Introduction

The rumen microbial ecosystem, comprising diverse communities of mainly anaerobic microorganisms, plays an essential role in meeting the nutritional needs of ruminants by converting lignocellulosic materials into readily available energy (volatile fatty acids) and valuable microbial protein1,2. However, ruminal fermentation is inevitably associated with partial losses of dietary energy and nitrogen, which reduces feed conversion efficiency and consequently animal productivity1,3,4. The use of feed additives is considered to be an effective strategy to improve rumen fermentation efficiency and thus reduce the negative environmental impact of ruminant livestock5,6.

Plant-based feed additives (PBFA) containing various plant secondary metabolites (PSM) have received increasing attention over the past few decades for their potential to improve rumen fermentation due to their natural origin, sustainability, and restrictions on the use of antibiotics in many regions of the world3,5,7. In this context, a wide range of essential oils, plant extracts, and their pure bioactive compounds have extensively been researched in vitro for their potential as alternatives to antibiotics in recent decades8,9,10,11. However, it is noteworthy that the majority of these experiments have focused on improving rumen efficiency by optimising ruminal fermentation of carbohydrates (e.g., by reducing ruminal methanogenesis, shifting the VFA pattern towards a higher C3/C2 ratio and improving fibre digestibility12,13,14) and proteins (by reducing ruminal protein degradability and improving the efficiency of microbial protein synthesis (EMPS)5,15,16). Therefore, there is relatively little information in the literature, especially from in vivo experiments, on the potential effects of phytochemicals on ruminal lipid metabolism. This is while, there is a growing public interest and concern regarding milk content of both harmful and beneficial fatty acids, primarily originating from the ruminal metabolism of dietary fatty acids.

Feeding lipids to ruminants, despite a possible depressive effect on DMI and milk fat content (at high inclusion levels, especially of unsaturated fatty acids (UFA))17,18, often has beneficial effects on their health, reproduction and productivity. These benefits are characterised by a lower incidence of ruminal acidosis19, a strengthened immune system20,21 and a moderate negative energy balance and improved energy efficiency19,22. The UFA-rich fat supplements are also considered to be one of the main sources of bioactive and healthy FA, such as n-3 FA and certain ruminal biohydrogenation (BH) intermediates, including c9,t11-CLA found in dairy and beef products23,24. The concentration of these bioactive FA in ruminant products may vary depending on the fat source, the type of diet and the microbial communities involved in the BH of UFA in the rumen25,26,27,28. Therefore, phytochemicals may influence the accumulation of these beneficial FA in ruminant products by altering the composition of the dominant microbial communities involved in ruminal metabolism of UFA29,30,31,32. It should be noted that despite the evaluation of a wide range of PBFAs for their effects on rumen fermentation in vitro9,11,33,34,35, there are few in vivo data on the promising candidates selected in the preliminary screening in vitro experiments. On the other hand, the use of common forms of PBFAs (essential oils, plant extracts and their pure compounds) in the diet is not standardised for in vivo experiments3, this is while their extraction and purification remain costly. Therefore, direct incorporation of PSM-rich plants into the diet may be more practical under field conditions.

Satureja (savory) is a genus of aromatic plants in the Lamiaceae family, with a wide distribution in many regions of the world. Satureja khuzistanica is an aromatic species of Satureja with a strong aroma and high PSM content (around 10% of DM, including the volatile and non-volatile fractions), used mainly to flavour foods in the Zagros regions36. Carvacrol and flavonoids have been found to be the main constituents of its essential oil and methanolic extract, respectively36,37.

In our previous in vitro experiment, both SP essential oil (SPEO) and dry extract (SPEX) improved rumen fermentation by enhancing EMPS, partially suppressing protozoa, and significantly reducing methanogenesis at a concentration of 300–450 mg/L5. Regarding the selective antimicrobial effects of savory phytochemicals on the rumen microbial ecosystem, we hypothesised that savory may modulate ruminal metabolism of dietary fatty acids, possibly enriching milk with the healthy fatty acids (CLA and n-3 fatty acids) that escape ruminal biohydrogenation. However, there is a particular lack of information in the literature on the effects of savory or its phytochemicals on ruminal metabolism of dietary fatty acids. Therefore, the principal objective of the current study was to evaluate simultaneously the effects of oil type and dietary inclusion of SP on rumen fermentation, total tract nutrient digestibility, milk performance and milk FA composition in dairy cows fed two diets supplemented with fish and soybean oils.

Materials and methods

Savory plant materials and the oils supplements

The savory plant material (including leaves, stalks, and flowers) was collected in April (during the blooming period) from the northern areas of Khuzestan province (31.4360° N, 49.0413° E) in western Iran. The plant samples were pooled, then air dried in the shed to a constant weight and stored in plastic bags. The essential oil and dry extract of savory were extracted and their chemical composition was determined by GC-MS and HPLC, respectively, as described in our preliminary in vitro assay5. The fish oil (extracted from Kilka fish, Clupeonnella engrauliformis) was provided by Ardmahikhazar Ltd. (Kiashahr City harbor, Gilan Province, Iran) and soybean oil was procured from Bazaargani Jahan soya company (Karaj City, Alborz province, Iran).

