Rapid Screening of Methane-Reducing Compounds for Deployment in Livestock Drinking Water Using In Vitro and FTIR-ATR Analyses

Abstract

Several additives have been shown to reduce enteric methane emissions from ruminants when supplied in feed. However, utilising this method to deliver such methane-reducing compounds (MRCs) in extensive grazing systems is challenging. Use of livestock drinking water presents a novel method to deliver MRCs to animals in those systems. This work evaluated 13 MRCs for suitability to be deployed in this manner. Compounds were analysed for solubility and stability in aqueous solution using Fourier transform infrared-attenuated total reflectance (FTIR-ATR) spectroscopy. Furthermore, aqueous solutions of MRCs were subjected to variations in temperature and starting pH of water used to assess solubility and stability of the MRCs in simulated water trough conditions, also using FTIR-ATR spectroscopy. In vitro batch culture fermentations were carried out using a medium-quality tropical grass feed substrate, to simulate pastures consumed by cattle in extensive grazing systems. Measurements were made of total gas and methane production, in vitro dry matter digestibility (IVDMD), and volatile fatty acid (VFA) concentration. Of the MRCs tested, 12 were found to be soluble and stable in water using the FTIR method employed, whilst the other could not be measured. Of the 12 soluble and stable MRCs, one containing synthetic tribromomethane (Rumin8 Investigational Veterinary Product) reduced methane production by 99% (p = 0.001) when delivered aqueously in vitro, without a reduction in IVDMD (p = 0.751), with a shift towards decreased acetate and increased propionate production and decreased total VFA production (p < 0.001). Other compounds investigated also appeared suitable, and the methods developed in this study could be used to guide future research in the area.

1. Introduction

To meet the targets outlined by the Paris Agreement, total methane emissions associated with animal production require a reduction of between 24 and 47% from 2010 measurements by 2050 [1]. The delivery of methane-reducing compounds (MRCs) to ruminants has been proposed as one means of reducing enteric methane emissions from the red meat industry [1,2], whilst also potentially facilitating production increases to meet future demand for animal-derived proteins [3].

Broadly, MRCs are classified into two modes of action: as either modifiers of the rumen microbiota or inhibitors of methane production. The former alters the population of methanogenic microbial species in the rumen, whilst the latter disrupts the biological pathways associated with methane production [4]. Several compounds have shown varying degrees of effectiveness in reducing enteric methane emissions when added to the diet of ruminants. Although there is some overlap, generally halogenated compound-containing organisms, such as macroalgae and 3-nitroxypropanol, act as rumen inhibitors, whilst plant secondary compounds including tannins, saponins, and essential oils, ionophores, organic acids, nitrates and lipids are classed as rumen modifiers [4,5]. Most of these compounds have primarily been developed and tested for deployment in dairy or feedlot applications [5]. In these intensive systems, achieving consistent and controlled delivery of methane-suppressing additives, primarily through feed rations, is relatively straightforward. However, the majority of beef cattle production in the world is pasture-based [6]. For example, 82% of cattle slaughtered in Brazil, the world’s largest beef exporter, is pasture-based [7]. Delivery of MRCs in rations does not readily translate to these systems, due to challenges linked with remoteness, environment, infrastructure, and animal behaviour. Common to both intensive and pasture-based production systems, however, is the availability of drinking water, often supplied via troughs. Delivery of MRCs via the drinking water provides an innovative approach to reducing enteric methane emissions from ruminants on a scale not possible with existing technologies.

Whilst the use of drinking water to deliver MRCs in extensive systems has been suggested previously [7,8], to date, very little research has been conducted on the viability of this method of delivery. Of principal concern initially is the capacity for the MRC to maintain both solubility and stability in aqueous solution. Whilst some commercially available MRCs consist of a single compound, others are complex mixtures containing proprietary formulations. Utilising instrumentation such as gas/liquid chromatography-mass spectrometry may not be possible for these compounds or requires time-consuming sample preparation. Where the components of a proprietary MRC product are unknown, damage may occur to high-end instrumentation when a non-compatible compound is introduced to a chromatography column, for example. Use of those techniques or others such as colorimetric assays would also necessitate the development of compound-specific methods. However, Fourier transform infrared-attenuated total reflectance (FTIR-ATR) spectroscopy is a method that has been established to determine both compound characterisation and concentration quickly and cheaply [9,10,11]. Therefore, the first aim of this study was to investigate the solubility and stability of currently available MRCs when mixed with water utilising FTIR-ATR spectroscopy. Compounds were selected for analysis from review articles that described existing available products [7,8,12], and their likely capacity for dissolution in water. These same review articles also highlight the potential negative impacts on animals and environment associated with some MRCs; however, this study focused only on the screening of these compounds for deployment in water, preferring to leave investigation of safety issues for future research.

In determining the suitability of an MRC for water delivery it is also important to determine if its methane-mitigating properties are retained, and whether the productivity of the animal may be compromised. Therefore, in vitro batch culture fermentations using rumen fluid and a medium-quality feed substrate were also carried out to evaluate the effects of aqueously delivered MRCs on fermentation parameters including dry matter (DM) digestibility (DMD), total gas and methane production, and volatile fatty acid (VFA) concentration in the culture media. It was hypothesized that (1) MRCs would remain both soluble and stable in aqueous solution under the treatment conditions tested; and (2) that rumen fermentation characteristics of MRCs delivered in water after in vitro batch culture fermentations would be comparable to fermentations where MRCs were introduced by other means. The overall objective was to utilise both the FTIR and in vitro analyses as rapid screening tools to determine the most suitable MRCs for water delivery to livestock.

2. Results

2.1. Identification of Key Peaks in MRCs Using FTIR-ATR

2.1.1. Agolin

Agolin is a commercially available methane-reducing feed additive mixture that contains several essential oils. Its primary constituents include linalool, eugenol, geranyl acetate and geraniol [13] (Figure 1). In addition to these essential oils, the sample used in this trial also contained benzoic acid and polysorbate 20, with the latter acting as the emulsifying agent.

The combination of so many compounds produces a complex FTIR spectrum, as demonstrated by the comparison of neat and water-diluted Agolin samples (Figure 2). There are two main differences between the neat and water-diluted samples, that being the reduction in intensities of the peaks at 2860 and 1041 cm⁻¹, and the reduction in the absorption band between 3500 and 3200 cm⁻¹ in the water-diluted samples. It is likely that this can be attributed to the impossibility of drying the neat Agolin sample prior to the measurement being taken, though an alteration or loss of constituent compounds cannot be definitively ruled out.

Despite this, several key peaks were identified as being suitable markers to enable continued analysis at low concentrations, those being observed at approximately 2922, 2860, 1732 and 1102 cm⁻¹. In polysorbate 20, peaks at 2920 and 2860 cm⁻¹ are attributed to symmetric and asymmetric methylene stretching, and those at 1095 cm⁻¹ belong to the stretching of the C₂-O-C₂ section [20]. Linalool exhibits symmetric and asymmetric methyl stretching around 2970 and 2920 cm⁻¹, as well as C-O stretching of the secondary alcohol at around 1114 cm⁻¹ [14,21]. Similarly, geraniol features distinctive peaks from a C–H asymmetric stretch from methylene at 2922 cm⁻¹ and a C-O stretch at 1099 cm⁻¹ [16]. Eugenol features absorbance bands at 2976 and 1118 cm⁻¹ from C-H stretching on the benzene ring and methyl groups [22]. Geranyl acetate may also contribute with a distinctive peak from C=O stretching at around 1740 cm⁻¹ as well as peaks at 2962 and 1070 cm⁻¹ from C-H and C-O stretching, respectively [17,23]. Finally, one of the most distinctive peaks of benzoic acid can be found from C=O stretching at around 1750 cm⁻¹ [19].

2.1.2. Beeocitrix+

The main components of Beeocitrix+ are ricinoleic acid (Figure 3) and green propolis. Propolis is a complex mixture containing hundreds of compounds, including phenolics, aromatic and amino acids, terpenes, hydrocarbons, waxes, and essential oils [24,25].

The FTIR spectrum of dissolved Beeocitrix+ contains peaks from O-H stretching and the scissoring of CH₂ and CH₃ groups between 3000 and 3500 cm⁻¹, and at 2926 and 2854 cm⁻¹, respectively. There is a small sharp peak at 1733 cm⁻¹ associated with C=O ester stretching and strong signals around 1562, 1405 and 1071 cm⁻¹ likely belonging to C=C stretching, C-H bending and O-H bending, respectively [24,26] (Figure 4).

2.1.3. Choline Chloride

In its neat form (Figure 5), choline chloride features distinctive FTIR peaks at 3220, 3026, 3006, 1639, 1481 and 1084 cm⁻¹ which are assigned to O-H stretching, C-H stretching, N-H stretching, CH₃ and CH₂ rocking [27,28,29].

After dissolution in water and air drying, there is a clear retention of water in the crystals evidenced by the broad peaks at 3350 and 1638 cm⁻¹ related to O-H stretching and bending. These new signals obscure the original O-H stretch/bend signals of choline chloride at 3220 and 1639 cm⁻¹ (Figure 6). The original peak at 3006 cm⁻¹ attributed to N-H stretching in choline chloride is no longer conspicuous after the addition of water, and given their proximity in the neat form, may have overlapped with the signal from C-H stretching to form the distinctive peak at 3028 cm⁻¹. The signals for CH₃ and CH₂ rocking remain clearly visible at 1478 and 1083 cm⁻¹, however. Thus, the key peaks used for further analysis of choline chloride were 3028, 1478, and 1083 cm⁻¹.

