Curli functional amyloid systems are phylogenetically widespread and display large diversity in operon and protein structure

Abstract

Escherichia coli and a few other members of the Enterobacteriales can produce functional amyloids known as curli. These extracellular fibrils are involved in biofilm formation and studies have shown that they may act as virulence factors during infections. It is not known whether curli fibrils are restricted to the Enterobacteriales or if they are phylogenetically widespread. The growing number of genome-sequenced bacteria spanning many phylogenetic groups allows a reliable bioinformatic investigation of the phylogenetic diversity of the curli system. Here we show that the curli system is phylogenetically much more widespread than initially assumed, spanning at least four phyla. Curli fibrils may consequently be encountered frequently in environmental as well as pathogenic biofilms, which was supported by identification of curli genes in public metagenomes from a diverse range of habitats. Identification and comparison of curli subunit (CsgA/B) homologs show that these proteins allow a high degree of freedom in their primary protein structure, although a modular structure of tightly spaced repeat regions containing conserved glutamine, asparagine and glycine residues has to be preserved. In addition, a high degree of variability within the operon structure of curli subunits between bacterial taxa suggests that the curli fibrils might have evolved to fulfill specific functions. Variations in the genetic organization of curli genes are also seen among different bacterial genera. This suggests that some genera may utilize alternative regulatory pathways for curli expression. Comparison of phylogenetic trees of Csg proteins and the 16S rRNA genes of the corresponding bacteria showed remarkably similar overall topography, suggesting that horizontal gene transfer is a minor player in the spreading of the curli system.

Figure 1. Phylogenetic Distribution of the Curli Systems and Operon Structure.

Taxonomic analysis was performed based on the NCBI taxonomy and visualized using MEGAN . The number of strains containing curli systems within each genus is indicated next to the taxonomic units. Note that these numbers are highly influenced by the number of sequenced strains within each phylogenetic group and therefore do not reflect the prevalence of curli systems within these groups. Genera highlighted in red represent genera where curli systems have been previously described. Organization of the curli operons is illustrated for each genus.

Figure 2. Organization of Curli Repeat Motifs.

Minimalistic curli repeats (X₆QXGX₂NX₁₀) are shown with arrows. Yellow arrows represent repeats within CsgA homologs, blue arrows repeats within CsgB homologs, and red arrows represent repeats in homologs, which cannot be reliably classified as either CsgA or CsgB homologs.

Figure 3. Comparison of Gammaproteobacterial CsgA and CsgB Repeat Regions.

Bold residues represent 50% (black), 80% (blue) and 100% (red) conserved residues.

Figure 4. Evolution of Curli Systems.

Comparison of phylogenetic trees based on the CsgF protein sequences and corresponding 16S rRNA genes. The trees based on aligned protein and nucleotide data were estimated using distance matrix and maximum likelihood and resulted in congruent tree topologies. Distance matrix trees are shown.

Figure 5. Curli Systems within Metagenomes.

CsgF homologs were identified within 10 large metagenomes covering a diverse range of habitats, see Table 2, using the curated CsgF HMM. The hits were aligned with the CsgF homologs identified within refseq and phylogenetic trees were estimated using distance matrix.