New temperate seaweed targets for mitigation of ruminant methane emissions: an in vitro assessment

ABSTRACT Methane is a potent greenhouse gas with a comparatively short (9 yr) atmospheric lifetime; therefore effective strategies for methane mitigation will contribute significantly to meeting the UN Sustainable Development Goal #13 of taking rapid action against climate change. Methane emissions from enteric fermentation constitute a large proportion of agricultural greenhouse gas emissions. Low inclusions of red seaweed from the genus Asparagopsis have demonstrated near elimination of enteric methane from ruminants; however, only a limited number of other seaweeds have been assessed for their anti-methanogenic potential. New Zealand red seaweed species Bonnemaisonia hamifera, Euptilota formisissima, Plocamium cirrhosum, Vidalia colensoi, and identified aquaculture target species Ecklonia radiata and Ulva sp. B were investigated as anti-methanogenic feed additives. Seaweeds were included at 0%, 2%, 6%, or 10% of feed organic matter (OM, ryegrass hay) during in vitro fermentation assays using rumen inoculant from non-lactating Friesian x Jersey dairy cows, using Asparagopsis armata as a positive control. Total gas, methane, hydrogen, volatile fatty acids, and organic matter degradation were measured over a 48 h incubation. Inclusion of all seaweeds except Ulva sp. B reduced the production of methane at either 6 or 10% OM. Bonnemaisonia hamifera was the best performing seaweed, reducing the production of methane by 17.1%, 95.4%, and 98.8% relative to the basal feed substrate control at inclusion levels of 2%, 6%, and 10% OM, respectively, with notable increases in the production of hydrogen. Euptilota formisissima and P. cirrhosum reduced the production of methane by up to 50.5 and 39.5%, respectively, at an inclusion level of 10%, with minimal effects on measured fermentation parameters. Bromoform, the primary bioactive component in Asparagopsis, was not detected in any of the new seaweeds tested. Our results therefore identify potential alternative anti-methanogenic seaweed targets that are less susceptible to the loss of volatile bioactives during processing.

Thus, the development of strategies to mitigate CH 4 emissions from ruminant production systems is critical for increasing the sustainability of these industries, especially in the context of New Zealand. Furthermore, the high global warming potential and short atmospheric life time of CH 4 means that the implementation of effective solutions for CH 4 mitigation will have rapid impacts, and therefore contribute substantially to meeting the UN Sustainable Development Goal #13 of taking urgent and decisive action against climate change (United Nations, 2016).
Over 900 species of seaweeds are currently recognized throughout New Zealand (Nelson et al., 2019), yet only a fraction of these species have been tested for their capacity to mitigate enteric CH 4 production (Machado et al., 2014;Maia et al., 2016). Species of brown seaweed including Cystoseira trinodis and Dicotya bartayresii and species of green seaweed from the genus Ulva have demonstrated significant antimethanogenic activity in vitro and reduced total gas production (Dubois et al., 2013;Machado et al., 2014). However, aside from Asparagopsis, the antimethanogenic potential of red seaweeds remains largely untested. Red seaweeds produce a large variety of halogenated secondary metabolites (Kladi et al., 2004), some of which are likely candidates for exhibiting antimethanogenic effects. Therefore, this study investigated the anti-methanogenic properties of New Zealand seaweeds that included examples of red (Bonnemaisonia hamifera, Euptilota formisissima, Plocamium cirrhosum, Vidalia colensoi), brown (Ecklonia radiata), and green (Ulva sp. B) seaweeds. Ecklonia radiata and Ulva sp. B were chosen specifically as there is an opportunity for rapid development and industry implementation due to both existing and developing aquaculture for securing supply, which aligns with meeting the UN Sustainable Development Goal #13 of taking rapid action against climate change (United Nations, 2015). We quantified the anti-methanogenic activity of seaweeds using in vitro fermentation assays, and measured changes in hydrogen (H 2 ), volatile fatty acids (VFAs), and organic matter degradation (OMdeg). The anti-methanogenic activity was benchmarked against Asparagopsis armata.

