Identification and Characterization of a GDSL Esterase Gene Located Proximal to the swr Quorum-Sensing System of Serratia liquefaciens MG1

Abstract

Serratia liquefaciens MG1 employs the swr quorum-sensing system to control various functions, including production of extracellular enzymes and swarming motility. Here we report the sequencing of the swr flanking DNA regions. We identified a gene upstream of swrR and transcribed in the same direction, designated estA, which encodes an esterase that belongs to family II of lipolytic enzymes. EstA was heterologously expressed in Escherichia coli, and the substrate specificity of the enzyme was determined in crude extracts. With the aid of zymograms visualizing EstA on polyacrylamide gels and by the analysis of a transcriptional fusion of the estA promoter to the promoterless luxAB genes, we showed that expression of the esterase is not regulated by the swr quorum-sensing system. An estA mutant was generated and was found to exhibit growth defects on minimal medium containing Tween 20 or Tween 80 as the sole carbon source. Moreover, we show that the mutant produces greatly reduced amounts of N-acyl-homoserine lactone (AHL) signal molecules on Tween-containing medium compared with the wild type, suggesting that under certain growth conditions EstA may be important for providing the cell with precursors required for AHL biosynthesis.

Serratia liquefaciens MG1 was originally isolated from a softened cucumber (21), suggesting that it is capable of causing rot of vegetables. However, more recent studies have shown that the strain is a good colonizer of tomato roots (42) and is unable to soften cucumber or other vegetables (unpublished results). However, like most Serratia species, strain MG1 secretes a broad spectrum of hydrolytic enzymes, including two proteases, a lipase, a phospholipase, a nuclease, and several chitinases (19, 33). Despite the fact that all of these extracellular enzymes are expressed in a similar growth phase-dependent manner, there is no common underlying regulatory mechanism. Expression and secretion of the phospholipase are tightly coupled to the synthesis and export of flagella via the flhDC operon, which encodes a master regulator that controls expression of flagellin, chemotaxis, and motility genes (20). Expression of nucleolytic activity is induced by the lexA-dependent SOS system (3, 19). Production of extracellular proteolytic and chitinolytic activity is, at least in part, regulated by the swr quorum-sensing system (reference 14 and unpublished results). This cell-cell communication system relies on two proteins: SwrI, which directs the synthesis of the diffusible signal molecules N-butanoyl-l-homoserine lactone (C₄-HSL) and N-hexanoyl-l-homoserine lactone (C₆-HSL) in a molar ratio of 10 to 1, and SwrR, which, after binding of the signal molecules, is thought to activate or repress transcription of target genes (14, 15). A global analysis by two-dimensional polyacrylamide gel electrophoresis (PAGE) showed that at least 28 proteins are under control of the swr regulatory system (18). One of the swr-regulated genes is lipB, which encodes a component of the Lip exporter (40). This type I protein secretion system is responsible for the transport of the S. liquefaciens MG1 lipase, metalloprotease, and S-layer protein. As a consequence of the involvement of the swr system in the regulation of lipB expression, the levels of extracellular metalloprotease and S-layer protein are significantly lowered in an swrI mutant.

In an attempt to identify further quorum-sensing-regulated functions in S. liquefaciens MG1, we sequenced the swr flanking DNA regions, as target genes of quorum-sensing circuits are often located in the vicinity of the respective regulatory genes (48). We show that a GDSL-type esterase, designated EstA, is encoded by a gene located upstream of swrR. A zymographic assay that allows the visualization of EstA on polyacrylamide gels was developed and was used to show that expression of the esterase is independent of the swr quorum-sensing system. Employing a gene replacement technique, we generated a defined estA mutant. This mutant grew more slowly on Tween-containing minimal medium and, most interestingly, was also defective in the production of N-acyl-homoserine lactones (AHLs) on this medium. These results suggest that the physiological roles of the esterase are (i) to provide access to certain carbon sources such as Tween and (ii) to supply the cell with precursors required for AHL biosynthesis under these growth conditions.

MATERIALS AND METHODS

Organisms and growth conditions.

