Identification and characterization of smallest pore-forming protein in the cell wall of pathogenic Corynebacterium urealyticum DSM 7109

Abstract

Background: Corynebacterium urealyticum, a pathogenic, multidrug resistant member of the mycolata, is known as causative agent of urinary tract infections although it is a bacterium of the skin flora. This pathogenic bacterium shares with the mycolata the property of having an unusual cell envelope composition and architecture, typical for the genus Corynebacterium. The cell wall of members of the mycolata contains channel-forming proteins for the uptake of solutes.

Results: In this study, we provide novel information on the identification and characterization of a pore-forming protein in the cell wall of C. urealyticum DSM 7109. Detergent extracts of whole C. urealyticum cultures formed in lipid bilayer membranes slightly cation-selective pores with a single-channel conductance of 1.75 nS in 1 M KCl. Experiments with different salts and non-electrolytes suggested that the cell wall pore of C. urealyticum is wide and water-filled and has a diameter of about 1.8 nm. Molecular modelling and dynamics has been performed to obtain a model of the pore. For the search of the gene coding for the cell wall pore of C. urealyticum we looked in the known genome of C. urealyticum for a similar chromosomal localization of the porin gene to known porH and porA genes of other Corynebacterium strains. Three genes are located between the genes coding for GroEL2 and polyphosphate kinase (PKK2). Two of the genes (cur_1714 and cur_1715) were expressed in different constructs in C. glutamicum ΔporAΔporH and in porin-deficient BL21 DE3 Omp8 E. coli strains. The results suggested that the gene cur_1714 codes alone for the cell wall channel. The cell wall porin of C. urealyticum termed PorACur was purified to homogeneity using different biochemical methods and had an apparent molecular mass of about 4 kDa on tricine-containing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

Conclusions: Biophysical characterization of the purified protein (PorACur) suggested indeed that cur_1714 is the gene coding for the pore-forming protein in C. urealyticum because the protein formed in lipid bilayer experiments the same pores as the detergent extract of whole cells. The study is the first report of a cell wall channel in the pathogenic C. urealyticum.

Keywords: Cell wall channel; Corynebacterium urealyticum; Lipid bilayer membrane; Mycolic acid; Porin.

Fig. 1

Extracts from whole C. urealyticum cells visualized on a 16.5% SDS-PAGE containing tricine. The SDS-PAGE was stained with silver. Lane 1: High molecular mass markers, Lane 2: whole C. urealyticum cell wall extract, heated at 95 °C for 10 min Lane 3: whole C. urealyticum cell wall extract (not heated) and Lane 4: Low molecular mass protein markers (4.6, 10, 15, 25 and 40 kDa). The arrows indicate the gel location of the putative pore-forming protein

Fig. 2

a Single-channel recording of a Diph-PC/n-decane membrane in the presence of detergent extract of whole C. urealyticum cells. The aqueous phase contained 1 M KCl, pH 6 and about 50 ng/ml protein extract. The applied membrane potential was 20 mV; T = 20 °C. b Histogram of the probability P(G) for the occurrence of a given conductivity unit observed with membranes formed of 1% Diph-PC dissolved in n-decane. The histogram was calculated by dividing the number of fluctuations with a given conductance unit by the total number of conductance fluctuations in the presence of detergent extracts of whole C. urealyticum cells. Two frequent conductive units were observed for 147 single events taken from 6 individual membranes. The average conductance of the steps corresponding to the left-side maximum was 1.75 nS and that of the right-side maximum was about 4.0 nS. The aqueous phase contained 1 M KCl, pH 6 and about 50 ng/ml protein extract, the applied membrane potential was 20 mV, T = 20 °C

Fig. 3

Analysis of the accessible genomes from C. glutamicum (ATCC 13032), C. efficiens (YS-314 DNA), C. diphtheriae (NCTC 13129), C. jeikeium (K411) and C. urealyticum (DSM 7109). The homologous genes of the chaperonin GroEL2 and a polyphosphate kinase PPK2 enclose a presumed conserved porin domain. The operon covering the genes CgporH and CgporA whose gene products form the main cell wall channel of C. glutamicum is assumed to exist in all strains except for C. jeikeium. Search within the genome of C. urealyticum indicated that this strain does not follow this rule and suggested instead that it contained 3 open reading frames (ORFs) between the genes coding for GroEL2 (cur_1716) and PKK2 (cur_1712). They may code for cell wall proteins similar to PorA. Possible terminator sequences of mRNA transcripts were predicted with TranstermHP (indicated by hairpins [78] or were identified manually (marked by asterisk))

