N -acylhomoserine lactones (AHLs) are used as signal molecules by many quorum-sensing Proteobacteria. Diverse plant and animal pathogens use AHLs to regulate infection and virulence functions. These signals are subject to biological inactivation by AHL-lactonases and AHL-acylases. Previously, little was known about the molecular details underlying the latter mechanism. An AHL signal-inactivating bacterium, identified as a Ralstonia sp., was isolated from a mixed-species biofilm. The signal inactivation encoding gene from this organism, which we call aiiD , was cloned and successfully expressed in Escherichia coli and inactivated three AHLs tested. The predicted 794-amino-acid polypeptide was most similar to the aculeacin A acylase (AAC) from Actinoplanes utahensis and also shared significant similarities with cephalosporin acylases and other N-terminal (Ntn) hydrolases. However, the most similar homologues of AiiD are deduced proteins of undemonstrated function from available Ralstonia , Deinococcus and Pseudomonas genomes. LC-MS analyses demonstrated that AiiD hydrolyses the AHL amide, releasing homoserine lactone and the corresponding fatty acid. Expression of AiiD in Pseudomonas aeruginosa PAO1 quenched quorum sensing by this bacterium, decreasing its ability to swarm, produce elastase and pyocyanin and to paralyse nematodes. Thus, AHL-acylases have fundamental implications and hold biotechnological promise in quenching quorum sensing.
In a process that has become known as quorum sensing (Winans and Bassler, 2002), bacterial cells communicate with each other by releasing, detecting and responding to small signal molecules. Many Gram-negative bacteria produce and monitor the local accumulation of N-acyl-homoserine lactones (AHLs), allowing them to regulate a wide range of their biological activities as a function of their population density (Eberhard et al., 1981; Zhang et al., 1993; Pearson et al., 1994). Cells produce AHL signals via the activity of AHL synthases that are very often encoded by luxI gene homologues. At low population densities, the signal molecule is diluted to its effective extinction as it is released into the environment. As the population grows and the local rate of signal synthesis increases, AHLs can accumulate to concentrations that allow their effective interactions with cognate transcriptional regulators, most often homologues of the LuxR protein of Vibrio fischeri. The signal-bound regulator serves to activate the expression of target genes. There are also examples of circuits in which AHLs serve to relieve transcriptional repression by DNA-bound LuxR homologues (Minogue et al., 2002).
Different bacterial species have been shown to use quorum-sensing mechanisms to regulate a broad range of biological functions, including bioluminescence, Ti plasmid conjugal transfer, production of virulence factors, antibiotics and other secondary metabolites, the ability to move by swarming and to become sedentary and form biofilms (Fuqua et al., 1996; de Kievit and Iglewski, 2000).
Quorum-sensing regulation is very often used by bacteria to facilitate their competition with, and exploitation of, other bacterial, fungal, plant and animal species. Because of this, it is perhaps of no surprise that their competitors have evolved diverse mechanisms to circumvent quorum sensing. Tapping into such naturally occurring, well-evolved mechanisms has aroused significant fundamental and biotechnological research interest, e.g. to target the activities of deleterious quorum-sensing bacteria. AHL signal analogues and inactivating enzymes have been identified. The red marine alga Delisea pulchra has been shown to produce halogenated furanones that are structurally related to homoserine lactones (Givskov et al., 1996) and, by binding to and increasing the turnover rate of a LuxR homologue, disrupt AHL-mediated swarming motility and surface colonization by the pathogenic bacterium Serratia liquefaciens (Manefield et al., 1999; Rasmussen et al., 2000; Manefield et al., 2002). Several varieties of pea and a number of other higher plants have been reported to produce compounds of unknown chemical structure that can activate some, but inhibit several other, quorum-sensing pathways (Teplitski et al., 2000). Chloroperoxidases produced by marine algae have been demonstrated to inactivate 3-oxoacylhomoserine lactones, but not other AHLs (Borchardt et al., 2001).
