Structural and mechanistic studies of pesticin, a bacterial homolog of phage lysozymes

Abstract

Yersinia pestis produces and secretes a toxin named pesticin that kills related bacteria of the same niche. Uptake of the bacteriocin is required for activity in the periplasm leading to hydrolysis of peptidoglycan. To understand the uptake mechanism and to investigate the function of pesticin, we combined crystal structures of the wild type enzyme, active site mutants, and a chimera protein with in vivo and in vitro activity assays. Wild type pesticin comprises an elongated N-terminal translocation domain, the intermediate receptor binding domain, and a C-terminal activity domain with structural analogy to lysozyme homologs. The full-length protein is toxic to bacteria when taken up to the target site via the outer or the inner membrane. Uptake studies of deletion mutants in the translocation domain demonstrate their critical size for import. To further test the plasticity of pesticin during uptake into bacterial cells, the activity domain was replaced by T4 lysozyme. Surprisingly, this replacement resulted in an active chimera protein that is not inhibited by the immunity protein Pim. Activity of pesticin and the chimera protein was blocked through introduction of disulfide bonds, which suggests unfolding as the prerequisite to gain access to the periplasm. Pesticin, a muramidase, was characterized by active site mutations demonstrating a similar but not identical residue pattern in comparison with T4 lysozyme.

FIGURE 1.

Structural model of pesticin from Y. pestis. A, Pst structure in schematic representation is displayed from two different views related to each other by a rotation of 180° around the y axis. The N-terminal (NT) 12 residues, including the TonB box (TBB) and residues 30–34, are not visible in the crystal structure due to disorder. Pst consists of three domains as follow: the translocation domain (T), marked in red; the receptor binding domain (R), marked in blue, and the activity domain (A), marked in orange. The secondary structure assignment is given for β-strands (β1–β8) and α-helices (α1–α11). Two independent Pst^AD1 and Pst^AD2 domains connected by α7 helix form the basis of the subdomains of the T4 lysozyme-like structure. The interface residues between the R (blue) and A domain (orange) are highlighted in the box below the structure. One particular hydrogen bond used for the construction of Pst^S89C/S285C is marked with dashed lines. The R domain includes a mixed α/β-fold, and the A domain is α-helical with three short conserved β-strands (β8–β10, see also Fig. 2A). The R domain displays a new fold that is drawn schematically below the structure (β-strands in blue and α-helices in red). The domain composition together with domain residue boundaries is given in the schematic bar below the structures. B, surface charge representation of Pst illustrated from the same orientations as in A. The active site (marked by a black star) is flanked by two patches of positively charged residues that are further surrounded by extended patches of negatively charged residues. C, activity domain in the same orientation as A is displayed from two sides related by a rotation of 180° around the y axis to illustrate conservation of residues. The structure is shown as a schematic model together with the surface representation of conserved (color coded in blue) and highly conserved residues (in yellow) (see also the alignment in supplemental Fig. S6B for comparison). Single residues lining the active site are more strongly conserved and marked with residue type and number. Residues Glu-178, Thr-201, and Asp-207 are particularly important as they form part of the active site (these residues are marked with large yellow dots).

FIGURE 2.

Comparison of the Pst activity domain with lysozyme topology structures. In the upper row the crystal structures of Pst structure alone and structure superpositions (A) with the domain of gp5 of bacteriophage T4 (PDB code 1WTH) (B), bacteriophage 21 lysozyme (PDB code 3HDF) (C), bacteriophage T4 lysozyme (PDB code 148L) (D), and HEWL (PDB code 1H6M) (E) are shown. In the bottom row, the secondary structural topology of the five proteins (based on the DSSP algorithm) are given with β-strands marked in blue and α-helices in red. A, two subdomains forming the activity domain of Pst are encircled by dashed lines. B, general structure motif of the lysozyme-like molecules excluding HEWL is marked by dashed lines. D, active site residues of the T4 lysozyme are marked in green (Glu-11, Asp-20, and Thr-26). One additional structurally related residue of Pst and T4L mutated in this work is shown.

FIGURE 3.

Structure and organization of the translocation domain in bacteriocins. A, surface representation of Pst shown in sand color with the T domain shown in schematic form as partially random coil in magenta with the N terminus (NT) marked. B, for comparison, the surface representation of colicin M (Cma) in the same schematic drawing as Pst showing the T domain wrapped around the R domain. C, model of the translocation of Pst through the FyuA receptor (the structure of FhuA, PDB code 1BY3, was taken as model for FyuA due to the similarity). The distance of the entire membrane-spanning protein is ∼9 nm, although the diameter of the membrane spanning part is 5–6 nm. Assuming the T domains of Pst are extended by an artificial force, the terminus could be elongated to ∼12 nm. If this terminus is reduced by five residues (∼1.5 nm), it would be sufficient to span the membrane and to attach to the periplasmic domain of TonB. A further reduction in length of this domain by another 10 residues (∼3 nm) would presumably not allow binding of the bacteriocin to TonB and thereby abolish uptake and in vivo activity. The model illustrates the effect of truncation of the T domain on Pst uptake. A schematic representation of the Pst constructs used to determine the importance of the T domain in uptake and activity is given on the right side. AA, amino acids; TBB, TonB box.