Experimental design, animals and treatments

This study was conducted and reported according to the ARRIVE guidelines. Animal care and sampling procedures were performed in accordance with the guidelines of the Bu-Ali Sina University Animal Care and Use Committee (ref: 1397/1139). A total of eight multiparous lactating Holstein cows (BW = 644.1 ± 33.2; parity = 2; milk yield = 36 ± 2.5; DIM = 100 ± 4.33; mean ± SD) were used in a replicated 4 × 4 Latin square design with 4 periods of 21 d to evaluate the effects of oil type (fish oil vs. soybean oil) and SP (0 vs. 370 g/d/head) on rumen fermentation, milk production and milk composition. As mentioned above, in our previous study, both SPEO and SPEX were tested in vitro within a dose range of 150–600 mg/L rumen inoculum. The results revealed that the optimal dosage for most rumen variables was 150 and 300 mg/L for SPEO and SPEX, respectively. Considering an extraction yield of 2.2% and 7.8% for SPEO and SPEX, respectively, and presuming an average rumen digesta volume of 90–100 L, a quantity of 370 g SP per animal was picked for daily provision. Each experimental period comprised of 14 days of dietary adaptation followed by 7 days of sampling. In each trial period, cows within the square were randomly assigned to one of the four dietary treatments using a 2 × 2 factorial design: (1) FOD = diet supplemented with 2% fish oil (DM basis); (2) SOD = diet supplemented with 2% soybean oil; (3) FODs = diet supplemented with 2% fish oil plus 370 g/d/head of dried savory plant (SP) as topdressing; and (4) SODs = diet supplemented with 2% soybean oil plus 370 g/d/head of SP. Cows were housed in individual pens (4 × 4 m2) within a barn, each equipped with a concrete feed bunk and drinking trough, and had free access to drinking water. Pens were bedded with clean straw and sand and refreshed daily. Diets were iso-energetic and iso-nitrogenous and were formulated to meet the NEL and MP requirements of multiparous dairy cows weighing 650 kg BW and producing 35 kg/d MY38 (Table 1). Diets were mixed as a total mixed ration (TMR) and offered add-libitum once a day with a target of 10% refusals. The SP was mixed thoroughly with 1 kg of TMR and top dressed at the feeding time.

Table 1 Ingredients and chemical composition of the experimental diets and savory plant (SP).
Full size table

Diet, feed ingredients, fecal and urinary sampling

Feed offered and orts were recorded on a daily basis to estimated daily feed intake. Samples of TMR diets, orts, and TMR ingredients were collected daily and stored at -20 °C until further analysis. These samples were later thawed overnight, samples of TMR diets and orts were pooled by treatment and by animal within period respectively, and those of TMR ingredients were pooled by week within each period. All these samples were then dried in a forced air oven at 55 °C for 48 h, ground to pass through a 1 mm sieve and subjected to proximate chemical composition analysis. Sub-samples of frozen TMR diets were freeze-dried and composited by treatment to determine their FA profiles. Faecal samples were collected twice daily (3 and 9 h after feeding) from the rectum on days 15 to 20 of each experimental period and stored frozen at -20 °C for subsequent analysis. Faecal samples were later thawed overnight and composited by cow within period. These samples were then oven dried at 55 °C for 48 h, ground through a 1 mm sieve and their chemical composition determined. Total tract digestibility of nutrients was determined using acid-insoluble ash as an internal marker, as explained by Van Keulen and Young39. Spot urine samples (100 mL) were collected three times daily for two consecutive days (d 15 and 16 of each experimental period at 2, 10, and 18 h after feeding in the first day and 6, 14, and 22 h post-feeding in the second day) to have 6 samples over 24 h and acidified with 2 M H2SO4 at a ratio of 100:6 (vol/vol) to achieve a pH ≤ 3.0. Acidified urine samples were then diluted with distilled water (1:10) and stored at -20 °C until analysed for their purine derivatives.

Ruminal fermentation characteristics and protozoa enumeration

Rumen fluid was collected 4 h post-feeding on the last day of each experimental period using a stomach tube (inner and outer diameter 10 and 14 mm, respectively; length: 3.6 m) connected to a vacuum pump (JB Industries Seri Platinum, DV-200 N, America). The first 50 ml of the samples was discarded because of possible contamination with saliva and ruminal pH was measured immediately using an Ultra Basic portable pH meter (Denver Instrument, Arvada, CO, USA). A portion of the rumen fluid was filtered through four layers of cheesecloth, and samples (10 mL) of the filtered rumen fluid were mixed with 2 mL of 25% orthophosphoric acid and stored at -20 °C until later analysis for VFA. Another set of 5 ml samples of rumen fluid were fixed with an equal volume of 0.2 N hydrochloric acid and stored at -20 °C for ammonia analysis. Samples of 5 mL of unfiltered rumen fluid were preserved in an equal volume of 50% formalin (18.5% formaldehyde) and stored at room temperature for subsequent protozoan enumeration. Protozoa were enumerated microscopically using a counting chamber (Hawksley and Sons Limited, Lancing, Sussex, UK) as described by Dehority40.

Body weight and body condition scores

Cows were weighed at the beginning and end of each experimental period and their body condition was scored by three experienced scorers according to Edmonson et al.41. The BW data were used to calculate the change in body weight of the cows during each experimental period.

Milk production and composition

Cows were milked twice daily at 08:00 and 20:00 and milk production was recorded at each milking. Milk samples were collected at 4 consecutive milkings (a.m. and p.m.) during the sampling week of each period in 50 ml tubes and preserved with bromo-2-nitropropane-1,3-diol (Valio Ltd., Helsinki, Finland) for subsequent analysis of milk constituents. A second set of unpreserved milk samples was collected as described above and stored at -20 °C. These samples were then pooled by cow within period and analysed for milk FA profiles.

FA analysis of TMR diets, SP, fish and soybean oils, rumen fluid, and milk fat

Total lipids from TMR diets, SP and supplemental oils were extracted in duplicate according to Folch, Lees and Sloane Stanley42. The extracted lipids were then methylated as described by Chouinard et al. (1999), evaporated under N2 and extracted with 1 ml of hexane and analysed for their FA profiles. Rumen fluid FAs were extracted and methylated according to the one-step methylation method of Sukhija and Palmquist43. The FA methyl esters (FAME) were dried under N2, reconstituted with 1 ml hexane and analysed by GC. Milk FAs were extracted in duplicate according to Hara and Radin44 and methylated as described by Chouinard, et al.45. The FAME were extracted with hexane and then analysed by GC.

Chemical analyses and calculations

The chemical composition of the samples (TMR diets, dietary ingredients, orts and faeces) was determined using standard AOAC methods46. Briefly, samples were subjected to wet chemical analysis for DM (ID no. 930.15), total ash, (ID no. 924.05), minerals (calcium and phosphorus, ID no. 985.01), ether extract (EE; ID no. 920.39), CP (ID no. 984.13) and ADF (ID no. 973.18). Amylase-treated NDF (aNDF) was determined according to Van Soest et al.47 using sodium sulphite and heat-stable α-amylase. The non-fibrous carbohydrate (NFC) content of the samples was calculated as 100 - (CP + NDF + EE + ash)38.