The strains used in this study are listed in Table 1. S. liquefaciens and Escherichia coli strains were routinely grown aerobically at 30 and 37°C, respectively, in modified Luria-Bertani (LB) medium (5) containing 4 g of NaCl liter⁻¹ instead of 10 g of NaCl liter⁻¹. The ability of S. liquefaciens strains to utilize different C sources was assessed in AB minimal medium (10) at the concentrations indicated in Table 2. Antibiotics were added for selection of plasmids, as required, at the following concentrations: ampicillin, 100 μg ml⁻¹; kanamycin, 50 μg ml⁻¹; and tetracycline, 10 μg ml⁻¹. Growth media for examination of swarming motility were LB and AB minimal media containing 0.25% Casamino Acids and either 0.4% glucose or 1% Tween 20. Plates were solidified with 0.7% agar. Exponential-phase cells were concentrated approximately 100-fold by centrifugation (10.000 × g for 2 min at 4°C) and then inoculated onto swarming plates. Following 12 h of incubation at 30°C, the plates were photographed.

TABLE 1.

Bacterial strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Relevant genotype and/or characteristicsa	Reference or source
Strains
S. liquefaciens
MG1	Wild type; Apr Tcr	21
MG44	swrI derivative of MG1; Apr Tcr Smr	14
MG3	flhDC derivative of MG1; Apr Tcr Smr	20
MG1-estA	estA derivative of MG1; Apr Tcr Kmr	This study
E. coli
JM109	F′ traD36 lacIq Δ(lacZ)M15 proA+B+/el4(McrA−) Δ(lac-proAB) thi gyrA96 (Nalr) endA1 hsdR17 (rK− mK+) relA1 supE44 recA1	52
MT102	F−thi araD139 ara-leuΔ7679 Δ(lacIOPZY) galU gal′K r− m+ Smr	Laboratory collection
BL21(DE3)	F−ompT hsdSB(rB− mB−) gal dcm (λIts857 ind1 Sam7 nin5 lacUV5-T7 gene1)	Invitrogen, San Diego, Calif.
Plasmids
p3403	pUC18 containing a 5-kb PstI fragment of chromosomal DNA of S. liquefaciens MG1; Apr	19
pCAROVIIIa	luxAB-based promoter probe vector; Apr Kmr	Laboratory collection
pGP704	Gene replacement vector; Apr	35
pGP704-Kmr	pGP704 with kanamycin resistance gene inserted into the XbaI- SacI sites of the multiple cloning site; Apr Kmr	This study
pMS470Δ8	tac promoter-based expression vector; Apr	4
pRK600	oriColE1 RK2-Mob+ RK2-Tra+; helper plasmid in triparental conjugations; Cmr	28
pTOM12	estA replacement vector derived from pGP704-Kmr; Apr Kmr	This study
pTO-estA	pMS470Δ8 containing estA; Apr	This study
pSL1	Promoter probe vector pCaroVIIIa with estA promoter as 0.52-kb DNA BamHI fragment inserted upstream of luxAB; Apr Kmr	This study
Primers
pestA-Sal	5′-TCTCAGGAGTCGACTCATGCC-3′
pestA-Xba	5′-CATATCGGATCTAGACGCCGAGG-3′
pestA-Sac	5′-CGATGAGAGCTCCGCCATGCG-3′
pestA-Eco	5′-GGGGGAATTCTATGTGAGGCG-3′
pestA-f	5′-CGACATATGCCTCTAAAAATAACGTGC-3′
pestA-r	5′-ATTAAGCTTCCGAGTCGTTCAAAAACGC-3′
pestA1	5′-GCTTGGATCCCTGCAGCAGGTTTTC-3′
pestA2	5′-GAAGGATCCCACTGCAAAGCAACGCC-3′

^aRestriction sites introduced by PCR primers are underlined.

TABLE 2.