2.1.4. Monensin Sodium Salt

Monensin, an ionophore, exhibited limited solubility in water, proving challenging to dissolve without the addition of ethanol, lengthy dissolution periods, or heating of the solution. This is unsurprising given its largely hydrophobic structure (Figure 7).

There is little detail on the specific functional groups of FTIR peaks of monensin spectra in the literature, but the spectra observed here closely match those found by Omar et al. [31] and Surolia et al. [32] (Figure 8). There is a broad band of three peaks from around 3000 to 2800 cm⁻¹, and a prominent peak at 1561 cm⁻¹, likely related to O-H stretching, C-H stretching and methyl group bending, respectively. Even at a concentration of 50 mg/L, the key peaks were difficult to identify. As its recommended dose rate is 3 mg/L, monensin was not subjected to any additional solubility/stability testing. The most suitable method for monensin analysis in aqueous solutions would likely be high-performance liquid chromatography [33].

2.1.5. Nitrates

Common to nitrate compounds is the nitrate ion (Figure 9).

Given that some of the nitrates used were in a hydrate form, as well as their application in aqueous solutions, the peaks associated with water between 3600 and 3000 cm⁻¹ and 1633 cm⁻¹ were ignored. The peaks of interest relate to N-O stretching in the region between around 1369 to 1338 cm⁻¹ [34,35]. More specifically, the nitrate peak of calcium nitrate tetrahydrate was found at 1313 cm⁻¹, whilst magnesium nitrate hexahydrate, potassium nitrate, and sodium nitrate displayed their nitrate peaks at 1352, 1360, and 1337 cm⁻¹, respectively (Figure 10).

2.1.6. Polygain

Polygain is a commercially available feed additive extracted from sugarcane (Saccharum officinarum). The components of interest for methane reduction investigation are the polyphenol compounds. Although the specific compounds in Polygain have not yet been profiled, literature on S. officinarum has shown it to contain myriad polyphenols, flavonoids, and organic acids [36,37]. The comparison of neat and water-diluted Polygain spectra (Figure 11) features a characteristic O-H stretch visible in the region between 3600 and 3000 cm⁻¹. There are also characteristic peaks at 2941, 1597, 1394, and 1047 cm⁻¹ which could potentially correspond to CH₂/CH₃ stretching, combinations of C=O stretching/C=C stretching, and O-H bending and C-O phenol vibrations, respectively, as has been established in other plant-derived polyphenols [38,39].

2.1.7. Rumin8 Investigational Veterinary Product (IVP)

Rumin8 IVP is a methane-reducing additive under development, with its active ingredient being synthetic tribromomethane, also known as bromoform (Figure 12).

Several key peaks are observed in the Rumin8 IVP, registering at 2923, 2859, 1741, 1457, 1150 and 1101 cm⁻¹ (Figure 13). Whilst the precise formulation of the Rumin8 IVP is proprietary, infrared spectra of tribromomethane in the literature and from chemical manufacturers identify prominent peaks around 3030 (C-H stretching), 2973, 2892, and 1145 cm⁻¹ [41,42,43]. Similar peaks at 2923, 2859, and 1150 cm⁻¹ observed in the Rumin8 IVP solution may be indicative of tribromomethane peaks, although slightly shifted when compared to spectra of pure tribromomethane.

2.1.8. Saponin

The saponin used in this study was an extract from the bark of Quillaja saponaria, which is actually made up of nearly 60 individual saponins, classed as triterpene bidesmosides [44] (Figure 14).

A characteristic O-H stretch is visible in the region between 3600 and 3000 cm⁻¹. There are also distinctive peaks at 2936, 1722, 1602 and 1039 cm⁻¹ (Figure 15). The FTIR spectrum of saponin has been well studied, and the aforementioned peaks match closely with those reported in the literature, associated with C-H stretching, C=O stretching, C=C stretching, and C-O-C stretching, respectively [45].

2.1.9. SilvaFeed

The primary component of SilvaFeed is an extract from the Quebracho tree spp. (Schinopsis lorentzii and Schinopsis balansae) which are rich in tannins (Figure 16).

Neat SilvaFeed was not able to be dried on the FTIR instrument, and as such the peaks at 3302 and 1619 cm⁻¹ are heavily influenced by absorbance associated with water. The diluted samples which were able to be dried on the instrument provided greater signal definition, and reveal distinct peaks between 3600 and 3000 cm⁻¹, and at 1605, 1517 and 1446 cm⁻¹ (Figure 17). The peaks between 3600 and 3000 cm⁻¹ are associated with O-H stretching, with the remaining peaks characteristic of C=C-C stretching of the aromatic rings [46].

2.1.10. Tartaric Acid

Tartaric acid is an organic acid (Figure 18).

The neat sample exhibits peaks associated with O-H stretching at 3399, 3329 and 3099 cm⁻¹, but the dried aqueous solution still retains some water molecules which overshadow those individual peaks, exhibiting a broad absorption between 3600 and 3000 cm⁻¹, and an additional peak at 1632 cm⁻¹ (Figure 19). The smaller peaks between 1500 and 1200 cm⁻¹ are also overshadowed in this manner. Therefore, the other three distinctive peaks at 1718, 1128 and 1081 cm⁻¹ were used as markers, associated with the COOH group, C=O, and C-O stretching, respectively [48].

The key peaks that were used in further solubility and stability analysis are summarised in

Table 1. Key peak wavenumbers of methane-reducing compounds derived from Fourier transform infrared-attenuated total reflectance spectra measurements used in solubility and stability analysis.

Compound 1	Key Peak Wavenumbers (cm−1)
Agolin	2922, 2860, 1732, 1102
Beeocitrix+	3500–3000, 2926, 2854, 1733, 1562, 1405, 1071
Choline chloride	3028, 1478, 1083
Monensin sodium salt	2966, 2929, 2879, 1561
Calcium nitrate tetrahydrate	1313
Magnesium nitrate hexahydrate	1352
Potassium nitrate	1360
Sodium nitrate	1337
Polygain	3600–3000, 2941, 1597, 1394
Rumin8 IVP	2923, 2859, 1741, 1457, 1150, 1101
Saponin	3600–3000, 2936, 1722, 1602, 1039
SilvaFeed	3600–3000, 1605, 1517, 1446
Tartaric acid	3600–3000, 1718, 1128, 1081

¹ Beeocitrix+, sodium monensin, and tartaric acid were not able to be measured at their recommended dose rates.

.

2.2. Establishment of MRC Solubility

Nitrates, tartaric acid, choline chloride, saponin, and sodium monensin, which are single-molecule products, have previously had their solubilities in water investigated and published by chemical manufacturers or chemical databases (

Table 2. Known solubility in water of single-molecule methane-reducing compounds used in this study.

Compound 1	CAS #	Solubility in Water	Reference
Calcium nitrate tetrahydrate	13477-34-4	1293 g/L at 20 °C	[49]
Sodium nitrate	7631-99-4	874 g/L at 20 °C	[50]
Choline chloride	67-48-1	140 g/L at 25 °C	[51]
Magnesium nitrate hexahydrate	13446-18-9	420 g/L at 20 °C	[52]
Potassium nitrate	7757-79-1	357 g/L at 25 °C	[53]
Saponin, from quillaja bark	8047-15-2	2000 g/L at 19.5 °C	[54]
Monensin sodium salt 1	22373-78-0	Sparingly soluble	[55]
L-(+)-Tartaric acid	87-69-4	1390 g/L at 20 °C	[56]

¹ Monensin sodium salt is listed as only sparingly soluble in water, but is readily dissolved in a small volume (50 µL) of ethanol with the resulting mix then added to water.

).

Where solubilities were not known, correlations between the absorbance intensities measured using FTIR and known concentrations of prepared solutions were used. Analysis revealed significant positive correlations (p < 0.001) between FTIR absorbance at key peak wavenumbers and the known concentration of the solution for all MRCs except for two key peaks of Beeocitrix+, which trended towards significance (p = 0.089 and 0.085) (

Table 3. Correlation analysis of key peak absorbance values from Fourier transform infrared-attenuated total reflectance spectra and known concentrations of methane-reducing compounds without known solubility values. Tests for significance were two-tailed with a threshold of p < 0.05.

Compound	Concentration	Key Peak Wavenumber (cm−1)	Equation	r-Value	p-Value
Agolin	20–200 mg/L	2922	y = 4 × 10−5x + 0.0004	0.976	<0.001
		2860	y = 4 × 10−5x + 5 × 10−5	0.990	<0.001
		1732	y = 2 × 10−5x + 6 × 10−5	0.956	<0.001
		1102	y = 5 × 10−5x − 1 × 10−5	0.989	<0.001
Beeocitrix+	0.5–10 g/L	3500–3000	y = 0.0084x + 0.0172	0.991	0.085
		2926	y = 0.0182x + 0.0281	0.998	0.043
		2854	y = 0.0142x + 0.0191	0.998	0.042
		1733	y = 0.0021x + 0.0036	0.990	0.089
		1562	y = 0.0144x + 0.0144	0.999	0.032
		1405	y = 0.0138x + 0.0115	0.999	0.029
		1071	y = 0.0112x + 0.008	0.999	0.033
Polygain	1–20 mL/L	3600–3000	y = 0.0641ln(x) + 0.0362	0.955	<0.001
		2941	y = 0.0338ln(x) + 0.0106	0.945	<0.001
		1597	y = 0.0913ln(x) + 0.0209	0.966	<0.001
		1394	y = 0.0666ln(x) + 0.0116	0.945	<0.001
Rumin8 IVP	5–40 mL/L	2923	y = 0.0018x − 0.0082	0.954	<0.001
		2859	y = 0.0012x − 0.006	0.954	<0.001
		1741	y = 0.0023x − 0.012	0.931	<0.001
		1457	y = 0.0007x − 0.0032	0.935	<0.001
		1150	y = 0.0017x − 0.0097	0.929	<0.001
		1101	y = 0.0014x − 0.008	0.935	<0.001
SilvaFeed	1.7–20 mL/L	3600–3000	y = 0.013x + 0.0225	0.982	<0.001
		1605	y = 0.0112x + 0.0056	0.986	<0.001
		1517	y = 0.0074x + 0.0031	0.985	<0.001
		1446	y = 0.0087x + 0.0032	0.987	<0.001

). For all compounds, the lines of best fit were linear, except for Polygain, which was best described with a logarithmic curve. This latter observation likely means that aqueous solutions of Polygain are potentially reaching saturation point at the higher concentrations tested. These strong relationships between absorbance and concentration indicated reliable solubility of the compounds across the analysed concentration ranges. Moreover, it suggested that the target in vitro fermentation dose rates for the aqueous solutions could be achieved.

