cAMP-independent Crp homolog adds to the multi-layer regulatory network in Porphyromonas gingivalis

Abstract

Introduction: Porphyromonas gingivalis encodes three CRP/FNR superfamily proteins: HcpR, PgRsp, and Crp^Pg, with Crp^Pg similar to cAMP-sensing proteins but not classified into known families. This study investigates the role of Crp^Pg in regulating the expression of factors essential for P. gingivalis virulence in A7436 and ATCC 33277 strains.

Methods: The role of Crp^Pg protein in P. gingivalis was determined using the Δcrp^Pg mutant strains to characterize their phenotype and to assess the impact of crp^Pg inactivation on gene expression using RNA-seq and RT-qPCR. Additionally, the Crp^Pg protein was purified and characterized.

Results: Key findings in the Δcrp^Pg mutant strain include up-regulated mfa1-5 and rgpA genes and down-regulated trxA, soxR, and ustA genes. While crp^Pg inactivation does not affect growth in liquid culture media, it impairs biofilm formation and enhances adhesion to and invasion of gingival keratinocytes. Crp^Pg binds directly to its own and mfa promoters without interacting with cyclic nucleotides or di-nucleotides. Its three-dimensional structure, resembling E. coli Crp in complex with cAMP and DNA, suggests that Crp^Pg functions as a global regulator independently of cAMP binding. The highest crp^Pg expression in the early exponential growth phase declines as cell density and metabolic conditions change over time, suggesting a regulatory function depending on the Crp^Pg protein amount.

Conclusions: By controlling the shift from planktonic to biofilm lifestyle, Crp^Pg may play a role in pathogenicity. Regulating the expression of virulence factors required for host cell invasion and intracellular replication, Crp^Pg may help P. gingivalis evade immune responses.

Keywords: CRP; CRP/FNR superfamily; Porphyromonas gingivalis; biofilm; gene expression regulation; virulence.

Figure 1

Theoretical structure-function analysis of P. gingivalis Crp^Pg. (A) A phylogenetic tree was generated using selected amino acid sequences of proteins from the CRP/FNR superfamily based on % identity (PID). Crp/Fnr protein homologs from Porphyromonas gingivalis are shown in blue. (B) Heat map showing % identity and similarity between protein sequences of selected CRP/FNR superfamily representatives. (C) The three-dimensional structure of the Crp^Pg protein (PDB: 2GAU) shows the typical Crp protein regions, including the N-terminal domain involved in ligand binding, the C-terminal DNA-binding domain, and the α-helix connecting both domains. Comparison of Crp^Pg monomer structure with structures of selected proteins from the CRP/FNR superfamily: Crp from Escherichia coli (Crp^Ec, PDB: 2GZW), Crp from Thermus thermophilus (Crp^Tt PDB: 4EV0), Vfr from Pseudomonas aeruginosa (Vfr^Pa, PDB: 2OZ6), CLP from Xanthomonas campestris (CLP^Xc, PDB: 3IWZ), CooA from Carboxydothermus hydrogenoformans (CooA^Ch, PDB: 2FMY), DdrI from Deinococcus radiodurans (DdrI^Dr, PDB: 8YZ7), and SdrP from T. thermophilus (SdrP^Tt, PDB: 2ZCW). Bu, Bacteroides uniformis; Pbm, Parabacteroides merdae; Tf, Tannerella forsythia; Pm, Porphyromonas macacae; Yp, Yersinia pestis; Mt, Mycobacterium tuberculosis; Bs, Bacillus subtilis; PgRsp, P. gingivalis redox-sensing protein; HcpR, Fnr-like protein from P. gingivalis.

Figure 2

Crp^Pg protein acts as a global regulator. (A) Summary of RNA-seq data of gene expression in P. gingivalis Δcrp^Pg mutant strain shown as a Volcano plot based on RNA-seq results presenting global changes in gene expression of the A7436Δcrp^Pg mutant strain in comparison to the A7436 wild-type strain. Gene expression with fold change < -1.5 (log₂-0.6) or >1.5 (log₂0.6) and adjusted p-value >0.05 (-log₁₀1.301) is considered as significant. Selected gene names are shown in the plot. (B) Groups of proteins encoded by genes whose expression has changed significantly. RNA-seq data are shown from three independent biological replicates. (C) EMSA was used to determine the binding of the Crp^Pg protein to the mfa and crp^Pg genes’ promoters. The crp^Pg and mfa promoter regions bound by Crp^Pg protein were verified using EMSA with increasing protein amount and biotin-labeled promoter regions. In addition, control using 100× fold excess of unlabeled promoter DNA, as well as non-specific DNA competitor (50 ng/μL poly; dI-dC) in all samples were used. The Crp^Pg-DNA complexes formed are shown as a shift. In addition, protein-DNA aggregates were formed in gel wells suggesting lower non-specific Crp^Pg binding to all DNA fragments examined. (D) Theoretical DNA fragments recognized by Crp^Pg were identified by comparing crp^Pg promoter fragment 4 and mfa promoter fragment 5 ( Supplementary Figure S4 ) with known sequences recognized by E. coli Crp^Ec, YeiL^Ec, Fnr^Ec, and D. radiodurans DdrI^Dr proteins. The consensus sequence is shown in red.