Sample collection
Samples of seven species of seaweed were collected from natural populations (Ministry for Primary Industries Special Permit number 560) or stock cultures. Bulk material (approximately 50 g fresh weight) of A. armata, B. hamifera, E. formisissima, P. cirrhosum, Vidalia colensoi (all gametophytes) and Ecklonia radiata (sporophytes) were collected from intertidal rocky shore or reef habitats from six sites located in the North Island, New Zealand (Table 1). Samples of A. armata and E. radiata were identified using morphological characteristics (Nelson, 2020). Bonnemaisonia hamifera, P. cirrhosum and E. formisissima were identified using DNA barcoding. DNA was extracted from 20 mg dried tissue of each sample using the CTAB method of Zuccarello & Lokhorst (2005). The rbcL locus was amplified and sequenced using various combinations of primers F57, F145, F762, R753, RrSs, R898, and R1442 (Nelson et al., 2010;Freshwater & Rueness, 1994). Sequences were trimmed and assembled using Geneious Prime 2020.2.3, and the consensus sequences were compared with sequences in GenBank using BLAST https:// blast.ncbi.nlm.nih.gov/). DNA sequences were submitted to Genbank under the following accession numbers: B. hamifera -MZ604718, P. cirrhosum -MZ604717, and E. formisissima -MZ604719. Ulva sp. B (Genbank accession number MW250819.1) was sampled from a mixed gametophyte/sporophyte stock culture cultivated at the Coastal Marine Field Station, University of Waikato, Tauranga, in nutrient enriched (Cell-Hi F2P, Varicon Aqua Solutions UK, 0.1 g l -1 , 12.3 mg nitrate-N l -1 and 1.1 mg P -1 ) filtered seawater, under a 12/12 h light/dark cycle at 18°C (Lawton et al., 2021). Wild harvested material was chilled on ice for transportation back to the laboratory prior to being frozen and stored at -20°C. All samples were then freeze-dried (48 h, 50 mBar, Buchi L-200 Freeze Dryer, OneLab, New Zealand) and milled to a fine powder using a domestic blender and stored at -80 °C until use.

Compositional analysis
Elemental analysis (% carbon, hydrogen, nitrogen, sulphur, bromine, chlorine, iodine; n= 2 as subsamples of homogenized biomass) for each species was determined through percentage elemental analyses performed by OEA labs (www.oealabs.com, Callington, UK), where carbon, hydrogen, nitrogen and sulphur were determined using gas chromatography-thermal conductivity detector (GC-TCD), and chlorine, bromine, and iodine were determined using ion chromatography (IC). The content of crude protein (CP) was estimated using the total content of nitrogen (wt%) in the biomass with nitrogen-protein conversion factors of 5.63 for Asparagopsis, 5.10 for remaining red species of seaweed, 4.49 for E. radiata, 5.14 for Ulva sp. B, and 6.25 for perennial ryegrass (RG299, basal feed substrate) (Angell et al., 2016). Analyses of crude fat (CF, AOCS 1 Official Procedure AM-5-04), acid detergent fibre (ADF, AFIA Method 1.9A(a)), neutral detergent fibre (NDF, NFTA method adapted for Ankom autoanalyser), soluble sugars (80:20 Ethanol:Water extraction and colorimetric determination, in house method), and starch content (Enzymic Hydrolysis of Starch, colorimetric determination of glucose, in house method) for seaweed and RG299 were performed by R J Hills Laboratories Limited (www.hills-laboratories.com), New Zealand. Results are reported "as received", i.e., not corrected for residual moisture (typically 5%), except for the content of CP for RG299. Polyphenols were quantified following Zhang et al. (2006) modified as per Magnusson et al. (2015) and scaled up for the use of cuvettes.

