Effects of monensin and cashew nut-shell extract on bacterial community composition in a dual-flow continuous culture system

Abstract The objective of this study was to evaluate the effects of including monensin and two doses of CNSE in a high producing dairy cow diet on ruminal bacterial communities. A dual-flow continuous culture system was used in a replicated 4 × 4 Latin Square design. A basal diet was formulated to meet the requirements of a cow producing 45 kg of milk per d (17% crude protein and 27% starch). There were four experimental treatments: the basal diet without any feed additive (CON), 2.5 μM monensin (MON), 100 ppm CNSE granule (CNSE100), and 200 ppm CNSE granule (CNSE200). Samples were collected from the fluid and solid effluents at 3, 6, and 9 h after feeding; a composite of all time points was made for each fermenter within their respective fractions. Bacterial community composition was analyzed by sequencing the V4 region of the 16S rRNA gene using the Illumina MiSeq platform. Treatment responses for bacterial community structure were analyzed with the PERMANOVA test run with the R Vegan package. Treatment responses for correlations were analyzed with the CORR procedure of SAS. Orthogonal contrasts were used to test the effects of (1) ADD (CON vs. MON, CNSE100, and CNSE200); (2) MCN (MON vs. CNSE100 and CNSE200); and (3) DOSE (CNSE100 vs. CNSE200). Significance was declared at P ≤ 0.05. We observed that the relative abundance of Sharpea (P < 0.01), Mailhella (P = 0.05), Ruminococcus (P = 0.03), Eubacterium (P = 0.01), and Coprococcus (P < 0.01) from the liquid fraction and the relative abundance of Ruminococcus (P = 0.03) and Catonella (P = 0.02) from the solid fraction decreased, while the relative abundance of Syntrophococcus (P = 0.02) increased in response to MON when compared to CNSE treatments. Our results demonstrate that CNSE and monensin have similar effects on the major ruminal bacterial genera, while some differences were observed in some minor genera. Overall, the tested additives would affect the ruminal fermentation in a similar pattern.

Monensin is a carboxylic polyether ionophoric antibiotic (Haney and Hoehn, 1968), which is produced by the fermentation of Streptomyces cinnamonensis (Anadón and Martínez-Larrañaga, 2014).Feeding monensin as a nonhormonal growth promoter in cattle (Russell and Strobel, 1989) is a common practice to improve ruminal fermentation.More specifically, monensin supplementation has been shown to inhibit H 2 -producing bacteria (Chen and Wolin, 1979), which decreases H 2 production and availability to methanogenic archaea that require H 2 to produce CH 4 (McGuffey et al., 2001).Additionally, monensin has also been reported to decrease butyrate production (Russell and Strobel, 1989), as well as the rate of ruminal biohydrogenation of unsaturated fatty acids in vitro (Fellner et al., 1997), and to increase the concentration of conjugated linoleic acid in milk fat in dairy cows (AlZahal et al., 2008).
CNSE is composed of a mixture of phenolic compounds (anacardic acid, cardanol, and cardol) with selective antimicrobial (Kubo et al., 1993), and antioxidative (Kubo et al., 2006) activities.Previous studies have shown that CNSE, and particularly anacardic acid, would kill lactate-producing ruminal bacteria, such as Streptococcus bovis, thus reducing the accumulation of lactic acid, preventing the development of ruminal acidosis and bloat in feedlot cattle (Nagaraja and Titgemeyer, 2007;Watanabe et al., 2010;Compton, 2021), similar to the responses observed when monensin is used.Previous studies feeding 4 g of CNSE per 100 kg of BW have reported an increase in propionate production and the abundance of propionate producing bacteria (Shinkai et al., 2012), while other studies feeding CNSE at 30 g per cow per d have found no effects (Branco et al., 2015).Therefore, not only there are inconsistencies regarding CNSE dosage, but also there is a knowledge gap on the effects of CNSE on ruminal bacterial communities.
The utilization of antibiotics as feed additives has sparked concerns regarding the development of antimicrobial resistance, posing a potential threat to human health (Gadde et al., 2017); however, the effectiveness of available alternatives has been inconsistent, thus underscoring the need for further investigation of substitute solutions, in light of the worldwide imperative to decrease antibiotic usage.We hypothesized that CNSE would have similar effects to monensin on ruminal bacterial communities.Our objective was to evaluate the effects of including monensin and two doses of CNSE in a high producing dairy cow diet on ruminal bacterial communities in a dual-flow continuous culture system.