FIGURE 4.

Structure and stability of the Pst-T4L chimera protein. A, chimera structure of the three-domain protein formed by the Pst N-terminal translocation and receptor binding domains (T and R, in red and blue) and the C-terminal domain, which represents the T4L protein (A in magenta). The structure is displayed from two points of view related by a rotation of 180° around the y axis. The secondary structure assignment essentially follows Fig. 1 for Pst, but the structural elements are individually assigned for the T/R domain (α1–α3 and β1–β6) and the A domain (α1′–α10′ and β1′–β3′). The domain organization (TBB, TonB box) also including the boundary residues are marked with numbers and represented in the bar-like scheme beyond the structure. B, CD spectra; C, melting curves of Pst (260 μg/ml), T4L (200 μg/ml), and Pst-T4L (260 μg/ml). The melting temperatures (T_m values) are indicated.

FIGURE 5.

Structure of the covalently linked disulfide derivative of Pst^S89C/S285C, location of cysteine mutations of Pst-T4L derivatives, and determination of redox state by SDS-PAGE. A, superposition of Pst (orange) and the double Cys mutant Pst^S89C/S285C in the color coding red (T domain), blue (R domain), and green (A domain). The root mean square deviation between the two structures is 0.55 Å. The left panel shows a zoom-in into the structure with the disulfide bond highlighted that connects Pst^R with Pst^A. B, location of the residues mutated to cysteines that were used for studies of the redox-dependent uptake of Pst-T4L. Disulfide bonds are displayed as dashed lines in the structure of the Pst-T4L activity domain (Pst-T4L^A). C and D, SDS-PAGE of purified double cysteine derivatives of Pst and Pst-T4L, respectively. Standard proteins were used as molecular weight markers (St; Fermentas). The gels were stained with Coomassie Blue. C, Pst^S89C/S285C oxidized with Cu²⁺ (ox, lane 3), reduced with 10 mm DTT (red, lane 1) or without addition (lane 2). D, Pst-T4L^I176C/L331C (lanes 1–3), Pst-T4L^S257C/Q289C (lanes 4–6), Pst-T4L^D294C/R321C (lanes 7–9), and Pst-T4L^I176C/R321C (lanes 10–12), oxidized with Cu²⁺ (ox, lanes 2, 5, 8, and 11), reduced with 10 mm DTT (red, lanes 3, 6, 9, and 12), or without addition (lanes 1, 4, 7, and 10).

FIGURE 6.

A, growth curves of E. coli DH5α transformed with plasmids that encode wild type Pst fused to the MalE signal sequence at 30 °C. Transcription was induced with 0.1% arabinose (■) and 0.001% arabinose (▾) or repressed by 0.2% glucose (glc) (●). Cells transformed with the mutated MalE′-Pst^E178A plasmid were not inhibited after induction with 0.1% arabinose (□). B, growth curves of E. coli SIP1332 fyuA⁺ transformed with plasmids that encode Pst^A fused to the MalE signal sequence at 37 °C. Transcription was induced with 0.1% arabinose (■), 0.01% arabinose (▴), or repressed by 0.2% glucose (●). Cells transformed with the mutated MalE′-Pst^A(E178A) plasmid were not lysed after induction with 0.1% arabinose (□) and 0.01% arabinose (Δ).

FIGURE 7.

Tracing the active center of Pst using point mutants. A, sequence and structural alignment of the activity domain of Pst and T4 lysozyme. In the sequence alignment (lines 1–3), the identical residues are marked with asterisk, strongly conserved residues are marked with colon, and weakly conserved residues are marked with period. In the structural alignment, only the identical positions (within a root mean square deviation of 3.5 Å between Cα positions) are marked (*). Residues mutated in Pst for activity assays are marked in green. Active site residues in T4 lysozyme are highlighted in red. B, superposition of the structures of Pst^A and T4 lysozyme (PDB code 168L) in ribbon representation in two orientations (left and right). Residues chosen for mutagenesis according to the structure and sequence alignment were mutated into alanines (see also A). These residues are marked as sticks together with residue and number for Pst and compared with identical T4 lysozyme positions (T4 lysozyme/Pst). Additional residues mutated in Pst are marked by sticks and residue numbers (Asp-207, Gln-301, and Asp-338).