Influence of the artificial sodium saccharin sweetener Sucram® on the microbial community composition in the rumen content and attached to the rumen epithelium in dairy cattle: A pilot study

The products of rumen microbial fermentations are considered essential for animal growth and performance. Changes in these microbial communities can have major effects on animal growth and performance. Saccharin-based artificial sweeteners can be included in livestock diets to increase palatability and encourage feed intake. Despite the importance of the rumen microbial fermentation, little or no research is available regarding how saccharin-based artificial sweeteners affect rumen content and rumen epithelial microbial communities. The aim of this study was to identify changes in both the rumen content and rumen epithelial microbial communities in response to the supplementation of Sucram®, a sodium-saccharin-based sweetener (Pancosma S.A./ADM Groups, Rolle, Switzerland) during standard, non-stress conditions using 16SrRNA gene amplicon sequencing. The rumen epithelial and rumen content microbiota of five Holstein-Friesian milking dairy cattle were compared before (baseline, BL) and after a 28-day supplementation of Sucram®. Illumina MiSeq-based 16S rRNA gene sequencing was conducted, and community analysis revealed significant changes in the abundance of specific phylotypes when comparing BL to Sucram® experimental groups. Sucram® did not have a significant effect on overall rumen microbial community structure between experimental groups. Statistically significant changes in microbial community composition following Sucram® supplementation were observed most consistently across a number of bacterial taxa in the rumen epithelium, while fewer changes were seen in the rumen content. Predicted genomic potentials of several significantly different OTUs were mined for genes related to feed efficiency and saccharin degradation. Operational taxonomic units (OTUs) classified as Prevotella and Sharpea were significantly (p<0.05) increased in samples supplemented with Sucram®, whereas a reduction in abundance was seen for OTUs classified as Treponema, Leptospiraceae, Ruminococcus and methanogenic archaea. This is the first study to report an effect of Sucram® on ruminant microbial communities, suggesting possible beneficial impacts of Sucram® on animal health and performance that may extend beyond increasing feed palatability.

integral for fiber-degradation and feed digestion [24,25]. Each of these distinct groups are key to 5 87 metabolic processes in the host, and the differences between the two should be considered when 88 conducting a rumen microbial analysis. 89 Until now, the effect of Sucram ® , or that of any other sweetener-based food additive, on 90 rumen microbiota composition has not been analyzed. This study aimed to provide preliminary 91 data to determine changes in the rumen content and rumen epithelium bacterial communities, in 92 response to supplementation of Sucram ® in dairy cows under standard, non-stressful, physiological 93 conditions. As a pilot study, targeting possible effects of Sucram ® on microbial community 94 composition, we did not aim for identifying possible effects of Sucram ® on feed intake or feed 95 efficiency. Given that the ruminant microbiome is critical to animal health and performance and 96 that there is very little general knowledge of microbial organisms inhabiting the rumen epithelium,   sweetener Sucram ® C-150 (Pancosma S.A./ADM Groups, Rolle, Switzerland) on rumen content 110 and rumen epithelium microbial communities, each animal was sampled before (baseline, BL) and 111 after 28 days of Sucram ® C-150 feeding. In this way, each animal acted as its own control (BL,112 pre-exposure to Sucram ® C-150) when analyzing potential effects of the compound on rumen 113 microbial communities. Animals were housed together under identical conditions at the ISU dairy 114 farm. All cows received the ISU dairy farm regular diet comprised of ground corn, soybeans, 115 cottonseed hulls, corn silage, baleage and alfalfa hay (53.8% dry matter (DM), 9.46% crude protein 116 (CP), and 13.91% neutral detergent fiber (NDF)). Details of the analysis and chemical composition 117 of the diet are given in Supplementary Table S1. The Sucram ® experimental group cows were 118 given 2 grams of Sucram ® suspended in 10 ml of 1x sterile phosphate buffered saline (PBS) (final 119 concentration: 0.2g/ml) per day and cow as per feeding protocols provided by Pancosma. The 120 Sucram ® C-150 containing solution was added directly through the fistula to ensure all cows 121 consistently received the same amount of sweetener. Our main aim was to identify if the presence 122 of Sucram ® has an influence or rumen microbial communities, and due to the administration 123 procedure of Sucram ® through the fistula, any sensory stimuli (i.e. taste or smell) would be limited 124 in this experimental setup. Consequently, Sucram ® would not have an effect on feed intake as it 125 wasn't mixed with the feed by-passing any sensory stimuli. Thus, we did not measure feed intake, 126 average daily gain or milk yield in response to administration of Sucram ® for this study.