We have reported previously on AHL-inactivating enzymes encoded by Gram-positive Bacillus species (Dong et al., 2000; 2002) and by a Gram-negative bacterium, Agrobacterium tumefaciens (Zhang et al., 2002). The enzymes involved in these bacteria have been characterized as AHL-lactonases, which serve to inactivate AHL signals by hydrolysing the lactone ring, yielding the corresponding acyl-homoserine (Dong et al., 2001; Zhang et al., 2002). Transgenic plants expressing AHL-lactonase exhibited significantly enhanced resistance to Erwinia carotovora infection (Dong et al., 2001). The expression of AHL-lactonase in E. carotovora and Pseudomonas aeruginosa has also been shown to quench quorum sensing by these species (Dong et al., 2000; Reimmann et al., 2002).
Another mechanism by which AHLs are degraded by bacteria has been reported (Leadbetter and Greenberg, 2000; Leadbetter, 2001). An isolate of the β-Proteobacterium Variovorax paradoxus was shown to be capable of using AHL signals as energy sources. When doing so, homoserine lactone was released into the medium as a major degradation product, whereas the fatty acid was metabolized as energy sources. However, little else is known about the genetic and enzymological details underpinning this AHL-acylase activity. Here, we report the cloning and expression of a gene encoding an HSL-releasing, AHL-acylase from another β-Proteobacterium, a Ralstonia isolate obtained from a mixed-species biofilm.
Isolation of AHL-inactivating strains from a biofilm sample
A bacterial biofilm was collected from an experimental water treatment system and found to inactivate AHL signals. Sixteen bacterial isolates of distinct colony morphology were obtained from the biofilm. Among them, several AHL-producing bacterial species, including a strain of P. aeruginosa, were also identified. Two of these isolates, strains XJ12B and XJ12A, were found to inactivate N-β-oxooctanoyl-l-homoserine lactone (3OC8HSL). Neither was observed to accumulate detectable levels of AHL signal molecules when grown either in liquid or on agar-solidified media, as determined using common bioassay methods. Analyses of their SSU rRNA-encoding genes indicate that strains XJ12A and XJ12B were both members of the β-Proteobacteria, sharing 97% and 96% identity, respectively, with the rRNA-encoding genes from strains of Ralstonia eutropha. Strain XJ12B exhibited a more pronounced signal-inactivating activity than XJ12A and was thus selected for further study. Strain XJ12B was capable of using both of two AHLs tested as energy sources for growth. In ammonium-replete, ‘MES 5.5’ medium incubated at 37°C, the isolate grew with a doubling time of 8.5 h and 10.5 h, respectively, in media containing either 3OC12HSL or C4HSL as sole energy source (data not plotted). The growth yields achieved with these two acyl-HSLs and with succinate were virtually identical to each other when normalized and compared as a function of their yields per mol carbon utilized (as opposed to yields per mol substrate). AHL-inactivating activity was found to be located primarily with cell debris after lysis, with little found associated with the cytosolic fraction or secreted into the growth culture fluid, suggesting that the relevant enzyme may be membrane or periplasm associated.
Cloning of the Ralstonia AHL inactivation gene
To identify the gene encoding AHL inactivation, a cosmid library was constructed using genomic DNA from Ralstonia XJ12B expressed in Escherichia coli. Sixteen hundred clones were screened for AHL-inactivating activity using 3OC8HSL as substrate. A single clone, p13H10, exhibited AHL inactivation. Subcloning localized the AHL inactivation gene to the 4 kb insert of plasmid p2B10, which was subsequently completely sequenced. Sequence analysis revealed a 2385 nucleotide open reading frame (ORF), which we called aiiD. A putative ribosome binding site, AGGAGA, was identified 6 bp upstream of the ATG start codon of aiiD. The peptide encoded by aiiD was predicted to be 794 amino acids, with a predicted molecular mass of 85 373 Da and, because of its 78 basic and 78 acidic residues, an isoelectric point at 7.48. The predicted peptide contains 301 hydrophobic and 174 polar amino acid residues.
