Connecting plant, animal, and human health using untargeted metabolomics.

doi:10.21203/rs.3.rs-3813988/v1

Download PDF

Article

Connecting plant, animal, and human health using untargeted metabolomics.

https://doi.org/10.21203/rs.3.rs-3813988/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Consumers of meat and milk products are interested in the connection between agricultural practices and the wellbeing of livestock. Consumers are also concerned about the impact of livestock products on their own health and wellbeing, with some turning to alternative plant-based protein sources. The connection between agricultural practice and consumer metabolism has previously been tenuous, but the development of untargeted metabolomic processes offers the opportunity to reconcile and connect opposing spectrums of the food-chain. We performed a cattle grazing study which compared the performance and metabolomic profiles of beef finished on three forage mixtures of either perennial ryegrass x white clover (PRG), a complex multispecies mixture (n = 22 species) grasses, legumes, and forbs (CMS), and adjacent monocultures of perennial ryegrass, chicory, plantain, lucerne and red clover, which were sown longitudinally across the paddock and all available to cattle throughout the four-month finishing period. Cattle were then processed in a commercial abattoir and the tenderloin collected from each animal for metabolomic analyses. The remaining meat was minced (5mm plate), homogenised and processed into 250 g meat patties. A double-blind randomized cross-over clinical trial was then performed with 23 individuals that consumed a single meal of the meat. Metabolomic analyses were performed using a combination of GCMS/MS, lipidomics and LC-qToF-on plasma samples collected from all participants before meat consumption, 3-, and 5-h post prandial, in addition to meat and pasture samples. Metabolomic profiles of plant, meat and plasma were altered by the three forage mixtures. Metabolomic profiles of plant were different across all three sward mixtures, although CMS and AMS shared more similarities then PRG. However, metabolomic and lipidomic profiles of meat and plasma indicated the greatest differences were observed between AMS and CMS and PRG, which indicates grazing management may provide the greatest opportunity to manipulate metabolomics as opposed to maximising the plant species diversity of a pastoral diet. Gamma-tocopherol (Vitamin E) was elevated in plant material of the AMS diet and the plasma of those consuming AMS meat, compared with PRG or CMS. Fatty acid metabolism was also altered in meat and plasma of the AMS diet as eicosapentaenoic acid was elevated compared with PRG or CMS. Additional differences were also detected in untargeted analyses although features were unable to be matched to the in-house metabolite library. While many features were not able to be identified, these results are the first evidence that metabolomic profiles of the human consumer reflect not only the beef finished from different forages, but also meat and plasma reflect metabolomic profiles of the different sward mixes.

Biological sciences/Ecology/Grassland ecology

Biological sciences/Ecology/Ecological networks

Consumers and producers of animal products are increasingly aware of the intricate connections between the health of the land, animals, and humans, as well as of our collective responsibilities to sustain and enhance Te Taiao, the natural world, for future generations (Gregorini and Maxwell, 2020). Consequently, land users, policymakers and wider society are calling for alternative approaches (Schiere & Gregorini, 2023) to pastoral systems; a call for functionally diversified-adaptive, and integrative agro-ecological systems that simultaneously operate across multiple scales and ‘scapes’ (Gregorini et al., 2022); FAO, 2020). This ultimately requires shifting our perspective from landscapes and foodscapes to healthscapes.

Previous research from the Lincoln University Pastoral Livestock Production Lab suggests that grazing phytochemically diverse swards improves animal welfare and environmental health (Beck & Gregorini, 2020, 2021; Garrett, Beck, et al., 2021; Marshall et al., 2021). More specifically, plants with particular secondary chemical compounds (e.g., tannins, terpenoids, phenols, carotenoids, etc., with known anti-oxidant, anti-inflammatory, anti-cancer, and heart-health promoting effects) reduce metabolic and physiological stress in grazing ruminants, increase animal performance and feed conversion efficiency (Beck & Gregorini, 2020; Garrett, Beck, et al., 2021; Garrett et al., 2022; Garrett, Marshall, et al., 2021; van Vliet et al., 2021), and enhance hedonic and eudemonic well-being (Beck & Gregorini, 2020).

Short-term, postprandial inflammation has been linked to several chronic diseases such as obesity, type-2 diabetes, cardiovascular disease, and cancer, as well as metabolic syndrome; all of which have been associated with the consumption of red meat and dairy products to (Chan et al., 2011; Micha et al., 2012). The latter explains in part, why some organisations or individuals in pursuit of health are venturing in foodscapes that are absent of animal products, while the scientific community is trying to resolve ongoing controversies on whether or not animal-source foods are causally related cardiometabolic disease (Barnard & Leroy, 2020). One of the main knowledge gaps relates to the large biochemical heterogeneity within the broad categories of meat and dairy coming from different production systems – and the diets that they feed to livestock (e.g., grazed pastoral species or ration-based diets). The question arises, are there ecological implications and connections between phytochemically diverse swards and the health of humans and soil (Provenza et al., 2019). If so, this expanded pool of nutrients, medicines, and prophylactics, known as plant secondary compounds, must be considered to understand the effects of consuming animal products on human health, as they have been reported to be concentrated in the meat and milk of livestock consuming those plants (Marshall et al., 2022; van Vliet et al., 2021).

Enhanced eudemonic and hedonic well-being of livestock coupled with better health suggest that phytochemically diverse swards, will improve not only animal welfare, but also the well-being (mental state and health) of animals and humans. Consequently, animal products that incorporate phytochemically rich grazing environments not only foster biodiversity and thus the health of soil, and livestock, but may affect human health as well. Furthermore, incorporating biochemical diversity in the grazing environment presents an opportunity to transform pastoral agriculture of livestock and counteract negative connotations of animal products. This transformation would lead to a change in the eating experience and thus hedonics (i.e., ‘healthy’ pleasure) of the consumer knowing that such a product not only is in tune with the land and animal, but integrally healthier for us and the land we all inhabit. “Ultimately, we are what we eat eats!” (Gregorini et al., 2017).

The objective of this work was to explore, for the first time, the effect of taxonomically and phytochemically rich functional swards-as opposed to ‘status quo’, simple mixes of perennial ryegrass (Lolium perenne) and white clover (Trifolium repens) on plant, animal, meat, and human health, by assessing the metabolomic profiles of plants, animals, meat, and humans (Kronberg et al., 2021). In other words, we are beginning to address the question: are there ecological implications and connections between phytochemically diverse and functional swards and animal and human health (Provenza et al., 2019)?

All procedures outlined were approved by the Lincoln University Animal Ethics Committee (AEC:2021-10) and Human Ethics Committee (HEC 2021-36).

Beef finishing trial

Experimental site, design, pasture, and herbage

This study was conducted at the Lincoln University sheep farm, Canterbury, New Zealand (-43.644601 ◦North, 172.451242 ◦East), to evaluate liveweight gain, oxidative stress, estimated dry matter intake (DMI) of growing cattle and untargeted metabolomics of their meat (i.e., beef) over 120-d between October-February of 2021–2022. The experiment was run as a complete block design involving three grazing treatments.

During the trial, 30, two-year-old, mixed sex (heifers = 15, steers = 15), Angus cattle (315 ± 35 kg liveweight) were grouped into triplicates based on their sex, liveweight, and date of birth, and randomly allocated to one of three forage treatments. Cattle were born in the spring of 2020 (October-November), and following weaning, grazed a traditional pastoral diet consisting of perennial and white clover at the Lincoln University Ashley Dene Station, Canterbury, New Zealand (-43.645105 ◦North, 172.332991 ◦East). The three dietary treatments consisted of a ‘status quo’ perennial ryegrass-based sward (PRG), a complex multispecies mixture (CMS, n = 24 forage species) including grasses (PRG, Italian ryegrass: Lolium multiflorum, tall fescue: Festuca arundinacea, meadow fescue: Festuca pratensis, prairie grass, Bromus wildenowii, timothy Phleum partense, and cocksfoot: Dactylis glomerata), legumes (white clover, red clover: Trifolium pratense, lucerne: Medicago sativa, lotus: lotus pedunculatus, lupins: Lupinus and vetch: Vicia sativa as well as crimson: Trifolium incartum, balansa: Trifolium michelianum, persian: Trifolium resupinatum and strawberry clovers: Trifolium fragiferum), in addition to brassicas (rape: Brassica napus and radish: Raphanus sativus) and herbs (chicory: Chicorium intybus and plantain: Plantago lanceolata). The third treatment consisted of a functionally diverse sward, as described by Garrett et al. (2022), containing five adjacent monoculture strips (AMS) of perennial ryegrass, red clover, plantain, chicory and lucerne that were sown longitudinally across the paddock of equal areas. Each group of animals was strip-grazed horizontally across each paddock behind a temporary electric fence which was moved every third day.

The total grazing area available consisted of 9.5 ha split across 10 paddocks ranging from 0.7–1.5 ha in size. The paddocks were randomly assigned a treatment forage mixture, which was established by direct drilling on the 23rd of April 2021. Two weeks prior to drilling, each paddock was fertilised with 250 kg/ha of diammonium phosphate (DAP), and an herbicide applied to remove the existing PRG sward. One application of urea (100 kg/ha) was applied to experimental areas prior to the beginning of the grazing experiment in October of 2021. Each treatment group of cattle grazed the paddocks (n = 10 per treatment) between the 10^{th of} October and the 16th of February before slaughter by a commercial abattoir in accordance with the Ministry of Primary Industries of New Zealand. Each treatment group was allocated the same amount of herbage dry matter ad libitum, which was calculated to exceed metabolic requirements for each animal based on their average liveweight (i.e., ~ 4% liveweight per day). Cattle were allocated a new area for grazing every third day using temporary electric fencing materials. The break size of each treatment was calculated by the available herbage mass (pre grazing herbage mass -post grazing herbage mass) assuming a target post-grazing herbage mass of 1700 kg DM/ha. Cattle were confined into the calculated break area by the use of a back fence and had free access to water at all times.

