Regulatory (pan-)genome of an obligate intracellular pathogen in the PVC superphylum

Abstract

Like other obligate intracellular bacteria, the Chlamydiae feature a compact regulatory genome that remains uncharted owing to poor genetic tractability. Exploiting the reduced number of transcription factors (TFs) encoded in the chlamydial (pan-)genome as a model for TF control supporting the intracellular lifestyle, we determined the conserved landscape of TF specificities by ChIP-Seq (chromatin immunoprecipitation-sequencing) in the chlamydial pathogen Waddlia chondrophila. Among 10 conserved TFs, Euo emerged as a master TF targeting >100 promoters through conserved residues in a DNA excisionase-like winged helix-turn-helix-like (wHTH) fold. Minimal target (Euo) boxes were found in conserved developmentally-regulated genes governing vertical genome transmission (cytokinesis and DNA replication) and genome plasticity (transposases). Our ChIP-Seq analysis with intracellular bacteria not only reveals that global TF regulation is maintained in the reduced regulatory genomes of Chlamydiae, but also predicts that master TFs interpret genomic information in the obligate intracellular α-proteobacteria, including the rickettsiae, from which modern day mitochondria evolved.

Figure 1

The 10 conserved TFs in Waddlia chondrophila and their orthologues in the PVC superphylum. Presence (blue) or absence (white) of the 10 TFs in the PVC superphylum. The % of identity, compared with W. chondrophila, is indicated in each square. The proteins are indicated on the top, organisms on the left and the family (black) and the phylum (blue) on the right. Organisms are ordered on the basis of the phylogenetic tree performed with the maximum likelihood method using the 158 core genes. The topology of the Chlamydiae phylum derived from the concatenation of the 158 core genes is similar to what is already known (Pillonel et al, 2015). Euo, HrcA and DnaA3 are only present in the Chlamydiae. PhoB, AtoC and NrdR are conserved in the PVC superphylum. ParB is not present in Simkania negevensis and NrdR is not present in Parachlamydia acanthamoebae.

Figure 2

Temporal expression of the 10 TFs during the developmental cycle of W. chondrophila. Samples (equal volume) of infected Vero cells were harvested at different times p.i. and analysed by qRT-PCR (a–c) or by immunoblotting using specific polyclonal antibodies to the TFs (d–f). The transcript abundance (%) during the developmental cycle (early (blue shading), mid (green) and late (yellow)) was determined for each TF (a–c). The data are shown as mean values±standard deviation of three independent experiments. Most of the genes show a mid-phase transcript profile (a, b) except for euo and atoC, which exhibit a late-phase transcript profile (c). The TF abundance (%, d and e) was quantified by normalization of the signal detected on the blot (f) according to the number of bacteria/well defined by qPCR. Most of the TFs accumulated steadily along the progression of the developmental cycle (d). Euo and DnaA1 exhibited a peak in abundance at 8 h and 24 h, respectively (e). Note that we could not unambiguously detect DnaA3 by immunoblotting and thus omitted it from this analysis.

Figure 3

Occupancy of the 10 conserved TFs on the W. chondrophila genome. (a–j) ChIP-Seq profiles of the 10 TFs. Black line on the graphs depicts the cutoff used to identify the total predicted targets of each TF, while the blue line denotes the median score used to select the predicted high confidence targets (blue) from the total predicted target sites (black). The coordinates below the graph (x axis) indicate the nucleotide (nt) position along the W. chondrophila genome and the y axis shows the relative abundance of the corresponding nt position in the precipitated sample. Owing to the poor quality of the DnaA1 and DnaA3 precipitates we elected not to predict targets.