Similarity of AiiD, the Ralstonia AHL inactivation enzyme, to other proteins
Among enzymes with a demonstrated function, AiiD was most similar to the aculeacin A acylase (AAC) from the high-GC Gram-positive organism, Actinoplanes utahensis (Takeshima et al., 1989; Inokoshi et al., 1992), sharing 40% identity at the peptide level. AiiD also shared 22–24% identity with several cephalosporin and penicillin acylases, including glutaryl-7-amonicephalosporanic acid acylase, 7-β-(4-carboxybutanamido) cephalosporanic acid acylase, and penicillin G acylase (Matsuda and Komatsu, 1985; Schumacher et al., 1986; Matsuda et al., 1987; Oh et al., 1987; Kim et al., 1999; Li et al., 1999; Lee et al., 2000). AiiD also shared significant similarity with deduced proteins of undemonstrated function from the published and unpublished genomes of diverse bacteria. At the amino acid level, AiiD shared 83% and 69% identity with acylase homologues found in the genomes of Ralstonia solanacearum and Ralstonia metallidurans respectively; 52% identity with a homologue from Deinococcus radiodurans; and 38–40% identities with homologues encoded by Pseudomonas aeruginosa, P. putida, P. syringae and P. fluorescens.
Further in silico analysis of aiiD predicted features similar to those of demonstrated acylases, which are very often post-translationally modified into two polypeptide subunits after the cleavage of signal and spacer peptides (Duggleby et al., 1995; Li et al., 1999; Hewitt et al., 2000). The N-terminus of AiiD encompasses a stretch of hydrophobic amino acids predicted to encode a signal peptide. The deduced propeptide of AiiD was aligned with the propeptides of several other known acylases, noting the expected or known locations of the signal sequence, α-subunit, spacer sequence and β-subunit regions (Fig. 1). Prominent at the amino-terminus of the presented alignment was a well-conserved glycine–serine pair. These residues are known to be essential to the post-translational modification of the acylase propeptide into two enzymatically active subunits (Duggleby et al., 1995; Li et al., 1999; Kim and Kim, 2001), and thus were targeted for site-directed mutagenesis in AiiD. Four variants of AiiD containing cysteine, glycine or threonine individual substitutions at these positions did not catalyse the inactivation of AHL signals, whereas expression of their propeptides was not significantly affected (Fig. 2A and B).
AiiD is an AHL-acylase
The aforementioned similarities between AiiD and known acylases suggested that the former might inactivate AHLs by functioning as an HSL-releasing AHL-acylase. To examine this possibility further, the predicted coding region of aiiD, including the putative signal peptide, was amplified via polymerase chain reaction (PCR) and cloned into a glutathione S-transferase (GST) fusion vector. Three different acyl-HSLs (3OC8HSL, 3OC10HSL and 3OC12HSL) were incubated with a purified preparation of AiiD. Within 3 h, all three signals became completely inactivated as determined by AHL bioassays. The enzyme exhibited demonstrable, but significantly less, activity on a short-chain AHL, e.g. 3OC6HSL (data not presented). Attempts to construct GST–AiiD fusion proteins devoid of the signal peptide failed to retrieve active enzyme, but this is not unusual. It was reported that the precursor polypeptide of penicillin G acylase lacking the signal peptide sequence was not processed to a mature enzyme and did not display enzyme activity (Schumacher et al., 1986). The signal peptide may be a prerequisite for enzyme processing and activity. Alternatively, accurate processing might only occur during or after secretion events, and not within the cytoplasm.
N- (3-oxodecanoyl)- l -homoserine lactone (3OC10HSL) was digested with purified AiiD. The reaction products were analysed by reverse phase high-performance liquid chromatography (RP-HPLC) and electrospray ionization mass spectrometry (ESI-MS). Before digestion, 3OC10HSL displayed a single peak with an HPLC retention time of 14.5 min and an M-H ion at an m/z (mass-to-charge ratio) of 268.1 ( Fig. 3A and B ). After incubation, a product emerged with a retention time of 3.1 min ( Fig. 3A ). ESI-MS analysis of the 3.1 min HPLC fraction showed a, M-H ion at an m/z of 185. This molecular weight is identical to that of 3-oxodecanoic acid ( Fig. 3B and E ). Also evident in the ESI-MS analysis was a daughter ion at an m/z of 141.0 ( Fig. 3B ). This corresponded to the expected decarboxylation product (M-H-CO 2 ) of the 185 ion, 3-oxodecanoic acid. This was subsequently confirmed by MS/MS analysis.