In vitro fermentation assay
The in vitro fermentations were conducted in three separate incubation runs using a published method (Muetzel et al., 2014). Briefly, 500 ± 15 mg of air-dried feed substrate perennial ryegrass (basal feed substrate) with seaweed biomass (B. hamifera, E. formisissima, P. cirrhosum, V. colensoi, E. radiata, or Ulva sp. B) included at 2%, 6%, or 10% OM on top of the basal feed substrate was weighed into 125 ml incubation bottles and covered with Parafilm until use. The inclusion level of 2% OM was selected as we expect a highly effective species to have a strong effect here, and a moderately effective species to exhibit some effect. We then evenly spread (increased by 4%-unit steps) the inclusion level up to 10% OM, as inclusion levels above this dose require amounts of seaweed that become impractical and prohibitive to implement for large-scale cattle herds. For the positive control, A. armata was added at a dose of 2% OM and a sample without added seaweed served as the negative control. A standard ryegrass used in every incubation was used as a run control to identify low activity incubations (Muetzel et al., 2014). Each of these treatments was incubated in two bottles (technical replicates). The prepared incubation bottles were randomized and prewarmed in a 39°C incubator prior to incubation. The rumen fluid donor animals were maintained according to the guidelines approved by the AgResearch Animal Ethics committee (AE 15320). Rumen fluid from a two nonlactating fistulated Friesian x Jersey dairy cows was collected into pre-warmed thermos flasks and filtered through one layer of cheesecloth and equal volumes were added at a level of 20% v/v to the 3.2 l of pre-warmed (39°C) in vitro buffer (6.0 mM Na 2 HPO 4 , 9.6 mM KH 2 PO 4 , 0.5 mM MgCl 2 , 64.5 mM NaHCO 3 and 17.8 mM NH 4 HCO 3 (Mould et al., 2005)). The medium was dispensed in 50 ml aliquots under a stream of CO 2 into the prepared incubation bottles, which were then capped with a butyl rubber stopper and placed in a rack in an incubator at 39°C. The bottles were connected to the gas system by perforation of the butyl rubber stopper with a syringe needle (23 g) connected to the gas system. The bottles were constantly shaken on a reciprocal shaker at 120 rpm and incubated for 48 h. For statistical analysis the incubation was repeated three times using rumen fluid from different donor animals. Technical replicates for each incubation were averaged and included in the final dataset. Results from the three incubations were analysed as a single dataset as there was no significant difference in TGP between the run controls (Supplementary figure S1).

In vitro fermentation with rumen fluid
Gas accumulation was measured automatically using a pressure sensor, described in Muetzel et al. (2014). Once a bottle had reached the threshold pressure of 9 kPa above ambient pressure, the gases were released into a gas chromatograph (GC-2010, Shimadzu, Kyoto, Japan) via a 6-port valve and a 20 µl sample loop. The GC was fitted with a HP-MolseivePlot column (30 m length × 0.53 mm ID), a thermal conductivity detector (TCD) and flame ionization detector (FID) (maintained at 105°C and 250°C, respectively) in series to simultaneously quantify CH 4 and H 2 production from each incubation bottle. The column temperature was maintained at 85°C using N 2 as a carrier with a flow rate of 13 ml/min. Each GC is attached to an array of 36 bottles and has a maximum analysis frequency of one sample per minute. Fermentation gases were quantified using a calibration of four standards containing 1, 5, 10, 20% CH 4 and 0.5, 2.5, 5, 10% H 2 in nitrogen respectively (BOC Gases New Zealand Ltd. (Muetzel et al., 2014)). Samples for VFA analysis (1.8 ml bottle) were collected from each incubation bottle after 48 h using a wide bore tip. Samples were centrifuged (21,000 × g, 10 min, 4°C) and 900 µl was combined with 100 µl of internal standard solution (19 mM ethylbutyrate in 20% v/v phosphoric acid). The samples were stored at -20°C for at least 16 h, thawed, and centrifuged again as described above. An 800 µl aliquot of the combined supernatant and internal standard solution was transferred into a 2 ml crimp gas chromatography vial and crimped immediately. VFAs were analysed by gas chromatography as described by Attwood et al., (1998) in a Shimadzu GC2010plus gas chromatograph equipped with a flame ionization detector using a Zebron ZB-FFAP 30.0 m × 0.53 mm I.D. × 1 µm film column (Tavendale et al., 2005) and an FID detector.