Herbage mass was estimated pre-and post-grazing by measuring all plant material collected within a 0.2 m² quadrant, randomly placed at two different locations of each strip of the AMS paddocks, and 6–10 locations within paddocks of PRG and CMS. The plant material collected was removed of dirt and oven dried 60℃ until achieving a constant weight. The average difference between pre- and post-grazing herbage mass (i.e. herbage disappearance) was used to estimate daily DMI from the area available for grazing. Botanical and chemical composition, dry matter percentage and the metabolomic profile of each diet were evaluated monthly, by collecting 10–20 random hand grab samples of herbage cut to grazing height (~ 4 cm above ground) from each treatment and within each monoculture of the AMS diet. After collection, pasture samples were mixed thoroughly and sub-sampled. One sub-sample was weighed and oven dried (60℃ for 48 hours) to determine DM%, one was used to determine botanical components, and two were stored at -20℃ until freeze dried and ground through a 1 mm mill (ZM200 Retsch) to analyse metabolomic profiles and chemical components (water-soluble carbohydrate, organic matter, dry matter, dry organic matter digestibility, and crude protein (WSC, OM, DM, DOMD and CP, respectively) by near infrared spectrophotometry (NIRS, Model: FOSS NIR Systems 5000, Maryland, USA; calibration > 90%). A sub-sample (5g) of dried herbage of each sward was pooled from each measurement to evaluate the metabolomic profile. Individual plants in the AMS treatment were also pooled (5g each) to evaluate metabolomic profile of the herbage offered to cattle.

Animal measurements

Each month, on the third day following the allocation of a pasture break, cattle were removed from their experimental pastures and herded to a crush where they were weighed (Tru-test, Texas, USA) and a sample of blood was collected from the coccygeal vein in two 10 mL vacutainers (with either lithium heparin or EDTA as the coagulant), which were immediately stored on ice. Following collection, blood was transported to a laboratory and the plasma was separated by centrifuging the samples at 3000 x g for 12 minutes. The plasma was pipetted into 1.5 mL Eppendorf containers and stored at -20°C until analysed. Plasma collected from Li Heparin coated vacutainers was analysed for total oxidative status (TAS: Cat No. NX2332), while those collected from EDTA coated vacutainers were analysed for non-esterified fatty acids (NEFA: Cat No. FA115) using a Randox RX Daytona clinical analyser (Randox Laboratories Ltd, Crumlin, UK) following their respective kit instructions. Blood samples and liveweights were collected on day one of the experiment and then each month after to allow for a 30-day acclimatisation to experimental diets.

The cattle grazed the experimental diets until 16 February 2022, when they were transported to a local commercial abattoir, weighed and then slaughtered. Following slaughter, meat quality/carcass (dressing out, carcass weight, temperature, pH, marbling, meat colour, fat colour, rib fat, and ossification) characteristics of each animal were evaluated. One cut from the pasoas major muscle of each animal was ground (5 mm mincing plate), sub sampled, homogenised by treatment, and processed in a commercial kitchen into beef patties weighing 250 ± 5g consisting of 100% beef (Food South, South Island Hub, Canterbury, New Zealand). Patties were blast frozen and stored at -20°C. The sub-sample of beef collected from each animal was storage at -20℃ until analysed using metabolomic methods.

Human post-prandial trial

Between May and July 2021, a three-way, double-blind, randomised cross-over trial was conducted at the Lincoln University (New Zealand) Exercise Science Laboratory (-43.644828, 172.465163, Canterbury, New Zealand).

Twenty-three individuals (n = 16 female, age = 58 ± 13 years and 7 = male, age = 53 ± 11.9 years old) participated in the experiment. On three occasion, each participant was offered two ~ 250 g (486 g raw weight) beef patties produced from the cattle that grazed the three treatments. During the trial, baseline, and post-prandial measurements (at 0-, 3-, and 5-h post-meal) of plasma glucose, cholesterol profiles, C-reactive protein (CRP), blood pressure and metabolomic profiles were collected to evaluate the response to a single meal of beef.

The trial was conducted using two cohorts of participants (n = 12 each) over 6 weeks (2, 3-week periods). During each 3-wk period, each participant was assigned one of the three treatments randomly (double-blind). A wash-out period of 7 days was used between visits (i.e., feeding), until each participant had consumed all three beef treatments. Prior to each visit, participants had a meal consisting of pasta (300g) and vegan pesto, which was prepared and delivered by Pasta Vera (Wigram, Christchurch, NZ) which was selected to prevent residual dietary effects on metabolomic and metabolic samples the following day and provide a simple standardised diet each night prior to consumption of the beef meal. Four participants were assessed in a day over 3 days each week and they were asked to abstain from smoking, alcohol, and eating after dinner the night prior to the consumption of beef. Each participant assessed and offered the beef meal individually with a 30-min gap between the arrival of each person for the collection of baseline measurements. Participants arrived to the laboratory for baseline measurements between 07:30 and 09:30 each day.

Patties were thawed at 4°C for 36h prior to cooking. During each visit, 486g of raw ground beef was cooked (mean = 341g cooked meat per person) using an electric fry pan (temperature controlled) with no added oil or salt until each patty reached an internal temperature of 71°C (approximately 5.5 min each side). The fry pan was thoroughly cleaned between each participant.

During the first measurement, each participant completed a physical activity survey (IPAQ-SF) and a body composition scan via an 8-electrode bioelectrical impedance (Accuniq B380) followed by blood pressure and carotid-femoral pulse wave velocity (tonometry) (Sphygmocor® EXCEL) tests on arrival. Resting venous blood samples were collected by a registered nurse from the antecubital vein using standard guidelines. Resting blood samples and blood pressure measurements were taken on arrival prior to meat consumption, and following baseline measurements, each participant was given up to 15 minutes to consume the meat allocated, or until they were satiated. Plasma samples for metabolite analyses were collected in an EDTA coated vacuette and centrifuged at 5000 RPM for 10 minutes at 4 ℃. Plasma was pipetted into 2mL plastic microtubules and stored at -20℃ until analysed. Participants were asked to remain sedentary (i.e., seated with no movement other than viewing an electronic device, or reading) without food (allowed to drink water) for five hours with blood pressure and blood sample collections repeated at 3 and 5 hours post prandial. After the final measurement, participants were offered a wide variety of snacks to eat prior to departure. To encourage participants to come to all three measurements, we offered a $50 voucher.

Metabolomic analyses

Sample extraction

Lipidomics

Samples were extracted for lipidomics based on the method of Huynh et al. (2019). Solvents were Optima LCMS grade from ThermoFisher (Auckland, New Zealand).

Pasture

Pasture samples were freeze dried, and ground using a bead mill (TissueLyser II, Qiagen) and a metal bead, shaking for 5 minutes at 30 Hz. 10 mg of powdered pasture was weighed into a 2 mL microcentrifuge tube, and 10 µL of internal standard mix (Splash Mix, Avanti Polar Lipids, Alabaster, Alabama, USA) and 190 µL of butanol:methanol (1:1 v/v) with 5 mM ammonium formate, added. A metal bead was added to the microcentrifuge tube and the sample shaken for 5 minutes at 30 Hz in a bead shaker (TissueLyser II, Qiagen). The samples were then sonicated for 60 minutes at 20°C before centrifugation (14 000 x g for 10 minutes at 20°C, Heraeus Megafuge 8R, Thermo Fisher). 150 µL of supernatant was aliquoted into an amber chromatography vial with 250 µL flat bottom insert, and 20 µL supernatant transferred to a tube for making a pooled quality control sample.

Meat

Homogenized sub-samples of meat were freeze dried, and ground using a bead mill (TissueLyser II, Qiagen, Hilden, Germany) and a metal bead, shaking for 5 minutes at 30 Hz. 25 mg of powdered meat was weighed into a 2 mL microcentrifuge tube, and 10 µL of internal standard mix (Splash Mix, Avanti Polar Lipids, Alabaster, Alabama, USA) and 990 µL of butanol:methanol (1:1 v/v) with 5 mM ammonium formate, added. A metal bead was added to the microcentrifuge tube and the sample shaken for 5 minutes at 30 Hz in a bead shaker (TissueLyser II, Qiagen). The samples were then sonicated for 60 minutes at 20°C before centrifugation (14 000 x g for 10 minutes at 20°C, Heraeus Megafuge 8R, Thermo Fisher, Auckland, New Zealand). 200 µL of supernatant was aliquoted into an amber chromatography vial with 250 µL flat bottom insert, and 50 µL supernatant transferred to a tube for making a pooled quality control sample.

Plasma

Ten µL of plasma was pipetted into a 2 mL microcentrifuge tube, and 5 µL of internal standard mix (Splash Mix, Avanti Polar Lipids, Alabaster, Alabama, USA) and 95 µL of butanol:methanol (1:1 v/v) with 5 mM ammonium formate, added. The sample was shaken for 5 minutes at 30 Hz in a bead shaker (TissueLyser II, Qiagen). The samples were then sonicated for 60 minutes at 20°C before centrifugation (14 000 x g for 10 minutes at 20°C, Heraeus Megafuge 8R, Thermo Fisher). 80 µL of supernatant was aliquoted into an amber chromatography vial with 250 µL flat bottom insert, and 20 µL supernatant transferred to a tube for making a pooled quality control sample.

Semi-polar metabolomics

Pasture

Pasture samples were freeze dried and ground as for lipidomics. 50 mg of powdered pasture was weighed into a 2 mL microcentrifuge tube and 980 µL of methanol:water (9:1 v/v), 20 µL of an in-house stable isotope internal standard mix, and a metal bead were added. The pasture was extracted in a bead shaker (5 min, 30 Hz, TissueLyser II, Qiagen) and tubes transferred to a -20°C freezer for 30 minutes. The extracts were centrifuged for 10 min at 14 000 x g at 4°C (Heraeus Megafuge 8R, Thermo Fisher). 180 µL of the supernatant was transferred to an amber chromatography vial with tapered insert for GC-MS/MS analysis, 180 µL transferred to an amber chromatography vial with flat bottomed insert for semi-polar metabolomics analysis, and 180 µL transferred to a 15 mL Falcon tube for a pooled QC sample.

Samples for GC-MS/MS were dried in a vacuum centrifuge concentrator (Christ RVC 2–18 CD plus, Martin Christ) and derivatised by methoxymation and silylation.

Meat

Meat samples were freeze dried and ground as for lipidomics. 25 mg of powdered meat was weighed into a 2 mL microcentrifuge tube and 980 µL of methanol:water (9:1 v/v), 20 µL of an in-house stable isotope internal standard mix, and a metal bead were added. The meat was extracted in a bead shaker (5 min, 30 Hz, TissueLyser II, Qiagen) and tubes transferred to a -20°C freezer for 30 minutes. The extracts were centrifuged for 10 min at 14 000 x g at 4°C (Heraeus Megafuge 8R, Thermo Fisher). 180 µL of the supernatant was transferred to an amber chromatography vial with tapered insert for GC-MS/MS analysis, 180 µL transferred to an amber chromatography vial with flat bottomed insert for semi-polar metabolomics analysis, and 180 µL transferred to a 15 mL Falcon tube for a pooled QC sample.