Homoserine lactone (HSL) standards were not resolved from the buffer fraction with the HPLC regime used in this study. In order to determine whether HSL was released as an AHL degradation product, samples of the digestion mixture were reacted with 5-dimethylamino-1-naphthalensesulphonyl chloride (DANSYL chloride). Dansylation increases the hydrophobicity of amino acids, making them more readily resolved during chromatographic separations. Fractionation of the dansylated digestion mixture revealed an HPLC peak with a retention time of 2.8 min, identical to that of a dansylated HSL standard (Fig. 3C). DANSYL chloride did not react with 3OC10HSL and showed a retention time of 2.1 min. The UV spectra of DANSYL chloride and dansylated HSL fractions are shown in Fig. 3D. The absorbance maxima for DANSYL chloride and its dansylated HSL derivative are 215 nm and 265 nm respectively. The absorbance maximum of the major dansylated product from the digestion mixture was identical to that of dansylated HSL. Taken together, these data demonstrate that AiiD is an AHL-acylase, i.e. hydrolyses the AHL amide bond, releasing HSL and 3-oxodecanoic acid into the reaction mixture (Fig. 3E).
AHL-acylase does not degrade penicillin, and diverse acylases do not degrade AHLs
Because AiiD is reasonably related to known penicillin acylase, we examined whether it might catalyse the degradation of the corresponding antibiotics. AHL-acylase did not release DANSYL-reactive material from penicillin G and ampicillin. Likewise, commercial preparations of porcine kidney acylase and penicillin acylase did not catalyse the release of fatty acid and HSL from AHLs (data not shown).
Expression of aiiD in P. aeruginosa PAO1 influenced AHL accumulation, extracellular product production and swarming motility
To examine the influence that AiiD might exert on quorum sensing by P. aeruginosa, aiiD was cloned into pUCP19 and transformed into P. aeruginosa strain PAO1. Transformants containing pUCaiiD failed to accumulate detectable levels of either 3OC12HSL or C4HSL (the two major AHLs produced by this species; data not shown for the butanoyl-HSL); growth was not otherwise affected in any obvious manner (Fig. 4A and B). In contrast, wild-type strain PAO1 and pUCP19-containing PAO1 controls showed the expected population density-dependent accumulation of 3OC12HSL (Fig. 4B). Because the expression of the AHL-acylase in P. aeruginosa appeared to preclude the accumulation of AHLs, we examined the effects of this on several traits known to be controlled by quorum sensing. The production of the virulence factors elastase (Fig. 4C) and pyocyanin (Fig. 4D) were dramatically reduced in PAO1 expressing the AHL-acylase. Although the expression of AHL-acylase did not completely impair the ability of this bacterium to exhibit quorum-regulated swarming motility (Köhler et al., 2000), this trait was significantly reduced by the activity of the AHL-acylase (Fig. 5).
Pseudomonas aeruginosa PAO1 expressing the aiiD gene is markedly attenuated in nematode killing
Because quorum sensing has been shown to regulate the cyanide-elicited paralysis and killing of Caenorhabditis elegans by P. aeruginosa (Darby et al., 1999; Gallagher and Manoil, 2001), we examined the impact of AHL-acylase expression on this trait. As expected, nearly all worms transferred to a bacterial lawn of wild-type strain PAO1 and PAO1(pUCP19) controls were killed within 4 h (Fig. 6). In contrast, over 80% of the worms survived the same period of exposure to a lawn of strain PAO1(pUCaiiD). Moreover, the remaining worms survived and were even observed to feed upon cells of PAO1(pUCaiiD) as a food source.
We have cloned and expressed a gene from a Ralstonia isolate that encodes a potent AHL-acylase. The gene, aiiD, thus encodes an important form of quorum-sensing signal degradation first reported in the bacterium Variovorax paradoxus (Leadbetter and Greenberg, 2000). Before this study, little else was known about the details of AHL-acylase activity at the molecular level. We have shown that the Ralstonia gene product, AiiD, cleaves the AHL amide, thereby releasing the corresponding fatty acid and free HSL. The degradation products do not exhibit any apparent reactivity with quorum-sensing bioassays; thus, AiiD serves potently to inactivate AHL signals. The latter is not an insignificant finding, as other AHL degradation products, specifically acyl-homoserines (the products of the chemical decomposition of AHLs under alkaline conditions and biological decomposition by bacterial AHL-lactonases), have been reported to hold some residual signalling activity (Zhu et al., 1998; Dong et al., 2001).
