on plasma glucose and nonesterified fatty acid concentrations agreed with the results reported by Socha et al. (2005), who found that these parameters were not affected when cows in early lactation were duodenally infused at 10.5 g/day with Met and 10.0 g/day with Lys. However, Sun et al. (2016) reported that RPMet supplementation at 15 g/day decreased the same parameters. Sun et al. (2016) also observed that total cholesterol was decreased by the addition of RPMet. In addition, Batistel et al. (2017) reported that plasma γ-glutamyl transpeptidase levels decreased when cows were supplemented with 15 g/day of RPMet. From these previous reports and our results, we infer that some plasma items might change when diets contain RPMet at 15 g/day or more, but supplementation at the rate we used did not affect plasma profiles.
With the exception of the result for caproate, the ruminal fluid profiles (Table 4) were in agreement with the results of Armentano, Bertics, and Ducharme (1997), who reported that all ruminal fluid parameters were unaffected upon supplementation with RPMet at 10.5 g/day. On the other hand, Chung et al. (2006) reported in an in vitro experiment that the contents of total VFAs and some individual VFAs were altered by the addition of rumen-protected Met. With the exception of the values for N, the apparent digestibility percentages (Table 5) were close to previous findings: dry matter and organic matter, (Ha & Kennelly, 1984;Miyaji, Matsuyama, Hosoda, & Nonaka, 2012), ether extract, NDFom and ADFom (Miyaji & Matsuyama, 2016;Miyaji et al., 2012). However, the results for apparent N digestibility were lower than in other reports. N digestibility has been reported as 67.6%-68.9% on a 15.5% CP diet (Miyaji et al., 2012) and 66.1% on a 13% CP diet (Ha & Kennelly, 1984). Our results support the suggestion by Ha and Kennelly (1984) that apparent N digestibility is elevated when dietary protein increases, because a higher protein content increases microbial fermentation in the rumen. From the results for the ruminal fluid profiling and digestibility, we can say that supplying RPMet did not affect ruminal fermentation.
The results for allantoin (Table 6) agree with those of Krober, Kulling, Menzi, Sutter, and Kreuzer (2000), who reported that urinary allantoin content did not differ between cows fed additional RPMet and those fed a control diet. Analysis of estimated metabolizable amino acids ( the results of both the metabolism trial and the production trial suggest that total metabolizable Met did not differ between the groups. The input and ouput results and the N partition ratios (Table 7) indicated that supplying RPMet on a 14.5% CP diet did not affect the N balance. In this connection, urinary N outputs in our cows were less than 159 g/day, although the milk yield was approximately 40.5 kg/ day. Our previous studies have also shown that a diet including 14.5% CP and 5.0% RUP results in milk yields of more than 40 kg/day and in a urinary N content of more than 155 g/day during early lactation (unpublished data). Our urinary N outputs were lower than that in a previous report: the output was 177 g/day and the milk yield was 41.0 kg/day when cows were fed a diet including 16.3% CP (Miyaji & Matsuyama, 2016). According to JFS NARO, 2006, the recommended CP level in accordance with the results of our production trial (multiparous cows, body weight 660 kg, DMI 24.5 kg/day, milk yield 40.5 kg/ day, milk fat 3.7%) is 15.5%; the level in our cows was thus about 1 percentage point lower. Thus a diet including approximately 14.5% CP-lower than recommended-can result in milk yields of 40 kg/day during early lactation. In contrast, dietary N that was not needed for amino acids for production is excreted as urinary N. We can therefore also conclude from the results for milk urea N (Table 2), plasma urea N (Table 3), and urinary N ( Table 7) that feeding RPMet on a 14.5% CP diet does not improve the N balance under these experimental dietary conditions.
Mastitis (Table 8) was the only infectious disorder with a (nonsignificantly) higher prevalence in Control animals than in the Treatment group. Osorio et al. (2014) and Sun et al. (2016) reported that dietary supplementation with RPMet improves immune function.
However, Batistel et al. (2017) found that the frequency of mastitis was unchanged when RPMet at 22 g/day was added, and Sun et al. (2016) found that the milk somatic cell count was unchanged when RPMet was fed at 15 g/day. Moreover, it is well known that mastitis is prevented mainly by using appropriate milking procedures.
Consequently, we infer that the incidence of health disorders -including mastitis -was not altered by RPMet, because the amount of total metabolizable Met was similar in the dietary groups.
In conclusion, we found here that dietary supplementation with RPMet did not dramatically affect milk production, plasma profile, rumen fermentation, nitrogen balance, or cow health when cows during early lactation were fed approximately 14.5% CP, because the amount of total metabolizable Met did not change. We also demonstrated that milk yields of more than 40 kg/day can be achieved despite a decrease in the dietary CP level by about 1 percentage point from the 15.5% recommended by JFS and in the absence of supplemental RPMet.

ACK N OWLED G EM ENTS
We thank Mitsuto Matsumoto (Japan Association for Technoinnovation in Agriculture, Forestry and Fisheries) for reading the manuscript, and Fuminori Terada (Tohoku University) for providing advice on the statistical analysis.