Samples for GC-MS/MS were dried in a vacuum centrifuge concentrator (Christ RVC 2–18 CD plus, Martin Christ GmbH, Osterode am Harz, Germany) and derivatised by methoxymation and silylation.

Plasma

One hundred µL of plasma was pipetted into a 2 mL microcentrifuge tube and 890 µL of methanol:water (9:1 v/v), and 20 µL of an in-house stable isotope internal standard mix. Plasma was extracted in a bead shaker (5 min, 30 Hz, TissueLyser II, Qiagen) and tubes transferred to a -20°C freezer for 30 minutes. The extracts were centrifuged for 10 min at 14 000 x g at 4°C (Heraeus Megafuge 8R, Thermo Fisher). 180 µL of the supernatant was transferred to an amber chromatography vial with tapered insert for GC-MS/MS analysis, 180 µL transferred to an amber chromatography vial with flat bottomed insert for semi-polar metabolomics analysis, and 180 µL transferred to a 15 mL Falcon tube for a pooled QC sample.

Samples for GC-MS/MS were dried in a vacuum centrifuge concentrator (Christ RVC 2–18 CD plus, Martin Christ) and derivatised by methoxymation and silylation.

Metabolomics analyses

Lipidomics

Lipidomics analyses were carried out using LC-qToF-MS (LCMS 9030, Shimadzu, Kyoto, Japan) in data independent analysis mode. Samples (2 µL) were injected onto a C18 column (Acquity CSH-C18, 100 x 2.1 mm, 1.7 µm). The gradient programme was as follows: starting, 10% B, 2.7 minutes, 45% B; 2.8 minutes, 53% B; 9 minutes, 65% B; 9.1 minutes, 89% B; 11 minutes, 92% B; 11.1 minutes, 100% B; 13.9 minutes, 100% B; 14 minutes, 10% B; 17 minutes, 10% B. The flow rate was 0.4 mL/min. The column oven was set to 60°C and autosampler set to 20°C. The mass spectrometer was run in data independent analysis mode with an electrospray ionisation source. This was set up as follows: scanning data between m/z 250 and 1300 in ToF mode with an event time of 0.03 seconds, and 41 windows of m/z 20 from m/z 300 to 1100, in MZ/MS mode with a collision energy of 23 and a collision energy spread of 6, and event time of 0.02 seconds. Analyses were carried out in positive ionisation mode.

Semi-polar metabolomics

Semi-polar metabolomics analyses were carried out using LC-qToF-MS (LCMS 9030, Shimadzu, Kyoto, Japan) in data independent analysis mode. Samples (4 µL) were injected onto a C18 column (Hypersil Gold C18, 100 x 2.1 mm, 1.9 µm particle size, ThermoFisher). The following chromatography gradient was used: Starting, 0% B; 1 minute, 0% B; 11 minutes, 100% B; 16.9 minutes, 100% B; 17 minutes, 0% B, 18 minutes, 0% B. Solvent A was 16 mM ammonium formate in water, and solvent B was 0.1% formic acid in acetonitrile, and the flow rate was 0.4 mL/min. Duplex column mode was used, with alternating positive and negative mode analyses on different columns. While one column was separating a sample, the other column was regenerating with the following programme: starting, 100% B; 12 minutes, 100% B, 15 minutes; 0% B, 17 minutes, 0% B. The column oven was set to 40°C and autosampler set to 4°C. The mass spectrometer was run in data independent analysis mode with an electrospray ionisation source. This was set up as follows: scanning data between m/z 55 and 1100 in ToF mode with an event time of 0.03 seconds, and 41 windows of m/z 20 from m/z 100 to 900, in MZ/MS mode with a collision energy of 23 and a collision energy spread of 15, and event time of 0.02 seconds. Analyses were carried out in both positive and negative ionisation modes.

GC-MS/MS

GC-MS/MS analyses were carried out on a GCMS TQ8040 instrument (Shimadzu) using the SmartMetabolites MRM method. This method covers around 350 metabolites spread across central metabolic pathways (e.g., energy metabolism, protein synthesis and breakdown, lipid metabolism). In brief, this targeted metabolomics method includes 450 multiple reaction monitoring transitions, covering approximately 400 individual metabolites (several metabolites have two chromatographic peaks). Analysis conditions were: injector temperature 250°C, interface temperature 280°C, and MS source temperature 200°C. The chromatography was carried out on a BPX5 column (30 m x 0.25 mm, 0.25 µm film thickness) with the following temperature gradient: start 60°C, 2 minutes 60°C, 20 minutes 330°C, 23 minutes 330°C. Two µL of derivatised sample was injected in split injection mode with a ratio of 1:30, and carrier gas linear velocity was held constant at 40 cm/second. Peaks were identified based on ratio of quantifier and qualifier ion and retention index.

Data processing

LC-MS data were processed using MS DIAL (Tsugawa et al., 2015), including peak deconvolution, database-based identification, and retention time and mass alignment. Data were normalised using LOWESS normalisation based on a quality control sample between every ten sample injections.

Lipidomics identifications were refined based on modelling likely retention times based on lipid standards. Lipid identifications that fell outside a +/- 10% retention time range were considered to be incorrectly identified.LC-MS metabolomics data was manually searched against an in-house library based on accurate mass, retention time and isotopic score using LabSolutions Insight Explore (Shimadzu).

Statistical analyses

Cattle performance, apparent DMI, and blood parameters (TAS, PUN and NEFA) were tested for normality and analysed by either linear-mixed models or generalised-linear mixed-models using the lme4 package in R studio (Pinheiro and Bates, 2018) in R (v.3.4.4., https://www.r-project.org/). Models were fitted by evaluating treatment, sampling date (LW, LWG, DMI, TAS, PUN and NEFA), and sex and their interactions as fixed effects and using the individual animal as the experimental unit. Baseline blood metabolite profile of individuals was used as a covariate.

Measurements of cardiovascular function and circulating concentrations of blood metabolites (glucose, CRP, cholesterol, high-density lipoprotein: HDL, and low-density lipoprotein: LDL) were tested for normality, and treatment effects were evaluated using either linear-mixed models or generalised-linear mixed-models using the lme4 package in R studio. Regression models were developed using treatment, time of sampling, sex, order of treatment eaten and their interactions as fixed effects, while individual nested within the day of sampling was considered as the random effect.

Statistical analyses of metabolomic data was performed in Metaboanalyst 5.0. plant and beef metabolomic and lipidomics were assessed using a one-way ANOVA. Plasma metabolomics were analysed using a combination of two-way ANOVA to evaluate time-series data (treatment × time interactions), and linear models to account for random error of individual people and to enable covariate analysis (sex, age, body mass index, amount of beef consumed). Due to the explorative nature of this study, differences in raw P-values were considered significant if P < 0.1, while the probability of FDR was considered as acceptable if P < 0.1. The FDR indicates the probability of a false discovery or a Type I error (the null hypothesis is excluded when in fact it is true) of multivariate testing. Due to the explorative nature of this study, we have also presented the FDR in cases where raw P-values were > 0.1 to ensure potentially meaningful results were not excluded and denoting them as ‘possible’ or ‘conjecture’ (Zhang et al. 2010; Kim et al., 2015).

Cattle performance and finishing

Sward composition

The botanical components of each sward are presented in Table 1. The PRG sward contained greater components of grasses (Italian and perennial ryegrass), which were at a greater reproductive state of growth compared with both CMS and AMS diets (40, 25, and 11% DM, respectively P < 0.01). The percentage of dead material also increased with the PRG diet compared with CMS and AMS (14, 4, and 2%, respectively P < 0.01). The proportion of weed species such as mellow (Modiola caroliniana), shepherd’s purse (Capsella bursa-pastoris), barley grass (Hordeum murinum) and stinging nettle (Urtica dioica) increased in the AMS swards compared with PRG or CMS (Table 1). The AMS sward contained a greater percentage of forbs (chicory and plantain) than CMS and PRG (27, 5, and 0%, P = 0.07, respectively). The clover content of the PRG diet was 16.3% DM, compared with 24.2% and 8.7% for CMS and AMS, respectively. The clover content of the swards did not differ statistically, although there was a high variation of clover between paddocks, explaining the high standard error reported for this variable (± 6.9%).

Table 1

Estimated dry matter intake (DMI, sward conditions and botanical components of the Perennial ryegrass based (PRG), complex multispecies mixture (CMS) or adjacent monoculture strips (AMS) swards.
Variable	PRG	CMS	AMS	SE	P-value
Variable	PRG	CMS	AMS	SE	Diet	Date
Herbage Mass (kg DM/ha)	5078	6144	4671	285.5	0.002	***
Post-grazing mass (kg DM/ha)	2436	2495	2350	129	0.72	***
Apparent Allocation (kg DM/cow/d)	25.6	20.2	23.0	2.43	0.33	0.02
DMI (/cow per day)	17.2	13.1	16.2	1.33	0.12	0.49
Botanical component
Grasses vegetative (% DM)	17.3	12.0	3.29	4.94	0.07	0.79
Grasses reproductive	40.1	25.0	11.4	14.0	**	0.21
Clovers	16.3	24.2	8.66	6.87	0.17	0.84
Other legumes	0.00	17.4	8.4	5.95	0.21	0.89
Herbs	0	4.52	26.9	15.94	0.07	0.99
Brassica	0	6.1	0	0.80	***	0.5
Weeds	12.3^ab	6.7^a	39.1^b	7.16	***	0.99
Dead	14.0^b	4.0^a	2.4^a	2.59	**	0.03
* P < 0.05, ** P < 0.01, P < 0.001

Pre- and post-grazing herbage masses differed across swards and interactions between date and type of sward were also identified (P < 0.05), reflecting seasonal variation in growth of individual components of each treatment. Apparent daily DM allocations and individual animal DMI were similar across treatments (Table 1), although the apparent DMI of the CMS diet was numerically less than PRG and AMS (13.1 versus 17.2 and 16.2 kg DM/cow, respectively). The chemical composition of each dietary sward-treatment and the individual monocultures within the AMS are presented in Table 2. All chemical components differed (P < 0.05) between individual plants but were not always different across overall sward-treatment. The DM content declined with chicory, but remained similar across all other plant species and were not different between diets. The NDF content of perennial ryegrass (47% DM) was greater within the individual monocultures of AMS but was reduced by chicory and red clover (30 and 32% DM, respectively), lowering the average NDF content of the AMS sward compared with PRG and CMS (38, 43 and 46% DM, respectively). The ADF content followed a similar trend to NDF, except that the ADF content of plantain was greater than ryegrass in the AMS diet (Table 2). The CP content was similar between the PRG and AMS swards but declined 15% with the CMS (Table 2, P < 0.001). While the CP content of chicory and plantain was low (< 12% DM), the CP of lucerne and red clover was between 17–18%. The WSC content of the PRG and CMS swards were similar (18.3%) but declined to 16% (P < 0.01) in the AMS, reflecting the lower WSC content in red clover and lucerne compared with chicory and ryegrass (Table 2). The WSC content of the ryegrass offered in the AMS was less than the PRG (15 vs 18.3%). The DMD of PRG and AMS were similar and declined in the CMS (Table 2).