Among the identified homologues of AiiD, the best characterized are the cephalosporin and penicillin acylases, which are members of the Ntn-hydrolase superfamily. Ntn-hydrolases are known to undergo post-translational processing resulting in a primary propeptide being cleaved into an active, two-subunit form (Brannigan et al., 1995; Oinonen and Rouvinen, 2000). The predicted AiiD polypeptide shares many of the known hallmarks for such modifications, including a readily identifiable signal peptide followed by an α-subunit, a spacer sequence and a β-subunit (Duggleby et al., 1995; Kim et al., 2000). When aligned with other acylases (Fig. 1), AiiD shares well-conserved glycine, serine, asparagine, histidine, valine and tyrosine residues that have been demonstrated to be of importance to both autoproteolytic processing and catalysis (Fig. 1, boxed residues; Duggleby et al., 1995; Kim et al., 2000; Kim and Kim, 2001; McVey et al., 2001). Our site-directed mutagenesis of AiiD confirmed that the conserved glycine and serine pair is essential for AHL-acylase activity. Thus, the peptide encoded by aiiD probably shares the post-translational modification and other conserved traits of most Ntn-hydrolases. The observation that the activity of AiiD seemed to co-fractionate with lysed cell debris is consistent with it possibly being membrane anchored, perhaps within the periplasm. This location is not inconsistent with the demonstration that HSL was either released or exported into the culture medium during the degradation of AHLs by V. paradoxus (Leadbetter and Greenberg, 2000). This localization of HSL-releasing activity, i.e. external to the cytoplasm, possibly helps the cell to avert HSL toxicosis. HSL is known to be deleterious to the health of both bacterial and mammalian cells (Jakubowski, 1997; 2000; Zakataeva et al., 1999).
We cannot yet say for certain that AiiD was evolved to degrade AHLs as its natural substrates, but we have many reasons to believe that this might be the case. Acylases as a group are capable of degrading an enormous diversity of compounds but, individually, they tend to show a high degree of substrate specificity (Kim et al., 2000). For example, penicillin G acylase hydrolyses benzylpenicillin to form phenylacetic acid and 6-aminopenicillanic acid (Duggleby et al., 1995). The glutaryl acylase from Pseudomonas sp. 130 (CA-130) is highly active on glutaryl 7-amino cephalosporanic acid, but shows surprisingly little activity on penicillins (Li et al., 1999). We note that, despite numerous attempts to modify the substrate specificities of penicillin G and glutaryl acylases for industrial purposes, to our knowledge, none has been reported as being successful (Fritz-Wolf et al., 2002 and references therein). Indeed, we have examined whether active AiiD preparations might degrade a commercially available substrate for one of these different acylases, i.e. penicillin G. It did not. Nor did commercial acylase preparations such as porcine kidney acylase and penicillin acylase detectably degrade AHLs when examined. The acylase homologue of demonstrated function that is most similar to AiiD (i.e. the acylase from A. utahensis) catalyses the degradation of the fungal antibiotic aculeacin A, yielding a hexapeptide and a long-chain fatty acid (Takeshima et al., 1989). However, its Km for this substrate was reported to be in the millimolar range, and little else is known about its substrate specificity. In comparison, AiiD was able not only to degrade AHLs, but to maintain them at low, nanomolar concentrations during their concomitant synthesis by P. aeruginosa; clearly, the AHL-acylase operates at physiologically relevant rates and AHL substrate concentrations (Fig. 2A and 4B). We also note that two residue positions that have generally been implicated in the substrate specificity of acylases (Kim et al., 2000), Leu-50β and Glu-57β in the A. utahensis enzyme, manifest themselves as Ile-50β and Ser-57β in the Ralstonia AHL-acylase (Fig. 1; shaded residues). Certainly, more extensive studies on the substrate specificity and affinity of the enzyme encoded by aiiD will help to resolve this issue, but the early indications are strong that it has probably evolved to serve as an AHL-acylase.