Fig. 4

Different preparations of PorACur visualized on a 16.5% SDS-PAGE containing tricine. The SDS-PAGE was stained with silver. Lanes 1 and 4, Low molecular mass protein ladder. Lane 2, Purified PorACur with C-terminal His₈-tag expressed in C. glutamicum ATCC 13032 ∆HA, Lane 3, Purified PorACur from preparative gel without tag expressed in C. glutamicum ATCC 13032 ∆HA, Lane 5, Purified CUR_1715 without GST-tag (around 9 KDa) expressed in E. coli BL21 DE3 (Omp8). The purified proteins with and without tags were reconstituted into lipid bilayer membranes. Only protein of lane 3 shows channel-formed activity in lipid bilayers. The channel-forming protein, named PorACur (coded by cur_1714 gene), was identified in the known genome of C. urealyticum by its similar chromosomal localization to known porH and porA genes of other Corynebacterium strains (see Fig. 3). The arrows indicate the location of the corresponding proteins

Fig. 5

Study of pore-formation in the presence of pure PorACur of C. urealyticum. a Single-channel recording of a Diph-PC/n-decane membrane in the presence of 10 ng/ml pure PorACur. The aqueous phase contained 1 M KCl and the applied membrane potential was 20 mV; T = 20 °C. b Histogram of the probability P(G) for the occurrence of a given conductivity unit observed with membranes formed of 1% Diph-PC dissolved in n-decane. The histogram was calculated by dividing the number of fluctuations with a given conductance step by the total number of conductance fluctuations in the presence of pure PorACur. Two frequent conductive units were observed for single events taken from different individual membranes. The average conductance of the steps corresponding to the left-side maximum was 1.75 ± 0.25 nS and that of the right-side maximum was 3 nS (65 current steps in total). The aqueous phase contained 1 M KCl, and10 ng/ml protein extract, the applied membrane potential was 20 mV, T = 20 °C

Fig. 6

Alignment of sequences of the cell wall protein PorACj of C. jeikeium K411 with that of C. urealyticum DSM 7109, PorACur, using Clustal W. The alignment was performed using Pole Bioinformatique Lyonnaise Network Protein Sequence Analysis (http://npsa-pbil.ibcp.fr). Amino acids identical in both proteins are highlighted in red (*), strongly similar amino acids (:) are given in green and weakly similar ones (.) in blue

Fig. 7

Secondary structure predictions for the putative pore-forming proteins of C. urealyticum. (a) Prediction for hypothetical protein CUR_1714 and (b) Prediction for hypothetical protein CUR_1715. The secondary structure predictions were performed using (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_seccons.html) and suggested that both proteins contained a high propensity forming α-helices

Fig. 8

Analysis of the secondary structure of PorACur. a The panel shows the hydrophobicity indices of the individual amino acids of PorACur according to [79]. b The secondary structure of PorACur was predicted using a consensus method at the Pole Bioinformatique Lyonnaise network (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_seccons.html) to form α-helices. Amino acid residues were arranged on the basis of heptameric repeats (a-g, rotation of 100 degrees per residue starting from a, in the clockwise direction) showing distinct separation in a hydrophobic domain that could be surrounded by lipid molecules (black) while the hydrophilic domain (grey) is suggested to represent the α-helices orientated to the water-filled lumen of the putative oligomeric PorACur channel

Fig. 9

a Models of octameric and hexameric channels of PorACur of C. urealyticum shown before and after unbiased MD simulations. b RMSD profiles of the unbiased trajectories with their respective starting structures as a function of time. c Cumulative charge moving through the octamer model at 100 mV and 1 M KCl solution

Fig. 10

Cladogram representing the phylogenetic relationships of porin proteins of Corynebacterium species. The tree was generated with the phylogeny.fr tool [80] using protein sequences downloaded from the NCBI protein database with the indicated identifiers. The multiple sequence alignment was calculated with MUSCLE using the custom mode with a maximum number of 16 iterations. The phylogenetic analysis was performed with PhyML and the approximate likelihood-ratio test for branch support. The substitution model was used in default settings. The tree was rendered with TreeDyn and default settings