2.3. Establishment of MRC Solubility/Stability across Simulated Water Trough Conditions

Apart from Beeocitrix+ and tartaric acid, which could not be measured at the target dose rate concentration, absorbances at all key peak wavenumbers were observed to some extent, indicating stability of the compounds across the treatment conditions (

Table 4. Solubility/stability analysis using retention of key peaks measured using Fourier transform infrared spectra from dried aqueous solutions containing methane-reducing compounds subjected to treatments (solutions held for 24 and 48 h periods at 4, 25, and 45 °C, and 24 and 48 h periods with water used to prepare solutions starting at 5.5 and 8.5 pH and held at 25 °C).

Compound 1	Key Peak Wavenumber (cm−1)	Key Peak Wavenumber Absorbance Values 4											p-Value
		Baseline	24 h 4 °C	24 h 25 °C	24 h 45 °C	48 h 4 °C	48 h 25 °C	48 h 45 °C	24 h acidic	24 h basic	48 h acidic	48 h basic
Agolin 2	2922	2.0 ± 0.10	1.5 ± 0.10	1.6 ± 0.30	1.8 ± 0.58	2.0 ± 0.09	3.8 ± 2.85	2.4 ± 1.39	1.7 ± 0.13	1.6 ± 0.33	1.4 ± 0.21	1.2 ± 0.48	0.205
	2860	1.4 ± 0.29	1.3 ± 0.04	1.4 ± 0.25	1.6 ± 0.60	1.7 ± 0.13	3.9 ± 3.44	1.7 ± 0.73	1.3 ± 0.08	1.1 ± 0.11	1.1 ± 0.23	1.0 ± 0.28	0.153
	1732	0.7 ± 0.42	0.9 ± 0.09	0.9 ± 0.40	1.4 ± 1.20	0.7 ± 0.16	0.7 ± 0.58	0.7 ± 0.16	0.6 ± 0.58	0.6 ± 0.20	0.9 ± 0.23	0.7 ± 0.32	0.717
	1102	1.3 ± 0.06	1.9 ± 0.29	2.0 ± 0.39	1.3 ± 0.33	1.3 ± 0.19	4.3 ± 3.60	2.1 ± 0.35	1.5 ± 0.40	1.5 ± 0.05	2.0 ± 0.10	1.7 ± 0.45	0.139
Choline chloride 3	3028	5.5 ± 0.03 d	5.6 ± 0.09 cd	5.6 ± 0.04 bc	5.6 ± 0.06 bc	5.6 ± 0.01 bc	5.6 ± 0.07 bc	5.6 ± 0.01 bc	5.7 ± 0.06 ab	5.5 ± 0.04 cd	5.8 ± 0.02 a	5.8 ± 0.03 a	<0.001
	1478	22.3 ± 0.07 a	19.5 ± 0.03 bcd	20.0 ± 0.04 b	19.9 ± 0.05 bc	19.4 ± 0.13 bcd	19.2 ± 0.15 cde	18.9 ± 0.45 de	18.9 ± 0.23 de	18.7 ± 0.07 e	18.1 ± 0.01 f	18.0 ± 0.16 f	<0.001
	1083	20.3 ± 0.09 a	16.4 ± 0.01 bc	17.1 ± 0.24 b	16.7 ± 0.08 b	15.8 ± 0.13 cd	15.6 ± 0.18 cd	15.1 ± 0.49 de	15.5 ± 0.41 d	15.2 ± 0.06 de	14.5 ± 0.01 e	14.4 ± 0.15 e	<0.001
Calcium nitrate 2	1313	5.4 ± 5.68	5.1 ± 5.43	13.2 ± 7.99	4.4 ± 0.12	9.8 ± 1.02	23.5 ± 16.63	2.6 ± 0.06	2.7 ± 0.93	6.0 ± 2.78	3.3 ± 0.62	7.6 ± 2.73	0.085
Magnesium nitrate 2	1352	4.4 ± 5.17 b	2.9 ± 1.97 b	2.4 ± 0.76 b	6.3 ± 1.88 b	2.8 ± 0.33 b	6.5 ± 3.65 b	6.4 ± 0.11 b	2.1 ± 0.84 b	2.3 ± 0.59 b	6.5 ± 2.89 b	63.5 ± 23.58 a	<0.001
Potassium nitrate 2	1360	5.3 ± 4.49 ab	13.9 ± 20.18 ab	1.7 ± 0.39 b	1.7 ± 0.18 b	2.0 ± 0.14 ab	1.5 ± 0.57 b	4.3 ± 4.53 ab	1.7 ± 0.43 b	33.8 ± 7.25 a	1.8 ± 0.67 b	2.5 ± 0.71 ab	0.047
Sodium nitrate 2	1337	15.6 ± 23.18	3.8 ± 2.66	4.9 ± 6.35	3.8 ± 3.33	4.8 ± 2.00	17.2 ± 24.11	32.3 ± 44.65	13.3 ± 1.24	8.2 ± 2.55	5.9 ± 0.19	19.0 ± 24.77	0.807
Polygain 3	3600–3000	5.7 ± 0.41	5.2 ± 0.54	5.7 ± 1.79	5.1 ± 0.31	7.1 ± 0.25	7.7 ± 3.00	9.2 ± 4.77	6.6 ± 0.25	6.1 ± 0.48	6.8 ± 0.06	84 ± 2.56	0.269
	2941	2.1 ± 0.15	1.9 ± 0.24	2.1 ± 0.58	2.2 ± 0.10	2.7 ± 0.02	3.6 ± 1.89	3.7 ± 2.02	2.2 ± 0.05	2.0 ± 0.13	1.9 ± 0.15	3.1 ± 1.11	0.199
	1597	4.9 ± 0.40	4.7 ± 0.69	4.6 ± 0.77	5.0 ± 0.13	6.5 ± 0.10	7.9 ± 3.70	9.0 ± 5.04	5.4 ± 0.03	4.8 ± 0.29	4.5 ± 0.39	7.1 ± 2.68	0.153
	1394	3.2 ± 0.28	3.1 ± 0.44	3.0 ± 0.56	3.4 ± 0.05	4.2 ± 0.02	5.5 ± 2.87	6.2 ± 3.61	3.5 ± 0.05	3.1 ± 0.19	2.9 ± 0.21	4.8 ± 1.92	0.145
Rumin8 IVP 3	2923	0.6 ± 0.12 ab	2.0 ± 1.59 a	1.2 ± 0.41 ab	1.9 ± 1.00 ab	1.2 ± 0.12 ab	0.8 ± 0.55 ab	0.4 ± 0.11 ab	1.0 ± 0.98 ab	0.3 ± 0.14 ab	0.4 ± 0.13 ab	0.3 ± 0.18 b	0.012
	2859	4 ± 0.08 abc	1.3 ± 1.02 a	0.7 ± 0.20 abc	1.2 ± 0.65 ab	0.8 ± 0.08 abc	0.5 ± 0.31 abc	0.3 ± 0.05abc	0.4 ± 0.32 abc	0.2 ± 0.07 bc	0.3 ± 0.08 abc	0.2 ± 0.10 c	0.005
	1741	0.6 ± 0.02 ab	2.5 ± 2.13 a	1.4 ± 0.56 ab	2.1 ± 1.15 ab	1.5 ± 0.27 ab	1.0 ± 0.66 ab	0.4 ± 0.07 b	0.4 ± 0.15 b	0.3 ± 0.20 b	0.2 ± 0.08 b	0.3 ± 0.21 b	0.002
	1457	0.1 ± 0.07 ab	0.6 ± 0.49 a	0.3 ± 0.11 ab	0.5 ± 0.31 ab	0.3 ± 0.02 ab	0.2 ± 0.17 ab	0.2 ± 0.16 ab	0.1 ± 0.10 ab	0.1 ± 0.01 b	0.0 ± 0.03 b	0.1 ± 0.04 b	0.013
	1150	0.4 ± 0.01 ab	1.5 ± 1.33 a	0.8 ± 0.27 ab	1.3 ± 0.78 ab	0.8 ± 0.01 ab	0.5 ± 0.43 ab	0.2 ± 0.13 b	0.5 ± 0.41 ab	0.2 ± 0.10 b	0.2 ± 0.07 b	0.2 ± 0.12 b	0.005
	1101	0.3 ± 0.04 abc	1.1 ± 0.89 a	0.5 ± 0.14 abc	1.0 ± 0.59 ab	0.6 ± 0.04 abc	0.4 ± 0.23 abc	0.4 ± 0.13 abc	0.3 ± 0.22 abc	0.2 ± 0.02 bc	0.2 ± 0.09 c	0.1 ± 0.08 c	0.005
Saponin 3	3600–3000	1.5 ± 1.19	2.4 ± 0.52	2.8 ± 0.82	2.7 ± 1.14	1.5 ± 0.09	1.4 ± 0.49	2.1 ± 0.23	1.9 ± 0.02	1.8 ± 0.19	2.0 ± 0.10	3.2 ± 1.71	0.499
	2936	0.6 ± 0.50	0.8 ± 0.24	0.7 ± 0.41	0.7 ± 0.09	0.6 ± 0.04	0.6 ± 0.22	0.9 ± 0.11	0.7 ± 0.06	0.8 ± 0.15	0.8 ± 0.05	1.4 ± 0.78	0.638
	1722	0.6 ± 0.47	0.8 ± 0.27	0.8 ± 0.25	0.7 ± 0.07	0.6 ± 0.00	0.5 ± 0.04	0.8 ± 0.14	0.6 ± 0.07	0.7 ± 0.11	0.7 ± 0.03	1.2 ± 0.69	0.711
	1602	1.0 ± 0.87	1.5 ± 0.49	1.5 ± 0.28	1.3 ± 0.20	1.1 ± 0.03	1.1 ± 0.34	1.5 ± 0.27	1.4 ± 0.04	1.4 ± 0.20	1.4 ± 0.05	2.5 ± 1.54	0.566
	1039	1.3 ± 1.17	1.7 ± 0.71	1.7 ± 0.35	1.5 ± 0.17	1.3 ± 0.04	1.4 ± 0.46	1.9 ± 0.32	1.5 ± 0.07	1.6 ± 0.20	1.6 ± 0.08	2.9 ± 1.73	0.730
SilvaFeed 3	3600–3000	3.7 ± 0.59 b	3.4 ± 0.46 b	3.2 ± 0.20 b	3.1 ± 0.39 b	5.1 ± 0.51 a	4.4 ± 0.05 ab	5.1 ± 0.35 a	3.7 ± 0.56 b	3.6 ± 0.30 b	3.2 ± 0.19 b	3.4 ± 0.22 b	<0.001
	1605	2.4 ± 0.40 bc	2.3 ± 0.39 bc	2.0 ± 0.15 bc	2.1 ± 0.19 bc	3.5 ± 0.47 a	3.1 ± 0.03 ab	3.6 ± 0.26 a	2.4 ± 0.25 bc	2.4 ± 0.34 bc	2.0 ± 0.06 c	2.2 ± 0.19 bc	<0.001
	1517	1.5 ± 0.28 cd	1.6 ± 0.27 bcd	1.2 ± 0.07 d	1.3 ± 0.09 d	2.3 ± 0.40 ab	2.2 ± 0.15 abc	2.5 ± 0.19 a	1.6 ± 0.10 bcd	1.6 ± 0.28 bcd	1.3 ± 0.02 d	1.4 ± 0.13 cd	<0.001
	1446	1.8 ± 0.27 bc	2.0 ± 0.45 abc	1.6 ± 0.09 c	1.6 ± 0.13 c	2.7 ± 0.39 ab	2.4 ± 0.08 abc	2.8 ± 0.27 a	1.8 ± 0.16 abc	1.9 ± 0.33 abc	1.5 ± 0.01 c	1.6 ± 0.13 c	<0.001