To test the effects of AHL-acylase expression on quorum-controlled virulence, the aiiD gene was transformed into P. aeruginosa PAO1. P. aeruginosa is an opportunistic pathogen that causes corneal, lung and burn wound infections. Two AHL quorum-sensing signals, i.e. N-(3-oxo-dodecanoyl)-l-homoserine lactone (3OC12HSL) and N-butanoyl homoserine lactone (C4HSL), play a key role in regulating the production of virulence factors, such as proteases, phospholipases, phenazine, rhamnolipid and cyanide (de Kievit and Iglewski, 2000). The same AHL signalling systems are also involved in the regulation of surface translocation and biofilm differentiation in P. aeruginosa (Davies et al., 1998; Köhler et al., 2000). Mutants defective in AHL signalling have been shown to be attenuated in virulence when tested on C. elegans and a mouse model of pneumonia (Tang et al., 1996; Darby et al., 1999). Here, the expression of the AHL-acylase encoded by the aiiD gene from Ralstonia in P. aeruginosa significantly reduced the accumulation of 3OC12HSL and C4HSL in the growth medium (Fig. 4B). This, in turn, resulted in the decreased production of quorum-controlled virulence factors, reduced cell surface translocation such as swarming motility and attenuated virulence on C. elegans (Figs 4C and D, 5 and 6). These results parallel previous reports that Bacillus-derived AHL-lactonases quench quorum-sensing signalling and attenuate virulence when expressed in E. carotovora, a plant pathogen causing soft rot disease in many plants, and in P. aeruginosa (Dong et al., 2000; 2001; Reimmann et al., 2002). Our results add to the already formidable evidence that AHL-mediated quorum-sensing systems are important to the regulation of virulence gene expression and pathogenesis of P. aeruginosa. The expression of AHL-acylases in quorum-sensing bacteria also provides another promising tool for exploring the details of the global control of gene expression mediated by AHL signalling.
It is now clear that AHL degradation enzymes can be of diverse mechanisms and are found widely distributed across many species. Following the discovery of the AHL-lactonase encoded by aiiA from a Gram-positive Bacillus isolate, many homologues of this AHL-lactonase have now been identified and expressed from numerous species closely related to it, as well as from A. tumefaciens, in which it has been shown to function as an essential component of an AHL signal turnover system (Dong et al., 2000; 2002; Lee et al., 2002; Reimmann et al., 2002; Zhang et al., 2002). We can speculate that AHL-acylases are also widely distributed. Certainly, this activity has already been shown to be in play during the utilization of AHLs as nutrients by V. paradoxus (Leadbetter and Greenberg, 2000). The physiological role of AHL-acylase in Ralstonia strains JX12B remains unclear. An obvious possibility is that the acylase may play a role during oligotrophic nutrient scavenging from the environment. Indeed, strain XJ12B proved capable of using two distinct AHLs, 3OC12HSL and C4HSL, as energy sources. Its ability to use them as sources of cellular N was not examined. It may also turn out to be the case that, in parallel to AHL-lactonase-mediated signal decay in A. tumefaciens, aiiD homologues could also serve as integral, modulating components of naturally occurring quorum-sensing circuits. Indeed, several homologues of aiiD can be identified in the genome sequences of AHL-producing, quorum-sensing species such as R. solanacearum and P. aeruginosa (Stover et al., 2000; Salanoubat et al., 2002). The potential roles for this class of AHL degradation enzymes in (i) the signal tuning of intra- and interspecies communications and (ii) other relationships occurring between quorum-sensing species and their competitors appears to be an area well worth investigating in the future.
Bacterial strains and growth conditions
Ralstonia spp. JX12A and JX12B were isolated from a biofilm from an experimental water treatment system (The National University of Singapore). E. coli strain DH5α was used as a host for DNA manipulations. A. tumefaciens strain NT1 ( traR ; tra::lacZ749 ) was used as a biosensor for AHL activity in the bioassay ( Piper et al., 1993 ). Chromobacterium violaceum strain CV026 was also used to detect N -butanoyl homoserine lactone (C4HSL) ( McClean et al., 1997 ). Ralstonia and E. coli were cultured in LB medium at 37°C; A. tumefaciens was cultured at 28°C in MM medium ( Zhang et al., 1993 ). For the examination of the growth of Ralstonia on either 1 mM 3OC12HSL or 2 mM C4HSL as sole energy source, ammonium-replete ‘MES 5.5’ medium and other previously described methods were used ( Leadbetter and Greenberg, 2000 ). Appropriate antibiotics were added when necessary at the following concentrations: ampicillin, 100 µg ml −1 ; tetracycline, 10 µg ml −1 (100 µg ml −1 for selection of PAO1 transformants); kanamycin, 50 µg ml −1 ; and carbenicillin, 200 µg ml −1 .