Ethical Approval
The University of Florida Institutional Animal Use and Care Committee approved all the animal care and handling procedures required for this experiment.

Experimental Design, Diets, and Treatments
A detailed description of the design, diets, and treatments can be found in our companion study (Sarmikasoglou et al., 2023a).A basal diet was formulated according to the NRC, (2001) recommendations for a lactating Holstein cow with 680 kg body weight and milk production of 45 kg per d with 3.5% fat and 3.0% protein.All diets were formulated to provide the same concentration of nutrients regardless of treatment (17.1 crude protein, 4.1 ether extract, 30.3 neutral detergent fiber, and 26.8 starch, % dry matter (DM), (Sarmikasoglou et al., 2023a).All feed ingredients were ground through a 2-mm screen in a Wiley mill (model No. 2; Arthur H. Thomas Co., Philadelphia, PA).The corn silage was dried for 72 h at 60 °C in a forced-air oven (Heratherm, Thermo Scientific, Waltham, MA) to allow for partial dryness (>90% DM) of the material before grinding.Subsamples (500 g) from each individual ingredient were ground through a 1-mm sieve for nutrient composition analyses.
Treatments were fed at the following concentrations: MON at 2.5 μM, final concentration in the fermenter, monensin sodium salt (M5273, Sigma-Aldrich Chemicals, Burlington, MA, USA), CNSE100 at 100 ppm and CNSE200 at 200 ppm of CNSE granule, which corresponds to 0.10 and 0.20 mg per g of DM, respectively (SDS Biotech K.K., Tokyo, Japan).The concentration of monensin was chosen based on previous dose response studies with protozoal cultures to avoid elimination of protozoa after the initial dosing so that its effect would not be limited to bacteria (Karnati et al., 2009;Sylvester et al., 2009).The concentrations of the two doses of CSNE were selected according to manufacturer guidelines and are comparable to those used previously (Watanabe et al., 2010).
Each fermenter was provided its respective experimental diet (106 g per d DM) divided equally between two feedings at 0800 and 1800 hours.The CNSE100 and CNSE200 were added as dry products to their respective diets and divided into two equal doses respectively.Regarding MON, it was used according to Sigma protocol and as established in previous studies (Sylvester et al., 2009;Capelari and Powers, 2017;Shen et al., 2017), because it is insoluble to water, MON was diluted using absolute ethanol.A stock solution (100×) was made before the experiment, stored at-20 °C, and pipetted into the fermenters immediately before both morning and evening feedings.The final ethanol concentration in the fermenters was less than 1.0 % v/v.Equal volume of absolute ethanol was pipetted to the rest fermenters to account for any effects from absolute ethanol, however potential effects of it to the physiology of the microbes cannot be excluded.

Dual-Flow Continuous Culture System Operation
A dual-flow continuous culture system, as described by Hoover et al., (1976) and modified by Paula et al., (2017) was used for this experiment.Ruminal fermentation is simulated in this system through continuous agitation (100 rpm), infusion of N 2 gas to displace oxygen, constant temperature (39 °C), and infusion of artificial saliva (Weller and Pilgrim, 1974) at 3.05 mL per min to individually regulate passage rates of liquid (11% per h) and solid (5.5% per h) effluents of digesta, and it has been extensively evaluated (Brandao and Faciola, 2019;Brandao et al., 2020).
This experiment consisted of 4 fermentation periods of 10 d each (40 d of in vitro fermentation total).On day 1 of each fermentation period the fermenters were inoculated with ruminal contents collected from 2 ruminally cannulated Holstein cows in mid lactation (108 ± 9 DIM on average) fed twice daily a total mixed ration with 38% corn silage, 19% ground corn, 13% soybean meal, 11% cotton seed, 9% citrus pulp, 8.5% mineral premix, and 1.5% palmitic acid supplement (on a DM basis) 3 wk before start and until completion of the experiment.Ruminal digesta was collected from the ventral, central, and dorsal areas of the rumen, strained through 2 layers of cheesecloth and transferred into prewarmed thermoses.Ruminal content from both cows was homogenized (50:50) and 1.82 L was inoculated to each fermenter.Fermenters were pre-warmed and under continuous flush of N 2 gas during inoculation.