Growth and AHL production of the S. liquefaciens wild-type MG1 and the estA mutant MG1-estA and activities of a P_estA transcriptional luxAB gene fusion on plasmid pSL1 in the wild-type background in minimal medium supplemented with different carbon sources^a

Carbon source	Growth rate (h−1)		Yield (OD600)b		C4-HSL production (%)c		Specific bioluminescenced (107) [MG1(pSL1)]
	MG1	MG1-estA	MG1	MG1-estA	MG1	MG1-estA
0.4% glucose	0.47	0.47	2.5	2.4	100	98	3.38 ± 0.04
0.4% fructose	0.34	0.34	2.3	2.3	97	84	NDe
0.4% saccharose	0.48	0.49	2.8	2.8	105	95	2.89 ± 0.06
0.4% citrate	0.34	0.33	1.6	1.6	88	86	ND
0.4% gluconate	0.32	0.32	1.6	1.6	86	92	ND
0.4% succinate	0.23	0.20	1.1	1.1	97	96	ND
1% glycerol	0.34	0.33	2.3	2.2	112	102	3.02 ± 0.05
1% Tween 20	0.26	0.19	1.8	1.3	77	18	2.13 ± 0.04
1% Tween 80	0.39	0.31	2.4	2.0	91	35	3.18 ± 0.06
1% olive oil	0.53	0.52	2.7	2.6	130	129	2.99 ± 0.04

^aAll values represent means from at least two independent experiments.

^bYield is defined as the maximum OD reached within 3 days of incubation.

^cSamples of spent culture supernatants were taken at an OD₆₀₀ of 1.0, and AHLs were quantified with the aid of the AHL monitor strain E. coli MT102(pSB403). The AHL concentration of the wild-type strain grown in minimal medium containing 0.4% glucose was arbitrarily set to 100%.

^dFollowing addition of 1 μl of n-decanal (Sigma) to 200 μl of sample, bioluminescence of MG1(pSL1) was quantified at various points during logarithmic growth in a Lumat LB 9506 luminometer (EG & Berthold, Bad Wildbad, Germany). Specific bioluminescence is defined as relative light units per OD₆₀₀ unit. Results are means and standard deviations for multiple samples taken during exponential growth, from at least two independent experiments.

^eND, not determined.

DNA manipulation and nucleotide sequencing.

All DNA manipulations, including purification, cloning, and electrophoresis, were performed by standard methods (41). The 5.12-kb PstI chromosomal S. liquefaciens MG1 DNA fragment containing the swr locus was sequenced by using the Thermo Sequenase fluorescence-labeled primer cycle sequencing kit (Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany) and a 4200 DNA sequencer (LI-COR Inc., Lincoln, Nebr.). Initial sequencing was performed with M13 universal sequencing primers, after which primers specifically designed from the sequence data obtained were used for successive reactions spanning the entire fragment length of both strands. Nucleotide sequence assembly and analysis were performed with the DNAMAN software (version 4.13; Lynnon BioSoft, Quebec, Canada). DNA sequences were compared to other sequences in GenBank by using the on-line BLAST search engine (1) at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). Multiple sequences were aligned with the Clustal W program (43).

Construction of a defined estA mutant.

A defined estA mutant was constructed by the gene replacement method described by Hoang et al. (24). First, two DNA fragments homologous to the 5′ and 3′ coding regions of estA were amplified by PCR with the primer pairs pestA-Sal/pestA-Xba and pestA-Sac/pestA-Eco (Table 1), yielding a 950-bp SalI-XbaI product and a 1,200-bp SacI-EcoRI product, respectively. The restriction sites introduced by the PCR primers were used to successively insert the two DNA fragments into the compatible sites of the gene replacement vector pGP704-Km^r (35). The resulting construct, which was designated pTOM12, was transferred to S. liquefaciens MG1 by triparental mating (9), and gene replacement mutants were selected on LB medium containing 100 μg of kanamycin ml⁻¹, 50 μg of tetracycline ml⁻¹, and 5% (wt/vol) sucrose. The genetic structure of the estA mutant, which was designated MG1-estA, was confirmed by Southern blotting with part of the estA gene as a probe.

Construction of a translational P_estA::luxAB fusion.

