Fusicatenibacter Is Associated with Kefir Drinking

Daily 16s rRNA-based microbiome sampling reveals that consumption of the fermented drink, kefir, is associated with a previously-unexplored genus Fusicatenibacter of the Firmicutes phylum within family Lachnospiraceae.


Introduction
Ke r is a fermented milk drink produced by the action of bacteria and yeasts and believed to have medicinal uses. A rigorous microbial analysis by Walsh et al. ( ) recently showed precisely which microbes are present in ke r, at various stages in the fermentation process. (See Figure ??). The grains themselves contain a combination of lactic acid bacteria (Lactobacillus, Lactococcus, Leuconostoc), acetic acid bacteria (Acetobacter), and yeast, clumped together with casein (milk proteins) and complex sugars in a matrix of a unique polysaccharide called Ke ran. The nutritional content apparently varies depending on fermentation time and other factors.
Numerous studies indicate that regular ke r drinking has positive e ects on health, and it is reasonable to assume that its known bacteriological properties would a ect the gut microbiome, but we are unaware of previous research that conclusively demonstrates that microbes in ke r make it successfully through the acidic environment of the stomach.
Although similar doubts have been expressed about another fermented dairy product, yogurt, careful research has shown several microbial strains that pass through the body successfully. (Uyeno, Sekiguchi, and Kamagata ( )) Furthermore, several of the microbes apparently persist in the gut and can be observed a full days after consumption.
We were interested to know if the same is true of ke r, and how it might alter the gut microbiome on a daily basis. We sequenced the S rRNA gene in near-daily samples of the microbiome of a single subject, a -year-old male in excellent health. Replicating the experiment in David et al. ( ), we carefully tracked diet, sleep, location, activity, and other variables. Most samples were from gut, but bimonthly samples were regularly taken of skin, nose, and mouth as well.
Because we had several hundred days worth of daily microbiome sampling before the subject rst encountered ke r, we also wanted to nd if any new microbes appeared (or disappeared) as a result. Finally, by continuing to test long after the ke r consumption began, we were able to see how long any such microbes remain in the gut.

Results
We found that ke r consumption was associated with a clear change in the abundance of several organisms, including one, Lactococcus, whose presence could be con rmed in the drink as well. To our surprise, we also found at least one new organism, a novel one that had not been observed in hundreds of previous samples taken from the same subject. Furthermore, the new organism, Fusicatenibacter, appears to remain in the gut after ending the ke r consumption, indicating a persistent alteration of the gut microbiome.
The high levels of Lactococcus in the subject's gut was not unexpected. Previous studies have shown it to inhabit ke r drinks and we also found it as the main genus of microbe found in our ke r samples as well. (see Table ) We found that, indeed, it survives passage through the stomach and presents in the subject's gut in high quantities. Because we have the subject's near-daily gut microbiome records for more than one year prior to his rst consuming ke r ( See Figure ), we con dentally attribute the newly-found abundance to the drink.
We also spotted a new new microbe, Fusicatenibactor that appears to exactly trace the ke r consumption. (Figure ). A gram-stain-positive, obligately anaerobic, non-motile, non-spore-forming, spindle-shaped bacterium, our literature search revealed nothing more of interest since its isolation in . Although it is seen regularly in the human gut, we * Replication les are available on the author's Github account (http://github.com/richardsprague). Kefir Taxa Abundance (%) Figure : Abundance levels of Genus Lactococcus in a -year-old male subject's gut. Ke r was consumed only on the dates indicated in blue. Samples were taken near-daily throughout the period, so abundance levels are zero unless otherwise indicated. We note that levels seem to dip when on days when the ke r is not consumed, such as during trips out of town in mid-April and mid-May. are unaware of any reports of its connection to diet. We believe this is the rst reported instance of its association with a speci c type of food. Fusicatenibacter is found at high abundance after drinking ke r. This chart shows abundance levels were zero since testing began more than two years previously.