¹ Beeocitrix+ and tartaric acid were not able to be measured at their recommended dose rates. ² Compounds have had absorbance values and standard deviation multiplied by 1000 for consistency of reporting. ³ Compounds have had absorbance values and standard deviation multiplied by 100 for consistency of reporting. ⁴ Entries in the same row with different superscript letters indicate significant differences based on a one-way ANOVA followed by post-hoc Tukey testing with a threshold of p < 0.05.

). However, there were likely some small changes to observed concentrations. It was anticipated that there would be some variability due to interactions between evaporative loss of solute and solvent, as well as the evenness of deposition of the compound on the ATR crystal upon drying. Solutions in treatments at 25 °C experienced a loss of approximately 0.1 and 0.5 mL of volume over 24 and 48 h periods, respectively, whilst those at 45 °C lost close to 0.2 and 0.6 mL, respectively.

There were no significant differences observed in absorbances for Agolin, Polygain, or saponin, indicating consistent concentration of the compounds across the treatment conditions (Table 4). In the case of nitrates, there was uneven crystallisation of the compounds on the ATR crystal upon drying, leading to inconsistent measurements, particularly for the K and Mg compounds (p = 0.047 and <0.001, respectively). Therefore, the use of a nitrate sensor would be a more suitable method to measure these concentrations in any future work. Absorbances for all peaks of choline chloride were significantly different (p < 0.001), trending downward over time and indicating a slight net loss of choline chloride from solution. Spectra from solutions of Rumin8 IVP were also highly variable, with low p-values for absorbances across all treatments (p = 0.002–0.013). Some of this variation could be attributable to tribromomethane’s high level of volatility. However, none of the absorbance values of the peaks likely associated with tribromomethane (2923, 2859 and 1150 cm⁻¹) were observed to be significantly lower than the baseline measurement. This would suggest that tribromomethane is retained in solution, and it is unclear if the variation in absorbance is driven by the tribromomethane or by other ingredients of the Rumin8 IVP. Finally, there were significant differences in the absorbances at the key peaks of SilvaFeed (p < 0.001), with the major differences observed as increases at the higher temperatures and longer time periods, indicating an increase of SilvaFeed in solution over time under those conditions.

Solutions of Beeocitrix+ and tartaric acid at much higher than recommended dose rates (100 g/L and 1 g/L, respectively) were also subjected to the same treatment conditions with single-point FTIR measurements. Tartaric acid exhibited a slight decrease in concentration over longer time periods and at higher pH levels whilst there were no significant differences observed in the majority of the Beeocitrix+ peaks (

Table 5. Solubility/stability analysis using retention of key peaks measured using Fourier transform infrared-attenuated total reflectance spectra from dried aqueous solutions containing Beeocitrix+ and tartaric acid subjected to treatments (solutions held for 24 and 48 h periods at 4, 25, and 45 °C, and 24 and 48 h periods with water used to prepare solutions starting at 5.5 and 8.5 pH and held at 25 °C). Significance based on two-tailed t-tests with a threshold of p < 0.05.

Compound	Key Peak Wavenumber (cm−1)	Baseline Absorbance	Treatment Absorbance Average	SEM	p-Value
Beeocitrix+ 1	3500–3000	3.3	2.8	0.003	0.660
	2926	5.3	5.2	0.005	0.030
	2854	3.8	3.7	0.004	0.040
	1733	0.5	0.8	0.007	0.375
	1562	3.3	3.2	0.003	0.229
	1405	2.9	2.7	0.002	0.051
	1071	2.2	2.0	0.002	0.176
Tartaric acid	3600–3000	0.2	0.2	0.001	0.107
	1718	0.6	0.5	0.006	0.001
	1128	0.5	0.5	0.012	0.141
	1081	0.6	0.5	0.008	<0.001

¹ Compound had absorbance values and standard deviation multiplied by 100 for consistency of reporting.

).

The retention of the absorbances at key peak locations for all the MRCs tested indicates that they would be suitable for deployment in animal drinking water.

2.4. In Vitro Batch Culture Fermentations

Despite the analysis in this study requiring multiple runs, comparison of control data provides insight to the potential correlation of observations across all fermentations. There were no significant differences between the control fermentations for in vitro dry matter digestibility (IVDMD) (p = 0.197), total gas (p = 0.106) or methane (p = 0.623) production, or the ratio of methane to total gas produced (p = 0.681). However, there were significant differences for the majority of the VFA metrics. This suggests that the fermentations can be reasonably compared with each other.

2.4.1. Fermentation One

There were significant differences across all metrics in this fermentation, primarily related to the 4 g/L nitrate doses including Ca and Mg (

Table 6. In vitro dry matter (DM) digestibility (IVDMD), total gas production, methane production, ratio of total gas to methane production, volatile fatty acid (VFA) concentration and proportion measured in vitro. Values presented as mean ± SEM.