The estA promoter region was PCR amplified with the primer pair pestA1 and pestA2 (Table 1). Following restriction with BamHI, the 0.52-kb DNA fragment was inserted into the promoter probe vector pCaroVIIIa cut with the same enzyme. The plasmid containing the insert in the orientation placing the estA promoter upstream of the promoterless luxAB genes of the vector was chosen. Next, this construct, which was designated pSL1, was transferred to the S. liquefaciens wild-type strain MG1 and the swrI mutant MG44.

Expression of EstA in E. coli.

For high-level expression of the esterase in E. coli, the estA gene was PCR amplified with the primer pair pestAf and pestAr (Table 1). The resulting 2,200-bp PCR product was digested with NdeI and HindIII and inserted into the expression vector pMS470Δ8 (4) cut with the same enzymes, yielding plasmid pTO-estA. The construct was transformed into the expression strain E. coli BL21. For overproduction of EstA, cultures were grown at 30°C to early exponential phase (optical density at 600 nm [OD₆₀₀] of 0.3), and then IPTG (isopropyl-β-D-thiogalactopyranoside) was added to a final concentration of 0.25 mM. Following 4 h of incubation in the presence of the inducer, cells were harvested by centrifugation (10.000 × g for 10 min).

Cell fractioning.

Preparation of crude cell extracts and membrane fractions were performed as described by Carinato et al. (8). Briefly, 30-ml LB cultures of the various S. liquefaciens strains and E. coli BL21 harboring pTO-estA were grown for 16 h at 30°C. The cells were then harvested by centrifugation (8.000 × g for 20 min), washed twice with 50 mM Tris-HCl (pH 8.0) and resuspended in 1 ml of the same buffer. Cells were disrupted by sonication (3 times for 30 s each; Branson Sonifier B-12, microtip) and spun at 40.000 × g for 30 min. The supernatant (intracellular protein fraction) was kept at −80°C. Sodium dodecyl sulfate (SDS) soluble membrane extracts were prepared by resuspension of the cell debris pellet in 1 ml of 3% (wt/vol) SDS in 50 mM phosphate buffer (pH 8.0), incubation at 100°C for 10 min, and centrifugation at 40,000 × g for 45 min. The supernatant was used as SDS-soluble membrane extracts. For the preparation of Triton X-100 membrane fractions, the cell debris pellet was resuspended in 1 ml of 2% (vol/vol) Triton X-100 in 50 mM phosphate buffer (pH 8.0), incubated for 30 min at room temperature, and centrifuged at 40,000 × g for 45 min. The supernatant, designated Triton-soluble membrane extract, was used for all enzyme assays in this study.

Protein assays, SDS-PAGE, and zymograms.

The method of Bradford (7) was used for quantitative protein determination with bovine serum albumin as a standard. SDS-soluble membrane proteins and proteins from the crude extract were analyzed by SDS-PAGE (30). After electrophoresis, the gels were stained with Coomassie brilliant blue. For the analysis of the esterase pattern, an activity staining (zymogram) was performed. After the SDS-PAGE, the proteins were renaturated by incubating the gel in 25% (vol/vol) isopropanol-50 mM Tris-HCl (pH 7.5) for 20 min, followed by incubation for 12 h in 50 mM Tris-HCl (pH 7.5) at 4°C. Finally, the renaturated SDS-polyacrylamide gels were overlaid with the fluorescent substrate methylumbelliferyl-butyrate (MU-butyrate) (0.01 M in dimethylformamide) in order to visualize esterase activity.

Esterase activity assays.