Table 2

Dry matter (DM), neutral detergent fibre (NDF), acid detergent fibre (ADF), organic matter (OM), crude protein (CP), dry matter digestibility (DMD), water-soluble carbohydrates (WSC) and metabolisable energy (ME) content of three dietary swards consisting of either Perennial ryegrass (PRG), a complex multispecies mixture (CMS), and adjacent monocultural strips of Ryegrass, Chicory, Plantain, Lucerne, and Red clover. Note: Chemical components of individual plants offered in the AMS diet are also presented in this table. Standard error of the mean (SE) is also presented along with P values of diet and date.
Plant	DM %	NDF %	ADF %	OM %	CP %	DMD	WSC %	ME (MJ/ kg DM)
PRG	23.5^a	43.2^b	27.7^ab	91.8^b	14.7^ab	70.8^b	18.3^b	10.8^b
CMS	21.6^a	46.1 ^b	30.1^b	92.2^b	12.5^a	66.9^ab	18.3^b	10.3^a
AMS average	19.8^a	37.9^a	28.7^ab	91.4^ab	14.6^ab	68.7^b	15.4^ab	10.5^a
AMS – Ryegrass	25.3^b	47.7^b	29.6^b	92.2^b	14.7^ab	67.6^b	15.0^ab	10.4^a
AMS – Chicory	13.9^a	30.8 ^a	26.6^a	90.1^a	13.2^ab	73.7^b	20.7^b	11.0^b
AMS – Plantain	19.9^b	35.5 ^a	31.5^b	92.0^b	10.7^a	63.9^a	14.9^a	9.8^a
AMS – Lucerne	20.8^b	37.8 ^a	27.3^a	92.0^b	7.0^b	70.5^b	13.1^a	10.8^b
AMS – Red clover	18.9^b	32.1 ^a	25.4^a	91.5^ab	18.9^b	72.6^b	14.7^a	11.1^b
SE	1.05	2.27	0.92	0.32	0.94	1.39	1.15	0.182
P-Value Plant	***	***	**	**	***	***	**	***
* P < 0.05, ** P < 0.01, P < 0.001. ^a−b Superscripts that differ within columns are different P < 0.05.

Animal Performance and Carcass Characteristics

Average liveweight of the cattle grazing CMS was 4% less compared with PRG and AMS (Table 3). The interaction between day and sward-treatment was significant (P < 0.001), reflecting an increase in average animal liveweight in January. Heifers were (P = 0.06) heavier than steers by 10.7 kg across all treatments. Average daily liveweight gain (LWG) of animal grazed on AMS was 8 and 15% greater compared with those grazing PRG and CMS, although, significant Diet x Day interactions were also observed (Table 3). Both TAS and NEFA declined (P < 0.1) in animals grazing CMS compared with PRG or AMS (Table 3). The plasma concentration of NEFA was 7% grater in heifers compared to steers (P < 0.05).

Table 3

*Average liveweight (LW), daily liveweight gain (LWG), total antioxidant status (TAS), and non-esterified fatty acids* (NEFA) of cattle fed either perennial ryegrass (PRG), complex multispecies mixture (CMS), or adjacent monoculture strips (AMS). Significant P-values of Treatment, Day, Sex and T×D are indicated * P < 0.05, ** P < 0.01, P < 0.001.
Variable		Treatment				P-value
	Sex	PRG	CMS	AMS	SE	Treat	Day	Sex	T × D
LW (kg)	H	382 ^a	372 ^b	386 ^a	3.17	0.03	***	0.06	***
	S	373 ^a	363^a	377 ^a	3.28
LWG (kg/day)	H	1.21^a	1.11^b	1.31 ^a	0.042	0.01	***	0.18	***
	S	1.13 ^a	1.02 ^b	1.23 ^a	0.043
TAS	H	1.03 ^ab	1.01 ^a	1.03 ^ab	0.011	0.08	***	0.31	0.06
	S	1.04 ^ab	1.02 ^a	1.03 ^ab	0.013
NEFA	H	0.22 ^b	0.16 ^a	0.20 ^b	0.019	0.06	**	*	0.16
	S	0.32 ^c	0.22 ^b	0.28 ^c	0.028
^a−c Superscripts that differ within rows (variables) are different P < 0.05.

Carcass characteristics are presented in Table 4. The slaughter liveweight of animals grazing on AMS was 10% less than those fed PRG, and greater than CMS (Table 4). Compared with PRG and CMS, the hot carcass weight of AMS cattle was 11 and 6% lighter, respectively (P = 0.03). Meat colour (redness of the meat) increased with both the CMS and AMS compared with PRG (P = 0.02). Ossification increased (P = 0.08) by 9 and 17%, respectively, in CMS and AMS animals compared with those fed PRG.

Table 4

Carcass characteristics and final liveweight of cattle finished on either a perennial ryegrass pasture (PRG), complex multispecies mixture (CMS) or adjacent monoculture strips (AMS). Significant P-values of Treatment, Day, Sex and T×D are reported.
Variable	PRG	CMS	AMS	SE	Treatment	Sex	T×S
Final Liveweight (kg)	440^b	408^ab	394^a	12.2	0.02	0.42	0.33
Dressing out %	53.7	54.3	53.2	0.48	0.25	0.32	0.62
Hot carcass weight (kg)	236^b	222^ab	209^a	7.02	0.03	0.64	0.27
pH	5.64	5.86	5.72	0.105	0.29	0.79	0.42
Temperature	7.63	7.47	7.54	0.206	0.75	0.30	0.20
Marbling	303	308	312	10.4	0.85	0.59	0.95
Meat colour	3.21^a	4.32^b	4.10 ^a	0.263	0.02	0.15	0.42
Fat Colour	1.83	1.20	1.40	0.258	0.23	0.90	0.60
Rib fat	4.04	3.58	4.20	0.658	0.80	0.47	0.81
Ossification	118^a	130^ab	142^ab	7.35	0.08	0.46	0.92
^a−c Superscripts that differ within rows (variables) are different P < 0.05

Human feeding trial

Meat intake and blood

The amount of meat consumed was not affected by treatment, although men consumed more than women (296 ± 21.7 versus 197 ± 6.6 g). Diet by time interactions were not significant (P > 0.1, Table 5). Plasma concentration of glucose, HDL, and CRP, diastolic blood pressure (DBP), triglyceride and cholesterol were not affected by treatment (P > 0.1, Table 6). Systolic blood pressure (SBP) increased (P = 0.04) in people who ate AMS beef compared with those that ate PRG. Diastolic blood pressure of those who ate CMS beef was intermediary (Table 6). Circulating concentration of low-density lipoprotein (LDL) tended (P = 0.06) to increase in people fed AMS beef compared with those fed PRG or CMS beef. The ratio of total cholesterol to HDL increased following the consumption of AMS beef compared with CMS beef, while PRG-raised beef was intermediary (Table 6)

Table 5

Total amount of beef consumed on average during each visit. P-values indicate significant differences of finishing diet: perennial ryegrass (PRG), complex multispecies mixture (CMS), and adjacent monoculture strips (AMS); Sex, and T×S.
								P-value
	Sex	PRG	SE	AMS	SE	CMS	SE	Treat	Sex	T×S
Beef Eaten (g/person)	M	315	22.5	314	23.2	261	19.3	0.11	***	0.16
Beef Eaten (g/person)	F	189	6.0	209	7.4	195	6.5	0.11	***	0.16

Table 6

Mean plasma concentration and standard error (SE) of glucose, low density and high density lipoprotein (LDL and HDL, respectively), systolic blood pressure (SBP), diastolic blood pressure (DBP), C-reactive protein, cholesterol, Triglyceride, and ratio of high density to low density lipoprotein (HDL ratio) measured in patients fed beef finished on perennial ryegrass (PRG), complex multispecies mixture, and adjacent monoculture strips (AMS). P-values indicate significance of finishing treatment, time, sex and T×T interactions (P < 0.01 **, P < 0.001 ***).
Variable	Sex	PRG	SE	AMS	SE	CMS	SE	P-Values
Variable	Sex	PRG	SE	AMS	SE	CMS	SE	Treat	Time	Sex	T×T
Glucose (mmol/L)	M	6.21^a	0.183	6.20 ^a	0.175	6.30 ^a	0.184	0.75	***	0.31	0.9
Glucose (mmol/L)	F	6.01^a	0.125	6.00 ^a	0.119	6.10 ^a	0.126	0.75	***	0.31	0.9
LDL	M	1.94^a	0.160	2.05^ab	0.177	1.83^a	0.144	0.06	***	0.02	0.94
LDL	F	2.45^c	0.181	2.62^cd	0.203	2.28^c	0.154	0.06	***	0.02	0.94
HDL	M	1.31 ^a	0.098	1.34 ^a	0.097	1.37 ^a	0.097	0.23	0.95	0.12	0.2
HDL	F	1.49 ^a	0.066	1.52 ^a	0.065	1.54 ^a	0.065	0.23	0.95	0.12	0.2
SBP (mmol/Hg)	M	126^a	3.35	131^b	3.57	128^a	3.46	0.04	0.04	0.80	0.25
SBP (mmol/Hg)	F	125^a	2.36	130^b	2.48	127^a	2.38	0.04	0.04	0.80	0.25
DBP	M	74.3 ^a	2.43	77.3 ^a	2.54	74.4 ^a	2.41	0.10	**	0.58	0.63
DBP	F	72.9 ^a	1.68	74.8 ^a	1.72	72.9 ^a	1.67	0.10	**	0.58	0.63
CRP(mmol/L)	M	1.78 ^a	0.342	1.81 ^a	0.326	1.52 ^a	0.329	0.37	0.47	0.87	0.18
CRP(mmol/L)	F	1.88 ^a	0.210	1.91 ^a	0.214	1.61 ^a	0.220	0.37	0.47	0.87	0.18
Cholesterol	M	2.07 ^a	0.052	2.15 ^a	0.054	2.08 ^a	0.051	0.31	0.58	0.05	0.99
Cholesterol	F	2.19 ^a	0.041	2.27 ^a	0.043	2.20 ^a	0.041	0.31	0.58	0.05	0.99
Triglyceride	M	1.86 ^a	0.223	1.87 ^a	0.223	1.74 ^a	0.224	0.21	***	0.95	0.46
Triglyceride	F	1.89 ^a	0.15	1.90 ^a	0.146	1.77 ^a	0.147	0.21	***	0.95	0.46
Cholesterol: HDL ratio	M	3.37^ab	0.198	3.42^b	0.199	3.20^a	0.179	0.05	0.51	0.86	0.34
Cholesterol: HDL ratio	F	3.41^ab	0.140	3.46^b	0.140	3.24^a	0.128	0.05	0.51	0.86	0.34
^a−d Superscripts within the same row that differ are different (P < 0.05).

