All animal procedures were approved by the Institution of Animal Care and Use Committee at Konkuk University (Approval no. KU21141). A total of 36 lactating Holstein cows (milk yield: 33.37±5.82 kg d⁻¹; parity: 2.03±1.17; days in milk: 198.64±62.09 d; methane production concentration from respiration, MPCR: 33.94±8.21 ppm m⁻¹; methane production concentration from eructation, MPCE: 133.74±66.75 ppm m⁻¹) were allocated into two groups considering parity, days in milk, milk production, MPCE, and MPCR. The cows were fed with or without 3% DM inclusion of Ulva sp. in the total mixed ration (TMR) for 4 weeks. In addition to the TMR, all cows were fed 1.84 kg DM of concentrate and 0.57 kg DM of rye silage.

All feeds were sampled every 2 weeks and analyzed in triplicate. Ether extract (Method 920.39) and crude ash (Ash; Method 942.05) were analyzed as described by AOAC [17]. The neutral detergent fiber (NDF) and acid detergent fiber were analyzed according to the procedure of Van Soest et al [18] using a heat-stable amylase with residual ash. The acid detergent lignin content was analyzed as described by Jung et al [19]. Nitrogen contents, including crude protein (CP), neutral detergent insoluble CP (NDICP), and acid detergent insoluble CP (ADICP), were analyzed using the procedure of AOAC [17]; Method 954.01). The CP, NDICP, and ADICP contents of Ulva sp. were estimated using a specific nitrogen-to-protein conversion factor of 5.13 [20]. Nitrogen in NDF (NDF_N) was calculated by subtracting NDICP from NDF. The net energy for lactation (NE_L) was calculated as described in NRC [21]. The chemical compositions of the diets are shown in Table 1.

Ulva sp. blooming in Korea were manually collected in Seongsan-eup, Seogwipo-si, Jeju-do, Republic of Korea, from September 2020 to August 2021. The collected Ulva sp. was roughly dried at 40°C–45°C for 48–96 h and then thoroughly air-dried at 55°C for 72 h. The dried Ulva sp. was pulverized and then mixed into the TMR.

Enteric CH₄ emissions were measured on the last week of the trial using a laser methane detector (LMD, mini-G[50A], Tokyo Gas Engineering), as described by Kang et al [22]. In brief, the exhaled CH₄ concentration (ppm m⁻¹) of each cow was measured 4 times a day (−2, −1, +1, and +2 h after morning feeding) for 6 min. The data were processed using an Automatic Multi-scale Peak Detection algorithm to separate the peaks into respiration and eructation. Then, MPCR and MPCE were analyzed and reported. The average of each of the 4 measurements was presumed to be the most representative daily CH₄ production from each cow. The MPCR and MPCE are expressed as ppm m⁻¹, and the methane intensity concentration from respiration (MICR) and the methane intensity concentration from eructation (MICE) are expressed as ppm m⁻¹ per kg milk yield per day.

Milk samples were collected during 2 consecutive milking events, pooled in equal volume with 2-bromo-2-nitropropane-1,3 diol (Broad Spectrum Bicrotabs 2; Advanced Instruments), and analyzed every 2 weeks using MilkoScan FT1 (FossAlle 1 DK – 3400).

Blood samples were collected from jugular veins every 2 weeks before feeding. Complete blood cell counts (CBCs) were analyzed using VetScan HM2 (Abaxis). Serum metabolites, including glutamic oxaloacetic transaminase (GOT), glutamic pyruvic transaminase (GPT), blood urea nitrogen (BUN), creatine, albumin, total protein, triglycerides, and uric acid, and plasma metabolites, including total cholesterol, low-density lipoprotein (LDL), high-density lipoprotein (HDL), lactate dehydrogenase (LDH), and glucose, were analyzed using a DRI CHEM 7000i biochemistry analyzer (FUJIFILM). Serum total antioxidant capacity (TAC) was analyzed using an assay kit (MBS2540515; MyBioSource) according to the manufacturer’s protocol using spectrophotometry (Synergy2; biotek).