Parameter 1	Control	Calcium Nitrate	Magnesium Nitrate	Potassium Nitrate	Sodium Nitrate	p-Value
IVDMD (% DM)	56.9 ± 1.31 a	36.9 ± 1.71 b	32.8 ± 3.98 b	59.3 ± 3.13 a	57.3 ± 2.75 a	0.002
Total gas production (mL/g DM)	114.8 ± 5.19 a	20.9 ± 1.44 c	20.5 ± 2.87 c	71.3 ± 0.56 b	82.8 ± 1.96 b	<0.001
Methane production (mL/g DM)	10.1 ± 1.00 a	N.D.	N.D.	4.7 ± 0.37 b	4.8 ± 0.43 b	<0.001
Total gas–methane	8.9 ± 1.27 a	N.D.	N.D.	6.6 ± 0.58 b	5.8 ± 0.38 b	<0.001
Total VFA (mM)	45.4 ± 0.96 a	18.8 ± 2.70 b	16.6 ± 1.54 b	38.9 ± 1.94 a	38.0 ± 5.67 a	0.004
Acetate (mM)	26.3 ± 0.15 a	17.6 ± 0.54 b	15.5 ± 1.44 b	28.8 ± 1.01 a	26.9 ± 2.47 a	0.003
Propionate (mM)	6.4 ± 0.04 a	2.1 ± 0.09 b	2.1 ± 0.25 b	5.3 ± 0.05 a	5.3 ± 0.43 a	<0.001
Butyrate (mM)	2.0 ± 0.06 a	0.0 ± 0.12 c	0.0 ± 0.01 c	0.5 ± 0.01 b	0.6 ± 0.15 b	<0.001
Valerate (mM)	3.9 ± 0.44 a	0.0 ± 0.61 b	0.0 ± 0.21 b	1.2 ± 0.50 ab	1.4 ± 0.67 ab	0.006
Total branched-chain VFA (mM)	6.6 ± 0.58 a	0.6 ± 1.34 ab	0.3 ± 0.05 b	3.1 ± 0.38 ab	3.8 ± 1.95 ab	0.045
VFA proportions
Acetate (%)	58.1 ± 1.57	95.3 ± 10.84	93.62 ± 0.03	74.1 ± 1.10	71.5 ± 4.19
Propionate (%)	14.2 ± 0.22	11.2 ± 1.12	12.6 ± 0.33	13.6 ± 0.55	15.0 ± 0.96
Butyrate (%)	4.5 ± 0.04	0.0 ± 1.11	0.0 ± 0.38	1.3 ± 0.05	1.6 ± 0.17
Valerate (%)	8.6 ± 0.79	0.0 ± 4.03	0.0 ± 0.90	3.0 ± 1.13	3.4 ± 1.24
Branched-chain (%)	14.6 ± 0.96	9.4 ± 3.74	1.7 ± 0.16	7.9 ± 0.57	9.4 ± 3.74
Acetate–propionate	4.1 ± 0.05 d	8.5 ± 0.16 a	7.4 ± 0.19 b	5.5 ± 0.14 c	5.1 ± 0.05 c	<0.001

¹ Entries in the same row with different superscript letters indicate significant differences where p < 0.05. N.D. stands for no data.

). This was reflected by substantially lower IVDMD (p = 0.002) and total gas production (p < 0.001) measurements. There was an apparent digestibility reduction of 35.1% for Ca and 42.4% for Mg, coupled with a reduction in total gas production of 81.7% and 82.1%, respectively. Gas production was so low as to prevent measurements of methane production in those samples. Conversely, there were no significant differences for IVDMD between the control and the 2 g/L nitrate doses using K and Na, ranging from 56.9–59.3%. However, there were decreases (p < 0.001) in total gas production, the ratio of total gas to methane, and methane production, the latter representing a decrease of 53.4% for K and 52.5% for Na nitrates compared to the control. Significant differences were also present in the VFA concentrations, with the high-dose nitrates much lower across the board. All nitrate doses had a much greater acetate to propionate ratio and exhibited a significant reduction in butyrate concentration relative to control. The methane reductions with minimal impact to IVDMD and VFA concentrations observed in vitro indicate that nitrate doses below 2 g/L in animal drinking water could be considered for application in future in vivo trials.

2.4.2. Fermentation Two

The only statistically significant observation (p = 0.017) in the second fermentation was a reduction in the total gas production of the SilvaFeed treatment, representing a reduction of 12.6% (

Table 7. In vitro dry matter (DM) digestibility (IVDMD), total gas production, methane production, ratio of total gas to methane production, volatile fatty acid (VFA) concentration and proportion measured in vitro. Values presented as mean ± SEM.

Parameter 1	Control	Agolin	Polygain	SilvaFeed	p-Value
IVDMD (% DM)	56.5 ± 3.46	59.3 ± 0.83	54.2 ± 1.44	57.8 ± 0.24	0.402
Total gas production (mL/g DM)	125.6 ± 3.38 a	124.7 ± 1.77 a	120.5 ± 0.17 ab	109.8 ± 1.61 b	0.017
Methane production (mL/g DM)	12.0 ± 3.20	12.0 ± 1.19	12.5 ± 0.55	9.5 ± 0.94	0.649
Total gas–methane	9.5 ± 2.29	9.6 ± 0.82	10.4 ± 0.44	8.6 ± 0.98	0.828
Total VFA (mM)	40.3 ± 1.37	35.9 ± 0.44	37.0 ± 3.48	35.7 ± 1.16	0.424
Acetate (mM)	27.4 ± 0.90	24.9 ± 0.92	25.3 ± 2.07	24.3 ± 0.54	0.427
Propionate (mM)	6.0 ± 0.13	5.4 ± 0.07	5.7 ± 0.39	5.5 ± 0.15	0.341
Butyrate (mM)	1.8 ± 0.08	1.8 ± 0.00	1.7 ± 0.15	1.7 ± 0.09	0.884
Valerate (mM)	2.1 ± 0.14	1.8 ± 0.18	2.0 ± 0.27	1.8 ± 0.03	0.569
Total branched-chain VFA (mM)	3.0 ± 0.13	2.0 ± 0.37	2.4 ± 0.60	2.3 ± 0.35	0.451
VFA proportions
Acetate (%)	68.0 ± 0.08	69.3 ± 1.72	68.4 ± 0.83	68.2 ± 0.71
Propionate (%)	14.9 ± 0.19	15.0 ± 0.01	15.3 ± 0.39	15.4 ± 0.07
Butyrate (%)	4.4 ± 0.05	5.0 ± 0.05	4.6 ± 0.04	4.9 ± 0.09
Valerate (%)	5.2 ± 0.17	4.9 ± 0.55	5.4 ± 0.22	5.1 ± 0.08
Branched-chain (%)	7.5 ± 0.06	5.7 ± 1.11	6.3 ± 1.04	6.5 ± 0.77
Acetate–propionate	4.6 ± 0.05	4.6 ± 0.11	4.5 ± 0.06	4.4 ± 0.03	0.417

¹ Entries in the same row with different superscript letters indicate significant differences where p < 0.05.

). Both Agolin and Polygain showed no significant differences in any metric when compared to the control. Though these results would appear to indicate that these compounds are not suitable for water deployment, there are some mitigating factors for these observations outlined in the discussion section.

2.4.3. Fermentation Three

Significant differences to all metrics were observed in fermentation three except for IVDMD, and were primarily associated with the action of choline chloride (

Table 8. In vitro dry matter (DM) digestibility (IVDMD), total gas production, methane production, ratio of total gas to methane production, volatile fatty acid (VFA) concentration and proportion measured in vitro. Values presented as mean ± SEM.

Parameter 1	Control	Beeocitrix+	Choline Chloride	Saponin	Tartaric Acid	p-Value
IVDMD (% DM)	59.6 ± 2.36	65.0 ± 1.70	58.5 ± 0.07	59.3 ± 1.52	65.4 ± 0.40	0.053
Total gas production (mL/g DM)	126.0 ± 1.64 bc	118.8 ± 0.53 cd	250.1 ± 1.76 a	129.9 ± 2.71 b	110.9 ± 0.19 d	<0.001
Methane production (mL/g DM)	13.8 ± 0.74 b	10.0 ± 3.48 b	45.2 ± 0.11 a	11.0 ± 1.14 b	10.7 ± 0.45 b	<0.001
Total gas–methane	11.0 ± 0.73 ab	8.4 ± 2.97 b	18.1 ± 0.08 a	8.5 ± 0.70 b	9.7 ± 0.39 b	0.021
Total VFA (mM)	36.8 ± 0.20 c	37.3 ± 0.29 bc	52.4 ± 0.12 a	39.6 ± 0.69 b	39.2 ± 0.57 bc	<0.001
Acetate (mM)	24.7 ± 0.13 d	25.2 ± 0.02 cd	41.3 ± 0.16 a	26.2 ± 0.26 b	25.8 ± 0.19 bc	<0.001
Propionate (mM)	6.6 ± 0.13 ab	6.5 ± 0.01 a	7.1 ± 0.11 b	6.7 ± 0.03 ab	6.8 ± 0.00 ab	0.023
Butyrate (mM)	2.0 ± 0.02 c	2.0 ± 0.02 bc	2.5 ± 0.02 a	2.1 ± 0.03 b	2.1 ± 0.07 b	<0.001
Valerate (mM)	1.8 ± 0.07 a	1.9 ± 0.11 a	1.3 ± 0.00 b	2.2 ± 0.12 a	2.2 ± 0.09 a	0.004
Total branched-chain VFA (mM)	1.6 ± 0.42 ab	1.4 ± 0.21 ab	0.2 ± 0.17 b	2.4 ± 0.31 a	2.3 ± 0.28 a	0.016
VFA proportions
Acetate (%)	67.2 ± 0.73	67.7 ± 0.58	78.9 ± 0.12	66.2 ± 0.50	65.9 ± 0.48
Propionate (%)	18.1 ± 0.46	17.5 ± 0.10	13.5 ± 0.18	16.8 ± 0.23	17.3 ± 0.25
Butyrate (%)	5.5 ± 0.09	5.5 ± 0.11	4.8 ± 0.03	.4 ± 0.16	5.5 ± 0.06
Valerate (%)	5.0 ± 0.16	5.0 ± 0.26	2.4 ± 0.01	5.6 ± 0.21	5.5 ± 0.16
Branched-chain (%)	4.3 ± 1.12	4.3 ± 0.53	0.4 ± 0.33	6.1 ± 0.68	5.9 ± 0.64
Acetate–propionate	3.7 ± 0.05 b	3.9 ± 0.01 b	5.8 ± 0.07 a	3.9 ± 0.02 b	3.8 ± 0.03 b	<0.001

¹ Entries in the same row with different superscript letters indicate significant differences where p < 0.05.