Experimental Procedure and Sampling
Each experimental period consisted of four 10 d of in vitro fermentation.The first 7 d of fermentation of each period were used for adaptation to experimental diets and stabilization of bacterial communities (Salfer et al., 2018).Experimental procedure and sampling for fermentation parameters and degradability of nutrients are described in detail in Sarmikasoglou et al., (2023a).
Data and samples for bacterial sequencing from both liquid and solid effluents and main fermentation variables (ruminal pH, volatile fatty acids [VFA], lactate, NH 3 -N, N metabolism, and degradability measurements) were collected on days 8, 9, and 10 of each period.
Samples for bacterial sequencing analysis were collected daily from both liquid and solid effluents of each fermenter at 3, 6, and 9 h post morning feeding.For the liquid fraction, 5 mL of liquid effluent were collected at each timepoint, totaling 45 mL per fermenter per period.For the solid fraction, 22 g of solid effluent were collected at each timepoint and strained through four layers of cheesecloth, totaling ~200 g of solid sample collected from each fermenter per period.Upon collection, samples were stored at -80 °C for subsequent desoxyribonucleic acid (DNA) extraction.
On day 10 of each experimental period, microbial samples were collected according to the modified method of Krizsan et al. (2010) and described in detail in Brandao et al. (2018).The obtained microbial pellets were freeze-dried, grounded by mortar and pestle, and later were analyzed for 15 N enrichment, total N and DM.All background, digesta, and microbial content samples were freeze-dried and ground with a mortar and pestle at least 24 h after the completion of freezedrying process.
Chemical analyses for feeds, nutrients, VFA and NH 3 -N as well as N metabolism and degradability of nutrients are described in detail in Sarmikasoglou et al., (2023a).

DNA Extraction, Polymerase Chain Reaction Amplification, and rRNA Sequencing
Total genomic DNA from ruminal samples were extracted using the Quick-DNA Fecal/Soil Microbe Miniprep Kit (D6010, Zymo Research Corporation, Irvine, CA, USA), following the manufacturer's instructions.Before storage in −80 °C, the extracted DNA concentration was measured using a Qubit Fluorometer (Invitrogen Waltham, MA, USA).The use of a DNA extraction kit was based on the quality required for downstream analysis, consistency of the method among different lab personnel, as well as the availability of equipment, and reagents.Authors acknowledge that previously DNA extraction kits have resulted in lower community DNA recoveries from rumen content samples (Henderson et al., 2013), and potential over-/under-estimation of certain taxa are possible.However, the same method is widely used in the field with abundant data to compare to.DNA sequencing procedures were performed according to Kozich et al. (2013).Amplification with polymerase chain reaction (PCR) was performed in a C1000 Touch Thermal Cycler (Bio-Rad Laboratories, Hercules, CA, USA).The V4 region of the 16S rRNA gene was amplified by dual-index universal bacterial primers (515F: 5ʹ-GTGCCAGCMGCCGCGGTAA-3ʹ; 806R: 5ʹ-GGACTACHVGGGTWTCTAAT-3ʹ; Caporaso et al., 2011) through an initial denaturation of 5 min under 95 °C, followed by 30 cycles of 30 s at 95 °C, 30 s at 55 °C, 1 min at 72 °C, and 5 min for final elongation at 72 °C.Forward and reverse primers, as well as small DNA fragment contaminants, were removed using a 1% low-melting agarose gel extraction kit.Amplicons were then purified and normalized using a SequalPrep plate kit (Invitrogen Waltham, MA, USA), and the DNA concentration was measured with a Qubit fluorometer (Invitrogen Waltham, MA, USA).Adapters were added to the amplicons, and the DNA library was constructed by equally pooling all the amplicons together and using quantitative real-time PCR for quality check.Sequencing was performed using a MiSeq reagent kit V2 (2 × 250 cycles run; Illumina) in an Illumina MiSeq platform at the Interdisciplinary Center for Biotechnology Research at the University of Florida (Gainesville, FL, USA).Sequences were deposited at the Sequence Read Archive of the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/sra)under access no.PRJNA901396.