The hair grown during the experiment was prepared according to the procedure described by Nejad et al [23]. The hair was ground using a bead beater (tacoPrep Bead Beater, ATPB-01; GeneReach Biotechnology) and analyzed using a cortisol ELISA kit (1-3002; Salimetrics) according to the manufacturer’s protocol.

The data sampled repeatedly, including milk production, CBCs, and blood metabolites, were analyzed using the MIXED procedure of SAS software (SAS Institute) with the sources of Ulva sp. inclusion, week, and their interaction as a fixed factor, and animal as a random factor. The covariance structure with the lowest AIC value was chosen among the CS, CSH, AR(1), ARH(1), TOEPH, and ANTE(1) statements. The other data, including methane production and intensity, hair cortisol, and TAC, were analyzed with the source of Ulva sp. inclusion as a fixed factor, and animal within a group as a random factor. All values are reported as LSMEANS. A significant difference was accepted when p<0.05, and the tendency was accepted when 0.05≤p<0.10.

The effects of 3 % DM inclusion of Ulva sp. on milk productivity are shown in Table 2. The milk yield, proportion of milk fat, protein, lactose, lactose yield, and somatic cell counts were not affected by Ulva sp. An interaction between treatment and week was detected for the milk fat yield (p = 0.04), energy-corrected milk (ECM; p<0.05), NE_L (p<0.05), milk urea nitrogen (MUN; p = 0.01), and milk acetone (p = 0.01). In detail, Ulva sp. inclusion increased the milk fat yield in week 2, but there were no significant differences in week 4 (Figure 1). Likewise, ECM, NE_L, and MUN increased with Ulva sp. inclusion in week 2, but there were no significant differences in week 4. Ulva sp. inclusion decreased milk acetone in weeks 2 and 4. However, Ulva sp. inclusion decreased milk beta-hydroxybutyrate (BHB) regardless of the week (p<0.05).

The effects of 3 % DM inclusion of Ulva sp. on methane emission are shown in Table 3. Ulva sp. inclusion did not affect MPCR or MPCE. However, it tended to decrease MICR (16.9%; p = 0.06) and MICE (14.2%; p = 0.06). Hair cortisol was also decreased numerically by Ulva sp. (17.6%; p = 0.10; Table 3). There was no effect of Ulva sp. inclusion on TAC.

The effects of 3 % DM inclusion of Ulva sp. on blood metabolites are shown in Table 4. An interaction between treatment and week was detected for GPT, but there were no significant differences between the control and treatment groups throughout the whole experiment (p = 0.01; Figure 1). Ulva sp. inclusion increased BUN in week 2, but there was no difference in week 4 (p<0.01). An interaction tendency was detected in glucose (p = 0.07) and HDL (p = 0.09). However, there were no significant differences between the control and treatment groups in each week. The main effects of Ulva sp. inclusion or interactions were not detected for GOT, creatine, albumin, total protein, triglycerides, uric acid, total cholesterol, LDH, or LDL.

The effects of 3 % DM inclusion of Ulva sp. on blood metabolites are shown in Table 5. An interaction between Ulva sp. and week was detected for white blood cell counts (WBCs; p = 0.04). In detail, Ulva sp. inclusion tended to decrease WBCs in week 2 (p = 0.08), but not in week 0 or 4 (Figure 1). The proportions of lymphocytes, monocytes, and granulocytes were not affected by the main effect of Ulva sp. inclusion or interactions. However, Ulva sp. inclusion tended to decrease the number of granulocytes (13.0 %; p = 0.10) throughout the 4 weeks. In addition, an interaction tendency was detected for the mean corpuscular volume (p = 0.07), but there were no significant differences between the control and treatment groups throughout the 4 weeks. Likewise, red blood cell counts, hematocrit, hemoglobin, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, platelet, mean platelet volume, mean platelet volume, and platelet crit were not affected by the main effect of Ulva sp. inclusion or interactions.