The degradability of organic matter (OMdeg, % degraded) was calculated using the equation OMdeg = 14.88 + 0.889 GP + 0.045 CP + 0.065 ASH where GP is the volume of gas (ml 200 mg -1 ) produced at 24 h (obtained from in vitro incubations), and CP and ASH are the total crude protein and ash content (mg g -1 dry weight (DW)), respectively, of substrate(s) used for in vitro incubations (Menke & Close, 1986).

Statistical analyses
Prior to data analyses, gas production data were checked for errors (i.e., system calibration errors and gas leaks) during the in vitro incubations, and bottles identified as errors (1x B. hamifera 6%, 1x D. compressa 2%, and 1x A. armata 2%) were excluded from analyses. The effect of seaweed inclusion level (fixed factor) on total gas, CH 4 , H 2 , total VFA, and individual VFA production and OMdeg was analysed using permutational analyses of variance (PERMANOVA) conducted in Primer v7 (Primer-E Ltd., UK) using Euclidean distances resemblance matrices, 9,999 unrestricted permutations of raw data, and Type III sum of squares (Anderson et al., 2008). Data for each species were analysed separately and compared to the ASP 2% OM positive control and the basal feed substrate control using planned contrasts. Monte Carlo P-values were used to assess significance (Anderson et al., 2008). Correlations between variables were examined using Pearson's correlation coefficients conducted in MS Excel (Version 1808).

Results and discussion
The inclusion of New Zealand native seaweeds as feed additives reduced the in vitro production of CH 4 in rumen fluid by up to 99% relative to the basal feed substrate control (Fig 1), with significant changes in other measures (H 2 , VFAs, and OMdeg) that are used to assess the efficiency of rumen fermentations. Bromoform, the main bioactive in Asparagopsis that inhibits methanogenesis, was not detected in the new seaweeds tested (Table 2), indicating that the bioactivity originates from alternative anti-methanogenic compounds.

Compositional analysis
The use of seaweed secondary metabolites as potential feed additives to reduce enteric CH 4 emissions has recently attracted much research interest, especially due to a growing preference for natural additives over antibiotics and/or chemical additives (Clark et al., 2010;Kobayashi, 2010). Many seaweed secondary metabolites are halogenated compounds and their presence are reflected in higher tissue contents of halides (e.g., bromine, chlorine, iodine).
In the current study, the organic matter content of seven seaweeds varied widely (44-80% DW; Table 3) while ash content was highly correlated with total halogen content (R 2 = 0.85) and chlorine content (R 2 = 0.88). Only a weak correlation was observed for bromine (R 2 = 0.24) and iodine (R 2 = 0.36), consistent with the differential tissue accumulation of bromine and/or iodine in the seaweeds studied here. A. armata had the highest content of bromine (7.1%) which was approximately double that of V. colensoi (3.9%) and four times that of B. hamifera (1.6%) ( Table 2); this is consistent with the characteristic brominated secondary metabolites reported for these species (Table S1). The content of bromoform, the main anti-methanogenic compound in A. armata, was 10.4 mg/g DW; which is similar to that measured for high-quality (6.6-7.8 mg g -1 DW) Asparagopsis spp. previously (Kinley et al., 2020;Roque et al., 2021). Despite the relatively high contents of bromine in V. colensoi and B. hamifera, bromoform was not detected in these or the other seaweeds tested in the current work, however these species contain other brominated compounds (Supplementary table S1). The iodine content of A. armata (1.0%), B. hamifera (0.8%), E. radiata (0.4%) and P. cirrhosum (0.1%) illustrate the need for measurement of this element for seaweed biomasses targeted at animal feed inclusions. While iodine is an essential element critical for animal function, and it is recognized that New Zealand cattle populations are commonly deficient in iodine (Anderson et al., 2007), excessive doses (tolerable upper limit for cattle: 50 mg iodine kg -1 dry matter (DM) day -1 (Weiss, 2005)) can result in negative effects on animal health and production (National Academies of Sciences, Engineering, and Medicine, 2016; Paulíková et al., 2002;Weiss, 2005). Polyphenols ranged from 2.1 to 55.3 mg g -1 DW and were highest for E. radiata (Table 3). Polyphenols can act as a H 2 sink during enteric fermentation highlighting the potential for producing a combination seaweed supplement to mitigate CH 4 emissions while simultaneously alleviating the increased partial pressure of H 2 in the rumen (Martinez-Fernandez et al., 2017).

