This study was conducted at Shengsheng Dairy Farm in Luoyang City, Henan Province, China. Animal management and experimental procedures were approved by the Animal Care Committee of China Agricultural University. A total of 96 lactating Holstein dairy cows (63 ± 25 d in milk; 34.4 ± 5.74 kg/d of milk production; mean ± SD) were randomly assigned to 1 of 2 dietary treatments and housed in two separate free-stall barns. The experiment lasted 75 d and consisted of 15 d for diet adaptation and 60 d for data collection. Diets were fed as total mixed ration (TMR; Table 1): diet containing 17.3% CP without RPM (control group; CON; n = 49) or diet containing 16.4% CP and supplemented with 15.0 g/d of RPM (treatment group; RPM; n = 47). The CON diet had a Lys:Met ratio of 3.39:1, while the RPM diet was 2.84:1. The RPM used in the study was Mepron^® provided by Evonik (China) Co., Ltd. Mepron^® is produced by coating methionine with a protective film. The TMR was fed three times daily at 0800, 1,200, and 1,700 h. Cows had free access to feed and water. The amount of feed offered daily was adjusted to achieve a 3–5% orts.

Cows were milked 3 times per day for the duration of the experiment. Milk samples were collected from each cow 3 times per day at 0700, 1,100, and 1,600 h and mixed well in a 4:3:3 ratio as one sample, then analyzed for fat, true protein, lactose, milk urea nitrogen (MUN), somatic cell count (SCC), solids-not-fat (SNF), total solid (TS) and freezing point (FP) at the Luoyang Dairy Cow Center (Luoyang, China). The TMR samples were collected 3 times per day during the experimental period and mixed well to determine the feed composition. Samples were dried at 65°C for 48 h until a constant weight was obtained. Then, dried TMR samples were ground through a 1-mm screen (KRT-34; KunJie, Beijing, China) and analyzed for DM using method 950.15 of Association of Official Analytical Chemists (15). Nitrogen was analyzed by the method 984.13 (15) and crude protein (CP) was measured by multiplying 6.25 by the nitrogen content. Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were measured according to the method described by Van Soest et al. (16). Ash (method 942.05) was measured according to the AOAC (17).

Fifteen Holstein dairy cows per group were randomly selected with similar milk yield and days in milk to collect blood, fecal grab and rumen fluid samples. On d 0, 30 and 60 of the experimental periods, blood samples were collected from the coccygeal vein into 10 mL heparin-coated tubes (Vacutainer; Becton Dickinson, Franklin Lakes, NanJing, China) before morning feeding. Blood samples were immediately centrifuged at 3,500 × g at 4°C for 15 min to obtain plasma samples and stored at −20°C for further analysis. Concentrations of total protein (TP), albumin (ALB), globulin (GLB), blood urea nitrogen (BUN), cholesterol (CHO), aspartate transaminase (AST), alanine transaminase (ALT), and alkaline phosphatase (ALP) were measured at a commercial laboratory (People's Liberation Army 534 Hospital, China).

Fecal grab samples were collected 4 times (0300, 0800, 1,300, and 1,800 h) each day on d 58, 59, and 60 from these selected cows. The samples of each cow were evenly mixed to make a simple composite and taken to about 200 g, adding 10% tartaric acid of 1/4 fecal weight to fix nitrogen. These fecal grab samples were held at 65°C in a forced-air oven until completely dried and were ground to pass a 1-mm screen (KRT-34; KunJie, Beijing, China). Fecal samples were analyzed for DM, NDF, ADF, CP, and ash according to the AOAC as described earlier. The acid-insoluble ash (AIA) was used as an internal marker to determine the apparent digestibility of nutrients. TMR, orts and fecal samples were analyzed according to the procedures by Vankeulen and Young (18). The apparent digestibility of nutrients was calculated as follows:

where Ad = AIA in the diets (g/kg); Af = AIA in the feces (g/kg); Nd = the concentration of a nutrient in the diet (g/kg); Nf = the concentration of a nutrient in the feces (g/kg).