127
Two sample types were taken for this trial: rumen content, and rumen epithelium biopsies, 128 taken from the dorsal part of the rumen wall, both collected directly through the fistula. Rumen    To compare alpha diversity between experimental groups, reads were randomly 167 subsampled to accommodate the sample with the lowest number of reads across data sets (20,000 168 sequences for rumen content samples and 20,000 sequences for rumen epithelial samples).

169
Measurements of Chao species richness, Shannon Diversity, and Simpson evenness were taken to 170 compare community structures between experimental groups. The means of the experimental 171 group alpha diversity measures were compared using a pooled t-test assuming equal variance.

172
Because the analysis compared cattle rumen samples of the same type (rumen content/rumen 173 content or rumen epithelial/rumen epithelial) and because animals of the same group at the same 174 farm, the samples were assumed to be highly similar in nature. Therefore, Bray-Curtis was selected 175 as the dissimilarity coefficient because of its ability to compare closely related samples. After 176 dissimilarity coefficients were assigned to each sample, experimental groups were compared using 177 the analysis of similarity (ANOSIM) package provided by mothur.

182
[32]) packages using the shared and taxonomy file generated in mothur. Sequences were randomly 183 subsampled to 20,000 sequences and Bray-Curtis dissimilarity measures were used to generate 184 distances between samples for the PCoA and CCA plots.  Table S2).  Table S4). BL 246 and Sucram ® bacterial communities of the rumen content were compared using ANOSIM, and no 247 significant differences between experimental groups were found (p = 0.527, R =-0.036). PCoA 248 plots generated with these data also provided no evidence of community clustering according to 249 experimental group and 10.5% of the total variation was due to experimental group (CCA, Fig. 3).

250
The lack of clustering and low amount of variation due to experimental group corroborates the 251 reported ANOSIM community comparison.

252
Significant differences in abundance of 11 OTUs were identified using LEfSe. However, 253 out of the 100 most abundant rumen content OTUs, none were found to be significantly different 254 in abundance between the experimental groups (Supplementary Table S5).  Mogibacterium (3.7%) (Fig. 2).  Table S7). Similar to the rumen content 281 dataset, no significant differences (p-value: 0.2, R-value: 0.025) were detected when comparing 282 entire bacterial communities of experimental groups using ANOSIM. This result was corroborated 283 by the lack of apparent clustering of experimental groups seen in the PCoA (Fig. 3). BL and 284 Sucram ® samples cluster separately in CCA (Fig. 3); however, only 5.2% of the total variation was 285 due to experimental group. OTUs found to be more abundant in the Sucram ® experimental group and 10 OTUs that showed 290 higher abundance in BL samples (Fig. 4, Supplementary Table S8). The 10 OTUs found to be 291 more abundant in the Sucram ® experimental group were classified as Prevotellaceae (OTUs 5, 11,    Figures S1 and S2). Very few of the putative genes for 320 saccharin degradation that were suggested by Deng et al. [39] were identified in any of the 321 genomes analyzed, and those that were, were identified within both experimental groups. As 322 mentioned above, the lack of knowledge pertaining to genes involved in saccharin degradation 323 makes interpreting this data difficult, but does offer a unique opportunity for future research.    In the present report, we show that the addition of Sucram ® did not induce significant 372 changes in overall species richness, evenness or diversity (Supplementary Tables S3 and S5), in 373 either rumen content or rumen epithelial microbial communities. This is somewhat unexpected, 374 considering the impact Sucram ® supplementation had on highly abundant phylotypes, especially 375 the highly abundant OTUs within the rumen epithelium data set (Supplementary Table S6).    We also found that OTU 44, which was classified as Treponema, was significantly   Additionally, to identify genes that may have implications on feed efficiency which 501 Sucram ® differentially influences, we mined known rumen genomes that matched some of the 16S 502 rRNA genes from this study with more than 97% sequence similarity for genes that may have  Comparisons of OTUs using the RGC as a reference are a prediction of genetic potential. supplementation were linked to methane production and reduced feed efficiency, whereas OTUs 534 increased following Sucram ® supplementation were associated with propionate and lactate 535 production potential and increased feed efficiency. Therefore, supplementation with Sucram ® may 536 foster a microbial community that decreases available metabolic hydrogen for methane generation