Bifunctional quorum-quenching and antibiotic-acylase MacQ forms a 170-kDa capsule-shaped molecule containing spacer polypeptides

Abstract

Understanding the molecular mechanisms of bacterial antibiotic resistance will help prepare against further emergence of multi-drug resistant strains. MacQ is an enzyme responsible for the multi-drug resistance of Acidovorax sp. strain MR-S7. MacQ has acylase activity against both N-acylhomoserine lactones (AHLs), a class of signalling compounds involved in quorum sensing, and β-lactam antibiotics. Thus, MacQ is crucial as a quencher of quorum sensing as well as in conferring antibiotic resistance in Acidovorax. Here, we report the X-ray structures of MacQ in ligand-free and reaction product complexes. MacQ forms a 170-kDa capsule-shaped molecule via face-to-face interaction with two heterodimers consisting of an α-chain and a β-chain, generated by the self-cleaving activity of a precursor polypeptide. The electron density of the spacer polypeptide in the hollow of the molecule revealed the close orientation of the peptide-bond atoms of Val20SP-Gly21SP to the active-site, implying a role of the residues in substrate binding. In mutational analyses, uncleaved MacQ retained degradation activity against both AHLs and penicillin G. These results provide novel insights into the mechanism of self-cleaving maturation and enzymatic function of N-terminal nucleophile hydrolases.

Figure 1

Amidohydrolysis of AHL (C₁₀-HSL) and penicillin G catalyzed by MacQ. As shown in Table 1, MacQ is capable of degrading wide varieties of AHLs as well as penicillin G.

Figure 2

Heterodimeric structure of MacQ. (a) Schematic representation of the single MacQ precursor polypeptide and its self-cleaving maturation products. The N-terminal sequencing and MALDI-TOF/MS analysis indicated that the α-chain, SP, and β-chain correspond to residues 5α-182α (29–206), 1SP-27SP (207–233), and 1β-581β (234–814), respectively. The parentheses refer to the residue number of precursor polypeptide of MacQ including signal peptide sequence. The signal peptide (residue 1–24) was predicted with the SignalP 4.1 program (http://www.cbs.dtu.dk/services/SignalP/). (b) Stereo-view ribbon diagram of MacQ heterodimer with the SP generated by self-cleaving activity of a MacQ precursor single polypeptide chain. The α-chain, SP, and β-chain are coloured in orange, magenta, and green, respectively. The bound PAA and the catalytic Ser1β are also shown as stick models in yellow and blue, respectively, and are labelled.

Figure 3

Capsule-shaped assembly of MacQ. (a) Ribbon diagram of molecular structure of MacQ formed face-to-face by two heterodimers shown in Fig. 2b. One heterodimer is shown as the same colour scheme to that in Fig. 2b, and the other heterodimer is shown in gray. Two perpendicular views are shown. (b) Cross section of the molecule in surface representation showing the interior space of the capsule-like molecule of MacQ and the substrate-binding pocket indicated by the red arrow. The bound PAA is also shown as a yellow stick model. (c) mFo-DFc omit map for the observed SP region contoured at 2.6σ level. The final refined model of the SP is shown as stick model and the residues are labelled. The bound PAA and the Ser1β are also shown in stick model. As shown here, the model for the residues Gly1SP-Ala7SP and Glu23SP-Gly27SP are not built due to the poor electron density.

Figure 4

Active site structure of MacQ. The colour scheme is similar to that in Fig. 2b. (a) Bound C₁₀ fatty acid covalently linked to the Ser1β. The side-chains of nearby residues created the hydrophobic C₁₀-binding pocket are also shown and labelled. The expected binding position of HSL moiety is roughly indicated as a red dotted circle. The residues of the SP near the active site are also shown as stick models and labelled. (b) mFo-DFc omit map for the bound C₁₀ and the Ser1β contoured at 2.5σ level. Two binding states of C₁₀ were observed: the covalently linked C₁₀ to Oγ of Ser1β (upper) and the noncovalently bound C₁₀ (lower). (c) Bound PAA. The expected binding position of 6-APA moiety is roughly indicated as a red dotted circle. (d) mFo-DFc omit map for the bound PAA and the Ser1β contoured at 2.5σ level. Two PAA molecules modelled with slightly different positions and conformations fit well onto the observed electron density map.

Figure 5

Mutational analyses of MacQ. (a) Cleavage site of MacQ and PvdQ and the amino acid sequence of S210A and SPΔ7 mutants around SP region. Deletion or substitutions are highlighted in yellow. The Gly21SP (Gly203 in the uncleaved form), the nearest SP residue to the active site, is indicated by the red asterisk. The seven-residue deletion is the minimum requirements to delete the Gly21SP (Gly203) in the cleaved form of MacQ with keeping the cleavage sequence Gly209-Ser210. (b) SDS-PAGE analysis of wild-type MacQ and mutants S210A and SPΔ7 under reducing (β-mercaptoethanol (βME) + ) and non-reducing (βME −) conditions. The results clearly show that both S210A and SPΔ7 mutants were not cleaved. (c) Overall structure of MacQ SPΔ7. The structure shows the same capsule-like molecule as that of wild-type MacQ, but consists of two uncleaved monomers. The residues 183–202, which correspond to the SP of the wild-type enzyme, are coloured in dark magenta. (d) Structure superimposition of wild-type MacQ heterodimer and SPΔ7 monomer. The α-chain, SP, and β-chain of wild-type MacQ are coloured in orange, magenta, and green, respectively. The SPΔ7 is coloured in gray, while the Gly183-Gly202 is coloured in dark magenta, as shown in (a). The conformation of the SP region of wild-type MacQ and the corresponding Gly183-Gly202 of SPΔ7 is completely different. (e) Structure superimposition of the SP/Ser1β of wild-type MacQ in ligand-free and C₁₀/PAA complex and the Gly183-Ser203 of the SPΔ7. A total of 12 chains for wild-type MacQ and 4 chains for SPΔ7 are shown. The SPΔ7 residues of 193–197 for chain B, 178–200 for chain C and 179–200 for chain D could not be built due to poor electron density. The carbon atoms are coloured in magenta for the SP of wild-type enzyme, and in gray for chain A, cyan for chain B, yellow for chain C, and light blue for chain D of the SPΔ7.

Figure 6

Comparison of molecular surface and active site cavities of wild-type MacQ and SPΔ7 mutant. (a) Top view and (b) side view molecular surface representation of heterodimeric wild-type MacQ and monomeric (uncleaved) SPΔ7. The colour scheme corresponds to Fig. 5d. The Ser1β of wild-type MacQ is located at the depth of the active-site cleft and is accessible from the solvent, while the corresponding Ser203 in SPΔ7 is buried inside due to the presence of uncleaved chain. (c) Active site cavities of the heterodimeric wild-type MacQ and monomeric (uncleaved) SPΔ7. The Ser1β/203 is represented as a stick model and coloured in red. C₁₀/PAA binding pocket and the putative HSL/6-APA binding position are indicated by asterisk in red and black, respectively. The red arrow represents the putative releasing route of the HSL/6-APA from the pocket in the vicinity of Ser1β. The corresponding pocket is not formed in the structure of SPΔ7 due to the presence of uncleaved polypeptide chain.