Bioinformatics and Analyses
Sequenced amplicons were processed using the DADA2 pipeline (version 1.16) in R (Callahan et al., 2016), and the taxonomy assignment was performed using the Bayesian RDP classifier trained with the RDP train set 18 database (Cole et al., 2014;Edgar, 2018).Briefly, paired-end raw reads were demultiplexed, and the quality profiles of the forward and reverse readings were separately inspected, filtered, and trimmed based on the relationship between error rates and quality scores.Amplicons were truncated at 30 bp to remove the forward primer.Reads with at least one ambiguous nucleotide were filtered out.The maximum error allowed in a read was set to 2. Forward and reverse readings were merged, chimeras removed, and an amplicon sequence variants table was created.The resulting tables were converted into a phyloseq object for downstream analyses.(McMurdie and Holmes, 2013).Before further data analysis, we calculated the coverage of the dataset according to Good (1953) to evaluate whether the number of sequences obtained for each sample was adequate to provide representativeness of the bacterial community (Supplementary Appendix 1).After rarefying the dataset at 10,671 sequences (equivalent to the sample with the smallest number of sequences), all samples had coverage > 99% and thus were considered representative.
To protect against amplicon sequence variants (ASVs) with small mean and large coefficient of variation, ASVs not seen more than 3 times in at least 20% of the samples were removed.Sequencing depth was normalized by the minimum library size (10,671 sequences per sample) to perform all microbiome analyses.

Statistical Analyses
Bacterial alpha diversity indices (Chao1 and Shannon) were calculated with R phyloseq package (McMurdie and Holmes, 2013).Bacterial community structure, using the Bray-Curtis dissimilarity, was visualized by Principal Coordinates analysis (PCoA), and the statistical differences among samples were measured by PERMANOVA using the vegan R package (Oksanen, 2007).Orthogonal contrasts were used to test the effects of (1) ADD-the control compared to all treatments with additives (CON vs. MON, CNSE100 and CNSE200); (2) MCN-treatment with monensin compared to those with CNSE (MON vs. CNSE100 and CNSE200); and (3) DOSE-the single dose compared to the double dose of CNSE (CNSE100 vs. CNSE200).The orthogonal contrasts (ADD, MCN, and DOSE) were used to test the effects of the treatments on phylum, and genus differential abundance using the R limma package (Ritchie et al., 2015).Significance was declared at P ≤ 0.05, and tendencies at 0.05 < P ≤ 0.10.Correlations of the main fermentation variables with genera affected by the above orthogonal contrasts and relative abundance > 0.10% were analyzed using the Pearson CORR procedure of SAS.Significance was declared at P ≤ 0.05.

Results and Discussion
Detailed ruminal fermentation data, including pH, total and individual VFA, N metabolism, nutrient degradability and flow, as well as protozoal numbers can be found in our companion study (Sarmikasoglou et al., 2023a).