Isolation of bacteria and AHL inhibition bioassay
A bacterial biofilm sample was suspended in sterilized water with shaking for 1 h before spreading onto YEB agar plates (Dong et al., 2000). Individual colonies were restreaked to ensure the purity of isolates. Bacterial isolates were cultured in LB medium in 96-well plates at 28°C with shaking overnight. An aliquot of culture was then mixed with an equal volume of fresh medium containing 20 µM 3OC8HSL. The reaction mixture was incubated at 28°C for 4–5 h, followed by 30 min sterilization under UV light. AHL activity was assayed as described previously (Dong et al., 2000).
Cloning and sequencing of aiiD and SSU rRNA-encoding genes
Genomic DNA from strain Ralstonia sp. XJ12B was partially digested with Sau3A. DNA fragments were ligated into the dephosphorylated BamHI site of cosmid vector pLAFR3 (Staskawicz et al., 1987). For subcloning, cosmid DNA of clone p13H10 encoding AHL inactivation activity was partially digested with Sau3A, and DNA fragments were cloned into the BamHI site of sequencing vector pGEM-3Zf(+). Cosmid and plasmid clones encoding AHL inactivation activity were identified using the bioassay method described by Dong et al. (2000). Site-directed mutagenesis of aiiD was performed using a QuikChange site-directed mutagenesis kit (Stratagene), according to the manufacturer's protocol. The TGSTM template generation system F-700 (Finnzymes) was used for inserting Mu transposon into plasmid DNA to provide primer binding sites for DNA sequencing. Plasmid clone p2B10, which contained the aiiD gene, was used as a template. Mutants that were unable to inactivate AHL were identified, and plasmids were subsequently purified for sequencing using primers supplied in the kit. A 1.3 kb fragment of 16S rDNA was amplified from genomic DNA of strain XJ12B by PCR with the forward primer 5′-TGACGAGTGGCGGACG GGTG and the reverse primer 5′-CCATGGTGTGACG GGCGGTGTG. The primer pair was designed based on the conserved bacterial 16S rDNA sequences. Sequencing was performed using an ABI Prism dRhodamine terminator cycle sequencing ready reaction kit (Perkin-Elmer Applied Biosystems). The nucleic acid sequence for the cloned aiiD has been submitted to GenBank (locus bankit477480).
Preparation of constructs for AiiD expression and enzyme purification
The coding region of the aiiD gene was amplified by PCR using a forward primer 5′-CGTGGATCCATGATGCAGG GATTCGCGCTGCGC-3′ and a reverse primer 5′-CGCG AATTCACCGGCAGCCCTCACTGCGACAAC-3′ containing BamHI and EcoRI restriction sites respectively. The PCR product was digested by restriction enzymes and fused in frame to the GST gene under the control of an IPTG-inducible lac promoter in vector pGEX-2T (Amersham Pharmacia) to generate the construct pGST-aiiD. Expression of pGST-aiiD in E. coli and the purification of AiiD, the fusion protein, was based on a previously described method (Zhang et al., 1998). Recombinant AiiD was released from the bound GST protein using thrombin, a site-specific protease. Purity was determined by SDS-PAGE analysis. To express AiiD in P. aeruginosa, a BamHI–EcoRI fragment containing the aiiD gene was released from the pGST-aiiD construct and then ligated into the vector pUCP19 (ATCC 87110; Schweizer, 1991) to generate plasmid pUCaiiD. Plasmid pUCaiiD and cosmid p13H10 were introduced into P. aeruginosa PAO1 via triparental mating with helper strain RK2013. The putative signal peptide of aiiD was identified using a web-based service (http:www.cbs.dtu.dkservices).