(i) Triglycerides with different chain lengths incorporated into agar plates (7 g liter⁻¹) were used to test for substrate specificity as described by Kugimiya et al. (29). Enzyme solutions were spotted onto the surface of triglyceride agar plates, and halo formation was visually analyzed after incubation at 37°C for 1 to 5 days. Chloramphenicol (20 μg ml⁻¹) was added to the plates to prevent cell growth. (ii). Standard photometric assays using p- and o-nitrophenyl fatty acid esters (Sigma) as substrates were performed at room temperature in 100 mM Tris-HCl (pH 7.0) containing 4 mM substrate (dissolved in dimethyl sulfoxide). Absorption coefficients of 2.4166 ml μmol⁻¹ cm⁻¹ for o-nitrophenol and 9.5946 ml μmol⁻¹ cm⁻¹ for p-nitrophenol at 405 nm were used to calculate the activity. One unit is defined as the amount of enzyme catalyzing the appearance of 1 μmol of nitrophenol per min. (iii) Specificities of the esterase were also analyzed after fractionation of proteins by native PAGE (12). Esterase activity of protein bands on gels was detected by incubation of gels in a solution of 5 ml of 100 mM phosphate buffer (pH 7.0), 200 μl of fast blue B stock solution (2% [wt/vol] in H₂O) and 500 μl of substrate stock solution (α- or β-naphthyl ester at 1% [wt/vol] in acetone). In this assay, esterase activity is indicated by the formation of dark red bands. Activity staining with HPTS (1-hydroxypyrene-3,4,8-trisulfonic acid) esters was performed by incubating the native polyacrylamide gel in a solution of 5 ml of 1 M sodium phosphate buffer (pH 7.0) and 50 μl substrate stock solution (HPTS ester in dimethyl sulfoxide) (2 mg ml⁻¹) for 5 min. After illumination with UV light, esterase activity gives rise to green fluorescent bands.

Quantification of AHL production.

AHL concentrations in spent culture supernatants of S. liquefaciens MG1 and the estA mutant were quantified with the aid of the AHL sensor strain E. coli MT102 harboring the plasmid pSB403 (51). This plasmid contains the Photobacterium fischeri LuxR together with a transcriptional fusion of the luxI promoter region to the promoterless bioluminescence gene cluster luxCDABE of Photorhabdus luminescens. The sensor strain responds specifically to the presence of AHLs with the emission of light. For routine analyses, strains were grown in minimal medium supplemented with different C sources to an OD₆₀₀ of exactly 1.0. One hundred microliters of filter-sterilized culture supernatants were incubated with 100 μl of exponential-phase culture of the AHL monitor strain E. coli(pSB403) or with a 200 nM concentration of the respective signal molecule 3-oxo-C₆-HSL in a microtiter plate. Following 2 h of incubation at 30°C, bioluminescence was measured with a Lambda Fluoro 320 Plus Reader (MWG Biotech, Ebersberg, Germany).

Nucleotide sequence accession number.

The nucleotide sequence of the entire 5.12-kb chromosomal DNA fragment of S. liquefaciens MG1 containing swrI, swrR, estA, and flanking regions has been deposited in GenBank under accession number AY168877.

RESULTS

We have previously cloned the C₄-HSL synthase gene of S. liquefaciens MG1, swrI, on an approximately 5-kb chromosomal PstI fragment (14). As quorum sensing-controlled genes are often clustered with their regulatory genes, we determined the sequence of the entire PstI fragment. Three complete and two partial open reading frames (ORFs) were identified on the DNA fragment (Fig. 1). The two genes coding for the LuxI and LuxR homologues SwrI and SwrR are divergently arranged (Fig. 1), as has been reported for those of Serratia marcescens SS-1 and Serratia sp. strain ATCC 39006 (25, 44). The third complete ORF, which is located upstream of swrR and is transcribed in the same orientation, spans 1,965 bases and is predicted to encode a protein that is highly homologous to members of family II of lipolytic enzymes. The protein, which was designated EstA, has 48% identity with ApeE from Salmonella enterica serovar Typhimurium (8) (Fig. 2), 42% identity with Lip1 from Photorhabdus luminescens (47), and 24% identity with EstA from Pseudomonas aeruginosa (50). The presence of a gene encoding a lipolytic enzyme upstream of swrR appears to be unique for S. liquefaciens MG1, as it is not found adjacent to the genes for the LuxR homologues of SS-1 or ATCC 39006. The N-terminal sequence (positions 1 to 28) of EstA strongly resembles a signal peptide, with an Ala-X-Ala motif at amino acids 25 to 28 that may serve as a signal peptidase cleavage site. On this basis the mature esterase has a calculated molecular mass of 69.35 kDa, which is consistent with the molecular mass determined by SDS-PAGE (see below).

FIG. 1.