). Addition of choline chloride drastically increased total gas and total methane production, ratio of total gas to methane, and all VFA except for branched chain VFA. The substantial increase in VFA when supplementing choline chloride (52.4 ± 0.12 mM) compared to control (36.8 ± 0.20 mM) from nearly the same amount of feed substrate indicates a potential for enhanced feed efficiency.

Addition of both saponin and tartaric acid resulted in significant reductions in the proportion of methane to total gas produced (22.7 and 11.8%, respectively) compared to the control, indicating that these compounds may be suitable for deployment in drinking water. Although there was some variability in the Beeocitrix+ measurement (8.4 ± 2.97), there was also significant reduction of 23.6% when comparing the proportion of methane to the total gas produced in the control.

2.4.4. Fermentation Four

The addition of both monensin and the Rumin8 IVP compounds resulted in methane production reductions (64.2 and 99%, respectively; p = 0.001), without a significant reduction in IVDMD (p = 0.751) (

Table 9. In vitro dry matter (DM) digestibility (IVDMD), total gas production, methane production, ratio of total gas to methane production, volatile fatty acid (VFA) concentration and proportion measured in vitro. Values presented as mean ± SEM.

Parameter	Control	Tribromomethane	Monensin	Rumin8 IVP	p-Value
IVDMD (% DM)	64.1 ± 0.35	65.6 ± 1.08	64.8 ± 0.04	65.4 ± 1.90	0.751
Total gas production (mL/g DM)	111.9 ± 3.40 a	63.1 ± 3.79 c	83.2 ± 3.05 b	64.3 ± 1.79 c	0.001
Methane production (mL/g DM)	12.3 ± 1.58 a	0.00 ± 0.02 b	4.4 ± 0.26 b	0.0 ± 0.01 b	0.001
Total gas–methane	11.0 ± 1.08 a	0.00 ± 0.02 c	5.3 ± 0.50 b	0.0 ± 0.01 c	<0.001
Total VFA (mM)	40.7 ± 1.23 a	25.5 ± 0.56 c	33.1 ± 0.37 b	31.2 ± 0.24 b	<0.001
Acetate (mM)	28.6 ± 0.71 a	14.9 ± 0.10 c	19.7 ± 0.15 b	19.9 ± 0.41 b	<0.001
Propionate (mM)	6.2 ± 0.11 b	7.6 ± 0.19 a	7.4 ± 0.06 a	7.7 ± 0.09 a	0.003
Butyrate (mM)	2.0 ± 0.05 b	2.9 ± 0.06 a	1.5 ± 0.02 c	2.9 ± 0.01 a	<0.001
Valerate (mM)	1.9 ± 0.11 b	2.4 ± 0.15 b	3.5 ± 0.04 a	2.3 ± 0.09 b	0.002
Total branched-chain VFA (mM)	2.0 ± 0.35 a	0.0 ± 0.06 b	1.0 ± 0.16 a	0.0 ± 0.01 b	<0.001
VFA proportions
Acetate (%)	70.2 ± 0.39	58.3 ± 0.89	59.6 ± 0.20	63.9 ± 0.82
Propionate (%)	15.2 ± 0.18	29.9 ± 0.11	22.4 ± 0.06	24.7 ± 0.47
Butyrate (%)	4.9 ± 0.27	11.4 ± 0.02	4.5 ± 0.02	9.4 ± 0.10
Valerate (%)	4.7 ± 0.14	9.3 ± 0.37	10.4 ± 0.22	7.5 ± 0.33
Branched-chain (%)	5.0 ± 0.70	0.0 ± 0.43	3.0 ± 0.46	0.0 ± 0.43
Acetate–propionate	4.6 ± 0.03 a	2.0 ± 0.04 c	2.7 ± 0.02 b	2.6 ± 0.08 b	<0.001

Entries in the same row with different superscript letters indicate significant differences where p < 0.05.

). The latter was also comparable to pure tribromomethane. There was also a reduction (p < 0.001) in the acetate to propionate ratio indicating increased use of hydrogen for propionate-synthesising microbes. However, the large methane reductions were accompanied by significant decreases (p < 0.001) in total VFA concentrations for both. This indicates that the compounds could be suitable for deployment in animal drinking water in future in vivo trials, but the dose rates may need to be optimised to ensure there is no negative impact on animal productivity.

3. Discussion

3.1. FTIR-ATR Analysis

With the exception of sodium monensin, FTIR-ATR spectroscopy was able to provide compelling evidence as to the solubility/stability of the MRCs tested in aqueous solution. However, there are limitations to the method that must be addressed. Inconsistencies in absorbances measured due to uneven drying or crystallization of the compound on the ATR crystal are possible [9], which can be controlled with increased replication. The main drawback of the technique is the inability of FTIR to discern specific compounds from mixtures, instead providing a unique spectral fingerprint of the entire mixture of compounds. There is a possibility that the active ingredient responsible for a compound’s methane-mitigating properties could be lost in the drying process, and the spectra are only reflective of the emulsifying agent/preservatives that remain. This can only be overcome with more powerful instrumentation. This is particularly pertinent in the FTIR analysis of the Rumin8 IVP. The volatility of tribromomethane and the proprietary formulation means that it cannot be definitively stated that the active compound of the Rumin8 IVP was retained across all the treatment conditions, and more detailed analysis of solubility/stability is required in future work. This would likely be best achieved using gas chromatography-mass spectrometry and solid-phase microextraction techniques. Despite this, as an initial screening tool FTIR appears to be effective, and by incorporating additional chemometric techniques such as principal component regression analysis validated by gas chromatography it could be used as a reliable and fast method for measuring concentrations in samples from water troughs where MRCs are deployed [10]. Future research would also need to incorporate longer-term stability analysis to replicate the conditions of storage for MRCs in an on-farm situation.

3.2. In Vitro Batch Culture Fermentations and Overall Suitabilty of MRCs for Water Deployment

The shortcomings of in vitro batch culture fermentations are well known, with the relatively short time frame of experiments providing only a superficial simulation of the complex rumen environment. In particular, it would be expected that MRCs which act as rumen modifiers would not be as effective in short fermentations without the necessary time to impact rumen microbiota. Nevertheless, as a quick and relatively inexpensive screening tool it is still a valuable method that allows for relative ranking of treatments to inform more time-consuming and costly in vivo experimentation [57].

The clear standout compounds in terms of methane reduction were sodium monensin and the Rumin8 IVP, delivering significant methane production reductions without impacting IVDMD. Reductions in VFA concentrations in vitro may be an indication that further investigation is required into optimizing dose rates. However, it should be noted that the reduction in total VFA production was less for the Rumin8 IVP than for pure tribromomethane. Results obtained using tribromomethane are consistent with other studies which have shown large methane reductions both in vitro and in vivo [58,59,60]. The decrease in the observed acetate to propionate ratio provides support for the mechanism of methane reduction being an inhibition of methane production, with a corollary increase in hydrogen which would be available for propionate-producing microbes [40]. The Rumin8 IVP, which utilizes synthetic tribromomethane, overcomes the issues of solubility, stability, cost-effectiveness, and supply associated with Asparagopsis algae as a source of tribromomethane. Despite a review article suggesting that the low levels of tribromomethane required for significant methane reductions make it safe for use [40], and a recent study showing that feeding 25 mg tribromomethane/kg DM sourced from Asparagopsis to feedlot cattle for 200 days increased weight gain with no effects on animal health and no residues in meat [61], further work is required to address concerns about its potential safety and negative impact on the ozone layer [62,63]. Tribromomethane itself is classified as a very short-lived substance and therefore has a relatively low ozone depletion potential overall [40]. Although natural contributions to atmospheric tribromomethane are mainly of oceanic origin, with the turnover of seaweed biomass estimated to produce 70% of the total global flux of tribromomethane [64], a full understanding of the environmental impact of tribromomethane and its metabolites on animals, the environment, and consumers is yet to be established [65].This, accompanied by tribromomethane’s high level of volatility, means that continued stabilization efforts and careful screening of animal products for human consumption, monitoring of animal health, and more precise measurements of tribromomethane in animal drinking water will be required in any future studies. It could be the case that a synthetic stabilised tribromomethane may pose less risk.

The results for methane production as well as acetate to propionate ratio using monensin are consistent with another in vitro study using lucerne as the feed substrate [66], indicating the mechanism of action is likely inhibition of microbes producing hydrogen, which then reduces the availability of hydrogen for methane production [67,68]. The current use of monensin in the cattle industry as an ionophoric antibiotic and improver of feed conversion efficiency make it a promising candidate for deployment in water. Despite this, a meta-analysis [69] of in vivo studies indicated that monensin only reduced methane by 10.7% in beef steers as reported by Beauchemin et al. [5]. Furthermore, the difficulty in dissolving it into water would necessitate the development of a pre-made product where the monensin is already dissolved.