Effects on Rumen Bacterial Community
A total of 2,488,377 high-quality 16S rRNA sequences were retained for analysis after filtering, denoising, merging, and removing chimeras with the DADA2 pipeline.Based on those sequences, a total of 2,181 ASVs (from kingdom, phylum, class, order, family, genus) were identified after taxonomy assignment.
The microbial community structure is reported in Figure 1.Principal coordinates analysis shows clustering in the microbiome of the liquid and solid fractions of fermentation but, no difference related to the treatments were observed regarding the community structure suggesting that the treatments did not affect the bacterial beta diversity.
Moreover, the PERMANOVA analysis indicated that the fraction type (liquid or solid) contributed to 20% of the variation in community distances (P < 0.01).Previous studies, aligning with our findings, have highlighted the presence of bacteria in the free floating fraction co-habiting with others attached to particles (Czerkawski, 1986;Henderson et al., 2013;Monteiro et al., 2022).The analyses of treatment effects on richness and diversity at the ASV level of the bacterial community are presented in Table 1.Regarding the bacterial richness (Chao1), and diversity (Shannon), no effects were detected in any of the orthogonal contrasts tested for either liquid or solid fraction, respectively.The absence of effects suggests that the microbial diversity remains unaffected by the inclusion of monesin and/or CNSE.Previously, monensin supplementation has been reported to not affect the ruminal bacterial richness at inclusion rates of 150 and 335 mg per d in vivo (Schären et al., 2017;Melchior et al., 2018), and 5 μM (0.35 mg per mL) in vitro (Shen et al., 2017), which comes into agreement with our findings.Moreover, previous studies have reported a decrease (Schären et al., 2017;Shen et al., 2017;(Sarmikasoglou et al., 2023b) or no effect (Melchior et al., 2018) in ruminal bacterial diversity by the inclusion of monensin.Lastly, contrary to our data, previous studies found that the CNSE supplementation (4-6 g per 100 kg BW) decreased both richness and diversity (Maeda et al., 2021).Overall, it seems that monensin and CNSE function in a similar way on the structure of ruminal bacterial community; however, further research needs to elucidate their potential to affect the ruminal microbiome structure.In order to better understand the possible effects of monensin and CNSE on ruminal microbiome, we analyzed their effects on relative abundance at phylum, and genus levels.

Effects on Bacterial Community Phyla
A total of 15 phyla were identified in both the liquid and solid fractions that exhibited a relative abundance greater than 0.10% (Table 2).The main phyla in both fractions were Firmicutes (liquid = 45.4%;solid = 53.1%),Bacteroidetes (liquid = 24.0%;solid = 26.2%),and Proteobacteria (liquid = 25.7%;solid = 13.8%), in accordance with those reported by previous studies evaluating the rumen core microbiome in cattle under predominantly forage-based diets (Fernando et al., 2010;Henderson et al., 2015;McCann et al., 2016).More specifically, greater relative abundances of Firmicutes are typically found in diets with greater forage content, while Bacteroidetes are usually predominate in diets with greater concentrate content (Clemmons et al., 2019).The major difference between fractions was the increase in the relative abundance of Proteobacteria in the liquid fraction at the expense of a small decrease in Bacteroidetes.Regarding the less-abundant phyla among those with relative abundance greater than 0.10%, Actinobacteria, Spirochaetes, Synergistetes, and Planctomycetes exhibited similar relative abundance in both fractions, while Fibrobacteres were more abundant in the solid fraction.
In the liquid fraction, the relative abundance of Bacteroidetes was increased (MCN, P = 0.03), when MON compared to CNSE, while it tended to increase (DOSE, P = 0.09) in response to CNSE200 when compared to CNSE100.Resistance of B. succinogenes, and B. ruminocola (both belonging to Bacteroidetes phylum) to monensin has been previously observed in pure culture (Chen and Wolin, 1979), when 2.5 μg per mL of monensin was supplemented.An increase in Bacteroidetes was also reported when 368 -518 mg per head per d of monensin was supplemented in vivo (McGarvey et al., 2019).
Concerning the solid fraction, we observed an increase in the relative abundances of Bacteroidetes (DOSE, P = 0.03), and Synergistetes (DOSE, P = 0.04), when CNSE100 compared to CNSE200.Resistance of Ruminobacter amylophilus, Succinivibrio dextrinosolvens, and Selenomonas ruminantium against CNSE at concentration of ≥ 50 μg per mL has been previously reported (Wakai et al., 2021), however none of those genera belong to Bacteroidetes, and/or Synergistetes.Therefore, our findings suggest that species from Bacteroidetes, and/or Synergistetes would exhibit similar effects and potential sensitivity to CNSE, when it dosed at concentrations greater than 50 μg per mL.Overall, future growth inhibition studies should focus on species from those phyla in order to assess if the development of tolerance is phylum wide or limited to species level.