Urine samples were pooled by cow within period on an equal volume basis and analysed for allantoin48, uric acid (enzymatic colourimetric uricase technique) and creatinine (Jaffe enzymatic method)49. Daily urine volume was calculated from urinary creatinine concentration and body weight as 29 (mg) × BW (kg) × [1/urinary creatinine (mg/L)]50. Endogenous purine derivative (PD; allantoin and uric acid) excretion rate (mmol/d) was calculated as 0.385 mmol/BW0.75 per day51. Estimated daily urine output was used to estimate daily urinary PD excretion. Urinary PD excretion was used to estimate duodenal microbial N flux51.

Ruminal VFA concentrations were determined according to Ottenstein and Bartley52 using a gas chromatograph (GC-2010, Shimadzu, Kyoto, Japan) equipped with a flame ionisation detector and a capillary column (CP-Wax 58 FFAP CB, 50 m×0.53 mm, film thickness = 2 μm). Hydrogen was used as the carrier gas at a flow rate of 30 ml/min. The oven temperature was set at 80ºC and held for 1 min, then increased to 230ºC at a rate of 30ºC/min and held at 230ºC for 5 min. Injector and detector temperatures were set at 220ºC and 240ºC respectively. Rumen ammonia concentration was determined by the phenol hypochlorite method according to Broderick and Kang53.

The morning and evening milk samples were pooled proportionally based on the milk yield of the a.m. and p.m. milkings and analysed for milk components (including fat, protein, lactose, SNF, TS, SSC, MUN, BHB and NEFA) using an FTIR spectrophotometer (Lactoscope model FTA, Delta Instruments, Drachten, the Netherlands) as described by Bach et al.54. Milk fat, true protein, lactose and TS yields (kg/d) were calculated from the averaged milk yield and the corresponding milk composition. Fat corrected milk was calculated using the NRC model38 and ECM was calculated using the Rico et al. equation55.

The FAs of the samples were analysed on an Agilent gas chromatograph (model 7890 A, Agilent Technologies Inc., Santa Clara, CA) equipped with a 105 m × 0.25 mm (i.d.) fused silica capillary column (Rtx®-2330, 0.2 μm film thickness) and a flame ionisation detector. The column temperature was maintained at 50 °C for 1 min, then increased to 198 °C at 20 °C/min and held at this temperature for 60 min. The injector and detector temperatures were set at 250 °C and 270 °C respectively. Helium was used as the carrier gas at a split ratio of 50:1 and a flow rate of 0.8 mL/min. Fatty acid peaks were identified using a mixture of methyl ester standards and heptadecanoic acid as an internal standard, and their proportions were expressed in grams per 100 g of FAME.

Statistical analysis

Data were averaged across days and analysed as a replicated 4 × 4 Latin square design using the MIXED procedure of SAS (SAS, version 9.4, SAS Institute Inc., Cary, NC). The model included fixed effects of SP, oil type, SP × oil type and period, and a random effect of cow. Statistical significance was declared at P ≤ 0.05 and tendency at 0.05 < P ≤ 0.10. Tukey’s multiple comparison test was used to compare least squares means between treatments.

Results

Nutrient intake, apparent total tract nutrient digestibility, and microbial nitrogen flux

The intake of all nutrients was only influenced by OT (Table 2), as cows fed SO diets had higher nutrient intakes than those fed FO diets (P < 0.01). The apparent total tract digestibility of DM, OM and EE were influenced by both OT and SP (P < 0.05). The total tract digestibility of DM, OM and EE was higher in cows fed fish oil-supplemented diets (FO diets) than in those fed soybean oil-supplemented diets (SO diets) (P < 0.001), and the inclusion of SP in the diet improved their digestibility (P < 0.05). Including SP in the diet had no significant effect on the digestibility of CP and NDF (P > 0.05); however, similar to other nutrients, their digestibility was higher in the FO diets when contrasted with the SO diets (P < 0.01). The ADF digestibility remained unaffected by treatment (P > 0.05); However, the total tract digestibility of NFC tended to be higher with FO supplementation compared to SO supplementation (P = 0.093), and the inclusion of SP in the diet improved its total tract digestibility (P = 0.018). Both endogenous PD (P = 0.079) and microbial PD (P = 0.060), as well as microbial N flux (P = 0.59) tended to be higher in cows fed SO diets compared to those fed FO diets. However, dietary SP had no significant effect on either endogenous or microbial PD or microbial N flux.

Table 2 Effect of oil type (OT) and savory plant (SP) on nutrient intake, apparent total tract nutrient digestibility and microbial N flux.
Full size table

Rumen fermentation and protozoan counts

Ruminal pH was influenced by both OT and SP (Table 3), as it was lower in cows fed SO than in those fed FO diets (P < 0.001) and increased with the inclusion of SP in the diet (P = 0.019). Similarly, total number of protozoa was also affected by both OT and SP, as feeding cows with FO diets resulted in a lower total number of protozoa compared with SO diets (P = 0.035), and the inclusion of SP in the diet reduced the total number of ciliates in the rumen (P = 0.035). Total VFA concentration remained unaffected by treatment; however, the VFA profiles were slightly altered by SP. There was a tendency for an interaction (P = 0.052) between OT and SP for the acetate proportion, as it tended to be reduced by the inclusion of SP in the fish oil-supplemented diet. The proportions of propionate and butyrate were not affected by treatment. Similarly, the acetate to propionate ratio remained unaffected by treatment. However, the proportion of BCVFA increased with SP inclusion (P < 0.05), especially in the fish oil-supplemented diet. There was an OT by SP interaction for valerate (P = 0.045), as its proportion tended to increase with inclusion of SP in the fish oil supplemented diet. Ruminal NH3 concentration was modified by both OT and SP, as cows fed SO diets had lower NH3 concentration than those fed FO diets (P < 0.01), and dietary SP reduced ruminal NH3 concentration (P < 0.01).