Metabolomics

Plant metabolomics

A total of 70 features were identified when C18 metabolomics were run in positive ionisation mode, 23 differed (raw P-value < 0.1) by treatment and three met the FDR criteria. The top 30 are presented in Fig. 1. Of metabolites to note, AMS increased the relative intensity of proline betaine, biochanin A an isoflavonoid, and baicalin a flavone glycoside (FDR < 0.1).

A total of 166 features were identified from GC-MS/MS analysis, of which, 49 differed by treatment based on raw P-values < 0.1 and 27 of these were below FDR < 0.1. These compounds were dicarboxylic acids, sugar acids, phenylacetic acids, TCA acids and amino acids. Other metabolites of interest are gamma tocopherol, which increased in the AMS diet (P < 0.0001, FDR < 0.0001). Of the lipids identified from lipidomics run in negative mode, 1651 features were identified, and of those, 45 were positively matched. Many were neutral glycosphingolipids, ceramides, and glycerophospholipids. All, except one ceramide, increased in AMS and CMS meat compared with PRG (Table 7). One hundred and eighty-six features were identified by lipidomics in negative mode. Of these, 65 differed with treatment (P < 0.1), and the FDR of 5 of these metabolites were less than the 10% threshold, while the FDR of the remaining metabolites were > 0.28. Of those that differed, all were increased in the AMS and CMS swards compared with PRG (Table 7). Some of these were sterols, triacyl- and diacyl-glycerides, and glycerolphosphocholines.

Table 7

Plant lipids detected positive and negative mode from three diets: Perennial ryegrass (PRG), complex multispecies mixture (CMS), and adjacent monoculture strips (AMS). Lipids are presented as abbreviated subclass and suggested fatty acid chain positions along with the mean relative intensities, P-value and false discovery statistic (FDR)
Lipid	Sub Class	PRG	CMS	AMS	SE	P-value	FDR
HexCer 34:1 O4\|HexCer 18:1 O3/16:0(2OH)	Neutral glycosphingolipids	8659^a	27668 ^b	13433^b	7587.0	0.001	< 0.01
Cer 34:1 O4\|Cer 18:1 O3/16:0(2OH)	Ceramide	16950^b	44339^a	53921^a	7600.0	< 0.001	0.02
Cer 42:1 O4\|Cer 18:1 O3/24:0(2OH)	Ceramide	90933^a	166521^b	176007^b	19303.1	< 0.001	0.02
PI 34:2\|PI 16:0 18:2	Phosphatidylinositol glycerophosphoinositol	63542^a	154153^b	180953^b	29914.1	< 0.001	0.08
CL 65:0\|CL 16:0 16:0 16:0 17:0	Cardiolipin/phospholipid	623^a	972^a	1265^b	152.4	< 0.001	0.08

Beef metabolomics

A total of 154 metabolites were detected in meat by GC-MS/MS targeted metabolomics. Twenty-eight metabolites were different at raw P < 0.1 between AMS, CMS, and PRG (Fig. 2). Following FDR correction only linoleic acid differed (FDR < 0.1), and it was greater in meat produced from AMS compared with PRG or CMS (Table 8). Several other metabolites increased in AMS beef in which raw P-values were significant, and FDR values suggest them as ‘possible’. These were hypoxanthine, pantothenic acid, ribulose-5-phosphate (R5P) and eicosapentaenoic acid (EPA), which were all elevated in AMS beef compared with PRG or CMS (Table 8). Of the other metabolites that differed, but did not meet the FDR criteria, several were medium- and long-chain fatty acids such as decanoic, lauric, stearic acid, palmitic, margaric, eladic, octanoic acid, oleic, caproic acid, palmitoleic FA, as well as amino acids such as beta alanine and histamine, and the dipeptide carnosine. Initially, these results suggest an impact of treatment on fat content or deposition, which covers both medium and longer-chain FA of metabolites with a P-value < 0.1 and suggests that treatment altered pathways associated with linoleic acid, beta-alanine, histidine, methionine, and taurine metabolism. Further evaluation of the fatty acid profiles of meat may be necessary to understand the relationship between treatments and fatty acid metabolism in cattle.

Table 8

Known metabolites and their mean relative intensity from the meat of cattle grazing either perennial ryegrass-and clover (PRG), complex multispecies (CMS) or adjacent monoculture strips (AMS). Metabolites were found to differ (P-value) *by treatment following False Discovery Rate correction (FDR < 0.1).*
Metabolites	Method	PRG	AMS	CMS	SD	P-value	FDR
Linoleic acid	GCMS	583^a	1377^b	680 ^a	497.6	0.00	0.00
Hypoxanthine	GCMS	13922^a	23796^b	16793^a	7961	0.002	0.14
Pantothenic Acid	GCMS	35.6^a	76.7^b	46.8^a	35.51	0.003	0.14
Ribulose-5-phosphate	GCMS	32.8^a	69.2^b	47.1^a	29.79	0.004	0.14
Eicosapentaenoic acid	GCMS	521.6^a	773.3^b	528.2 ^a	249.61	0.09	0.48
Benzoate	Semi-polar Neg	130^b	164^b	83^a	60.5	0.01	0.09
Histidine	Semi-polar Neg	1419^b	1218^a	1296^ab	203.3	0.01	0.09
Carnosine	Semi-polar Neg	43012^b	38502 ^a	39079^ab	4782.9	0.02	0.09
3-Hydroxy benzaldehyde	Semi-polar Neg	130^ab	164^b	95^a	57.9	0.02	0.09
Beta alanine	Semi-polar Pos	1944^b	1717^a	1700^a	647.9	0.00	0.02
Taurine	Semi-polar Pos	557^a	1273^b	593^a	538.0	0.00	0.05
Palmitocarnitine	Semi-polar Pos	2545^a	4451^a	3211^b	1435.1	0.00	0.06
Hypoxanthine	Semi-polar Pos	11387^a	14826^b	13019^a	2551.3	0.00	0.06
Methionine	Semi-polar Pos	614^b	440^a	470^a	193.3	0.00	0.05
Tyrosine	Semi-polar Pos	421^b	318^a	334^a	103.8	0.01	0.08
Lauroylcarnitine	Semi-polar Pos	995^a	1985^b	1223^a	732.1	0.01	0.09
^{a−c: Superscripts within rows that are differ are different P < 0.05.}

A total of 1053 features were detected in positive mode – this includes unknowns and tentative IDs (untargeted discovery analysis). Two hundred and sixty-seven metabolite features differed between diets. While we were unable to confidently identify many of these features, many were in a retention time range that corresponds to smaller lipids and larger polyphenols (tannins, flavonoids, hydrolysable tannins) as well as alkaloids. Furthermore, these metabolites appear to be greater in AMS beef compared with PRG or CMS.

Multivariate analysis of metabolites which positively matched the metabolite library yielded 18 that were significantly different between treatments (raw P < 0.05), 8 of which met the FDR cut-off (Table 8). The heat map in Fig. 3 presents the treatment differences of these metabolites. Based on Euclidean differences, semi-polar metabolomic profiles of AMS and PRG were different, while CMS shared similarities with both PRG and AMS. Most of the metabolites were amino acids such as histidine, taurine, leucine, tyrosine, hypoxanthine, and methionine.

One hundred and nineteen metabolite features were identified when the C18 column was run in negative mode, including unknown and tentative IDs. Four features differed between diets following false discovery correction (Table 8). Metabolites such as histidine and carnosine were also identified by semi-polar metabolomics run in negative ionisation. Relative intensities of benzoate and 3-hydroxybenzaldehyde, which are metabolites related to phytochemicals, increased with AMS compared to PRG. Relative intensities of carnosine and histidine declined with AMS and CMS compared with PRG. These metabolites are associated with beta-alanine metabolism.

Lipidomic Analyses

Three thousand two hundred and twenty-three lipid features were detected in positive mode, including unknowns and tentative IDs (untargeted discovery analysis). Based on FDR < 0.1, 258 lipid features differed; however, only 223 were able to be positively matched to the in-house library (Table 9) and from these only two features had raw P values < 0.1, and FDR were < 0.3 and > 0.1.. Lipids from a wide range of lipid classes differed between beef coming from different swards treatment, though most appear to be phospholipids and variations on phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine and phosphatidylglycerol. Diacylglycerols were generally elevated in AMS beef than CMS or PRG, while triacylglycerol lipids were often reduced compared with CMS or PRG (Fig. 4). While most lipids that were affected by treatment-swards were not identified, hierarchical clustering suggest that lipidomics of CMS and PRG beef showed more similarities than the lipidomic profile produced from the AMS beef (Fig. 4).