Genetic organization of the swr gene region present on the 5.12-kb chromosomal PstI fragment of S. liquefaciens MG1. Genes are represented by arrows indicating the direction of transcription. tag, DNA-3-methyladenine glycosylase gene; estA, outer membrane esterase gene; swrR LuxR-type transcriptional regulator gene; swrI, AHL synthetase gene; glyS, glycyl-tRNA synthetase alpha subunit gene.

FIG. 2.

Multiple protein sequence alignment of members of the GDSL family of lipolytic enzymes. Shown on grey background are identical amino acids; the asterisks indicate the putative catalytic triad residues, the GDSLS consensus motif is underlined, and blocks of sequence conserved in the GSDL esterase family (2) are boxed. In the C-terminal half boxes indicate putative amphiphatic β-barrels, sufficient to traverse the hydrophobic core of the outer membrane (predicted by the computer program PSIPRED http://bioinf.cs.ucl.ac.uk/psipred/). The two amino acids that are fully conserved in all members of the autotransporter family (Pro 526 and Gly 622) are shown in boldface. SL EstA, outer membrane esterase from S. liquefaciens (accession no. AY168877); ST ApeE, outer membrane esterase from S. enterica serovar Typhimurium (accession no. AF047014) (8); PL, Lip1, a secreted esterase from P. luminescens (accession no. X66379) (46).

Upstream of swrI we identified a partial ORF, designated glyS, the translation product of which shows strong homology to various glycine tRNA synthetase alpha subunits. The best identity was 76% to GlyS of E. coli (6). The partial ORF upstream of estA, tentatively designated tag, showed strong homology to DNA-3-methyladenine glycosylases, with the highest homology (74% identity) being to Tag of Haemophilus influenzae (17).

Cellular localization of EstA.

To characterize EstA in better detail, we decided to overexpress the protein in E. coli. To this end we constructed plasmid pTO-estA, which contains estA under control of the strong IPTG-inducible tac promoter of plasmid pMS470Δ8 (Fig. 3A). Following induction of E. coli BL21(pTO-estA) with 2.5 mM IPTG for 4 h, the cells were harvested and fractionated. Measurements of esterase activities with p-nitrophenyl butyrate as a substrate revealed that virtually all of the enzymatic activity is present in the membrane fraction (Fig. 3B). The activity of the intracellular protein fraction amounted to only 20% of that of the membrane fraction, and no significant activity could be detected in the culture supernatant. When membrane fractions were analyzed by SDS-PAGE, a strong band of approximately 69 kDa, corresponding very well with the calculated mass of EstA, was observed (Fig. 3C, lane 3). The amount of this protein was greatly reduced in the crude extract, and it was completely missing in the culture supernatant (Fig. 3C, lane 2, and data not shown). We also overlaid renaturated SDS-polyacrylamide gels with the fluorescent substrate MU-butyrate in order to visualize the activity of the esterase (Fig. 3C, lanes 4 and 5). This zymographic analysis identified the 69-kDa protein in the membrane fraction as a lipolytic enzyme. In conclusion, these data provide strong evidence that EstA of S. liquefaciens MG1 is a membrane-bound enzyme.

FIG. 3.

Expression of estA in E. coli BL21. (A) Schematic map of the expression cassette present on pTO-estA. (B) Distribution of esterase activity in cellular compartments of E. coli BL21 containing pTO-estA. Bars: 1, culture supernatant; 2, Triton X-100-soluble membrane fraction; 3, intracellular protein fraction. Each bar represents the average from at least three independent experiments. Error bars represent the standard errors of the means. The activity of the membrane fraction was arbitrarily set to 100%. (C) SDS-PAGE (lanes 2 and 3) and zymogram (lanes 4 and 5) of cellular compartments of E. coli BL21 containing pTO-estA. Lane 1, 10 μl of protein standard (Sigma); lanes 2 and 4, 3 μl of intracellular fraction; lanes 3 and 5, 5 μl of SDS-soluble membrane fraction. Enzymatic activities with MU-butyrate as a substrate were visualized as described in the text.

Substrate specificity of EstA.

