Structural characterization of the late competence protein ComFB from Bacillus subtilis

Abstract

Many bacteria take up DNA from their environment as part of the process of natural transformation. DNA uptake allows microorganisms to gain genetic diversity and can lead to the spread of antibiotic resistance or virulence genes within a microbial population. Development of genetic competence (Com) in Bacillus subtilis is a highly regulated process that culminates in expression of several late competence genes and formation of the DNA uptake apparatus. The late competence operon comF encodes a small protein of unknown function, ComFB. To gain insight into the function of ComFB, we determined its 3D structure via X-ray crystallography. ComFB is a dimer and each subunit consists of four α-helices connected by short loops and one extended β-strand-like stretch. Each subunit contains one zinc-binding site formed by four cysteines, which are unusually spaced in the primary sequence. Using structure- and bioinformatics-guided substitutions we analyzed the inter-subunit interface of the ComFB dimer. Based on these analyses, we conclude that ComFB is an obligate dimer. We also characterized ComFB in vivo and found that this protein is produced in competent cells and is localized to the cytosol. Consistent with previous reports, we showed that deletion of ComFB does not affect DNA uptake function. Combining our results, we conclude that ComFB is unlikely to be a part of the DNA uptake machinery under tested conditions and instead may have a regulatory function.

Figure 1. Late competence protein ComFB

(A) comF operon in B. subtilis encodes for three proteins: ComFA, a DNA-helicase and two proteins of unknown function, ComFB and ComFC. The second and third ORFs overlap by four nucleotides. (B) Alignment of ComFB proteins from several firmicute species. α-Helices of B. subtilis ComFB fold are labeled. Cysteine residues are denoted with asterisks. Residues of the alternative dimerization interfaces, found in crystals, are marked with closed and open arrowheads respectively. Ba, Bacillus amyloliquefaciens; Bl, Bacillus lichenomorphis; Bp, Bacillus pumilis; Bs, Bacillus subtilis; Cp, Clostridium phytofermentans; Ct, Clostridium thermocellum. (C) SDS-PAGE analysis of the purified histidine–tagged ComFB protein. (D) Recombinant ComFB protein dimerizes in solution. ComFB protein was analyzed via SEC on a Superdex75 10/300 column. Molecular weight of purified ComFB monomer is 13.5 kDa. SEC peak position corresponds to ∼35 kDa as defined by calibration with the SEC standards of known molecular weight (indicated on the top bar of the graph).

Figure 2. 3D structure of ComFB protein

(A) Monomer of ComFB in cartoon representation showing positions of the four α-helices. (B) ComFB dimer (chains A and B) in two different orientations, showing N-terminal tail, β-like stretch (arrowhead) and helix H2 forming the interface. (C) Most of the structure variation among the four ComFB molecules in the asymmetric unit originates from the position of helix H3 and the proceeding H3–H4 loop as is illustrated in the present study by superimposing the four molecules using all Cα atoms except those of residues 50–65.

Figure 3. Zinc-binding site of ComFB

(A) Simulated-annealing omit map contoured at 1.5σ, showing a significant extra density flanked by four cysteine residues of ComFB. (B) XAFS spectra of native ComFB crystals that indicate presence of Zn atoms. (C) Anomalous difference map (green mesh; 10σ) and refined ComFB structure. The distances from the cysteine sulfur atoms to the Zn²⁺ cation are indicated.

Figure 4. Dimerization interface of ComFB

(A) Polar interactions at the dimer interface with a buried surface area of ∼1550 Ų. (B) SEC analyses of mutant ComFB proteins show that all three tested variants differ from wild-type protein. Bottom of the panel shows SYPRO Ruby stained SDS-PAGE gels of the fractions collected during SEC runs that identify ComFB as a major species in each of the preparations.

Figure 5. ComFB is expressed upon development of competence in B. subtilis although its absence does not affect transformation efficiency

(A) Deletion of comFB does not significantly affect transformation efficiency of B. subtilis cells. Strains deleted for comFC or comF(BC) were complemented with comFC under the comFA promoter and tested in transformation efficiency assays. WT is the wild-type PY79 B. subtilis strain. The average transformants/cfu for three experiments is plotted. Error bars indicate ±1 S.D. (B) ComFB–GFP and P_comK-driven mKate co-expression. A lacA::P_comK-mKate construct was introduced to a strain with a ycgO::P_comFA–comFB–gfp background. Cells were grown to competence in 1× MC media and examined by fluorescence microscopy. ComFB–GFP expression is highly correlated with mKate expression, verifying that only competent cells express ComFB–GFP. Scale bar is 10 μm. (C) ComFB–GFP expression. Strains carrying ycgO::P_comFA–comFB–gfp were grown in 1× MC media and normalized samples were harvested at specified time points. Western blots of the samples, probed with α-GFP antibodies, verify that ComFB–GFP is detectable beginning around 4–5 h time points, corresponding with the onset of competence. The location of free GFP is indicated and confirms that the levels of cleavage products are low. A strain carrying an amyE::P_veg–gfp construct was used as a positive control for free GFP, whereas wild-type (PY79) strain served as a negative control.

Figure 6. Cysteine motifs in proteins of comF operon

Fragments of ComFA and ComFC alignments illustrating presence of classic four-cysteine motifs in ComFA and cysteine-rich N-terminus of ComFC protein. Ban, Bacillus anthracis; Bl, Bacillus lichenomorphis; Bs, Bacillus subtilis; Bt, Bacillus thuringiensis; Sp, Streptococcus pneumoniae.