The candidate phyla radiation (CPR) is a diverse monophyletic group of bacteria that has been found in many environments, from the human body to oceans, freshwater, soil and deep subsurface sediments. 

Most of the information about CPR comes from the reconstruction of genome sequences from metagenomic surveys. Based on the analysis of concatenated protein sequences, CPR has been estimated to contain anywhere between 15% and 50% of the total phylum-level bacterial diversity on Earth, with the most recent estimates placing this figure at not more than 26.3%. The diversity of 16S rRNA sequences has led to the division of CPR into 65 or more different phyla. However, a recently proposed approach to taxonomy that uses 120 concatenated protein sequences and normalises for lineage-specific rates of evolution changes this picture completely. This ‘Genome Taxonomy Database’ reclassifies the CPR as a single phylum and proposes the name ‘Patescibacteria’.

So, what do CPR or Patescibacteria look like? How do they grow? Why are they so difficult to culture? The first clues lie in the genome sequences themselves. These sequences reveal a widespread lack of metabolic pathways that are essential for independent cellular growth. Yet there are genes that are enriched in the CPR including type IV pili and a system for the uptake of extracellular DNA. These systems hint at an obligate symbiotic or parasitic lifestyle, where CPR attach to host cells using adhesins such as type IV pili and scavenge key nutrients such as nucleotides.

If the genomes are small, are the cells small too? In 2015, Jillian Banfield and colleagues answered this question by passing groundwater through an ~0.2 µm filter and examining the filtrate. Metagenomic analyses indicated that the vast majority of microorganisms in the filtrate were CPR bacteria. Cells were visualised by cryo-electron tomography and shown to possess periodic surface layers (S-layers) and, in many cases, pili-like structures of varying length and sizes. Occasionally, these pili appeared to link small cells to larger cells including spirochaetes.

Perhaps the most important breakthrough came from Xuesong He and Floyd Dewhirst at the Forsyth Institute in Boston, USA, and their collaborators, also in 2015. Recognising that genome sequences of CPR bacteria contained an unusual base substitution in the 16S rRNA gene, this group predicted that the TM7 phylum (now Saccharibacteria) within the CPR would be resistant to streptomycin. Using streptomycin enrichment, they were able to obtain co-cultures of a saccharibacterium (now known as ‘Nanosynbacter lyticus TM7x’) with Actinomyces odontolyticus. This was the first time that any CPR bacterium had been cultured and provided a unique opportunity to explore the biology of the species. In keeping with the Banfield study, the cultured TM7x were only 200–300 nm in diameter and were obligate epibionts, relying on their host for key nutrients. Further characterisation has shown that TM7x can infect naive host strains of A. odontolyticus, rapidly killing the majority of host cells but leaving a reservoir of uninfected cells. The association quickly evolves into a stable relationship where TM7x only infects and kills a relatively small subpopulation of A. odontolyticus, with little impact on the overall rate of growth of the host.

Armed with a better understanding of CPR and the confidence that they can be cultured, the Forsyth group has developed a protocol for isolating Saccharibacteria in association with host strains by a combination of filtration, ultracentrifugation and co-culture. At an inaugural symposium on ‘The Uncultivable Bacteria’ at the Forsyth Institute, this technique was employed on willing workshop participants, and used to culture a range of novel Saccharibacteria isolates. Other groups are hot on their heels in the race to culture the uncultured members of the human oral microbiota. In 2019, Mircea Podar and colleagues reported a ‘reverse genomics’ approach to engineer antibodies against Saccharibacteria and other uncultured bacteria. They were able to culture three novel Saccharibacteria isolates as well as another previously uncultured lineage termed SR1. The Saccharibacteria were all obligate epibionts of Actinobacteria.

So far, the culture of CPR bacteria has been restricted to the human oral microbiota. We are beginning to build a picture of the Saccharibacteria, although many basic questions about their cell structure and biology remain unanswered. Importantly, it is not yet known how these organisms contribute to human health or disease or to biogeochemical processes in the environment. One key challenge is to transfer our knowledge about isolation procedures to capture CPR from other environments such as the oceans or soil. A first step will be to identify hosts that support growth and enable laboratory culture. This will be critical if we are ever to appreciate the full range of microbiology on Planet Earth.