ABOVE: Researchers explored how purple sulfur bacteria that clump together to form pink berries protect themselves from infectious agents. Scott Chimileski, Marine Biological Laboratory

The vast majority of bacteria are single-celled creatures that compete with one another for survival. However, some bacteria team up to build multicellular communities. While loner bacteria boast impressive genetic diversity, cooperating bugs are genetically similar to ensure that no members carry mutations that favor their own growth at the expense of the collective.1 Low genetic diversity increases the risk that the population could be wiped out by a bacteriophage outbreak, and multicellular bacteria need to evolve rapidly-adapting defenses to develop resilience. Now, researchers have found a mutagenesis system in bacterial communities that resembles the human immune system in more ways than one.2 Their findings, reported in Proceedings of the National Academy of Sciences, challenge the view that bacteria lack intricate defense systems to stave off invaders. 

Scientists knew about this particular mutagenesis system, known as diversity-generating retroelements (DGR), for some time but had not unraveled its purpose.3 “No one really understood what bacteria or archaea were doing with them,” said Elizabeth Wilbanks, a microbiologist at the University of California, Santa Barbara and coauthor of the study. 

The innate arm of the human immune system carries toll-like receptors (TLR) that detect common signals from infectious agents, like components of bacterial membranes or fungal cell walls.4 In a similar fashion, DGR encode proteins carrying domains that detect common bacterial threats, such as the nucleotide oligomerization domain-like receptor (NLR) domain, which recognizes foreign DNA coming from bacteriophages.5 

“The [DGR proteins are] not evolutionarily related but [are] structurally analogous to antibodies,” said Aravind Iyer, a computational biologist at the National Institutes of Health, who was not involved with the work. “The inference would be that it is performing something similar by analogy.” 

B cells churn out antibodies and fine-tune their structures to bind more strongly to their antigens by hypermutating the antibody gene, while DGR hypermutate specific sites in target genes using an error-prone polymerase.6 This system might allow bacteria to rapidly reshape immune sensors to adapt to a variety of infectious agents. 

Struck by its similarity to human immunity, Wilbanks and her team further investigated this elusive apparatus. To probe whether DGR might protect multicellular bacteria from infectious threats, the team chose pink berries, which are millimeter sized aggregates of bacteria living on the floor of salt marshes in Woods Hole, Massachusetts that tend to evolve slowly. “What’s really nice about the pink berries is they’re in this sweet spot in that they’re wild, but they’re not so ephemeral and changing that we can’t ask mechanistic questions,” Wilbanks said.

Found in only two percent of prokaryotic genomes, DGR are considerably rare, and among species carrying them, only two percent harbor more than one version of the system.7 The team sequenced the genome of the purple sulfur bacterium Thiohalocapsa PB-PSB1, the dominant species in pink berries, and inspected it for DGR. This bacterium hoarded nine varieties. This treasure trove of mutagenesis systems allowed the species to hypermutate 21 sequences across 15 target genes, making them capable of producing a staggering 10282 variations, a near-boundless variety greater than the number of protons in the universe.

To gauge how frequently other bacterial species carry DGR in their genomes, Wilbanks and her team ran a phylogenetic screen on publicly available genome sequences. They found that only multicellular bacteria carried multiple DGR, barring one single-celled exception, suggesting that these systems are crucial for sustaining multicellular life. “It tells us from a convergent-evolution standpoint that evolving a sophisticated immune response is very important very early on to survive as a multicellular form,” said Wilbanks.

     Pink berries on a marsh pool floor next to mussel shells (left) and in a petri dish besides a ruler (right).
Researchers scope the floor of marsh pools to collect aggregates of multicellular bacteria called pink berries, which have a striking genetic similarity.
Elizabeth Wilbanks

To confirm that DGR were hypermutating their targets, the team searched for variation in the target genes. By sequencing individual genomes across hundreds of pink berries, they saw great diversity at the target sites. In stark contrast, they saw high similarity across the rest of the genomes, sharing more than 98.9 percent of their sequences. This revealed that the genes targeted by DGR mutate much faster than the rest of the genome, intimating that they need to quickly adapt to changing environmental conditions or disease outbreaks.

While their findings suggest that DGR arm bacteria against invaders, much remains to be uncovered. Iyer said, “The most important thing would be to see what kind of viruses are being recognized and what molecules within those viruses.” 

How they limit pathogen spread also remains to be explored. When human cells become infected, they often sacrifice themselves via apoptosis, which entombs the pathogen in small apoptotic bodies and limits its spread to other cells.8 Similar to this system, the immune sensors encoded by the target genes in DGR are involved in programmed cell death pathways, so infected bacteria might similarly sacrifice themselves for the good of the colony. “It becomes beneficial in a system where you can protect the rest of the collective from getting infected,” Wilbanks suggested.

Although growing evidence suggests that DGR staves off pathogens, scientists have not directly demonstrated this experimentally. As an alternative to immune defense, these systems might allow bacteria to sense whether close kin are present nearby. Some fungi species fuse their cells together and can’t afford to mix up species or strains, so they use NLR domains to check if a nearby cell is a fellow clone before fusing.9 Wilbanks thinks that bacterial brethren in the pink berries might employ DGR machinery for similar audits. “They could be being used to detect subtle strain-level variation in deciding who they want to form an aggregate with,” she said. “That is another really intriguing possibility for what these could be used for.”


1. Wucher BR, et al. Breakdown of clonal cooperative architecture in multispecies biofilms and the spatial ecology of predationProc Natl Acad Sci USA. 2023;120(6):e2212650120. 
2. Doré H, et al. Targeted hypermutation of putative antigen sensors in multicellular bacteriaProc Natl Acad Sci USA. 2024;121(9):e2316469121. 
3. Kaur G, et al. Bacterial death and TRADD-N domains help define novel apoptosis and immunity mechanisms shared by prokaryotes and metazoanseLife. 2021;10:e70394.
4. Fitzgerald KA, Kagan JC. Toll-like receptors and the control of immunityCell. 2020;180(6):1044-1066.
5. Kibby EM, et al. Bacterial NLR-related proteins protect against phageCell. 2023;186(11):2410-2424.e18.
6. Di Noia JM, Neuberger MS. Molecular mechanisms of antibody somatic hypermutationAnnu Rev Biochem. 2007;76(1):1-22.
7. Roux S, et al. Ecology and molecular targets of hypermutation in the global microbiomeNat Commun. 2021;12(1):3076.
8. Kvansakul M. Viral infection and apoptosisViruses. 2017;9(12):356.
9. Daskalov A, et al. Molecular mechanisms regulating cell fusion and heterokaryon formation in filamentous fungi. Microbiol Spectr. 2017;5(2):5.2.02.