In vitro fermentation
Enteric fermentation is a multistep digestive process where carbohydrates (e.g., hemicellulose, cellulose, and starch) from ingested plant feed are degraded by microbial enzymatic activity into VFAs, mainly acetate, propionate, and butyrate, that serve as the major energy source for ruminants (Kumari et al., 2020). During this process, CO 2 and H 2 are produced as gaseous byproducts, and are the substrate for the production of CH 4 by methanogenic archaea (Bhatta & Enishi, 2007;Buddle et al., 2011). Importantly, differences in basal feed substrate composition, the addition of antimethanogenic supplements, and feed digestibility can influence the total gas production, and proportions of these gases, with flow on effects to other fermentation parameters (e.g., VFAs, OMdeg).
Several studies have demonstrated that the effects of Asparagopsis on rumen fermentation are dependent on seaweed quality, based on the concentration of bromoform, and on the composition of ruminant basal feed (Kinley et al., 2020;Maia et al., 2016;Roque et al., 2012Roque et al., , 2021. In this study, inclusion of high-quality A. armata at 2% OM resulted in a 16.5% reduction in total gas production compared to the basal feed substrate control (ryegrass hay) with complete inhibition of CH 4 production and a significant increase in H 2 production (Fig 1;  Supplementary table S2). This indicates that either a component of the seaweed or the CH 4 inhibition decreased overall fermentation and therefore potentially animal production in vivo. When all other seaweeds were included at 2% OM the total gas production was Table 3. Composition (%DW) of organic matter (OM), ash, crude protein (CP), crude fat (CF), acid detergent fibre (ADF), neutral detergent fibre (NDF), soluble sugars, starch, and polyphenols (PP) (mg/g DW) for perennial ryegrass (negative control), and seaweed species Asparagopsis armata (ASP, positive control), Bonnemaisonia hamifera (BNM), Euptilota formisissima (EPT), Plocamium cirrhosum (PLC), Vidalia colensoi (VDA), Ecklonia radiata (ECK), and Ulva sp. B (ULVA).  However, no significant decrease in CH 4 was observed for these treatments. Seaweed inclusions of 6 and 10% OM led to decreases in total gas production of 7.9-17.4%. Inclusion of all seaweeds tested, except Ulva sp. B, reduced the production of CH 4 (↓8.9-98.9%) at 6 and 10% OM, with B. hamifera reducing the production of CH 4 at 2% OM (Fig 1; Supplementary table S2). Indeed, B. hamifera showed the best reduction in CH 4 giving an inclusion level-dependent reduction in the production of CH 4 compared to the basal feed of 17.1% (not statistically significant), 95.4% and 98.9% at inclusion levels of 2%, 6% and 10% OM, respectively. At inclusion levels of 6 and 10% OM an increase in the production of H 2 from 0.9 mg l to 24.8 ml g -1 and 21.1 ml g -1 , respectively, was observed for B. hamifera. At inclusion levels of 6% and 10% OM, E. formisissima and P. cirrhosum reduced the production of CH 4 by 36.7% and 50.5% (E. formisissima), and 16.4% and 39.5% (P. cirrhosum), respectively. Both seaweeds led to small increases in the production of H 2 at the highest level of inclusion, consistent with the metabolic hydrogen ([H]) being redirected into alternate biochemical pathways (Janssen, 2010;Leng, 2014;Ungerfeld, , 2018. A recent meta-analysis of in vitro batch and continuous rumen fluid fermentation highlighted that the mean amount of metabolic hydrogen [H] incorporated into H 2 in 100% methanogenesis inhibited fermentations averaged 10% and 6%, respectively . Specifically, the production of one CH 4 molecule from CO 2 requires 8[H] and measured changes in the production of H 2 in methanogenesis inhibited fermentations only accounted for a small proportion of the predicted production of H 2 . In this regard, in