On the last day of the experiment, rumen fluid samples were collected from the selected 15 cows per group. Samples were collected before morning feeding and 2, 4, 6, 8, 12 h after morning feeding by a 200 mL esophageal tube. Rumen fluid was filtered through four layers of gauze, centrifuged at 1,800 × g for 15 min and 1 mL of the supernatant collected and acidified with 4.5 mL of 0.2 mol/L HCL for later analysis of ammonium nitrogen. Meanwhile, l mL 25% metaphosphate acid was added to 4 mL of the supernatant and stored at −20°C for later analysis of microbial crude protein (MCP), ammonia N (NH₃-N) (analyzing in the Evonik (China) laboratory, Beijing) and volatile fatty acids (VFA), including acetate, propionate, butyrate, valerate, isobutyrate, and isovalerate (using Shimadzu GC-7A gas chromatograph, Japan).

All data were analyzed using SAS (SAS version 9.2, SAS Institute Inc., Cary, NC, USA) with cow as the experimental unit. Milk yield, milk composition, apparent digestibility of nutrients, blood parameters, and rumen fermentation products were analyzed using the PROC MIXED procedure of SAS. Milk yield, milk composition and blood parameters were calculated by averaging the samples collected in d 0, d 30 and d 60 and then analyzing. The model of rumen fermentation products included the fixed effects of treatment, time, and time × treatment interaction, and cow within treatment as a random effect. Degrees of freedom were counted using the Kenward-Roger approximation option of the MIXED procedure. In order to explicate the repeated measures within-subject, the covariance structures were executed for each repeated variable based on the best fit determined by the Akaike information criterion. A significant difference between the treatment and control group was declared at p < 0.05 and tendencies were considered when 0.05 ≤ p < 0.10. A highly significant difference was indicated at p ≤ 0.01.

The results of the milk yield and composition are shown in Table 2. Milk yield, fat, protein, lactose, TS, MUN, SCC, and FP were similar between the CON and RPM groups. There was a trend for lower SNF with decreased CP feeding and RPM supplementation (p = 0.09). Our findings implied that decreasing dietary CP to 16.4% with RPM supplementation and a relatively constant fermentable carbohydrate failed to affect the milk yield and composition negatively. The results are consistent with other trials that found milk production performance was similar when dietary AA was well balanced, and differences in CP of the diets ranged between 13 and 18.6% (2, 9, 11). Moreover, several trials have demonstrated that RPM supplementation increased milk (7, 19) and milk protein (20) yields. In contrast, it has been reported that milk yield or milk composition, including fat, lactose and SNF, were significantly decreased when cows were fed 11% CP as opposed to 13, 15, and 17% CP (9). The diets of 11, 13, 15, and 17% CP were supplemented with RPM, which provided 4.2, 8.1, 10.3, and 12.4 g/d of absorbed Met respectively.

Although certain ranges of difference in CP might not influence milk yield, several studies have highlighted the high sensitivity of SNF to dietary CP content. As has been shown by researchers who fed cows tat 17.3 or 16.1% CP (11), reducing dietary CP might significantly reduce the milk SNF yield, which is a key factor that could affect milk payment. Results from the current study differed from the report of Broderick et al. (11) because cows were predicted to be in negative balance in their study, whether energy or mean N. However, in the current study, diets were nutritionally balanced in consideration of normal milk yield and DMI.

The role of EAA in MUN has been discussed widely. MUN is a key indicator that can estimate whether a diet provides excessive protein (21). A previous study found that MUN concentration was lower in cows fed supplementing rumen-protected essential amino acids while reducing dietary protein (22). Compared with feeding an 18% CP diet, MUN values decreased in lactating cows fed either 16.4 or 15.6% CP diets supplemented with RPM (2). An earlier study reported no difference in MUN concentration in cows fed diets that differed by 1% in dietary CP content (23). The similar content in MUN in our study could be explained by the small differences (0.9%) in dietary CP between the two experimental groups. The SCC level, an indicator of mammary health, often increases when infection occurs in the mammary gland (24, 25). The SCC values were within the normal ranges (100,000–314,000 cells/mL) (26) between the two groups.

The effects of supplementing RPM on blood parameters are shown in Table 3. The BUN concentration was lower (p < 0.001) in RPM cows compared to the CON group. Other blood parameters, such as ALT, AST, TP, ALB, GLB, CHO and ALP, did not exhibit any differences between groups. This result was not anticipated, but it is in agreement with published data (2).