Direct comparison between this and previous in vitro studies investigating nitrates is difficult, as most existing work has measured dosage rates as a proportion of feed and added the whole nitrate compound to this feed or directly to the fermentation vessel. However, the resulting decrease in methane production for Na and K nitrates in the current study appear consistent with other in vitro results, with methane reductions ranging from 55.1–76.1% [70,71,72]. However, an in vivo study by Tomkins et al. [73] reported marginal effects of nitrates on enteric methane production, and that the risk of nitrate toxicity was also a major factor. Nitrate in excess can lead to the development of Methemoglobinemia due to the intermediate nitrite forming a strong bond with haemoglobin, which affects oxygen transportation in the blood [74]. The authors emphasised though that spreading the intake of nitrates over a larger number of intake events each day would reduce the risk of toxicity to animals. The potential for conversion of nitrate to nitrite from rumen microbes introduced to the drinking water would also need to be considered in safe dosing rates. Water delivery of nitrates may disperse consumption across the day, but the ad libitum nature of water-based supplementation limits the ability to control intake to safe levels. The observed increase in acetate to propionate ratio and butyrate production when supplying nitrates are also results reported in several other in vitro and in vivo studies [72,73,75,76].

The most common explanation for the methane reduction associated with nitrate supplementation is that the NO₃⁻ anion acts as a hydrogen sink in the thermodynamically favourable processing of nitrate to nitrite, and then nitrite to ammonia reductions, competing with methanogenic species for hydrogen [77,78]. It has also been suggested that nitrate is toxic to protozoa, fungi, and some bacteria [79] and intermediate nitrite is toxic to methanogens [72]. Given the large effect of high nitrate doses on rumen fermentation in the current study, it appears that both mechanisms might be responsible for the reduction in methane production. In their meta-analysis of in vivo studies where nitrate had been used as an anti-methanogen, Feng et al. [77] reported dosage rates ranging from 4 to 27 g nitrate/kg of DM. This would suggest that the dose rates utilised in aqueous solution in this study are safe, yet the dose rate of 4 g/L resulted in negative impacts on DM digestibility and VFA parameters. There is the possibility that nitrates delivered in aqueous form may affect the rumen microbiome at a different rate, potentially because of already being available in NO₃⁻ anion form, subsequently able to act more quickly than when incorporated as a complete compound in feed.

When looking at SilvaFeed, reductions in total gas ranging from 4.3 to 8.6% have been reported in two other in vitro studies on condensed Quebracho tannins, with dose rates of 2 and 8% of dried tannin powder added as a portion of total feed substrate [80,81]. These studies also reported methane reductions of 9.9 and 16.9%, respectively, and an in vivo trial using Quebracho at 2% of DM intake (DMI) reported methane reductions of up to 29% [82]. Although not statistically significant (p = 0.649), this study observed a not dissimilar decrease in methane of 20.8% using condensed tannin compared to the control, even though the dose rate was much smaller than that of the other studies mentioned. While further investigation is required, this suggests that methane reductions may be possible using condensed tannins supplied through animal drinking water, in quantities lower than are currently being added to feed.

Although studies utilising Polygain sugarcane extract are limited, two in vivo trials on dairy cattle [83] and lambs [84] reported methane reductions of 34.7 and 49.2%, respectively, both at a dose rate of 0.25% of DMI. In the case of the dairy cattle, animals were both on pasture and concentrate, and the Polygain was added to concentrate at 10 g/kg, with an average daily Polygain consumption of 50 g. The dose rate applied in this study would be equivalent to those studies where water consumption was around 33 L/day. Prathap et al. [84] speculate that the potential mechanism for methane mitigation in sugarcane polyphenols is the binding of protein and a subsequent increase in rumen undigestible protein. The low protein content of the Rhodes grass used in this study may be a reason for the lack of methane reduction compared to other studies and suggests that although water delivery of Polygain may be feasible, methane reductions may also be diet-dependent. The response of Agolin at the dose rate used was expected, given it has been established that Agolin, in its capacity as a rumen modulator, requires a four-week modulation period to impact rumen fermentation, and thereby methane production, in vivo [13]. Despite this, the proven methane-reducing capacity of Agolin reported in the literature and the stability and solubility results in this study indicate that long-term in vivo testing of Agolin in animal drinking water could still be feasible.

Although there are few studies on the supplementation of choline chloride as a methane mitigant, the large increases in gas, methane, and VFA concentration observed in this study are similar to those reported by Li et al. [85], who supplemented choline chloride in vitro in the same concentration investigated in the current study, but by treating the buffer instead of a separate addition of aqueous solution, making the actual dose lower in this study. In that study, the in vitro fermentation was carried out over 15 days and resulted in an approximately 50% increase in total methane production after 48 h, but a steady decrease over the rest of the fermentation until methane could not be detected at the end of the experiment [85]. This indicates that choline chloride could potentially be effective as a methane mitigant and feed conversion improver when supplemented through drinking water, but longer-term in vivo experimentation would be required.

The small reductions in the ratio of methane to total gas produced when adding saponin and tartaric acid appear to support other studies which observed reductions in methane production using those compounds [86,87,88,89,90,91,92]. Conversely, an in vitro study by Reis et al. [91] indicated the unsuitability of tartaric acid as a methane mitigator, as they observed no difference in methane production. In the current study, the quantity of tartaric acid used was kept towards the lower concentrations tested in previous studies [90,91], to ensure that the pH of the water was not too low, which might impact water intake of animals. Beauchemin et al. [5] also noted that methane production decreases utilising tannins are also highly dependent on diet and the adaptability of the methanogenic community present in the animal to degrade it. Thus, more investigation would be required for both tartaric acid and saponin to support its use in animal drinking water.

Similarly, reductions in the methane to total gas production observed using Beeocitrix+ closely match those calculated from data provided by Ehtesham et al. [93] and Morsy et al. [94], with Iranian (14.9%), Brazilian red (9.3%), and Egyptian brown (16.5%) propolis delivering proportional in vitro methane reductions using dairy cow and sheep rumen fluid. Santos et al. [95] also reported significant in vitro methane reductions using Brazilian green propolis and Holstein cow rumen fluid. The dose rates in the latter study ranged from 1–56 g/kg of feed substrate, much higher than the 10 mg/L used in this study, indicating that a larger aqueous dose rate may be required for significant methane reductions. The primary action of propolis is believed to be as a bacterial and protozoal inhibitor [96], but the polyphenols contained therein may also act in a similar fashion to those of Polygain mentioned earlier, with the low protein substrate used in this study potentially also contributing to the non-significant effect on methane production.

Whilst the delivery of MRCs in cattle drinking water may be feasible, there are still several challenges that would need to be addressed before sustainable deployment could be achieved on-farm. There are existing technologies capable of delivering supplements to animal drinking water that could also deliver MRCs [97]. This should enable consistent dosing to be achieved. However, long-term solubility and stability studies on the MRCs being delivered in this fashion would need to be conducted to ensure minimal human interventions are required in remote locations. At the time of writing, the costs per animal for doses of MRCs are also quite high, and in lieu of a rebate scheme or a lowering of these costs through economies of scale, this may be a barrier to entry for many producers. Finally, a thorough investigation of the environmental and animal impacts of the water supplementation of MRCs also needs to be conducted. Of particular importance are the consideration of the impacts of MRCs on soil microbiota, whose overall biomass and relative abundance can be impacted by compounds that have an antibiotic effect [98], as well as the effect on other animals that may access livestock drinking water.

4. Materials and Methods

4.1. Selected Compounds

A total of 13 compounds were selected for analysis. Nitrates (calcium tetrahydrate, magnesium hexahydrate, potassium, and sodium), saponin (Quillaja saponaria), choline chloride, monensin sodium salt, tartaric acid, and pure tribromomethane (used as a positive control for in vitro fermentation) were all sourced from Sigma-Aldrich (Bayswater, Australia). Also included were a synthetic tribromomethane (Rumin8 IVP; Rumin8, Perth, Australia), a polyphenolic plant secondary compound from sugarcane (Saccharum officinarum; Polygain; TPM, Keysborough, Australia), Quebracho tannins (Schinopsis lorentzii/balansae; SilvaFeed; SilvaTeam; San Michele Mondovì, Italy), an essential oil blend (Agolin Ruminant L; Alltech technology, Nicholasville, KT, USA), and green propolis (Beeocitrix+; Beeotec; Santa Rita do Sapucaí, Brazil). Aqueous solutions were prepared to simulate dose rate concentrations of the MRC based on manufacturer recommendations and/or previous studies (

Table 10. Concentrations of methane-reducing compounds used in Fourier transform infrared-attenuated total reflectance (FTIR-ATR) spectroscopy and in vitro fermentations.

Compound	Fermentation	Concentration
Nitrate, potassium and sodium	1	2 g NO3/L 2
Nitrate, calcium and magnesium	1	4 g NO3/L 2
Agolin	2	30 mg/L 1
SilvaFeed	2	1.7 mL/L 3
Polygain	2	1.5 mL/L 4
Choline chloride	3	27.92 g/L 5
Saponin	3	0.5 mg/L 6
Tartaric acid	3	15 mg/L 7
Beeocitrix+	3	10 mg/L 4
Rumin8 IVP	4	10 mL/L 4
Sodium monensin	4	3 mg/L 6
Tribromomethane	4	45 mg/L

¹ Castro-Montoya et al. [99]. ² Tomkins et al. [73]. ³ Menci et al. [100]. ⁴ Manufacturer recommendation. ⁵ Li et al. [85]. ⁶ Castro-Montoya et al. [101]. ⁷ Reis et al. [91].

). All solutions were prepared using Milli-Q water. Chemical structures were reproduced using PubChem Sketcher v2.4.