Table 3 Effect of oil type (OT) and savory plant (SP) on rumen fermentation and protozoal count.
Full size table

Milk yield, milk composition, and body condition

Milk yield, FCM4%, ECM and ECM/DMI were only influenced by OT (Table 4), as cows fed SO diets proportionally produced more standardised and non-standardised milk with higher efficiency than those fed FO diets (P < 0.05). However, dietary SP did not affect milk yield and efficiency (P > 0.05). Concentrations and yields of milk fat and true protein, as well as milk lactose yield, were higher in cows fed SO diets than in those fed FO diets (P < 0.05). Milk component yields were not affected by SP (P > 0.05), but milk fat (P = 0.062) and protein (P = 0.087) contents tended to decrease with inclusion of SP in the diet. Milk MUN and SCC were not changed by OT but decreased with SP inclusion in the diet (P < 0.01). Milk NEFA and BHB concentrations were influenced by both OT and SP, as their concentrations were lower in the milk of cows fed SO diets than in that of cows fed FO diets (P < 0.01), and the inclusion of SP in the diet had a depressing effect on these variables (P < 0.01). Body condition score (BCS) did not change with treatment, but BW was influenced by OT, with SO diets resulting in higher BW compared to FO diets (P = 0.018). Similarly, body weight change (BWC) was significantly influenced by OT, with cows fed SO diets showing approximately 30% greater weight gain than cows fed FO diets (P < 0.01). However, SP also tended to influence BWC (P = 0.089) as it improved the weight gain of cows by approximately 11% compared to the control.

Table 4 Effect of oil type (OT) and savory plant (SP) on milk yield, milk composition, and body condition of dairy cows.
Full size table

Dietary, rumen and milk FA composition

Oleic acid (C18:1c, OA), palmitic acid (c16:0, PA) and docosahexaenoic acid (C22:6, DHA) were the main FA in fish oil (about 65% of total FA), and linoleic acid (C18:2c, LA), OA and PA were the main FA in soybean oil (about 87% of total FA) (Table 5). Myristic acid (C14:0) and the odd-chain FA (C15:0, C17:0) were significantly higher in fish oil than in soybean oil, and the very long chain n-3 FA (C20:5, C22:6) were absent in soybean oil. Alpha linolenic acid (C18:3c, ALA) PA, LA and OA were the main FA in SP fat.

Rumen fluid FA profiles were modified by both OT and SP (Table 6). The proportion of short and medium chain FA (S-MCFA) was higher in cows fed SO diets than in those fed FO diets (P < 0.01), and the inclusion of SP in the diet reduced the proportion of this group of FA (P < 0.01). The odd-chain FA (including C15:0 and C17:0) were only affected by OT, and their ruminal proportions were higher in cows fed FO than in those fed SO (P < 0.01). The proportion of both C16:0 and C16:1 was not affected by SP, but was influenced by OT, as feeding cows with FO diets increased these FA proportions in the rumen compared to SO diets (P < 0.01). The proportion of C18:0 was significantly increased by SO versus FO diets (P < 0.01) and decreased by inclusion of SP in the diet (P < 0.01). Conversely, the proportions of both cis and trans isomers of C18:1 were lower with SO than with FO diets (P < 0.01), and their proportions increased with the inclusion of SP in the diet (P < 0.01). Feeding SO to cows increased the ruminal proportion of ΣC18:2c but decreased that of ΣC18:2t compared to FO diets (P < 0.05). However, the inclusion of SP in the diet increased the proportion of both ΣC18:2c and ΣC18:2t in the rumen (P < 0.05).

Table 5 Fatty acid composition (g/100 g FA) of total mixed ration (TMR), fish oil (FO), soybean oil (SO), and savory plant (SP).
Full size table
Table 6 Effect of oil type (OT) and savory plant (SP) on ruminal fatty acid composition (g/100 g FA).
Full size table

Both CLA isomers (c9,t11-CLA and t10,c12-CLA) were influenced by OT, as their ruminal proportions were higher in cows fed FO diets compared to cows fed SO diets (P < 0.01). Inclusion of SP in the diet tended to increase the proportion of both C18:2(c9,t11-CLA) and C18:2( t10,c12-CLA) in the rumen (P ≤ 0.1). Among the n-3 FAs, C20:5 (EPA: eicosapentaenoic acid) and C22:6 (DHA: docosahexaenoic acid) were absent in the rumen of cows fed SO diets, but the proportion of C18:3c (ALA) was higher in this group compared to cows fed FO diets (P < 0.001), and the inclusion of SP in the diet increased the proportion of ALA as well as DHA in the rumen (P < 0.05). The FO diets increased the ruminal proportion of UFA at the expense of SFA compared to the SO diets (P < 0.01). Similarly, the FO diets increased the ruminal proportion of MUFA (P < 0.01), but the proportion of PUFA was unaffected by OT. Inclusion of SP in the diet significantly increased the ruminal proportions of both MUFA and PUFA at the expense of SFA (P < 0.01).

The FA profiles of milk fat were also modified by both OT and SP (Table 7). Feeding SO diets resulted in a greater proportion of S-MCFA (C4:0 - C12:0) compared to FO diets (P < 0.05). The proportion of S-MCFA in milk fat decreased with SP inclusion in the diet (P < 0.05). The proportion of milk OCFA was greater in cows fed FO diets compared to those fed SO diets (P < 0.05), and SP incorporation in the diet enhanced the proportion of most OCFA in milk fat (P < 0.05). The proportion of C16:0, especially C16:1, in milk fat was significantly higher in cows fed FO diets than in cows fed SO diets (P < 0.01), and the inclusion of SP in the diet decreased the proportion of these fatty acids in milk (P < 0.01). Milk fat from cows fed FO diets had a greater C18:0 proportion than that from cows fed SO diets, and the inclusion of SP in the diet reduced this FA proportion in milk fat (P < 0.01). Feeding SO diets to cows decreased most of the C18 unsaturated FAs (C18:1c, C18:1t, C18:2t and both isomers of CLA in milk but increased those of C18:2c and C18:3c compared to FO diets (P < 0.05). However, the addition of SP to the diet had a positive effect on the enrichment of milk fat with C18 unsaturated FA (except for C18:2t) (P < 0.05). The proportion of both isomers of CLA in milk fat was significantly higher in cows fed FO diets than in cows fed SO diets (P < 0.01), and including SP in the diet increased C18:2 (c9,t11-CLA) and C18:2 (t10,c12-CLA) by 23% and 63%, respectively. Similar to rumen fluid, EPA and DHA were absent in milk from cows fed SO diets, but ALA was higher in milk from cows fed SO diets than in milk from cows fed FO diets (P < 0.05). Incorporating SP into the diet enhanced milk n-3 FA (except EPA) content (P < 0.05), resulting in a 25% and 98% increase in milk ALA and DHA content, respectively, in cows fed SP-supplemented diets. Oil type had no effect on the milk SFA content, but milk UFA content (mainly MUFA) was higher in cows fed FO diets compared to those fed SO diets (P < 0.01). The proportion of both MUFA and PUFA in milk fat was increased by the inclusion of SP in the diet (P < 0.01).

