Bacteroidota: starve them of fibre and they eat your gut lining

Ask most people to sort bacteria and you get two piles: the good ones in the yogurt advert, and the bad ones on the door handle. Bacteroidota does not go in either pile, and that is not a diplomatic dodge — it is what the published record says. The same phylum that digests the part of your dinner you cannot, and that helps teach your immune system what to tolerate, is also the group that shows up in most anaerobic infections when it gets somewhere it should not be. So let us take the advantages and the disadvantages in order, with the study named each time.

What it is, in one sentence

Bacteroidota is a phylum — a very large branch of the bacterial family tree — of Gram-negative, non-spore-forming bacteria whose members specialise in breaking down complex sugars. Its gut members are strict anaerobes; the phylum also includes soil and water lineages that live with oxygen.

That specialisation is the through-line. A 2019 review of the phylum's biology describes the shared toolkit: members package their carbohydrate-active enzymes into polysaccharide utilisation loci, gene clusters that switch on only when the matching sugar is actually present, and they push those enzymes outward through a secretion system found in this phylum and essentially nowhere else ⁴. It is an efficient design. Do not make the enzyme until the meal arrives.

Where it lives in you

Mostly in the large intestine, and in enormous numbers. When Eckburg and colleagues examined 13,355 ribosomal RNA gene sequences from the colonic mucosa and stool of three healthy adults in Science in 2005 — one of the first attempts to read the gut without trying to grow anything — Bacteroidetes came back as one of the two divisions that dominate the healthy colon ². Later work found the community sorts into a small number of recurring configurations rather than a smooth gradient: a 2011 Nature analysis identified three robust clusters across cohorts from several countries, one of them driven by Bacteroides ¹².

The phylum is not only intestinal. Its relatives live in the mouth, the airways and the vagina, and outside the body it is a dominant phylum of soil, where the same enzyme machinery takes apart plant and fungal glycans instead of dietary fibre ⁴.

What it does for us

It eats what you cannot. Your own genome encodes a thin set of carbohydrate enzymes. Theirs is enormous. When Xu and colleagues published the complete genome of Bacteroides thetaiotaomicron in Science in 2003, the 4,779-member proteome turned out to include an elaborate apparatus for acquiring and hydrolysing otherwise indigestible dietary polysaccharides, plus a large sensing system for working out which sugar has arrived ³. The fermentation products of that work — short-chain fatty acids — are the fuel your colonic lining runs on.

It helps build the immune system. This is the finding that changed how the field talks about commensals. Mazmanian and colleagues showed in Cell in 2005 that a single molecule from Bacteroides fragilis, polysaccharide A, directs the maturation of the host immune system: colonising germ-free mice with the bacterium corrected systemic T-cell deficiencies and T-helper imbalances and guided lymphoid organogenesis, while a mutant lacking that polysaccharide did not ⁷. One bacterial sugar, doing developmental work the mouse could not do alone.

The caveat, said out loud: that was germ-free mice. It is a clean demonstration of a mechanism, not evidence that any particular human intervention reproduces it.

It keeps the mucus barrier honest — if you feed it. Desai and colleagues, writing in Cell in 2016, put mice carrying a defined human gut community on fibre-free diets; deprived of plant polysaccharides, the bacteria turned to the host's own mucus glycoproteins as a nutrient source, the colonic mucus layer eroded, and the animals became far more vulnerable to a mucosal pathogen ⁵. The organisms did not become malicious. They ate the only substrate left.

What it does against us

Outside the gut, it is a serious pathogen. Hannah Wexler's review in Clinical Microbiology Reviews is blunt about the arithmetic: Bacteroides species are found in most anaerobic infections, with an associated mortality of more than 19 %, and B. fragilis — which makes up only about 0.5 % of the human colonic flora — is the most commonly isolated anaerobic pathogen, because of its virulence factors rather than its abundance ⁶. When the gut wall is breached by surgery, perforation or trauma, a quiet resident becomes an abscess.

It carries formidable antibiotic resistance. The same review notes that Bacteroides species have the most antibiotic resistance mechanisms and the highest resistance rates of all anaerobic pathogens, with increasing clinical resistance to cefoxitin, clindamycin, metronidazole, carbapenems and fluoroquinolones ⁶. That is a reason for stewardship, not for fear of your own colon.

One strain type has a cancer file. Some B. fragilis carry a gene for a secreted toxin. Wu and colleagues showed in Nature Medicine in 2009 that enterotoxigenic B. fragilis, but not the non-toxigenic form, triggered colitis and strongly induced colonic tumours in a mouse model of intestinal neoplasia, through a Th17 inflammatory pathway that could be blocked by neutralising interleukin-17 ⁸. In humans, Boleij and colleagues compared mucosal biopsies from 49 colorectal neoplasia patients with 49 controls and found the toxin gene on 85.7 % of left-side and 91.7 % of right-side case samples, against 53.1 % and 55.5 % of control biopsies ⁹.

The caveat, and it is a large one: that is an association in a small study, half the controls carried the gene too, and carrying a toxin gene is not the same as developing cancer. The mouse work shows a mechanism exists; the human work shows a correlation worth studying. Neither shows causation in people.

In the mouth, one member rewrites the whole room. Porphyromonas gingivalis belongs to this phylum and is the founding example of the keystone-pathogen hypothesis, set out in Nature Reviews Microbiology in 2012: a low-abundance organism that provokes disease not by outgrowing everything else but by remodelling an otherwise benign community into a damaging one ¹⁰.

The name: Bacteroidota, formerly Bacteroidetes

Both names are correct; one is current. Until recently the rank of phylum had no standing in the formal code of prokaryotic nomenclature. Once it did, Oren and Garrity published formal descriptions giving valid standing to 42 phylum names in 2021, and Bacteroidetes was regularised as Bacteroidota ¹. NCBI Taxonomy carries the phylum under its current name, with "CFB group bacteria" listed as an older informal label; the taxonomy identifier is in the entity panel of this page. If you are searching the literature, search both spellings — most of the classic papers above were published under the old one.

The "lean-gut ratio" headline, and why to distrust it

You have probably met the claim that a high Bacteroidetes-to-Firmicutes ratio marks a lean gut. It has not held up. Sze and Schloss pooled ten independent data sets in mBio in 2016 and found that while some diversity measures were weakly associated with obesity status, the Bacteroidetes-to-Firmicutes ratio and the individual phylum abundances were not; machine-learning models trained on one data set and tested on the others achieved a median classification accuracy between 33.01 % and 64.77 %, and the difference in Shannon diversity between obese and non-obese people was 2.07 % ¹¹.

That is the honest summary of a decade of headlines. The phylum matters enormously to how your gut works. It is not, on the current evidence, a dial you can read off a stool test and act on.