To determine the specificity of EstA, we tested the ability of the enzyme to hydrolyze a variety of substrates, including o- and p-nitrophenyl saturated fatty acid esters, triglycerides, and other ester compounds. The recombinant enzyme exhibited high activity when p- and o-nitrophenyl butyrate esters were used as substrates, while virtually no activity was observed with nitrophenyl esters with chain lengths longer than six carbons (Fig. 4). Interestingly, ApeE from S. enterica serovar Typhimurium, which is most closely related to EstA, was shown to hydrolyze p-nitrophenyl esters of all straight-chain fatty acids from C₆ to C₁₆ (8). In general, p-nitrophenyl esters were better substrates than o-nitrophenyl esters. The ability of the enzyme to hydrolyze triglycerides was semiquantitatively tested on triglyceride agar plates (Table 3). EstA hydrolyzed triglycerides with short-chain fatty acids, including tributyrin (C₄), tricaproin (C₆), and tricaprylin (C₈), while tricaprin (C₁₀) was a poor substrate and no detectable activity was observed with trimyristin (C₁₆) and triolein (C₁₈). The hydrolytic action of the enzyme on various α- and β-naphthyl and HPTS ester substrates was qualitatively examined by activity staining of native gels. As with the other substrates, EstA hydrolyzed only short-chain naphthol esters with a maximum of six carbons (Table 3). Like other members of the GDSL family of esterases, EstA exhibited significant activity towards polar HPTS esters with fatty acid chains longer than eight carbons.

FIG. 4.

Hydrolysis of o- and p-nitrophenyl fatty acid esters by EstA. Recombinant expressed EstA was used in these assays. C2 to C8 represent the carbon chain length of the fatty acid moiety of the nitrophenyl ester. C2, nitrophenyl acetate; C3, nitrophenyl propionate; C4, nitrophenyl butyrate; C6, nitrophenyl valerate; C8, nitrophenyl caproate. Values from parallel experiments (minimum of three) were within ±5%.

TABLE 3.

Substrate specificity of EstA^a

Side chain length (no. of C atoms)	Activityb on:
	α-Naphthyl ester	β-Naphthyl ester	HPTS substrates	Triglyceride
2	++	+++	−	ND
3	++	+++	ND	ND
4	++	+++	−	+
5	ND	++	ND	ND
6	+	+	ND	+
8	ND	ND	+	ND
9	−	ND	+	ND
10	ND	ND	ND	+
12	−	ND	++	+/−
14	−	ND	ND	−
16	−	ND	ND	ND
18	ND	ND	+++	−

^aThe ability of EstA to hydrolyze various chromatogenic and fluorogenic substrates was tested with the aid of qualitative colorimetric assays and by staining of native polyacrylamide gels as described in Materials and Methods. Membrane protein fractions of E. coli BL21(pTO-estA) were used for these assays.

^b+++, strong activity; ++, moderate activity; +, weak activity; −, no activity, ND, not determined.

Physiological function of EstA.

To study the physiological function of EstA in S. liquefaciens MG1, we constructed a defined estA mutant. This was accomplished by allelic exchange of the wild-type estA gene with the recombinant estA::npt allele present on the suicide plasmid pTOM12 (for details, see Materials and Methods). The resulting mutant, which was designated MG1-estA, was greatly impaired in its ability to hydrolyze p-nitrophenyl butyrate (Fig. 5A). More importantly, a zymographic analysis of Triton X-100 soluble membrane proteins clearly showed that the mutant no longer produced the 69-kDa esterase (Fig. 5B).

FIG. 5.

(A) Hydrolysis of p-nitrophenyl butyrate by the S. liquefaciens wild-type MG1, the estA mutant MG1-estA, the swrI mutant MG44 in the absence or presence of 250 nM C₄-HSL, and the flhDC mutant MG3. Triton X-100-soluble membrane fractions were assayed. Each bar represents the average from at least three independent experiments. Error bars represent the standard errors of the means. The activity of the wild type was arbitrarily set to 100%. (B) SDS-PAGE and zymogram of the SDS-soluble membrane fractions of the S. liquefaciens wild-type MG1, the estA mutant MG1-estA, the swrI mutant MG44 in the absence or presence of 250 nM C₄-HSL, and the flhD mutant MG3.