Table 7 Effect of oil type (OT) and savory plant (SP) on milk fatty acid composition (g/100 g FA).
Full size table

Discussion

Based on the chemical analysis of SP phytochemicals in our previous in vitro study, the yields of SPEO and SPEX were 2.2% and 7.8% (DM basis), respectively, and carvacrol and flavonoids were the major compounds, accounting for approximately 72% and 70% of the SPEO and SPEX fractions, respectively5. Extensive research (mainly in vitro) on carvacrol and flavonoids, or plants rich in these phytochemicals, has shown that they have great potential to modulate rumen fermentation57,58,59,60,61, making SP a promising candidate for improving rumen function and thus animal productivity.

Nutrient intake, apparent total tract nutrient digestibility and microbial nitrogen flux

The lower DMI and other nutrients intake with FO compared to SO diets was consistent with previously reported literature data18,62,63,64. The depressive effects of fish oil on feed intake have mainly been attributed to the stimulatory effects of its high EPA and DHA65 on the secretion of CCK66, gut peptides67 and hepatic oxidation68. The lack of effect of SP on DMI in the current study was consistent with previous findings on the fact that additives rich in carvacrol or phenolics have no effect on DMI69,70,71,72 .

A lower intake of nutrients in FO diets compared with SO diets may explain the higher total tract digestibility of nutrients observed with FO diets. It is well known that a low feed intake can improve the overall digestibility of nutrients by reducing the passage rate and thereby increasing the average residence time of feed particles in the gut73,74. In addition, previous research has shown that EPA and DHA in fish oil have a stimulatory effect on duodenal secretion of CCK, which has a direct inhibitory effect on intestinal motility, reducing the rate of passage75. On the other hand, CCK is one of the main factors that induce the secretion of hepatic and pancreatic enzymes, thereby improving intestinal digestion of nutrients76.

In order to accurately estimate nutrient digestibility and to avoid contamination of the samples by dust and soil, feed samples were collected immediately before distribution of the diets to the animals, following preparation of the TMR diets. In addition, faecal samples were collected directly from the rectum of the cows to minimise the risk of contamination. Except for CP and cell walls, the total tract digestibility of other nutrients was improved in response to the inclusion of SP in the diet. The improvement in DM and OM digestibility was consistent with our previous results from in vitro experiment where SPEX improved ruminal DM and OM digestibility at low and medium doses5. In our in vitro experiment, the improvement in ruminal digestibility of OM by SPEX was accompanied by a shift in the VFA pattern towards propionate at the expense of acetate, indicating an improvement in ruminal digestibility of NFC rather than NDF5. These results suggest that part of the improvement in total tract NFC digestibility by SP in the current study may be of ruminal origin, particularly with regard to a numerical shift in the VFA profiles towards propionate, especially in the FO diets.

There is no comparable data in the literature on the effect of savory on nutrient digestibility; however, research on carvacrol and flavonoids (the major compounds in the essential oil and extract of SP, respectively) has shown a positive effect from these phytochemicals on nutrient digestibility11,77. The mechanisms by which phytochemicals may improve nutrient digestibility are not fully understood in ruminants. However, there is evidence that certain phytochemicals may improve ruminal digestibility by interacting with microbial enzymes and feed particles12,78. Other research has shown that essential oils can improve nutrient digestibility and absorption by altering intestinal morphology, inducing the secretion of digestive enzymes and inhibiting intestinal pathogens77,79,80,81.

The tendency for greater endogenous PD with SO vs. FO diets is primarily related to a higher body weight in cows fed SO diets, as this variable is estimated based on the metabolic body weight of the cows. The tendency towards higher microbial PD and N flux with SO vs. FO diets is indicative of improved microbial protein synthesis in cows fed SO diets. This improvement may be attributed to an increased DMI leading to a higher rate of passage82, resulting in an increased amount of microbial protein entering the small intestine83.

The absence of SP effect on microbial N flux was consistent with our in vitro findings that microbial biomass remained unaffected by either SPEO or SPEX. However, in our in vitro experiment, the microbial protein (MP, including only the bacterial fraction, as the protozoal fraction was excluded), as well as EMPS were substantially improved by both SPEO and SPEX5. Therefore, it is probable that the positive effect of SP on the bacterial fraction of microbial protein was partially offset by its negative impact on the protozoan population, as protozoa contribute between 10 and 30% of the duodenal flux of microbial protein84. These results should be considered with caution, as estimating microbial protein synthesis from PD is subject to several limitations that can affect the accuracy and precision of the estimates. The diet type and feed intake effects, endogenous PD, the linearity of the relationship between PD excretion and microbial protein, the method of spot urine sampling and the possible contribution of dietary PD to the purines absorbed from the rumen are the limitations that may impact the accuracy of this method85,86. Diurnal variation of urine creatinine and PD has been reported as issue for urine spot sampling method; however, multiple urine sampling (at least 6 samples over 24 h) with equally spaced sampling interval has been shown to reduce variations in urine metabolites and lead to valid estimations of microbial N flux87,88. In this method, urine PD is also adjusted to endogenous PD estimated from metabolic body weight. Nevertheless, indirect estimation of microbial protein synthesis is still less accurate than direct measurement of microbial N flux by duodenal sampling in fistulated animals.