Table 9

Lipid features (retention time, mass: charge m/z, lipid class and mean relative intensities) identified in positive mode that differed significantly with dietary treatment of a monotonous Perennial ryegrass with white clover (PRG), complex multispecies mixture (CMS) and adjacent monoculture strips (AMS). P-values are indicated along with FDR statistics. Tukey’s HSD indicate the Treatment groups which differ P < 0.05. Lipid features and their classes are indicated along with the retention time and frequency that they were removed from the column.
Retention time		Metabolite	Class	MON	AMS	CMS	SD	P-value	FDR	Tukey's HSD
8.236	742.5723	PE O-37:2\|PE O-19:1¹	EtherPE²	154	115	167	50.71	2.89E-08	2.28E-06	CMS-AMS; PRG-AMS
7.888	728.5596	PE O-36:2\|PE O-18:1	EtherPE	1089	870	1155	251.01	1.62E-05	0.000641	CMS-AMS; PRG-AMS
7.01	828.5768	PC 36:3\|PC 18:0	PC³	757	1482	822	441.21	4.10E-05	0.001079	CMS-AMS; PRG-AMS
8.992	756.5909	PE O-38:2\|PE O-20:1	EtherPE	357	309	355	124.43	8.33E-05	0.001645	CMS-AMS; PRG -AMS
6.86	773.5361	PG 36:2\|PG 18:1	PG⁴	163	348	172	152.77	0.000621	0.009807	CMS-AMS; PRG -AMS
3.734	450.2995	LPE O-17:1	LPE⁵	57	122	78	51.19	0.001287	0.016946	PRG -AMS; PRG -CMS
7.398	804.5769	PC 34:1\|PC 16:0	PC	9844	8860	9836	1588.85	0.003168	0.031729	CMS-AMS; PRG -AMS
7.174	672.4988	PE O-32:2\|PE O-16:1	EtherPE	183	165	240	84.45	0.004202	0.031729	CMS-AMS
6.739	786.5296	PS 36:2\|PS 18:1	PS⁶	467	911	551	332.64	0.00431	0.031729	CMS-AMS; PRG-AMS
7.282	740.525	PE 36:3\|PE 18:0	PE⁷	647	1212	635	425.31	0.004415	0.031729	CMS-AMS; PRG-AMS
7.938	742.5399	PE 36:2\|PE 18:0	PE	2092	4242	2330	1559.63	0.004418	0.031729	CMS-AMS; PRG-AMS
8.033	730.5762	PE O-36:1\|PE O-18:1	EtherPE	329	278	348	109.58	0.005563	0.033452	CMS-AMS; PRG-AMS
4.079	464.3148	LPE O-18:1	LPE	966	2108	1072	1006.30	0.005912	0.033452	CMS-AMS; PRG AMS
7.054	816.5762	PC 35:2\|PC 17:1	PC	389	402	415	96.35	0.005928	0.033452	CMS-AMS; PRG-AMS
6.717	747.5185	PG 34:1\|PG 16:0	PG	850	1665	755	736.44	0.007229	0.037123	CMS-AMS; PRG-AMS
8.202	730.5413	PE 35:1\|PE 18:0	PE	357	340	316	153.16	0.007519	0.037123	PRG-AMS
6.073	745.5052	PG 34:2\|PG 16:0	PG	69	206	73	107.00	0.011568	0.053756	CMS-AMS; PRG-AMS
6.306	788.5459	PC 33:2\|PC 15:0	PC	181	295	193	91.00	0.012619	0.055384	CMS-AMS; PRG-AMS
7.773	712.5295	PE O-35:3\|PE O-17:1	EtherPE	359	526	506	184.78	0.015836	0.065843	PRG-CMS
7.077	740.5246	PE 36:3\|PE 18:1	PE	702	1400	863	597.01	0.017307	0.068362	PRG-AMS
6.938	714.5089	PE 34:2\|PE 16:0	PE	208	337	247	125.39	0.020425	0.076836	PRG-AMS
6.494	721.5024	PG 32:0\|PG 16:0	PG	65	118	73	45.36	0.023137	0.079866	PRG-AMS
6.718	720.4974	PE O-36:6\|PE O-16:1	EtherPE	3318	3296	3741	1403.00	0.023252	0.079866	CMS-AMS
6.717	802.5616	PC 34:2\|PC 16:0	PC	3675	6182	4075	1456.00	0.026928	0.088638	CMS-AMS; PRG-AMS
^{1 Metabolite suggests the lipid class, followed by suggested positioning of fatty acid substitutions. 2Ether−linked phosphatidylethanolamine: EtherPE. 3Phosphatidylcholine: PC. 4 Lysophosphatidylethanolamine: LPE. 5 Phosphatidylserine: PS. 6 Phosphatidylglycerol: PG. 7 Phosphatidylethanolamine: P}

A total of 887 features were detected in negative mode, 304 lipids differed by diet (P < 0.1, FDR < 0.5), and 60 features were highly significant (P < 0.05, FDR < 0.1; Fig. 4). From negative lipidomic output, 79 features were able to be positively matched to the in-house library and of these 24 features differed by treatment (P < 0.1, FDR < 0.1), a further 9 features were different by treatment, but FDR were > 0.1 and < 0.2 (P < 0.1). Figure 5 presents differences in peak intensities of features. The AMS treatment reduced ether linked phosphatidylethanolamines and increased several phosphatidylcholine (PC), phosphatidylglycerol (PG) and lypophosphatidylcholine (LPC) features in meat, compared with either CMS or PRG meat.

Human Metabolomics

Using GC-MS/MS, 81 metabolites were identified. Correcting for variation of individuals (e.g., sex, age, BMI), and the time post-prandial, gamma tocopherol increased in plasma following consumption of the AMS beef compared with CMS or PRG. In addition, 50 metabolites varied across time of sampling but there was no time × treatment interactions (P > 0.1). When sex, and BMI were included as covariates in the statistical analyses, a total of 7 metabolites differed by treatment with a raw P-value < 0.1. Across the 3- and 5h samples gamma tocopherol (raw P-value = 0.001, adj P-value = 0.09), 3-aminopropionic acid (raw P-value = 0.01, adj P-value = 0.46), indol-3-acetic acid (raw P-value = 0.06, adj P-value = 0.88) increased in plasma following consumption of the AMS meat compared with PRG or CMS. Docosahexanoic acid (raw P-value = 0.03, adj P-value = 0.81) and kynurenine (raw P-value = 0.08, adj P-value = 0.88) increased following consumption of the PRG meat.

Untargeted C18 positive metabolomics yielded 21 metabolites, 7 of which (carnitine, acetoacetate and phosalone features) were significant (raw P-values < 0.1) and of those, only one unknown compound had an FDR < 0.1. Targeted C18 positive metabolomics identified 41 metabolites, of which only glycochenodeoxycholate declined (raw P-value = 0.05) in participants fed the AMS beef compared with those fed PRG or CMS, although differences exceeded the FDR threshold. Untargeted C18 metabolomics yielded 23 features, 10 of which differed by treatment (raw P-value < 0.1), and one unknown metabolite met FDR criteria < 0.1. From targeted C18 metabolomics, 5 features were different (P-values < 0.1). These were 3-hydroxymethylglutarate (HMG), indoxyl sulphate, arginine, oxoadipate, and D-sedoheptulose (Fig. 6, FDR = 0.38). Compared with people fed PRG and CMS beef, AMS participants had greater relative abundance of HMG, arginine and oxoadipate, while indoxyl sulphate and D-sedoheptulose declined compared with participants that consumed PRG or CMS meat (Fig. 6).

Lipidomics analyses yielded 1200 metabolite features in positive ionisation, of which 21 differed with treatment (adjusted P-value < 0.1). In general, phosphatidylcholine lipids declined following the consumption of AMS and CMS beef compared with PRG (Fig. 7). When run in negative mode, 371 features were found, of which 61 were affected by treatment (raw P < 0.1). Only six lipids met the FDR criteria (FDR = 0.07). They were variations of phosphatidylcholines and cardiolipins that seem to be enhanced in plasma of those fed the AMS beef (Fig. 7).

We hypothesised that taxonomical and phytochemical diversity of plants species and their spatial arrangement across the paddock (i.e., functionality of plant diversity; Garret, 2022), would alter metabolomic and lipidomic profiles of beef and the post-prandial metabolomic profiles of human consumers. Our results suggest that metabolomic profiles of beef and human consumers are indeed altered by the type of sward that cattle were grazing. While changes in the metabolomic profile of human plasma was more subtle due to the limited detection of metabolites [(in house metabolite library of AgResearch (Lincoln, New Zealand)], approximately 30% of the metabolites detected were affected in their relative abundances based on raw P-values, and several unknown compounds differed by treatment-sward. These effects have never been reported, making this study the first of its kind, supporting the original hypothesis of Gregorini et al. (2017) that states “We are what we eat eats!”.

Metabolomic profiles of the three treatments-swards differed and this affected cattle metabolism. Enrichment ratios suggest that pathways associated with linoleic fatty acid metabolism and eicosapentaenoic acid were upregulated in cattle grazing AMS compared with PRG or CMS, indicating increased presence of polyunsaturated fatty acids in the AMS swards. Eicosapentaenoic acid is an omega-3 fatty acid usually found in fatty fish such as salmon but also in pasture-raised land animals and has several health benefits including antidepressant activities and prevention of coronary heart disease thrombosis – blood clots, and arterial plaque formation (Russell & Bürgin-Maunder, 2012). Eicosapentarnoic acid is a long-chain fatty acid (FA), that is synthesised from alpha linolenic acid (ALA), the main omega-3 FA in plants, by the enzyme delta-15-desaturase. Alpha-linolenic acid is also an essential omega-3 FA with pleiotropic actions that have anti-inflammatory, anti-depressant, neuroprotective and cardio-protective outcomes (Blondeau et al., 2015). While ALA is reported to reduce circulating LDL-C, (Blondeau et al., 2015) point out that many of the positive responses have been reported from chronic ALA supplementation, which is supported by others who have reported subtle changes to plasma, EPA and omega-3 FA after 4 weeks consumption of either grain- or pasture-finished beef, or 6 months consumption of eggs and chicken meat enriched with algae-sourced omega PUFA (McAfee et al., 2011; Stanton et al., 2020). Therefore, the short-term nature and potentially low dosage of the current study may not have captured this response as responses to omega FA in human consumers.