the current study 100% inhibition of methanogenesis with A. armata resulted in only 21% of [H] being redirected into the production of H 2 . This is consistent with previous in vitro and in vivo studies with Asparagopsis (Kinley et al., 2020;Machado et al., 2018;Roque et al., 2021;Stefenoni et al., 2021) and other antimethanogenic products (Hristov et al., 2015;Machado et al., 2016a;Martinez-Fernandez et al., 2016Mitsumori et al., 2012). Among the seaweeds tested here, inhibition of methanogenesis by 95.4% and 98.9% by inclusion of B. hamifera resulted in a similar proportion of [H] being redirected into the production of H 2 (i.e., ↑23% and 19%) as for the inclusion of A. armata. However, the lower rates of inhibition of methanogenesis by 36.7% and 50.5% with the inclusion of E. formisissima and P. cirrhosum, respectively, resulted in no significant increase in the production of H 2 . This result could indicate that other hydrogen utilizing pathways like propionate, valerate, or formate formation, reductive acetogenesis, or increased microbial biomass production (Leng, 2014; incorporated the excess metabolic [H]. Ingested OM is degraded during ruminal fermentation by an assortment of rumen microbes which generates VFAs, the primary source of energy contributing to the animal's nutrition (Russell et al., 1992). Thus, both OMdeg and VFA production are indicators of fermentation efficiency and reductions in these parameters are generally considered undesirable. In this regard, previous in vitro and in vivo studies have found that increasing inclusion levels of Asparagopsis can lead to a decrease in OMdeg (Machado et al., 2016b(Machado et al., , 2018 and VFA production (Li et al., 2016;Machado et al., 2016bMachado et al., , 2018. In the current study, with the exception of Ulva sp. B, OMdeg was decreased by 8.9% and 4.4-13.5% compared to the basal feed substrate (RG299) with the inclusion of A. armata (2% OM) and other seaweeds (≥6% OM), respectively (Fig 1; Supplementary table S2). Compared to basal feed substrate control fermentations, A. armata also significantly decreased the production of total VFAs by 22.4%, and B. hamifera decreased the production of total VFAs by 20.1 and 25.4% at inclusion levels of 6 and 10% OM, respectively (Table 4). These data are consistent with the decrease in gas production observed for the treatments and is also consistent with earlier work involving different seaweed species (Machado et al., 2014). The molar proportions of butyrate, propionate, and valerate increased for A. armata and B. hamifera at inclusion levels of 6 and 10% OM, while the proportions of acetate, isobutyrate, and isovalerate decreased (Table 4). Plocamium cirrhosum and E. formisissima induced similar changes at inclusion levels of 6 and 10% OM, however the effects were less pronounced. Conversely, V. colensoi, E. radiata, and Ulva sp. B had little or no effect on the production of individual VFAs. The shift from a high acetate: propionate ratio to one which favours the production of propionate is commonly observed with the use of CH 4 inhibitors (Janssen, 2010;Mitsumori et al., 2012), and is due to competition for hydrogen between methanogenesis and propiogenesis pathways (Hungate, 2015;Janssen, 2010;Mitsumori et al., 2012;Moss et al., 2000). This shift in VFA production has important implications on animal production systems with increased propionate production contributing to gluconeogenesis (Aschenbach et al., 2010), while reduced acetate production may lead to a reduction in milk fat percentage (Urrutia et al., 2019). However, the net changes in total VFAs and subsequent theoretical changes in [H] do little to explain the discrepancy between the measured H 2 and the theoretical H 2 production from fermentations with significantly inhibited methanogenesis.
