The circadian clock and darkness control natural competence in cyanobacteria

Abstract

The cyanobacterium Synechococcus elongatus is a model organism for the study of circadian rhythms. It is naturally competent for transformation-that is, it takes up DNA from the environment, but the underlying mechanisms are unclear. Here, we use a genome-wide screen to identify genes required for natural transformation in S. elongatus, including genes encoding a conserved Type IV pilus, genes known to be associated with competence in other bacteria, and others. Pilus biogenesis occurs daily in the morning, while natural transformation is maximal when the onset of darkness coincides with the dusk circadian peak. Thus, the competence state in cyanobacteria is regulated by the circadian clock and can adapt to seasonal changes of day length.

Fig. 1. T4PM and competence proteins comprise the machinery for natural competence in cyanobacteria.

a Fitness values and confidence levels associated with barcoded transposon mutants for natural transformation. Loci that are essential for or strongly contribute to natural transformation are characterized by a low fitness with a high T-value, whereas improved transformation following knockout of a locus generates a high fitness with a high T-value. Fitness values were calculated for 1885 loci, from three independent experiments, representing a total of 82,495 distinct mutants obtained under selective conditions and 90,872 under control conditions. With the exception of a few selected loci representing selected functional categories, only those whose loss of function had a strong and significant fitness effect are displayed on this plot. The data points were labeled with S. elongatus locus tag (Synpcc7942_) numbers and gene names in parentheses. b Depiction of the natural competence machinery, based on the T4PM architectural model of Chang et al., in which each protein is labeled with S. elongatus locus tag number. T4PM or competence protein homologs that are not required for natural transformation are in parentheses. c Transformation assays performed on insertion knockout mutants of selected loci. Strains of S. elongatus that carry a chromosomal insertion in loci known to not affect natural competence, pilA2 or neutral site 3 (NS3), served as positive controls for transformation, and those same strains to which no eDNA was added served as negative controls. The assays were performed with two independent clones for each strain and yielded similar results. Source data are provided as a Source Data file.

Fig. 2. The S. elongatus T4PM contains minor pilins required for natural competence.

a Heatmap of protein sequence homologies between T4PM and competence proteins of S. elongatus and other strains of cyanobacteria. Gray cells indicate no homolog in the corresponding strain. Strains organized according to 16S rRNA gene phylogeny, inferred by Maximum Likelihood. Nodes supported by a bootstrap ≥ 70% (n = 500) are marked with a red asterisk and strains that are known to be naturally competent are marked with a blue plus sign. b Genomic organization and RB-TnSeq insertion loci for two sets of genes essential for transformation. For each these coding regions and flanking sequences, the number of sequencing reads for insertion-mutant barcodes in selective (top boxes) and non-selective control (bottom boxes) conditions is shown for one representative experiment. c Transformation assays performed on deletion and complemented strains of the pilA3 and pilW, and rntA and rntB loci. For each locus, both open reading frames were deleted then complemented separately and together. WT S. elongatus PCC 7942 served as a positive control and the fully complemented strains to which no DNA was added served as negative controls. The assays were performed with three independent clones for each strain and yielded identical results. Source data are provided as a Source Data file.

Fig. 3. Natural competence is under circadian control and induced in the dark.

a Amplitude of circadian-expressed genes reported at peak time, calculated from S. elongatus circadian transcriptomics^,. Most of the T4PM genes (including pilB, M, N, O, P, Q) are circadian controlled and expressed in the morning while a subset of genes needed for natural transformation such as pilA3, pilW, and rntA and rntB, and the competence genes including comEA, comEC, comF, and dprA, are circadian with highest expression levels at dusk. A red circle indicates a gene whose loss of function adversely affects natural competence and a blue circle indicates a gene whose loss of function positively affects natural competence. b Expression change for genes after a 60-min shade pulse given 8 h after the onset of the day, plotted according to RB-TnSeq fitness values as described for Fig. 1. c Transmission electron micrographs of cells grown in LD and incubated at different time points (Zeitgeber time, ZT time) for 6 h after removal of their pili. Electron micrographs were chosen as representatives of 5–10 pictures of cells selected randomly and prepared from two biologically independent cultures for each time point. d Transformation efficiency over a 24-h circadian cycle. Entrained cells of WT S. elongatus released in LL were incubated with eDNA at different circadian time points (CT time). Efficiencies were calculated as the number of antibiotic-resistant colonies per colony forming unit (CFU) without selection upon transformation of three biologically independent cultures (circles, triangles, and plus signs) and plotted as mean values ± standard error of means (SEM) on a square-root scale to compress large values. A similar experiment using a circadian reporter strain of S. elongatus and incubated with eDNA carrying a different antibiotic-resistance gene that integrates at a different neutral site yielded similar results (Supplementary Fig. 4). Source data are provided as a Source Data file.

Fig. 4. The timing of natural competence is controlled by the phosphorylation state of KaiC and requires SigF2-dependent expression of DprA.

a Transformation efficiency at 4 time points over a 24-h circadian cycle of ΔkaiC complemented with a mutant kaiC-SE phosphomimetic allele that mimics the dusk phosphorylation state of KaiC. Efficiencies were calculated as the number of antibiotic-resistant colonies per CFU without selection upon transformation of three biologically independent cultures (circles, triangles, and plus signs) and plotted as mean values ± SEM on a square-root scale. b Transformation assays performed on knockout and complemented strains of sigF2. WT S. elongatus PCC 7942 with and without added eDNA served as controls. The assay was performed using three independent clones for each strain, which yielded similar results. c Expression analysis of selected T4PM and competence genes in a sigF2 knockout strain, a sigF2 complemented strain, and a WT control measured by RT-qPCR on cultures collected 2 h after circadian dusk with cells that went into the dark at ZT 12 or were maintained in light. Fold-changes were calculated as 2^−∆∆Ct in dark relative to light conditions for three biologically independent cultures (circles, triangles, and plus signs) of each strain and plotted as mean values ± SEM. The expression of sigF2 by RT-qPCR was not determined (ND) in the sigF2 knockout and sigF2 complemented strains. Source data are provided as a Source Data file.

Fig. 5. Natural competence is under circadian control, requires external coincidence with darkness, and responds to changes in day length.

Transformation efficiency over a 24-h circadian cycle of cultures entrained with (a, b) 12-h day length, and (c, d) 16-h day length under (a, c) constant light, and (b, d) light-dark conditions. Efficiencies were calculated as the number of antibiotic-resistant colonies per CFU without selection upon transformation of three biologically independent cultures (circles, triangles, and plus signs) and plotted as mean values ± SEM on a square-root scale. White and black bars are either overlaid or shown side-by-side to make both values visible. Source data are provided as a Source Data file.