Rumen fermentation and protozoan counts

A lower ruminal pH in cows fed SO diets was consistent with a numerically higher total VFA in this group compared with cows fed FO diets. This could be partly attributed to a high DMI in cows fed SO diets. Additionally, there is evidence that fish oil can increase ruminal pH independently of DMI67 and this has been linked to its high n-3 fatty acid content, likely through their action on inhibiting rumen fermentation and VFA production67,89,90. However, the increase in ruminal pH was not consistent with the total VFA concentration when SP was included in the diet, which remained unchanged. Given the lack of SP effect on the proportion of major VFAs in the rumen, the reduction in ruminal concentration of minor VFAs with strong acidity (e.g. lactate) by SP may have increased ruminal pH. In this connection, previous research has revealed that flavonoids have the potential to reduce ruminal lactate concentrations by inhibiting lactate-producing bacteria91 or by increasing the abundance of lactate-consuming bacteria91,92.

Despite a lower coefficient of OM digestibility with SO vs. FO diets, total VFA was numerically higher in cows fed SO diets, which could be mainly a result of a higher DMI in cows fed SO diets. The positive impact of SP on total tract digestibility of OM was not reflected in total VFA. This could be primarily due to a post-ruminal improvement in OM digestion through the stimulatory effect of SP on pancreatic and intestinal enzyme secretion, as discussed previously. These results were inconsistent with those of our in vitro experiment, where SPEX improved both ruminal OM digestibility and total VFA concentration at low and medium doses5. This discrepancy may be attributed to different conditions of the in vitro and in vivo experiments, and probably a lower ruminal concentration of the SP bioactive compounds due to their gradual release from the plant material and rapid passage from the rumen. In an in vitro experiment, Castillejos et al.93. found a dose-dependent effect of savory (Satureja montana) essential oil on TVFA, as TVFA increased at doses of 5 and 50 mg/L, but remained unaffected at 500 mg/L of savory essential oil. In their experiment, the acetate proportion increased at 50 and 500 mg/L, and the propionate proportion increased at 500 mg/L of savory essential oil. The tendency of SP to reduce the acetate proportion in the fish oil-supplemented diet was accompanied by a numerical increase in the propionate proportion. These results were in line with those observed in vitro with low and medium doses of SPEX5. A reduced proportion of acetate with FODs in the current study may be a result of a synergistic interaction between the SP phenolic compounds and fish oil EPA and DHA in partially inhibiting cellulolytic bacteria94,95. The BCVFA concentrations in the rumen are generally influenced by ruminal protein degradation and deamination activity on the one hand, and by their uptake by rumen bacteria (mainly cellulolytics) on the other hand96,97. A greater proportion of BCVFA by SP, especially in the fish oil-supplemented diet is thought to be linked to reduced uptake of BCVFA by cellulolytic bacteria, taking into account the depressant effects of SP on ruminal NH3.

In the current study, a lower ammonia concentration with SO diets was associated with a significant improvement in microbial N flux in these diets compared with FO diets. These results suggest that at least part of the reduction in ruminal ammonia concentration with SO compared with FO diets can be attributed to the improvement in microbial protein synthesis as a result of the increased nutrient supply, and probably a higher passage rate (improving EMPS by reducing microbial maintenance requirements) with SO diets98. In addition, a high passage rate in cows fed SO diets may also have shifted the site of protein digestion from the ruminal to the post-ruminal, thereby reducing the concentration of ruminal ammonia.

The depressive effect of SP on ruminal ammonia in the current study was consistent with that observed from both SPEO and SPEX on ammonia concentration in our in vitro experiment5. Consistently, Talatapeh et al.99. reported that inclusion of summer savory (Satureja hortensis) essential oil at 200 mg/d in the diet of growing goat kids reduced ruminal ammonia concentrations. Similarly, Castillejos et al.93 found a depressive effect from savory (Satureja hortensis) essential oil on rumen ammonia when tested at concentrations of 5, 50, and 500 mg/L in vitro. The reduction in ruminal ammonia might be related to different factors, including the suppressive impact of SP on rumen protozoa11,100, the improvement in EMPS5, as well as the increased protein bypassing the rumen due to the protective effect of SP phenolic compounds101,102,103,104.

The reduction in protozoan numbers in cows fed FO diets could be due to a lower DMI in this group, as well as the toxic effects of EPA and DHA to rumen microorganisms95,105,106. Despite the role of protozoa in the ruminal digestion of feed, their partial suppression is generally associated with improvements in energy and protein utilisation efficiency100,107. In the current study, the reduction in total protozoan counts by SP was consistent with our in vitro findings, where partial removal of rumen protozoa by SPEO and SPEX was associated with reduced ruminal ammonia concentration and methane production, as well as an improvement in EMPS5. The negative effect of SP on rumen protozoa can be attributed to its essential oils and flavonoids, both of which have shown anti-protozoal activity11,108.

Milk yield, milk composition, and animal body condition

Despite the improvement in total tact digestibility of nutrients, supplementation of fish oil in the diet resulted in a significant reduction in DMI and therefore in digested nutrient intake, as digested DMI, OMI, CPI, EEI and NFCI decreased by 5.1%, 5.5%, 5.5%, 3.6% and 8.0%, respectively in cows fed FO compared with cows fed SO diets. This could be the primary and most important cause of the 6.2% reduction in milk production in cows fed FO diets in comparison to those fed SO diets. Likewise, a lower digested nutrient intake could also be the major cause of the lower concentrations and yields of milk components in cows fed FO versus SO diets. The negative effect of FO diets on milk production efficiency could be due to a shift in the partitioning of nutrients, especially fatty acids, from the mammary glands to the adipose tissue109. This is supported by a higher weight gain in cows fed FO compared with SO diets. These results are consistent with previous findings indicating that diets inducing milk fat depression (MDF) may prioritise weight gain at the expense of milk yield110,111. In the few data available in the literature directly comparing fish and soybean oils on milk performance; the majority have indicated a negative effect from fish oil on milk yield111,112 and milk components111,113,114 compared to soybean oil. In the current study, while it is probable that the including 2% of both fish and soybean oils in the diet have caused some degree of MFD, the milk fat content was significantly lower with FO than with SO diets. In addition to the shift in partitioning of dietary FA towards adipose tissue, a lower milk fat content with FO compared with SO diets may also be partly caused by a high level of EPA and DHA in fish oil, as well as a higher production of certain BH intermediates known to inhibit mammary gland de novo FA synthesis, such as t10,c12-CLA and t10-C18:163,115,116. A higher concentration of milk BHB and NEFA with FO vs. SO diets could be an additional evidence of the fish oil negative impact on FA de novo synthesis in the mammary glands117,118.The reduction in milk protein content with FO diets in the current study may be linked to the depressive impact of fish oil on microbial protein synthesis113, as evidenced by the 16% reduction in duodenal microbial N flux in cows fed FO versus SO diets.