The AMS also increased the relative abundance of pantothenic acid (PA; vitamin B5) in beef. Vitamin B5 is an essential nutrient involved in the synthesis of coenzyme A – needed for FA synthesis and degradation. Supplementation of PA reduced cholesterol levels by 17% after 4 weeks in people suffering from dyslipidaemia (high LDL, TGL, and total cholesterol) (Chen et al., 2015). The long-term supplementation of PA is inversely correlated to CRP levels and has been suggested to reduce low-grade inflammation associated with early stages of heart disease (Chen et al., 2015). In addition, 3-hydroxybenzaldehyde (3HBA) also increased in beef produced with the AMS. Benzaldehydes are commonly found in nature and 3HBA has a OH group at the meta position of the phenol ring and it is a potent intracellular antioxidant (Bortolomeazzi et al., 2007; K.-Y. Chen et al., 2021). Research in mice has identified that 3HBA is vaso-protective by preventing angiogenesis, and adenosine diphosphate - a platelet agonist – thrombosis formation (Kong et al., 2016), in addition to antimutagenic (Gustafson, 2000), cytotoxic (Tseng et al., 2001), and antibacterial functions(Chen et al., 2021) and can even influence flavour (Bortolomeazzi et al., 2007). Vitamin B5 and 3HBA have vaso-protective, anti-inflammatory and ‘heart-healthy’ roles were upregulated by the AMS beef, which could have human health benefits compared with traditional -status quo- PRG beef, highlighting the need for longer-term human health studies.

Our results suggest that human metabolism following a single meal of beef produced from different treatment-swards is altered. Gamma tocopherol (vitamin E) was elevated in the AMS sward, and this increased the circulating concentrations of gamma-tocopherol in people following the consumption of AMS beef. Gamma tocopherol is the second most common form of vitamin E, a lipophilic molecule (second to α-tocopherol) but as 1 of 8 isomers, it is the most effective anti-inflammatory, due to its ability to inhibit cyclooxygenase and 5-lipoxygenase activity, and antioxidant through capture of lipophilic electrophiles (Jiang et al., 2001). The content of gamma tocopherol was observed to be numerically greater in relative abundance of the chicory plant compared with all other plants or plant mixes. eported gamma tocopherols are present in Chicorium spinosum (spiny chicory), which belongs to the same family Ateraceae as the C. intybus used in the current study. Research of γ-tocopherol has also indicated that it is superior to α-tocopherol in its ability to reduce cancer cell growth, trapping of reactive N-species through formation of 5-nitro-γ-tocopherol (Jiang et al., 2022).

The human results also presented five different (P-values < 0.1) features between treatments. These were 3-hydroxymethylglutarate (HMG), indoxyl sulphate, arginine, oxoadipate, and D-sedoheptulose, with their FDR locating them as ‘hypothetically possible’. Compared with people eating PRG and CMS beef, AMS participants had greater relative abundance of HMG, arginine and oxoadipate, while indoxyl sulphate and D-sedoheptulose declined (Fig. 6). Therefore, we conjecture that these differences can add to the positive effects of AMS to humans. Indoxyl sulphate, is produced from the breakdown of tryptophan by colon microbes – and may reflect exogenous intake of phenols and indoles (Evenepoel et al., 2009; Schepers et al., 2010). In patients with chronic kidney disease, it is considered a uremic toxin, but as it is also reported to vary considerably between individuals and increases with the consumption of meat when compared to a vegetarian diet (Patel et al., 2012). Accumulation of indoxyl sulphate is also associated with additional side effects in the kidneys, bones, and cardiovascular system (Evenepoel et al., 2009; Schepers et al., 2010). Arginine is an amino acid required help build muscle and rebuild tissue. The body also convert this amino acid into the chemical nitric oxide, which is a vasodilator (Preli et al., 2002). These results not only need further investigation, but also highlight the need for longer-term human health studies.

Metabolomic profiling offers an insight into changes in biochemical pathways and small metabolites fundamental to basic metabolism. While changes in the relative abundance of metabolites indicate that some pathways may be upregulated or downregulated in response to the inclusion of functionally diverse AMS swards, further analysis in other laboratories with broad metabolomic and lipidomic libraries are needed to identify, quantify and compare our current results. The untargeted LCMS/MS analyses revealed 267 metabolites in human plasma differed by treatment, while we were unable to identify these lipids, they still prove our hypothesis that sward composition influences consumer metabolism. There is a need to continue developing diverse pasture systems based on metabolomic profiling of livestock which consume such feeds and the human consumer.

Beef finishing and performance

Several interactions between treatment and date were identified across plant and animal measurements. While cattle were allocated similar amounts of herbage and estimated DMI was similar across treatment, DMI of the CMS cattle was numerically less than PRG or AMS and this most likely explains the lower LW and LWG achieved from this treatment. Furthermore, CMS herbage contained lower quantities of CP, greater content of NDF, ADF and DMD tended to be the less than AMS or PRG, which likely explain the reduced performance of CMS cattle. Plasma concentrations of NEFA declined 29% compared with AMS or PRG. The plasma concentrations of NEFA reflect fatty acids produced in the liver from adipose reserves within the animal. Therefore, the lower concentration of NEFA in the CMS treatment may reflect lower caloric intake and rumen digesta outflow, reducing nutrient supply to the host animals and thereby average LWG.

Treatment by time interactions for livestock performance (LW and LWG) were observed across all treatments reflecting seasonal variation of plant growth and chemical components such as CP and ME content. The AMS diet contained less CP and was generally lower in WSC and DMD compared with PRG counterparts and reduced the nutritional value of the diet during late summer compared with PRG and CMS. However, the chemical composition and growth of the AMS sward was favourable for liveweight growth and performance during late Spring and early Summer (November-January), which maintained average LWG and LW of the cattle grazing this sward when averaged across the experimental period. Further evaluation of the seasonal variation of growth and chemical composition of diverse diets requires further evaluation at the farm system-level to understand performance outcomes and profitability at greater scale.

Carcass characteristics and beef quality were assessed using Silver Fern Farms Eating Quality System. Meat colour increased in cattle that grazed the AMS swards, compared with those grazing PRG or CMS. Meat colour of the chilled rib eye muscle area was scored against a set of colour reference standards with scores increasing relative to the darkness of the meat colour. Meat colour reflects myoglobin levels and ultimate pH as postmortem glycolysis decreases muscle pH, resulting in a brighter colour of meat that is preferred by consumers. If the ultimate meat pH is high, the meat proteins will associate with more water in the muscle and therefore fibres will be tightly packed resulting in meat which is ‘tough’. However, the ultimate pH in AMS cattle was similar to PRG or CMS cattle. In addition, the numerical difference between the treatments was minor (0.2–0.3) and did not alter the overall grade of the carcass, thus this difference may not be meaningful in terms of marketability, but nonetheless requires further study.

The results of this study indicate that animal performance can be enhanced using AMS when used strategically throughout the finishing period. Further development of the plant species and their level of inclusion in the sward along with consideration of their seasonality relative to animal finishing and requirements are needed to further evaluate and develop a more consistent response in both performance and metabolomic profiles. It is also worth noting that the limited number of animals used in this study is likely also a factor in the response and studies with a greater number of animals and replication across mobs (as opposed to the individual animal used currently) will provide a better indication of the animal response.

This research investigated the hypothesis that plant species metabolomic profiles would influence not only the metabolomic profiles of cattle, but the metabolomic profiles of human consumers. Based on a range of targeted and non-targeted approaches, the metabolomes of cattle that grazed a functionally diverse sward -AMS- differed from those that grazed either a PRG or a CMS mixture containing a greater relative abundance of metabolites which may have several health benefits, particularly regarding cardiovascular health. Moreover, the metabolomic profiles of the swards where cattle were finished were reflected in the post-prandial plasma of human consumers which could have promising benefits to human health, even if this needs further confirmation in follow-up studies. For instance, gamma-tocopherol was elevated in the AMS sward which directly increased the relative abundance in the plasma of the human consumer. These are the first to show a clear metabolomic linkage between grazing management and the consumer.

Author Contribution

P.G, A.F, F.P, S.V, M.H, F.L, and C.E were responsible for conceptualisation and experimental design. A.F, K.G, C.J.M, and C.E performed experimental proceduresA.F performed statistical analysesA.F and P.G prepared the main manuscript textAll authors reviewed the manuscript

Acknowledgements

This research was supported by Lincoln University, Silver Fern Farms, Fertiliser NZ Ltd, and Craigmore Sustainables Ltd. Many thanks to Agricom for provision of the seed mix and agronomic expertise in the establishment of forages used to for finishing cattle. Also, thanks to Jasmine Tanner and Dr David Baird for their statistical advice in the experimental design and analysis of metabolomic data.

Animal products reflect the history of our landscape, foodscapes and agricultural systems manifested though soil and plant chemistry, and thereby our health and that of the planet (Gregorini et al., 2017)