Figure 3

Characterization of Crp^Pg protein. (A) Potential cAMP-binding pocket with labeled predicted amino acids. Crp^Pg structure with cAMP (yellow) was generated using eDock (Zhang et al., 2020). (B) Analysis of cAMP binding using its fluorescent analog 2-Aza-ϵ-cAMP. Protein was incubated with 50 µM 2-Aza-ϵ-cAMP (2.5× excess) and with cAMP as a competitor (12.5× excess). After removing the unbound cyclic nucleotide, the fluorescence of the sample was measured. Fluorescence of buffer alone and initial solution of 2-Aza-ϵ-cAMP and 2-Aza-ϵ-cAMP with cAMP were used as controls. Heme-binding PgRsp protein was used as a negative cAMP-binding control. (C) SDS-PAGE-based analysis of the cAMP influence on Crp^Pg dimer formation using crosslinking with 0.2% formaldehyde. (D) Comparison of the three-dimensional structure of Crp^Pg (blue) with DdrI^Dr (dark green) indicating the amino acids located in the binding pocket. (E) Comparison of the three-dimensional structure of the Crp^Pg (blue) protein with the Crp^Ec protein in the inactive state (apo form; green) and in the active state (Crp^Ec-cAMP complex; purple) suggests that the Crp^Pg protein shows more similarity in the N-terminal domain with the active version of the protein (shown by arrows). (F) Amino acid sequence alignment of P. gingivalis Crp^Pg protein and Crp homologs from T. forsythia (Crp^Tf), P. macacae (Crp^Pm), E. coli (Crp^Ec), P. aeruginosa (Vfr^Pa), D. radiodurans (DdrI^Dr), T. thermophilus (SdrP^Tt). The predicted cAMP-binding amino acids are marked in purple in Crp^Pg, and experimentally determined cAMP-binding amino acids in other proteins are marked in orange and grey. Amino acids blocking the cAMP-binding pocket are marked in red with an asterisk (*), and homologous amino acids in other Crp homologs are marked with a red frame. (G) Analysis of the cAMP-binding pocket and its location in the Crp^Pg protein dimer revealed steric hindrance between the glutamic acid (E142) and the cAMP-binding site.

Figure 4

Analysis of P. gingivalis Crp^Pg expression. (A) Influence of the growth phase on the crp^Pg gene expression in A7436 and ATCC 33277 (33277) wild-type strains. Bacteria were grown in iron and heme-replete conditions (Hm) and collected at the indicated time points. The optical density of the culture was determined at 600 nm (OD₆₀₀). Gene expression was examined using RT-qPCR. Transcript levels determined after 10 h and 24 h were shown concerning those determined after 4 h (set as 1). (B) The recombinant Crp^Pg-HA protein production was examined in the Δcrp^Pg +Crp^Pg-HA complemented strain at the indicated time points using Western blotting with anti-HA antibodies and subsequent densitometric analysis. A7436Δcrp^Pg mutant strain was used as a control. **P<0.01.

Figure 5

Phenotypic characterization of Δcrp^Pg mutant strains generated in P. gingivalis A7436 and ATCC 33277 (33277) wild-type strains. (A) The growth of bacteria in liquid culture media containing iron and heme (Hm) or without heme and supplemented with the iron chelator 2,2-dipyridyl (DIP) was monitored over time by measuring the optical density at 600 nm. (B) Expression of cell-associated HmuY protein was determined by Western blotting with anti-HmuY antibodies. (C) The relative gingipain activities of the whole P. gingivalis cultures were measured using lysine-specific (Kgp) and arginine-specific (Rgp) substrates. The activity of the wild-type strains was set as 100%. (D) Biofilm formation of P. gingivalis on an abiotic surface and S. gordonii or P. intermedia pre-coated plates. Biofilms formed were determined using crystal violet staining, concerning the Kgp activity of whole bacterial cells embedded in biofilm structures (Kgp activity) or the production of HmuY protein by whole bacterial cells forming the surface of biofilm structures (HmuY amount). The biofilm formation assay was performed 4 independent times, every time with 2 biological replicates for S. gordonii and P. intermedia biofilm, and 3 biological replicates for P. gingivalis. (E) The interaction of P. gingivalis with host cells. The ability to invade and adhere to host cells was analyzed using a P. gingivalis-gingival keratinocytes co-culture model. The number of viable bacteria was shown as colony-forming units per ml (CFU/ml). Adhesion – live bacteria attached to keratinocytes; invasion – live bacteria that invaded keratinocytes; interaction – the total number of live bacteria that invaded and adhered to keratinocytes. (F) Gene expression was determined in P. gingivalis cultured for 4 h in the medium collected after 24-h keratinocytes culture (DMEM^KER) in comparison to the fresh medium (DMEM), the latter set as 1. Gene expression was examined using RT-qPCR. (G) Protein pattern of fresh DMEM and DMEM^KER determined with SDS-PAGE. Results are shown as mean ± SE (A) or mean ± SD (C-F). *P<0.01, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 6

Analysis of Crp^Pg influence on P. gingivalis glycosylation. (A) Protein production was examined in whole P. gingivalis wild-type and Δcrp^Pg mutant cell lysates by SDS-PAGE. After electrophoresis, proteins were visualized using CBB G-250. The arrows indicate ~100 and ~120 kDa protein bands in 33277 and 33277Δcrp^Pg which were analyzed with MS. (B) Glycosylation patterns were analyzed using lectin blotting with JACALIN (specific to 3-substituted GalNAcα, and tolerates substitutions at the position 3 of GalNAc via either α or β linkage, including GalNAc, GlcAc, Gal, and longer oligosaccharides), SNA (specific to sialic acid linked to the terminal galactose with α-2,6 or α-2,3 linkage), and MAL II (specific to α2-3-sialylated Galβ1-3GalNAc in O-glycans and tolerates substitutions at the 6-position of GalNAc; e.g., sialylation, sulfation, GlcNAc) lectins (Bojar et al., 2022). (C) Analysis of cell-associated RgpB protein production was examined by Western blotting with anti-RgpB antibodies. RgpB – RgpB protein modified with LPS, Rgp – catalytic domains of RgpA and RgpB. Protein loading after the transfer was verified by Ponceau S staining.