Products of AiiD-catalysed AHL degradation
AHL signal digestion was carried out in 1 ml of 1× PBS buffer containing 3 mM 3OC10HSL and 40 µg of purified AiiD. Incubation was at 30°C for 3 h with gentle shaking. After incubation, the digestion mixture was extracted three times with equal volumes of ethyl acetate; thereafter, the combined organic phase was evaporated to dryness in a rotary evaporator. For HPLC analysis, the sample was redissolved in 0.2 ml of methanol and introduced onto a Waters SymmetryTM C18 reverse-phase column (4.6 × 250 mm). Fractions were eluted isocratically with 50:50 methanol–water (v/v) at a flow rate of 1 ml min−1. ESI-MS and tandem mass spectrometry were performed on a Finnigan/MAT LCQ ion-trap mass spectrometer. The sample dissolved in 50:50 methanol–water (v/v) was introduced into the mass spectrometer by loop injection. For DANSYL chloride derivatization, 100 µl of either the digestion mixture or a 2 mM homoserine lactone (HSL) standard was reacted with an equal volume of DANSYL chloride (2.5 mg ml−1 in acetone) at 40°C for 4 h. After evaporation to dryness, 50 µl of 0.2 M HCl was added to the sample to hydrolyse any excess DANSYL chloride. The sample was then diluted as necessary for HPLC fractionation under the same conditions detailed above. The action of AHL-lactonase on penicillin G and ampicillin was examined by incubating the active purified preparation with 2 mM respective antibiotics, followed by examination of the reaction mixture for the expected released products. The action of porcine kidney acylase I (Sigma) and penicillin acylase (Sigma) on AHLs was examined by dissolving each of the commercial preparations in 1× PBS and incubating them with 2.5 mM AHL, with monitoring for the release of HSL.
Assay for virulence factor production and swarming motility
Elastolytic activity in culture fluids was determined using Elastin Congo red (ECR) assays (Pearson et al., 1997). Briefly, P. aeruginosa cells were inoculated in fresh brain–heart infusion (BHI) medium and incubated at 37°C with shaking. At different time points, culture supernatants were sampled, filtered (0.45 µm pore-size filter) and stored at −70°C until analysis. Samples (50 µl) from each supernatant were combined in a tube with 1 ml of 100 mM Tris buffer (pH 7.2) containing 1 mM CaCl2 and 20 mg of ECR. The mixtures were incubated for 18 h at 37°C with rotation, after which 0.1 ml of 0.12 M EDTA was added, and they were placed on ice. Insoluble ECR was removed by centrifugation, and the OD495 of the clarified mixture was determined.
Pyocyanin production was determined according to the method of Essar et al. (1990) with minor modifications. Cell-free culture fluid (5 ml) was extracted with an equal volume of chloroform for 2 h. The chloroform was decanted and subsequently extracted with a one-fifth volume of 0.2 M HCl. The pyocyanin partitioned to the HCl aqueous phase, which was removed and quantified by measuring OD520.
Swarming motility of P. aeruginosa was assayed on semi-solid agarose (0.35% w/v; Bio-Rad Laboratories) containing 1% (w/v) and 0.5% (w/v) tryptone and NaCl respectively (Rashid and Kornberg, 2000).
Nematode paralysis assay
The assay procedure for C. elegans paralytic killing has been described (Darby et al., 1999; Gallagher and Manoil, 2001). In brief, a 2-day-old P. aeruginosa PAO1 colony was suspended in 150 µl of BHI broth at an optical density of 0.2 (OD600); alternatively, the overnight bacterial culture in BHI was diluted 100 times and spread onto a 3.5-cm-diameter plate containing ≈ 4 ml of BHI agar. The plates were incubated for 48 h at 37°C. Nematodes (strain N2) were collected in M9 buffer transferred onto the bacterial lawn. The plates were sealed and incubated at room temperature for at least 4 h. The numbers of nematodes experiencing paralytic killing were scored during microscope observations. E. coli strain OP50 was used as a control to evaluate background levels of worm death. At least five independent assays were carried out with worms of various ages. C. elegans was routinely maintained in NGM agar plates (Sulston and Hodgkin, 1988) containing E. coli strain OP50.
We thank A. Kerr and M. Tate for critical review of the manuscript. This work was supported by the Agency for Science and Technology and Research, Singapore (L.-H.Z.) and by the US Department of Agriculture, Soils and Soil Biology Program (no. 2001-01242; J.R.L.).