Barnard, N. D., & Leroy, F. (2020). Children and adults should avoid consuming animal products to reduce the risk for chronic disease: Debate Consensus. The American Journal of Clinical Nutrition, 112(4), 937–940. https://doi.org/10.1093/ajcn/nqaa237
Beck, M. R., & Gregorini, P. (2020). How Dietary Diversity Enhances Hedonic and Eudaimonic Well-Being in Grazing Ruminants. Frontiers in Veterinary Science, 7. https://doi.org/10.3389/FVETS.2020.00191
Beck, M. R., & Gregorini, P. (2021). Animal Design Through Functional Dietary Diversity for Future Productive Landscapes. Frontiers in Sustainable Food Systems, 5. https://doi.org/10.3389/FSUFS.2021.546581
Blondeau, N., Lipsky, R. H., Bourourou, M., Duncan, M. W., Gorelick, P. B., & Marini, A. M. (2015). Alpha-Linolenic Acid: An Omega-3 Fatty Acid with Neuroprotective Properties—Ready for Use in the Stroke Clinic? BioMed Research International, 2015. https://doi.org/10.1155/2015/519830
Bortolomeazzi, R., Sebastianutto, N., Toniolo, R., & Pizzariello, A. (2007). Comparative evaluation of the antioxidant capacity of smoke flavouring phenols by crocin bleaching inhibition, DPPH radical scavenging and oxidation potential. Food Chemistry, 100(4), 1481–1489. https://doi.org/10.1016/j.foodchem.2005.11.039
Chan, D. S. M., Lau, R., Aune, D., Vieira, R., Greenwood, D. C., Kampman, E., & Norat, T. (2011). Red and Processed Meat and Colorectal Cancer Incidence: Meta-Analysis of Prospective Studies. PLoS ONE, 6(6). https://doi.org/10.1371/JOURNAL.PONE.0020456
Chen, K.-Y., Chen, Y.-J., Cheng, C.-J., Jhan, K.-Y., Chiu, C.-H., & Wang, L.-C. (2021). 3-Hydroxybenzaldehyde and 4-Hydroxybenzaldehyde enhance survival of mouse astrocytes treated with Angiostrongylus cantonensis young adults excretory/secretory products. Biomedical Journal, 44(6), S258–S266. https://doi.org/10.1016/j.bj.2020.11.008
Chen, Y., Zhao, S., & Zhao, Y. (2015). Efficacy and tolerability of coenzyme A vs pantethine for the treatment of patients with hyperlipidemia: A randomized, double-blind, multicenter study. Journal of Clinical Lipidology, 9(5), 692–697. https://doi.org/10.1016/j.jacl.2015.07.003
Evenepoel, P., Meijers, B. K. I., Bammens, B. R. M., & Verbeke, K. (2009). Uremic toxins originating from colonic microbial metabolism. Kidney International, 76, S12–S19. https://doi.org/10.1038/ki.2009.402
Garrett, K., Beck, M. R., Marshall, C. J., Maxwell, T. M. R., Logan, C. M., Greer, A. W., & Gregorini, P. (2021). Varied diets: implications for lamb performance, rumen characteristics, total antioxidant status, and welfare. Journal of Animal Science, 99(12). https://doi.org/10.1093/JAS/SKAB334
Garrett, K., Marshall, C. J., Beck, M. R., Fleming, A., Maxwell, T. M. R., Logan, C. M., Greer, A. W., & Gregorini, P. (2021). From the get-go: Dietary exposure in utero and in early life alters dietary preference in later life. Applied Animal Behaviour Science, 244. https://doi.org/10.1016/J.APPLANIM.2021.105466
Garrett, K., Marshall, C. J., Beck, M. R., Maxwell, T. M. R., Logan, C. M., & Gregorini, P. (2022). A diverse diet as an alternative to ryegrass can improve the total antioxidant status of dams at lambing. Frontiers in Sustainable Food Systems, 6. https://doi.org/10.3389/FSUFS.2022.885436
Gregorini, P., Gordon, I. J., Kerven, C., & Provenza, F. D. (2022). Editorial: Grazing in Future Multi-Scapes: From Thoughtscapes to Landscapes, Creating Health From the Ground Up. Frontiers in Sustainable Food Systems, 6, 213. https://doi.org/10.3389/FSUFS.2022.880809/BIBTEX
Gregorini, P., Villalba, J. J., Chilibroste, P., & Provenza, F. D. (2017). Grazing management: Setting the table, designing the menu and influencing the diner. Animal Production Science, 57(7), 1248–1268. https://doi.org/10.1071/AN16637
Gustafson, D. L. (2000). Vanillin (3-methoxy-4-hydroxybenzaldehyde) inhibits mutation induced by hydrogen peroxide, N-methyl-N-nitrosoguanidine and mitomycin C but not 137Cs gamma-radiation at the CD59 locus in human-hamster hybrid AL cells. Mutagenesis, 15(3), 207–213. https://doi.org/10.1093/mutage/15.3.207
Huynh, K., Barlow, C. K., Jayawardana, K. S., Weir, J. M., Mellett, N. A., Cinel, M., Magliano, D. J., Shaw, J. E., Drew, B. G., & Meikle, P. J. (2019). High-Throughput Plasma Lipidomics: Detailed Mapping of the Associations with Cardiometabolic Risk Factors. Cell Chemical Biology, 26(1), 71–84.e4. https://doi.org/10.1016/j.chembiol.2018.10.008
Jiang, Q., Christen, S., Shigenaga, M. K., & Ames, B. N. (2001). γ-Tocopherol, the major form of vitamin E in the US diet, deserves more attention. American Journal of Clinical Nutrition, 74(6), 714–722. https://doi.org/10.1093/AJCN/74.6.714
Jiang, Q., Im, S., Wagner, J. G., Hernandez, M. L., & Peden, D. B. (2022). Gamma-tocopherol, a major form of vitamin E in diets: Insights into antioxidant and anti-inflammatory effects, mechanisms, and roles in disease management. Free Radical Biology and Medicine, 178, 347–359. https://doi.org/10.1016/j.freeradbiomed.2021.12.012
Kong, B. S., Im, S. J., Lee, Y. J., Cho, Y. H., Do, Y. R., Byun, J. W., Ku, C. R., & Lee, E. J. (2016). Vasculoprotective Effects of 3-Hydroxybenzaldehyde against VSMCs Proliferation and ECs Inflammation. PLOS ONE, 11(3), e0149394. https://doi.org/10.1371/journal.pone.0149394
Kronberg, S. L., Provenza, F. D., van Vliet, S., & Young, S. N. (2021). Review: Closing nutrient cycles for animal production – Current and future agroecological and socio-economic issues. Animal, 15, 100285. https://doi.org/10.1016/j.animal.2021.100285
Marshall, C. J., Beck, M. R., Garrett, K., Barrell, G. K., Al-Marashdeh, O., & Gregorini, P. (2021). Nitrogen balance of dairy cows divergent for milk urea nitrogen consuming either plantain or perennial ryegrass. Animals, 11(8). https://doi.org/10.3390/ANI11082464
Marshall, C. J., Garrett, K., Van Vliet, S., Beck, M. R., & Gregorini, P. (2022). Dietary and Animal Strategies to Reduce the Environmental Impact of Pastoral Dairy Systems Result in Altered Nutraceutical Profiles in Milk. Animals, 12(21). https://doi.org/10.3390/ANI12212994
McAfee, A. J., McSorley, E. M., Cuskelly, G. J., Fearon, A. M., Moss, B. W., Beattie, J. A. M., Wallace, J. M. W., Bonham, M. P., & Strain, J. J. (2011). Red meat from animals offered a grass diet increases plasma and platelet n -3 PUFA in healthy consumers. British Journal of Nutrition, 105(1), 80–89. https://doi.org/10.1017/S0007114510003090
Micha, R., Michas, G., & Mozaffarian, D. (2012). Unprocessed red and processed meats and risk of coronary artery disease and type 2 diabetes - An updated review of the evidence. Current Atherosclerosis Reports, 14(6), 515–524. https://doi.org/10.1007/S11883-012-0282-8
Patel, K. P., Luo, F. J.-G., Plummer, N. S., Hostetter, T. H., & Meyer, T. W. (2012). The Production of p-Cresol Sulfate and Indoxyl Sulfate in Vegetarians Versus Omnivores. Clinical Journal of the American Society of Nephrology, 7(6), 982–988. https://doi.org/10.2215/CJN.12491211
Preli, R. B., Klein, K. P., & Herrington, D. M. (2002). Vascular effects of dietary l-arginine supplementation. Atherosclerosis, 162(1), 1–15. https://doi.org/10.1016/S0021-9150(01)00717-1
Provenza, F. D., Kronberg, S. L., & Gregorini, P. (2019). Is grassfed meat and dairy better for human and environmental health? Frontiers in Nutrition, 6. https://doi.org/10.3389/FNUT.2019.00026/FULL
Russell, F., & Bürgin-Maunder, C. (2012). Distinguishing Health Benefits of Eicosapentaenoic and Docosahexaenoic Acids. Marine Drugs, 10(12), 2535–2559. https://doi.org/10.3390/md10112535
Schepers, E., Glorieux, G., & Vanholder, R. (2010). The Gut: The Forgotten Organ in Uremia? Blood Purification, 29(2), 130–136. https://doi.org/10.1159/000245639
Schiere, J. B. (Hans), & Gregorini, P. (2023). Complexity, Crash and Collapse of Chaos: Clues for Designing Sustainable Systems, with Focus on Grassland-Based Systems. Sustainability, 15(5), 4356. https://doi.org/10.3390/su15054356
Stanton, A. V., James, K., Brennan, M. M., O’Donovan, F., Buskandar, F., Shortall, K., El-Sayed, T., Kennedy, J., Hayes, H., Fahey, A. G., Pender, N., Thom, S. A. M., Moran, N., Williams, D. J., & Dolan, E. (2020). Omega-3 index and blood pressure responses to eating foods naturally enriched with omega-3 polyunsaturated fatty acids: a randomized controlled trial. Scientific Reports, 10(1), 15444. https://doi.org/10.1038/s41598-020-71801-5
Tseng, T.-H., Tsheng, Y.-M., & Lee, Y.-J. (2001). Cytotoxicity effects of di- and tri-hydroxybenzaldehydes as a chemopreventive potential agent on tumor cells. Toxicology, 161(3), 179–187. https://doi.org/10.1016/S0300-483X(01)00346-8
Tsugawa, H., Cajka, T., Kind, T., Ma, Y., Higgins, B., Ikeda, K., Kanazawa, M., VanderGheynst, J., Fiehn, O., & Arita, M. (2015). MS-DIAL: data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nature Methods, 12(6), 523–526. https://doi.org/10.1038/nmeth.3393
van Vliet, S., Provenza, F. D., & Kronberg, S. L. (2021). Health-Promoting Phytonutrients Are Higher in Grass-Fed Meat and Milk. Frontiers in Sustainable Food Systems, 4, 299. https://doi.org/10.3389/FSUFS.2020.555426/BIBTEX

No competing interests reported.

Download PDF

Editorial decision: Revision requested
26 Apr, 2024
Reviews received at journal
25 Apr, 2024
Reviewers agreed at journal
09 Apr, 2024
Reviewers agreed at journal
09 Mar, 2024
Reviews received at journal
16 Feb, 2024
Reviewers agreed at journal
06 Feb, 2024
Reviewers invited by journal
06 Feb, 2024
Editor assigned by journal
06 Feb, 2024
Submission checks completed at journal
01 Feb, 2024
First submitted to journal
27 Dec, 2023

You are reading this latest preprint version

Connecting plant, animal, and human health using untargeted metabolomics.

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Beef finishing trial

Experimental site, design, pasture, and herbage

Animal measurements

Human post-prandial trial

Metabolomic analyses

Sample extraction

Lipidomics

Semi-polar metabolomics

Pasture

Meat

Plasma

Metabolomics analyses

Lipidomics

Semi-polar metabolomics

GC-MS/MS

Data processing

Statistical analyses

Results

Cattle performance and finishing

Human feeding trial

Meat intake and blood

Metabolomics

Plant metabolomics

Beef metabolomics

Lipidomic Analyses

Human Metabolomics

Discussion

Beef finishing and performance

Conclusion

Declarations

Author Contribution

Acknowledgements

References

Additional Declarations

Status:

Version 1