In our previous study, Ulva sp. bloomed in Korea linearly decreased the production and proportion of methane after 48 h of incubation [7]. Likewise, Ulva sp. harvested from the tropical ocean [4] or cultivated in the lab [3,13] also decreased rumen in vitro CH₄ emission. In the current in vivo study, MICR and MICE tended to decrease as a result of Ulva sp. inclusion, which is in line with our second hypothesis that Ulva sp. inclusion may reduce rumen methane emitting potential. To the best of our knowledge, this is the first study to report the methane-mitigating potential of blooming Ulva sp. in lactating cows. However, methane production, which was reported as MPCR and MPCE, was not affected by the treatment. This could be because LMD data can be influenced by various environmental factors, such as wind speed and direction, relative humidity, and pressure [24]. Meanwhile, the persistency of the anti-methanogenic effects of Ulva sp. is still unclear because this study was conducted for 4 weeks, and methane emission was not repeatedly measured. Ulvan, a water-soluble sulfated polysaccharide that represents 8%–29% of the dry weight of Ulva sp., has been suggested to be a major bioactive compound of Ulva sp. [10,11]. With this value, the estimated ulvan contents fed to cows was 0.24%–0.87% of dietary DM. However, rumen microbes can adapt to some feed additives. Guan et al [25] reported that 33 mg kg⁻¹ of monensin administration decreased 30% of methane emissions in beef cattle, which was restored within 2 months. Wu et al [26] reported that rumen microbes started to adapt to citrus essential oil in terms of methane emissions after 3 weeks of administration. Bacteroidetes phylum has a wide range of carbohydrate-active enzymes, which can degrade algal-derived sulfated polysaccharides, such as fucoidan [12,27]. Therefore, the possibility of rumen microbial adaptation to ulvan remains, which could be related to the temporary effect of Ulva sp. on the milk fat yield described later. A long-term study is required to examine the microbial adaptation to Ulva sp. with repeated measurements of CH₄ emissions and rumen microbial analysis.

Milk yield, milk protein, and lactose yield and proportion were not affected by Ulva sp., which is consistent with our first hypothesis. The milk fat yield, ECM, and NE_L increased as a result of Ulva sp. inclusion in week 2 but did not persist to week 4. It is difficult to ascertain whether the increased milk fat yield, NE_L, and ECM were due to the additional energy consumption with Ulva sp., as the energy content of Ulva sp. is comparatively lower than that of the TMR due to its high ash amounts. Therefore, we suggest that the increased milk fat yield, ECM, and NE_L in week 2 were partially due to the saved energy by Ulva sp. such as methane mitigation. Up to 12% of feed gross energy could be partially redirected to milk production via CH₄ mitigation [28]. Kinley et al [2] reported that 0.10% OM of Asparagopsis taxiformis reduced 38% of CH₄ production with an increase in average weight gain. This is partially supported by the reduced milk acetone in this study. Milk acetone is an indicator of hyperketonemia in dairy cows [29]. Although the acetone levels in both groups were far below the hyperketonemia level (1.4 mM), reduced milk acetone showed the potential of Ulva sp. in ketosis prevention by revising the energy balance in lactating cows. However, our previous study reported that the proportion of acetate linearly decreased with the Ulva sp. inclusion rate and that the proportion of propionate linearly increased [7], making it difficult to explain the increased milk fat yield and the maintained milk and lactose yield. Likewise, Maia et al [3] reported that a 25 DM% addition of cultivated Ulva sp. increased the acetate proportion with CH₄ mitigation in the rumen in an in vitro batch culture system. Furthermore, MICR and MICE decreased up to 15% by Ulva sp. in this study. However, MPCR and MPCE were not changed. In addition, ECM increased around 3.3 kg/d by Ulva sp., which is hard to explain only with the CH₄ reduction. If the increased milk fat yield was not via the reduced methanogenesis, then it could be due to its effects on the rumen lipid metabolism and de novo milk fat synthesis in the mammary gland. Considering the low-fat content of Ulva sp. (0.21% DM), it is difficult to presume that the fatty acid profile of Ulva sp. directly affected the milk fat yield. The antimicrobial property of ulvan [10, 11] in Ulva sp. could possibly impact rumen biohydrogenation microbes such as Butyrivibrio fibrisolvens and Butyrivibrio proteoclasticus, and decrease alternative biohydrogenation pathway represented by trans-10, cis-12 C18:2 and trans-10 C18:1. Huang et al [30] reported that phlorotannin extract from Sargassum sp. inhibited rumen biohydrogenation and elevated carbohydrate-utilizing bacteria in the rumen. Likewise, algal bioactive compounds such as ulvan could impact rumen lipid metabolism, resulting in the milk fat yield. Therefore, the impact of ulvan on rumen lipid metabolism is worthwhile to be investigate in future studies. If the effect of Ulva sp. on milk fat yield was via CH₄ reduction or altering rumen lipid metabolism, it disappeared in week 4. It indicates that its effects on the rumen microbiome were temporary probably due to the microbial adaptation. Longer in vivo trials might be required focusing on the rumen microbiome in the future.