Despite improving nutrient digestibility, the inclusion of SP in the diet had no significant on milk yield, although cows fed diets supplemented with SP produced slightly (820 g) more milk than control cows. In addition to the slow release of bioactive compounds from SP in the rumen (as discussed earlier), the lack of SP effect on milk production could also be due to a shift of nutrients from milk production to adipose tissue restoration (as SP tended to increase body weight gain). There is evidence that phytochemicals from various plant sources, particularly flavonoids, have the potential to stimulate insulin secretion or sensitivity119,120. Insulin has been reported to play an important role in the partitioning of energy between different tissues in dairy cattle, and its high levels may redirect nutrients from the mammary glands to body reserves, thereby improving body weight gain35,121. Flavonoids have also been shown to promote adipogenesis directly by regulating genes involved in adipogenesis122. In an in vivo experiment by Talatapeh et al.99. , the inclusion of summer savory (Satureja hortensis) essential oil at 200 and 400 mg/d in the diet of growing goat kids had no effect on their feed intake and growth performance.

A tendency to lower milk fat and protein content with SP may be partly explained by a dilution effect of a slight improvement in milk production, as milk fat and protein yields were not affected by the inclusion of SP. However, an increased concentration of milk PUFA by SP could also contribute to a decrease in milk protein123 and fat63 content. A lower concentration of milk BHB and NEFA in cows fed SP compared with the control could be an indication of the positive effect of SP on energy balance, as SP tended to increase body weight gain. Milk SCC as an indicator of mammary gland health and milk quality124 and MUN as an index of animal protein nutritional status and N-utilisation efficiency125 were both reduced by the inclusion of SP in the diet. Higher SCC is typically associated with mammary infections and reduced milk production and quality126,127. Flavonoids have been shown to reduce SCC11,127, mainly by suppressing infectious bacteria128 and improving the immune system129,130. A lower MUN in cows fed SP compared with the control diets was consistent with the lower ruminal ammonia concentration in the first group. These results are consistent with those of Hristov, et al.58 who reported a depressant effect of oregano leaf on MUN. Previous research has shown that lower MUN is associated with reduced urinary N excretion, which ultimately improves N use efficiency131,132. These findings demonstrate that SP has the potential to improve milk quality by reducing SCC and MUN.

Dietary, rumen and milk FA composition

The FA composition of fish oil and soybean oil were consistent with those previously reported in the literature133,134,135,136, and their inclusion in the diet changed the FA profiles of the experimental diets proportionally based on the major FA of fish or soybean oil. Changes in milk FA profiles in response to both OT and SP were consistent with those observed for rumen digesta FA profiles. A lower concentration of milk S-MCFA with FO versus SO diets could most likely be related to the negative effect of fish oil on de novo FA synthesis in the mammary glands, as described previously62,63. Similar to fish oil, the negative effect of SP on milk S-MCFA could be mediated mainly by increasing the ruminal concentration of PUFA and certain BH intermediates (t10,c12-CLA, and C18:1 isomers), known to exerting negative effects on mammary gland de novo FA synthesis137,138. There are no comparable data in the literature on the effect of phytochemicals on milk short and medium chain fatty acids, but the polyphenol-rich by-products of olive oil extraction139 and a commercial phytonutrient product composed of cinnamaldehyde, eugenol and capsicum oleoresin140 slightly reduced milk short and medium chain fatty acids. The presence of OBCFA in milk is of great importance, as they not only possess several health-promoting properties141,142,143, but can also provide information on rumen fermentation and microbial communities, as they are either directly produced by the rumen microbiome or indirectly derived from rumen-derived precursors144. In the current study, FO diets promoted rumen OCFA and milk OBCFA concentrations, these results are consistent with previous studies reporting a positive effect of fish oil or DHA on milk OBCFA63,145,146. Much of the increase in rumen and milk OBCFA, especially OCFA, is expected to stem from fish oil, which is proportionally rich in OCFA compared with soybean oil. However, it has been speculated that fish oil may enhance the microbial biosynthesis of OBCFA by increasing the supply of their precursors145.

In the current study, the inclusion of SP in the diet did not affect milk OCFA, but increased milk BCFA concentrations, which is consistent with the increased proportion of ruminal BCVFA in cows fed these diets. The BCVFA are considered to be precursors of BCFA147. Despite the results of earlier research on the close linkage between milk OBCFA and rumen microbial communities148,149, recent findings on the possibility of mammary de novo synthesis of OBCFA, especially certain odd-chain fatty acids150,151, the different rate of transfer of ruminal OBCFA to the mammary glands, and their derivation from even-chain FA144,152,153 have challenged the strict direct connection between milk OBCFA and rumen microbial populations. However, considering the strong correlation reported between milk iso C16:0 and microbial crude protein148, the increased concentration of milk iso C16:0 by SP in the present study may suggest a favourable impact of SP on microbial protein synthesis. This is supported by our previous in vitro findings, where SPEO and SPEX significantly improved MP and EMPS5. A higher concentration of C14 and C16 and their unsaturated isomers with FO compared with SO diets could be a result of a higher supply of these FA by fish oil.