The decline in BUN in our study accords with a previous study in which cows were fed a low-protein diet supplemented with RPM while reducing protein level (2, 27). Lower dietary CP and rumen degradable protein (RDP) are most likely to reduce the BUN concentrations in lactating dairy cows (28). However, reducing dietary protein should be accompanied by balancing dietary AA, especially the Lys to Met ratio, to enhance protein utilization in dairy cows (22), as we did in the current study by supplementing RPM in the diets. Low BUN has been associated with improved protein utilization (29), and reduced BUN concentrations could stimulate rumen urea transfer and increase urea transport rates (30). In addition, some previous studies reported that reduced BUN and MUN could promote reproductive performance in dairy cows (31, 32).

The results of the apparent digestibility of nutrients in dairy cows are shown in Table 4. The total apparent digestibility of DM, CP, NDF, ADF and ash were not affected by the treatment, despite the difference in CP levels.

Feed digestibilities of OM, NDF, and ADF have been associated with dietary CP content (11). Low CP (<15% of DM) and RDP (<10% of DM) concentrations can reduce NDF digestibility (11, 33). Earlier studies showed that a diet with 9% RDP lowered OM digestibility (34). Inadequate RDP could have a negative effect on utilizing NH₃-N, then low NH₃-N efficiency could reduce fiber digestion and microbial growth (35). However, control and RPM diets used in our research might have provided adequate RDP, 11.7 and 11%, respectively, resulting in no impact on nutrient digestibility. Our results are consistent with a previous study which showed that the apparent digestibility of DM, OM, NDF, CP and starch were not different whether or not RPM was added and RDP was adequate (10% of DM) (33). In line with the above-mentioned results, more urea was recycled to the rumen when dietary CP was higher than 16%, and ruminal NH₃-N concentrations in Table 5 were above the minimum requirement (≥ 5 mg/dL) of ruminal NH₃-N concentration for rumen microbial growth (36) to further support the findings of this study. Although a large proportion of protected methionine escaped ruminal degradation, a small fraction of methionine was still released into the rumen. Salsbury et al. (37) observed that unprotected supplemental methionine enhanced ruminal bacteria growth rate in vitro. So, supplemental rumen-protected methionine may change the community composition of the rumen microbiota and their metabolism, and feed digestibility will be expected to increase with the activity increasing of rumen bacteria. However, our results showed no difference on feed digestibility between RPM and CON group. They indicated that a small fraction of methionine released from the rumen-protected supplement may not have affected the growth of major bacterial species in the rumen (10). No effect on nutrient digestibility by supplemented RPM was observed when diet CP was higher than 16%.

Effects of experimental diets on rumen fermentation products are shown in Table 5. The results showed that MCP in the RPM group was higher than in the CON group (p = 0.006). All rumen fermentation products were significantly affected by time (p < 0.05), but no differences were found between treatments. Treatment × time interaction effects (p < 0.05) were observed for ruminal butyrate and isovalerate concentrations, with cows fed RPM having higher butyrate and isovalerate concentrations 2 h after feeding, as shown in Figure 1. The mean concentration of NH₃-N was numerically higher in the RPM diet; however, the effect was not statistically significant.

Unprotected methionine has been demonstrated to promote rumen bacterial growth and protein synthesis in vitro (37, 38). Although we are unsure if rumen bacterial growth rates were altered in our study, we did observe a higher ruminal MCP concentration in cows fed RPM with reduced dietary CP. Similar results were also observed in a study by other researchers, who found that MCP content could be increased in vitro by supplementing RPM (0.81 g/kg DM) in a low-protein diet (13).

Although no effects were found in VFA concentration during the whole experimental period, cows fed RPM had higher ruminal butyrate and isovalerate concentration 2 h after feeding, consistent with a previous animal study (39) and an in vitro study (13). Ruminal VFA production is dependent on the type of feed substrate, microbial composition, and extent of fiber degradation (40). Using a rumen simulation technique, Abbasi et al. (13) incubated rumen inoculum under high or low dietary CP conditions with or without RPM. These authors found that the abundance of Ruminococcus albus was highest under low dietary protein with high RPM. Moreover, an in sacco study utilizing methionine analogs [2-hydroxy-4-(methylthio) butanoic acid (HMB) and its isopropyl ester (HMBi)] reported higher VFA concentrations and ruminal abundance of F. succinogenes and R. flavefaciens, two important cellulolytic bacteria (41). The increase in ruminal butyrate and isovalerate concentration at 2 h after feeding might imply that the RPM positively affected the microbiota that ferment these VFAs. Future studies should include rumen microbial analysis to confirm these findings in animal studies.