Milk urea nitrogen increased as a result of Ulva sp. inclusion in week 2. This could be due to the high nitrogen uptake in the treatment group, which resulted in increased BUN. Our previous study reported that Ulva sp. has a high RUP (54.64% of CP) compared to alfalfa hay (28.57% of CP), which might be due to the high NDICP content [7]. We suggested that the high NDICP content might have originated from the high amounts of glycoproteins in the cell wall of Ulva sp. [31]. The increased MUN and BUN imply that the high RUP of Ulva sp. can be digested and absorbed in lactating cows. Gaillard et al [32] also reported that the total amino acids of Ulva sp. collected from northern Norway have 27.4%, 51.2%, and 78.6% ruminal, small intestinal, and total tract degradability, and they concluded that Ulva sp. can be considered as relevant protein sources for ruminants. However, they did not result in high albumin or total protein in the serum and milk in the present study. This is consistent with the study by Rjiba-Ktita et al [33], who reported that the 20–40 DM% inclusion of Ulva sp. collected from Tunisia did not alter the average daily weight gain, total tract CP digestibility, or the nitrogen balance in lambs. To understand the discrepancy between the increased MUN and BUN, maintained albumin, and the total protein in serum and milk, additional parameters, including urinary and fecal nitrogen excretion, should be measured in future studies.

Despite the significant interaction between the treatment and week affecting the serum GPT level, there were no significant differences between the control and treatment group throughout the experiment. Likewise, serum GOT and creatine were not affected by Ulva sp. inclusion, indicating that Ulva sp. did not impact liver and kidney function [34,35]. Additionally, the plasma LDH level was not affected by Ulva sp., indicating that Ulva sp. did not affect overall tissue damage and inflammation in lactating cows [36]. Meanwhile, Ulva sp. tended to decrease WBCs in week 2 by reducing the number of granulocytes, which might be related to the immunomodulating effect of ulvan [11]. Kim et al [37] found that 80% ethanol extraction from Ulva sp. increased nitric oxide and TNF-α in RAW 264.7 macrophages, which also could be resulted from the immunomodulating property of ulvan. However, it is difficult to assert that the possible immunomodulating property of Ulva sp. affected the dairy productivity in this study, as all cows maintained a healthy status without any clinical symptoms. More straightforward clinical studies should be conducted in the future to demonstrate the immunomodulating function of Ulva sp. in lactating cows.

An antioxidant capacity of Ulva sp. has been suggested [38]. We hypothesized that the antioxidant capacity of Ulva sp. may contribute to decreasing the stress level in lactating cows. Ulva sp. inclusion showed a tendency to decrease hair cortisol levels by 17.6%, which shows the potential stress alleviation effects of Ulva sp. in lactating cows. This is partially consistent with the study by Ellamie et al [39], who reported that the heat-stress-alleviating effects of Sargassum latifolium in sheep alleviated anemia and dyslipidemia and improved the antioxidant defense system and inflammatory response. However, TAC and uric acid were not affected by Ulva sp. in this study, which is inconsistent with our third hypothesis. Therefore, the detailed mechanism of the stress-relieving effect of Ulva sp. in lactating cows is still unclear, and it should be reinvestigated in the future.