4.2. Solubility and Stability Testing Using FTIR-ATR

4.2.1. FTIR-ATR Instrument Set-Up and Baseline Key Peak Determination of Compounds

Two separate FTIR instruments were used to reduce analysis time. Data from spectra were only compared within the same MRC, so there was no need to use the same instrument for all analysis. Rumin8 IVP, Beeocitrix+, Agolin, SilvaFeed, Polygain, tartaric acid, and monensin were analysed using a Bruker Alpha FTIR spectrophotometer (Bruker Optics Gmbh; Ettlingen, Germany) fitted with a platinum diamond ATR single reflection module. Nitrates, choline chloride, and saponin were analysed using a PerkinElmer Spectrum 100 FTIR spectrophotometer (PerkinElmer Inc.; Shelton, CT, USA) fitted with a Diamond/ZnSe universal ATR. The FTIR spectral data were visualised and analysed in Spectragryph v1.2.16 (Friedrich Menges; Oberstdorf, Germany).

For baseline neat product analysis, where the compound came in crystal or powder form, the reflection module was covered with the sample (approximately 100 mg) and pressure applied to achieve uniform contact between the ATR interface and sample before the spectrum was measured. In instances where the compound came in liquid form, a 40 µL sample was applied to the reflection module, dried where possible, and the spectrum measured. Air was used as a reference background. The Bruker Alpha spectrometer utilised OPUS software v7.5 (Bruker Optics Gmbh; Ettlingen, Germany), and the FTIR spectra were recorded between 4000 and 400 cm⁻¹ as the average of 24 scans at a resolution of 4 cm⁻¹. The PerkinElmer spectrometer utilised Spectrum IR software v10.6.2.1159 (PerkinElmer Inc.; Shelton, CT, USA), and the FTIR spectra were recorded between 4000 and 650 cm⁻¹ as the average of 8 scans at a resolution of 4 cm⁻¹. Spectra from FTIR analysis of the neat compounds was used to identify key peaks from approximately 1200 to 3500 cm⁻¹. Initial baseline spectra from compounds dissolved in water were also obtained, with a 40 µL sample applied to the reflection module and allowed to evaporate to dryness before measuring the spectrum.

4.2.2. FTIR-ATR Solubility and Stability of Compounds

Initially, where the solubility of an MRC was unknown, aqueous solutions of various concentrations were prepared (n = 3), and a 40 µL aliquot was applied to the reflection module and allowed to dry before measuring the spectrum. Correlation between the intensity of key peak absorbance and known concentration of solution was used to establish solubility. To investigate solubility and stability in conditions likely to be encountered in production systems, aqueous solutions of the MRCs at the dose rate concentration (Table 1) were prepared and a 40 µL aliquot was applied to the reflection module and allowed to evaporate to dryness before measuring a baseline spectrum. Exactly 5 mL of the same solution was pipetted into six 10 mL test tubes, with the caps left off. Two of these tubes were then held at 45 °C (±1 °C) in an oven (Memmert; Schwabach, Germany), two at ambient room temperature (approximately 25 °C), and two refrigerated at 4 °C (±1 °C). After 24 h, a 40 µL aliquot of the sample was withdrawn from the region between the 4 and 5 mL mark on the tube, and then applied to the reflection module, allowed to dry, and its spectrum measured. Measurements were taken from the remaining tubes after 48 h to simulate the time taken in the field until the water in a trough was completely consumed by animals. This process was repeated where acidic and basic (pH 5.5 and 8.5) water was used to prepare the solutions and tubes were held at room temperature. The retention of key peak absorbances after treatment conditions when compared to the baseline were used to indicate stability and solubility.

4.3. In Vitro Batch Culture Fermentations

4.3.1. Experimental Design and Sample Collection

A total of four in vitro batch culture fermentation runs were completed. The method used was a modified version of that described by Kinley et al. [60]. Briefly, fermentations were conducted using individual 250 mL septa port vessels (Simax; Křížová, Czech Republic) including blank, control, and MRC treatments (n = 2). To each fermentation vessel, approximately 1 g of a medium-quality Rhodes grass (Chloris gayana) hay (931 g organic matter, 37 g crude protein, and 699 g neutral detergent fibre/kg DM, 2 mm particle size) substrate was added, as well as 100 mL of pre-warmed (39 °C) Goering and van Soest [102] buffer. The reaction vessels were placed into a shaking water bath (Julabo SW22; Julabo GmbH; Seelbach, Germany) set at 39 °C. Following this, a further 20 mL of liquid was added to the fermentation vessels. Blanks and controls received 20 mL of water, whilst the treatments had 20 mL of aqueous MRC solution at recommended dose rate concentration added (Table 10). This was designed to simulate the water intake of an animal, where rumen volume was 100 L, and water consumption was 20 L.

A further 2 mL of reducing solution was added, before the addition of 25 mL of well-mixed rumen fluid collected [103] from seven to eight grass-fed steers of known history slaughtered at an accredited abattoir (JBS Australia, Rockhampton, Australia). The fermentation vessels were then capped with an Ankom RF gas production module (Macedon, NY, USA), flushed with CO₂ as per manufacturer instructions, and gently stirred. The shaking water bath was then set to 85 RPM, and the fermentations allowed to run for a period of 72 h. The vessels were gently stirred twice a day to cycle any substrate that may have accumulated on the sides. The Ankom RF gas production system measured the total gas produced over the fermentation. The parameters for the unit were configured as follows: maximum pressure—3 psi, live interval—60 s, recording interval—20 min, valve open time—250 ms. Cumulative pressure was used as part of the natural gas law to determine the volume of gas produced, which was expressed as mL/g of substrate DM. The venting gas from the Ankom RF gas production system modules was captured in FlexFoil PLUS gas sample bags (SKC; Dorset, UK), connected to the Ankom modules using Teflon tubing and an Ankom vent valve adapter.

At the end of each fermentation run, vessels were chilled on ice to minimize fermentation before the in vitro fluid was vacuum filtered through size 1 Duran glass fritted crucibles that had previously been filled with approximately 1 cm of fine sand. The crucibles were then oven-dried at 105 °C until constant mass was achieved, and in vitro DM digestibility was determined gravimetrically. A 4 mL subsample of the fermentation fluid was added to 1 mL of 20% metaphosphoric acid with internal standard (11 mM 4-methylvaleric acid) and stored at −20 °C prior to VFA analysis.

4.3.2. Sample Analysis

Methane concentration was measured using gas chromatography on an Agilent Technologies 6890 N GC system (Santa Clara, CA, USA), fitted with a Supelco custom 80/100 HAYESEP Q 3ft x 1/8 in stainless steel column and a flame ionisation detector. Column temperature was 65 °C, injector was 100 °C, and FID was 100 °C. Nitrogen was the carrier gas at 10 mL/min, and injection volume was 250 µL. Subsamples were obtained by extraction from the foil gas sample bags using a gas-tight syringe before direct injection into the GC instrument.

Gas chromatography-mass spectrometry (GC-MS) was used to measure VFA concentration by a method modified from that described by Kinley et al. [60]. Thawed fermentation fluid samples were centrifuged (4 °C, 2000× g, 20 min), and 1.5 mL of the supernatant was passed through a 0.22 µm syringe filter into a sample vial. Analysis was completed on a Shimadzu (Nakagyoku, Kyoto) QP2010 Plus System fitted with an autoinjector/autosampler (AOC-20i/s) and an Agilent (Santa Clara, CA, USA) HP-INNOWAX column (30 m × 0.25 mm × 0.25 µm). The injection temperature was 250 °C, the ion source temperature was 230 °C, and helium was used as the carrier gas at a column flow rate of 1.27 mL/min. The column temperature started at 80 °C, was held for 2 min, ramped to 230 °C at 10 °C per min, then held for 5 min. A split injection mode was used at a ratio of 30 and an injection volume of 0.2 µL. Scanning between 30 and 450 m/z was used as the acquisition mode. A multi-acid standard was used in combination with the internal standard for calibration curve and concentration calculations.

Nitrogen content of the feed substrate was determined using a LECO TruMac Series Carbon and Nitrogen Analyser (Sydney, Australia) and converted to protein content using a conversion factor of 6.25. Neutral detergent fibre was measured using an Ankom (Macedon, NY, USA) model 200 fibre analyser as per the manufacturer’s instructions and reported as g/kg DM.

4.4. Statistical Analysis

All statistical analysis was completed using IBM SPSS statistics (v26). For correlation analysis of MRC solubility, where the line of best fit resulted in linear relationships, Pearson correlation was used; when the relationship was logarithmic, Spearman’s correlation was used, both with two-tailed tests. For solubility/stability analysis of aqueous solutions one-way ANOVA was completed, except where single-point analysis was conducted on Beeocitrix+ and tartaric acid, which was completed using two-sided t-tests. For in vitro analysis, individual fermentation vessels were treated as the experimental unit, with the MRC treatment as the fixed effect. One-way ANOVA was used along with post-hoc Tukey testing. The threshold for significance in all analyses was p < 0.05

5. Conclusions

This study evaluated the suitability of 13 commercially available MRCs for stability, solubility and efficacy when delivered via drinking water under in vitro conditions. Nearly all compounds examined in this study appeared to retain their structure and dissolve readily into aqueous solution at the tested concentrations and temperature and pH conditions using FTIR spectroscopy as a rapid analysis tool. Sodium monensin was the only compound unable to be analysed using FTIR. In the in vitro batch culture fermentation, the synthetic tribromomethane product Rumin8 IVP delivered the greatest reduction in methane production without impacting IVDMD; however, a clear shift in acetate to propionate ratio was accompanied by decreased total VFA production. In terms of pure methane reduction, the Rumin8 IVP appears to be the most promising compound for water deployment, but further research is required to avoid potential impacts to animal productivity. The other compounds investigated also appeared to be suitable for further in vivo analysis, and this study could be used to guide that research or the screening of emerging compounds in the future.