Despite the adverse effects that PUFA or their ruminal BH intermediates may have on milk fat content, the enrichment of animal products with these bioactive FA is of particular interest to nutritionists because of their numerous beneficial effects on human health154. In this context, the manipulation of ruminal BH of PUFA by phytochemicals is an effective strategy to improve the quality of animal products through their enrichment with bioactive FA. In the current study, ruminal and milk concentrations of C18:1c and C18:3c in both groups of cows fed FO and SO diets were proportional to their respective dietary contents. Therefore, the ruminal metabolism of these FA does not appear to be affected by oil type. However, despite reducing milk yield and milk fat content, fish oil supplementation improved the milk health index by increasing the concentrations of C18:2, c9,t11-CLA as well as EPA and DHA compared with soybean oil. These findings are consistent with previously reported data on the beneficial effects of fish oil in fortifying milk with these fatty63,111,113,155,156. A higher ruminal and milk concentrations of both isomers of CLAs and C18:1t (VA: vaccenic acid) with FO compared to SO diets, suggest that fish oil probably has a general inhibitory effect on ruminal BH of FA. These results were consistent with those of Vargas-Bello-Perez, et al.24, who reported a fortifying effect from fish oil on milk CLAs and VA. The negative effect of fish oil on FA BH has been attributed to its EPA and DHA toxic effects on the rumen bacteria responsible for BH, particularly those involved in the reduction of CLA to VA157,158 and those that reduce VA to C18:0159,160. Inhibition of the last step of BH, resulting in a higher ruminal concentration of VA, provides more VA to the mammary glands to produce more CLA through Δ9-desaturase161, this appears to be one of the major factors in the current study that significantly increased milk CLA content in cows fed FO than those fed SO diets.

The inclusion of SP in the diet of dairy cows altered rumen and milk FA composition, mainly by decreasing mammary de novo FA synthesis and increasing the concentration of UFA and their BH intermediates. In general, the rumen metabolism of UFA involves two major stepwise processes: lipolysis and FA biohydrogenation162. The first step of UFA metabolism (lipolysis) starts with the hydrolysis of ester bonds of FA by lipases produced by limited species of rumen bacteria, mainly Anaerovibrio lipolytica and, to a lesser extent, Butyrivibrio spp. and Clostridium spp163,164. The lipolysis of esterified UFA is crucial for their subsequent biohydrogenation, as the first step of UFA biohydrogenation (isomerisation from cis to trans geometry) requires UFA with a free carboxyl group25. The free UFAs then undergo extensive biohydrogenation, mainly by Butyrivibrio spp., which limits their transfer to animal products164,165. The main pathway of C18:2 n-6 rumen biohydrogenation involves isomerisation to cis-9,trans-11 CLA, which is further hydrogenated to trans-11 C18:1 and finally to C18:0166. The major biohydrogenation pathway of C18:3 n-3 involves isomerisation to cis-9,trans-11,cis-15 conjugated linolenic acid, which is further hydrogenated to trans-11,cis-15 C18:2, then trans-11 C18:1 and ultimately to C18:0165,166. The last step of hydrogenation of trans-11 C18:1 to C18:0 has been found to be the rate-limiting step of C18 n-6 and n-3 PUFA biohydrogenation in the rumen25,167,168. Most of the cis-9-C18:1 is directly hydrogenated to C18:0, with a small amount isomerised to trans-11-C18:1 and finally reduced to C18:0165. Other very long chain n-3 PUFA such as EPA and DHA are also extensively hydrogenated to their corresponding saturated FA, mainly through the initial step of hydrogenation of the double bond closest to the carboxyl group164,169. Phytochemicals may modulate the ruminal metabolism of UFA by influencing microorganisms involved in both the lipolysis and BH processes of FA, promoting the transfer of intact dietary UFA or BH intermediates to animal products170,171.

Several in vitro and in vivo studies have revealed that phytochemicals, especially polyphenols from various plant sources, can alter BH products by influencing different steps of the ruminal BH process172. However, essential oils have shown more pronounced inhibitory effects on ruminal BH of UFA than tannins and saponins171. The increased concentration of PUFA (including C18 and longer PUFA) by polyphenolic compounds has been found to be mainly caused by a general inhibition of BH173,174 or by limiting the lipolytic activity of bacteria involved in the first step of BH (A. lipolytica)175. Most of the phenolic compounds tested, especially condensed and hydrolysable tannins, have been shown to favour the concentration of BH intermediates (CLA and VA) by inhibiting the reductive activity of bacteria, especially those involved in the final step of BH94,176,177,178. In certain experiments, an increased concentration of BH intermediates (mainly rumenic acid (RA)) has been associated with a selective stimulatory effect of condensed and hydrolysable tannins on B. Fibrosolvense179,180. A more pronounced increase in RA (28 and 23% increase for ruminal and milk contents, respectively) compared with VA (14 and 12% increase for ruminal and milk contents, respectively) by SP in the current study could indicate either an increased abundance of bacteria involved in the isomerisation of LA to RA or a decrease in the reduction of RA to VA.

The increase in ruminal and milk concentrations of LA, ALA, EPA and DHA by SP could be attributed to its inhibitory effect on the lipolytic activity of rumen bacteria. In the current study, a more pronounced increase in milk C18 PUFA, as well as milk C22:6 and C22:5 compared to their corresponding ruminal levels suggests that SP may have a stimulatory effect on mammary FA desaturase activity.

Conclusions

Results from this study revealed that fish and soybean oils, both rich in UFA, have different impacts on milk production and milk quality. Milk production was higher with SO diets and milk healthy fatty acids content was higher with FO diets. Including SP in the diet improved rumen fermentation through reducing ruminal NH3 and protozoa and balancing rumen pH. The inclusion of SP in the diet, despite improving total tract nutrient digestibility, had no significant effect on milk performance but tended to improve cows’ weight gain, implying that SP may have the potential to improve animal productivity in feedlot production systems. The inclusion of SP in the diet also modified the ruminal metabolism of dietary fat by reducing the proportion of S-MCFA in favour of UFA, especially the health-promoting FA (CLA and n-3 FA). There was no interaction between savory plant and oil type for fairly all the variables. However, the simultaneous incorporation of SP with SO in the diet led to the highest milk yield accompanied by the lowest ruminal ammonia, and the combinational inclusion of SP and FO in the diet resulted in the highest digestibility and more pronounced enrichment of milk with healthy fatty acids. Overall, these findings suggest that SP has a great potential as a novel feed additive that could improve ruminant performance and their products quality that merits further research.