In theory, based on the gross energy of VFAs, diverting the flow of [H] from the production of CH 4 into the production of VFAs and other nutritionally beneficial [H] sinks may increase feed conversion efficiency and Table 4. Effect of seaweed species Asparagopsis armata (ASP, positive control), Bonnemaisonia hamifera (BNM), Euptilota formisissima (EPT), Plocamium cirrhosum (PLC), Vidalia colensoi (VDA), Ecklonia radiata (ECK), and Ulva sp. B (ULVA) (mean ± SE, n = 3) at inclusion levels of 0% (perennial ryegrass, negative control), 2%, 6%, and 10% (OM-based) on total (mmolg -1 ) and individual volatile fatty acid (VFA) (% total) production at the end of in vitro incubations (48 h data). the sustainability of ruminant production systems (Johnson & Johnson, 1995); however, the true effect of CH 4 inhibition is yet to be resolved (Ungerfeld, 2018). In vitro and in vivo studies investigating the effects of the inhibition of methanogenesis on ruminant fermentation frequently find a reduction in the total VFAs (Li et al., 2016;Machado et al., 2016bMachado et al., , 2018; consistent with high partial pressure of H 2 leading to inhibition of microbial dehydrogenases and reduction in fermentation efficiency, feed digestibility, and DM intake (Janssen, 2010;Leng, 2014). Potentially, such a reduction in fermentation efficiency in vitro could translate to reductions in animal productivity in vivo, however few studies have concurrently measured both total VFAs and animal productivity when methanogenesis has been inhibited (Bayat et al., 2018;Kinley et al., 2020;Melgar et al., 2016), due to challenges in quantifying the constant flux of VFAs in dynamic biological systems. Studies that have, both in vitro Machado et al., 2016a, 2016b, Chagas et al., 2019 and in vivo (Kinley et al., 2020;Li et al., 2016), show differences in VFA production that depend on the level of inhibition. While the high partial pressure of hydrogen in the in vitro experiments in the current study presumably led to the reduction in total VFAs, these results cannot be used to make predictions on animal production as the partial pressure of hydrogen in vivo is much lower due to diffusion and eructation. However, when methanogenesis is inhibited, animal productivity is mixed, with both increased (Haisan et al., 2017;Kinley et al., 2020) and decreased (Roque et al., 2012) productivity demonstrated. For example, inclusion of A. taxiformis at 0.1% and 0.2% in the feed of Brahman-Angus cross steers led to reductions in the production of CH 4 of 40% and 98%, with concurrent weight gain improvements of 53% and 42%, respectively (Kinley et al., 2020). Inclusion of A. armata at 0.5% OM in the feed of lactating Holstein cows reduced CH 4 production by 26.4% with no change in milk yield or quality measures, but with reduced DM intake, i.e., an increase in feed conversion efficiency (Roque et al., 2012). Higher inclusion levels of A. armata (1% OM) resulted in higher CH 4 inhibition (CH 4 production ↓62.7%) but also impacted on animal productivity (milk production ↓11.6%). These examples highlight that in vitro fermentations can be used to identify potential CH 4 inhibitors but they are poor predictors of animal productivity responses. In the current study, we identified new seaweed targets worthy of further investigation to identify the precise nature of their active constituents and their activity in vivo.

Conclusion
The current study identified several new species of red seaweeds that reduced enteric CH 4 production in a dose-dependent manner. Inclusion of B. hamifera into the basal feed substrate (ryegrass) resulted in near elimination of enteric CH 4 production at ≥6% OM inclusion, while E. formisissima and P. cirrhosum also demonstrated significant anti-methanogenic activity at ≥6% OM inclusion. A. armata, B. hamifera, E. formisissima and P. cirrhosum had similar effects on measured fermentation parameters to those recorded in previous studies investigating the effects of Asparagopsis spp. and other halogenated methane analogues on ruminant fermentation. Importantly, the chemistries of the new seaweed species identified here are unique compared with that of Asparagopsis as they do not contain bromoform, the primary bioactive in Asparagopsis. These species therefore might provide alternative anti-methanogenic seaweed targets that are less susceptible to the loss of volatile bioactives during processing.