Effects on Bacterial Community Genera and Correlations Between Genera and Ruminal Fermentation Variables
To better understand changes in bacterial phyla, we conducted downstream analysis to the genus taxonomic level.Also, the genera with relative abundance > 0.10%, that had significant abundance in any of the contrasts tested, were correlated with the ruminal pH, VFA, lactate, NH 3 -N, N metabolism, and degradability measurements.
Based on the obtained taxa, a total of 107 out of 139 observed genera in both the liquid (Figure 2; Supplementary Appendix 2) and solid (Figure 3; Supplementary Appendix 2) fractions had a relative abundance greater than 0.10%.The total abundance (absolute) of ruminal bacteria was not affected (P > 0.05) by the inclusion of any of the additives tested in both fractions (data not shown).However, in order to investigate the effects of them in specific ruminal microbial populations we analyzed specific general.Prevotella exhibited the greatest relative abundance in both the liquid (34.0%) and the solid (30.7%) fractions.Dominance of Prevotella in rumen bacterial communities has been also previously reported in dual-flow continuous culture (Salfer et al., 2018;Dai et al., 2019;Monteiro et al., 2022) and in vivo (Henderson et al., 2015;Clemmons et al., 2019).Prevotella sensitivity to monensin has been previously observed, however some species such as P. bryantii seem to exhibit tolerance (Callaway andRussell, 1999, 2000;Ferme et al., 2008).Concerning the CNSE, previous reports supplemented 50 to 200 μg per mL of CNSE in vitro and reported that both in genus level (Sarmikasoglou et al., 2023b), and strain level (Watanabe et al., 2010) the relative abundance of Prevotella was decreased.
In our study the relative abundance of Prevotella in liquid fraction tended to decrease (MCN, P = 0.07) when CNSE compared to MON (Supplementary Appendix 2), which suggests that the CNSE potential mode of action would be strain independent within Prevotella genus.
The relative abundance of Sharpea in the liquid fraction decreased (MCN, P < 0.01), when MON was compared to cashew treatments, while increased by CNSE100 (DOSE, P = 0.01), between the cashew treatments (Figure 2; Supplementary Appendix 2).The Sharpea spp.are anaerobic Gram-positive bacteria that belong to Firmicutes phylum and have been shown to produce lactate, formate, ethanol and acetate during their fermentation (Kumar et al., 2018).In general, greater abundance of Sharpea has been associated with low CH 4 -producing microbiomes in sheep (Kittelmann et al., 2014).Previous studies, found a decrease in Sharpea when haylage was supplemented with monensin at 33 mg per kg feed DM (Kim et al., 2014), which comes in accordance with our findings.Regarding CNSE, a previous study observed a decrease in Sharpea, when MON compared to CNSE in a serum vial culture (Sarmikasoglou et al., 2023b).Currently, no previous studies have evaluated the effects on Sharpea in response to CNSE; however, our results from this study and our companion study (Sarmikasoglou et al., 2023b) suggest that CNSE would promote the growth of Sharpea.Based on the correlations per treatment we observed a negative correlation (r = −0.73;P = 0.04) on CNSE100, between valerate concentration, and the relative abundance of Sharpea (Figure 5A).Regarding CNSE200, positive correlations were observed between isobutyrate (r = 0.74; P = 0.04), as well as isovalerate (r = 0.72; P = 0.04), and the relative abundance of Sharpea, respectively (Figure 6A).Lastly, no significant correlations were observed between MON and the fermentation variables tested.Previously, Sharpea has been shown a positive correlation with propionate and valerate concentrations in high milk yield (>34.5 kg per d) and high milk protein (>3.20%) content from Holstein dairy cows (Xue et al., 2019).In our study, Sharpea relative abundance has been shown a negative correlation with valerate under the presence of CNSE100, which would suggest that CNSE100 could potentially inhibit the Sharpea-enriched microbiome in a fermentation shift towards the production of intermediates, such as lactate, and end products, such as butyrate.Moreover, the relative abundance of CNSE200 was not correlated with valerate, which suggests that Sharpea would exhibit a non-linear correlation with it.However, further studies are needed to validate this those hypothesis.Regarding the isoforms, no previous studies have reported any correlation of isoacids with Sharpea in presence of cashew nut-shell.However, this study indicates that in presence of CNSE200, amino acid degradation would increase resulting in greater amounts of isobutyrate, and isovalerate.Ammonia concentration did not increase with CNSE200, which could indicate greater ammonia utilization by the microbial population.
Mailhella relative abundance in the liquid fraction was decreased (MCN, P = 0.05), when MON compared to the cashew treatments (Figure 2; Supplementary Appendix 2), suggesting a potential sensitivity of this genus to ionophores.Moreover, a positive correlation (r = 0.73; P = 0.04) was observed on MON (Figure 4A), between isovalerate concentration, and the relative abundance of Mailhella, while no significant correlations were observed between each of the cashew treatments and the fermentation variables tested (Figures 5A and 6A).