Discussion
Although we were pleased to see that one of the microbes in the original drink made it through the body and could be found in the gut, this result alone merely con rms what intuition would suggest: microbes that go in the mouth can successfully navigate the entire gastrointesintal tract. Other researchers have reported similar persistence of the same microbe Lactococcus after yogurt consumption, another fermented drink. Interestingly, our subject, an occasional yogurt eater, showed virtually none of this microbe in the years of measurement before drinking ke r. We speculate that there may be something uniquely robust about the particular species of Lactococcus found in this sample, one that may not be found in the commercially-available yogurt previously consumed by the subject.
It is interesting to note that the subject was drinking homemade ke r, fermented overnight in his kitchen, and thus exposed to the same environmental microbes that would have surrounded the subject himself. We hypothesize that the known high variance in microbial environments may play a signi cant role in which microbes appear in the gut. Commercially-purchased ke r is produced in sanitized industrial environments which, while enabling a consistent product and protective against pathogens, may inevitably result in di erences in microbial strains.
We do not understand why a novel microbe, Fusicatenibacter would appear in the gut in such large quantities immediately after the rst drink. We con rmed with the lab that this microbe was unlikely to result from contamination. Although it had not been found in this subject previously, the lab reports that it is found regularly in samples from other people. Analysis of the plates on which the subject's samples were processed indicated no irregularities; in fact, the wells directly adjacent to the subject's sample did not show any of this microbe, though that was present in other samples processed in the same run.
A literature search reveals nothing of clinical or other apparent interest about this microbe, a Clostridium that appears within the family Lachnospiraceae of phylum Firmicutes. We can nd no apparent link to health or other conditions documented by other projects. Since it persists and makes up from -% of the subject's post-ke r microbiome, we think it must have found a role in the microbial ecosystem.
Note that the subject remained in excellent health before and after the ke r consumption. We could detect no signi cant di erences in blood chemistry or other quantitative health metrics. A review of his activity, sleep, and diet reveals no other signi cant di erences that might compound the microbiome changes that occurred after beginning ke r.

Methods
Samples were collected on a daily basis, following instructions from commercially-available kits from uBiome, Inc. Fecal samples were lightly mixed and swabbed throughout to lessen distribution anamolies within the sample. The swabs were stirred into a lysis bu er and then transported at room temperatures to the uBiome lab. Published accounts of the uBiome protocol indiciate that genomic DNA was extracted by a liquid-handling robot, ampli ed up to times using PCR, with primers inserted at the V subunit of the rRNA gene (( F: GTGCCAGCMGCCGCGGTAA and R: GGACTACHVGGGTWTCTAAT) using Illumina NextSeq platform rendering x bp pair-end sequences. Samples were barcoded with a unique combination of forward and reverse indexes allowing for simultaneous processing of multiple samples. De-multiplexing of samples was performed using Illumina's BCL FASTQ algorithm. Acquired reads were ltered using an average Q-score > . Primers and any leading bases were subsequently trimmed from the reads, and forward and reverse reads were appended together. To e ectively cluster real biological sequences and to identify reads that contain errors as a product of sequencing, reads were clustered using the Swarm algorithm (Mahé et al. ) using a distance of nucleotide. The most abundant sequence per cluster was considered the real biological sequence and was assigned the count of all reads in the cluster.
The representative reads from all clusters were subjected to chimera removal using the VCHIME algorithm (Rognes et al.
). Reads passing all above lters were aligned using an in-house uBiome database of S sequences derived from the NCBI-nr database (Benson et al. ). Decreasing sequence identities were used to map reads to di erent taxonomic rankings: > % sequence identity was used for the assignment to a species, > % sequence identity for the assignment to a genus, > % for assignment to a family, > % for assignment to an order, > % for assignment to a class, and > % for assignment to a phylum. The relative abundance of each taxonomic group was calculated by dividing the abundance of the taxonomic group to all sequences that map to any sequence in the bacterial domain.
Bionformatics was performed in R using Phyloseq (McMurdie and Holmes ( )). Source code is available on our Github page.

Conclusions
Interesting associations with microbes of unknown function are commonly found in all microbiome experiments, and the volume of data collected ensures that some "statistically-signi cant" results will be found simply due to random chance. Although the function of Fusicatenibacter is unknown, the association in this subject was so pronounced, and so clearly associated with the start of ke r drinking that we felt other researchers may bene t from learning about our experience. Despite the n= oddity of this experiment, we have seen the organism in other ke r-drinking subjects, and we present this work in the hopes that it may be useful others working to understand the e ect of ke r and other fermented drinks on human health.