Figure 4

ChIP-Seq validation for HrcA and PhoB. (a) Binding of His₆-HrcA on promoters identified by ChIP-Seq (P_groES1, P_groES3, P_hrcA, P_phoH and P_{wcw_1080}). The promoters were amplified by PCR using specific primers coupled to Cy5. DNA fragments were incubated in the absence (−) or in presence of an increasing concentration of His₆-HrcA (50, 250 and 1250 nm) and analysed by EMSA. (b) EMSA analysis showing the binding of His₆-HrcA to a synthetic fragment harbouring the HrcA-triple consensus. As a negative control, we used an analogous triple-repeat fragment harbouring the predicted NrdR consensus motif. (c, d) In vivo binding of HrcA on the triple consensus and on its own promoter (P_hrcA) using lacZ reporter gene in E. coli. HrcA was expressed from an arabinose-inducible promoter on pBAD22 (Guzman et al., 1995). As a negative control, the expression empty vector was used. Data are means±standard deviation of three biological triplicates. Panel (c) absolute values, panels (d) and (e) normalized value according to the empty plasmid (set as 100%). When HrcA was expressed, a decrease of 40% of the LacZ activity was observed, suggesting that HrcA is able to bind and to repress the expression of lacZ. (e) In vivo binding of PhoB (expressed from pBAD22) on three promoters (P_{wcw_0193}, P_{wcw_1016} and P_{wcw_1714}) identified by ChIP-Seq and on the PhoB-triple consensus using lacZ reporter gene in E. coli. A decrease of the LacZ activity was observed for the four lacZ promoter-probe plasmids.

Figure 5

Euo consensus binding site analysis. (a) Identification of 50-bp region which includes the consensus and is necessary for the binding of His₆-Euo. PCR probes shifted by 50 bp were designed for the P_queF and P_ftsY promoters and all fragments were used for in vitro binding assay (EMSA) with His₆-Euo. A total of 80 ng of DNA fragments were incubated in the absence or in the presence of an increasing concentration of His₆-Euo, protein–DNA complexes were detected using GelRed. Results are presented in the left column. + or − indicates whether or not shifted bands were observed. (b) EMSA analysis of His₆-Euo binding to the Euo-triple consensus amplified with a specific primer coupled to Cy5. As a negative control, we used the triple repeat of the predicted NrdR consensus binding site (see Figure 4b). (c) Mutations in the consensus poorly affect the binding of His₆-Euo tested by EMSA (GelRed detection, see Supplementary Information). His₆-Euo concentrations used in the gel shift assay are indicated in (c).

Figure 6

Discovery of a minimal Euo target box. (a) Several minimal Euo boxes (green) are present in P_queF, P_ftsY and P_ndh. Euo consensus binding site is also represented in light green. (b) Consensus based on these Euo boxes showing the conservation of the TTT. The TTT was replaced by GGG (shown in red) and the His₆-Euo binding was tested by EMSA (c–e). The synthetic DNAs carrying the mutations were amplified using a specific primer coupled to Cy5 and used for EMSA. Mutation (red boxes) of the four Euo boxes (green) in P_queF and P_ndh completely abolished the binding of His₆-Euo as no band-shift was observed (c, d). Mutations in the two first Euo boxes in P_ftsY completely abolished the binding of His₆-Euo (e).

Figure 7

Developmental control by Euo. (a) LacZ-based promoter-probe plasmids were co-transformed into E. coli with a plasmid carrying euo under the control of an IPTG-inducible promoter (pSRK-Gm) (Khan et al., 2008). As a control, we used an empty expression vector. The LacZ activity was determined and the data represent mean values±standard deviation of three biological triplicates. When Euo expression is induced, the LacZ activity strongly decreased for the P_queF, P_rpoB, P_rhs9 and P_{wcw_1705} and slightly decreased for the P_{wcw_0066}, P_hrcA and P_gltT. (b) Transcriptional interference assays as in (a) with plasmids expressing the point mutant versions of Euo as indicated in the panel. Mutations are located in the two TorI-like wHTH arranged in tandem (individually and in combination) and cloned in a pBAD22 vector under the control of an arabinose-inducible promoter. These plasmids were co-transformed into E. coli with the plasmid carrying the P_queF-lacZ fusion. Euo strongly decreased (by 60%) the LacZ activity for the P_queFpromoter, while the triple mutant only decreased the activity by 30%. (c–e) Temporal expression of Euo target genes was assessed by qRT-PCR at different time p.i. The transcript abundance (%) during the developmental cycle was determined for each Euo target gene. Data are mean values±standard deviation of three independent experiments. Most of the genes exhibited a peak of expression during the mid-phase of the developmental cycle and were considered as mid-phase genes. Wcw_1215 is an early gene since the peak of transcript abundance was at 3 h p.i. whereas katA is a late gene since the expression remained stable during the late phase.