The relative abundance of Ruminococcus in the liquid fraction was decreased (MCN, P = 0.03), when MON was   (Patel et al., 1981), thus alleviate the elevated NADPH levels from methanogenesis inhibition (Chalupa, 1977).Except Coprococcus spp., the E. oxidoreducens, and Streptococcus bovis have been also identified as phloroglucinol degraders in the rumen (Patel et al., 1981;Krumholz and Bryant, 1986b).These reports would justify the similar pattern in the relative abundance between Eubacterium and Coprococcus in our study (Figure 2; Supplementary Appendix 2).Lastly, Coprococcus has been reported in greater abundance in the rumen of lower CH 4 emission dairy cows by promoting the acrylate pathway for propionate production (Shabat et al., 2016), and also has been involved in the degradation of nitrotoxins in the rumen (Majak and Cheng, 1981).In our study, the CNSE is composed of a blend of phenolic compounds that seems to stimulate the growth of Coprococcus, thus acting similarly to flavonoids.Therefore, CNSE seems to promote the growth of species associated with the degradation of phenolic compounds, and further the utilization of NADPH, which in case of CH 4 inhibition seems to be important.Based on the correlations per treatment, we found a positive correlation on CNSE200, between acetate (r = 0.82; P = 0.01), and the relative abundance of Coprococcus (Figure 6A).This finding would further indicate that CNSE would promote the growth of species associated with the degradation of phenolic compounds.Overall, further research should focus on the effects of anacardic acid, cardanol, and cardol to ruminal Coprococcus and assess the metabolic pathways that this genus utilizes in presence of those phenolic compounds.
Syntrophococcus relative abundance in the solid fraction was increased (MCN, P = 0.02), when MON was compared to the cashew treatments (Figure 3; Supplementary Appendix 2).Also, no correlation effects were detected (Figures 4B, 5B,  and 6B).Species of Syntrophococcus genus are strictly anaerobic and are known to produce acetate as the end product of their fermentation (Krumholz and Bryant, 1986a).In general, acetogens are outcompeted by the methanogens in the rumen, however they are important during the initial rumen development (Morvan et al., 1994) and their association with ruminal protozoa communities (Schiel-Bengelsdorf and Dürre, 2012).Previous studies reported that Syntrophococcus preferentially grows in a pH = 6.00 to 7.60, increases its growth under high concentrate diets (60% to 70% grain), and has a positive correlation with caproate (Neubauer et al., 2018).Previous studies, found a decrease (Schären et al., 2017) or no effects (Kim et al., 2014) in Syntrophococcus, when MON supplemented in vivo.To our knowledge, no previous studies have evaluated the effects on Syntrophococcus in response to CNSE; however, our results suggest that CNSE would not affect growth of Syntrophococcus.Overall, seems that monensin exhibits a growth promoting effect to Syntrophococcus, while the CNSE seems to not affect it.
The relative abundance of Catonella spp. in the solid fraction was decreased (MCN, P = 0.02), when MON was compared to the cashew treatments, while tended to increase by CNSE100 (P = 0.09) between the cashew treatments (Figure 3; Supplementary Appendix 2).Also, no correlation effects were detected (Figures 4B,5B,and 6B).Catonella belongs to the family of Lachnospiraceae, and has been reported that its abundance decreases drastically in presence of soybean-or fish-oil in the diet (Ngu et al., 2022).In addition, some studies have highlighted the potential positive correlation of Catonella with increased mean pulmonary arterial pressure, and subsequently the development of brisket disease in cattle (Gaowa et al., 2020).Currently, no previous studies have evaluated the effects of monensin, and CNSE to Catonella relative abundance, however, our results suggest a potential toxic effect of monensin to the genus.
The relative abundance of Rarimicrobium in the solid fraction decreased (DOSE, P = 0.04) in response to CNSE200, when compared to CNSE100, while tended to decrease (MCN, P = 0.10) when MON compared to the cashew treatments (Figure 3; Supplementary Appendix 2).Based on the correlations per treatment, no correlation effects were detected (Figures 4B,5B,and 6B).The Rarimicrobium spp.are strictly anaerobic, non-spore forming, and nonmotile gram-negative bacteria that belongs to the phylum Synergistetes (Jumas-Bilak et al., 2015).Scarce knowledge about the members of Synergistetes exists in the literature with some studies to highlight their importance in the degradation of plant-derived toxins, and their minor participation in the fermentation of macronutrients (Allison et al., 1992).Our results indicate a potential toxic effect of phenolic compounds included in CNSE to specific species of Synergistetes, suggesting directions of future research.
Lactobacillus relative abundance (ADD, P = 0.07) in the liquid fraction, and the relative abundances of Acutalibacter (ADD, P = 0.09) from the solid fraction tended to decrease in response to the additives tested (Supplementary Appendix 2).Furthermore, the relative abundance of Mitsuokella (ADD, P = 0.09) from the solid fraction tended to increase in response to the additives tested (Supplementary Appendix 2).Regarding the contrast between monensin and the cashew treatments, the relative abundance of Mediterranea (MCN, P = 0.07), and Alterileibacterium (MCN, P = 0.09) in the solid fraction tended to decrease in response to cashew treatments (Supplementary Appendix 2).Concerning the comparison between the cashew treatments, the relative abundance of Dialister (DOSE, P = 0.07), and Alterileibacterium (DOSE, P = 0.05) from the solid fraction tended to increase in response to CNSE100 (Supplementary Appendix 2).Overall, except for Lactobacillus, the physiology and ecology of the remaining aforementioned genera is rudimentary, thus any responses from those genera do not allow for conclusive statements regarding the tested treatments.

Conclusions
This study is one of the few that evaluated the effects of MON and CNSE on the ruminal microbiome.Most of the effects of MON and/or CNSE on the ruminal microbiome were observed on lower abundance genera, mostly in the liquid fraction of the fermentation.The relative abundance of Sharpea, Mailhella, Ruminococcus, Eubacterium, and Coprococcus in the liquid fraction, and the relative abundance of Ruminococcus, and Catonella in the solid fraction were decreased, while the relative abundance of Syntrophococcus was increased in response to MON when compared to CNSE treatments.Our results contribute to the current knowledge towards a better understanding of the effects of CNSE on ruminal fermentation and indicate no major differences on the ruminal microbiome pattern between MON and CNSE.Further research would be needed to compare the effects of MON, and CNSE on milk production and milk composition.

Figure 5 .
Figure 5. Pearson correlation analysis on CNSE100, between the main fermentation variables and the relative abundance of the bacterial genera affected by the inclusion of the additives tested, in liquid (A), and solid (B) fractions.* P ≤ 0.05.

Figure 4 .
Figure 4. Pearson correlation analysis on MON, between the main fermentation variables and the relative abundance of the bacterial genera affected by the inclusion of the additives tested, in liquid (A), and solid (B) fractions.* P ≤ 0.05.

Table 1 .
Effects of CNSE and MON in richness and diversity of the ruminal bacterial community in dual-flow continuous culture