The mutant grew equally well as the parental strain in minimal medium supplemented with various different carbon sources (Table 2; Fig. 6A). However, in minimal medium with Tween 20 or Tween 80 as the sole carbon source, the mutant grew with significantly reduced growth rates and yielded less biomass compared to the parental strain, indicating that EstA is involved in the utilization of certain fatty acid esters. To exclude the possibility that EstA is required for detoxification of a growth-inhibitory compound present in Tween 20, the strains were grown in media containing glucose plus Tween 20. The presence or absence of Tween 20 did not affect growth rates but slightly increased the yield of the wild-type strain (Table 2; Fig. 6A).

FIG. 6.

Growth and AHL production of the S. liquefaciens wild-type MG1 and the estA mutant MG1-estA in minimal medium supplemented with either 0.2% glucose (G), 1% Tween 20 (T20), or a mixture of 0.2% glucose and 1% Tween 20 (G + T20) as sole carbon sources. (A) Growth was measured spectrophotometrically at 600 nm. The growth curves are averages from three independent experiments. Error bars represent the standard errors of the means. (B) When the cultures reached an OD₆₀₀ of 1.0, samples of spent culture supernatants were taken and AHLs were quantified with the aid of the AHL monitor strain E. coli MT102(pSB403). Error bars represent the standard errors of the means.

Previous studies have shown that LuxI-type proteins catalyze the synthesis of AHLs from the appropriately charged acyl carrier protein (acyl-ACP), as the major acyl chain donor, and S-adenosylmethionine, which provides the HSL moiety (23, 27, 36, 39, 46).

The enzymatic action of EstA generates fatty acids and may therefore influence the cellular acyl-ACP pool. This, in turn, may have an effect on AHL biosynthesis. To address this issue, we determined the AHL concentrations in spent culture supernatants of the wild type and the estA mutant grown in different media to an OD₆₀₀ of exactly 1.0. Growth with most of the carbon sources tested yielded similar amounts of AHLs (Table 2; Fig. 6B). However, when Tween 20 or Tween 80 was used as the sole carbon source, the mutant produced greatly reduced amounts of AHLs compared to the wild type. Importantly, the presence of Tween 20 did not affect the bioassay used for quantification of AHLs (Fig. 6B). These results suggest that EstA is required for AHL production when the cells are growing on fatty acid esters.

Expression of EstA is regulated neither by quorum sensing nor by FlhDC.

To determine whether estA is controlled by the swr quorum-sensing system we measured the esterase activities of the S. liquefaciens wild-type MG1 and the swrI mutant MG44 grown in the presence or absence of 250 nM C₄-HSL. As shown in Fig. 5A, the abilities of the two strains to hydrolyze p-nitrophenyl butyrate were virtually indistinguishable. Furthermore, the presence or absence of C₄-HSL in the medium did not affect the esterase activity of the swrI mutant. We also overlaid renaturated SDS-polyacrylamide gels of Triton X-100-soluble membrane proteins of the various cultures with MU-butyrate in order to visualize the esterase (Fig. 5B). In agreement with the enzymatic measurements, we observed that the 69-kDa esterase band is present in similar amounts in samples prepared from cultures of MG1 and MG44. From these data we conclude that the swr quorum-sensing system has no significant effect on the expression of the EstA outer membrane esterase.

In S. liquefaciens MG1, expression of the PhlA phospholipase is controlled by the FlhDC master regulator of the flagellar regulon (20). We therefore tested the FlhD mutant MG3 for esterase activity. Both enzymatic measurements and a zymographic analysis showed that FlhDC is not involved in the regulation of EstA expression (Fig. 5).

Transcriptional regulation of estA.

A transcriptional fusion of the estA promoter region to the promoterless luxAB luciferase genes was constructed as described in Materials and Methods. Measurements of bioluminescence in the wild-type and swrI mutant backgrounds revealed that the estA promoter is not regulated by the swr quorum-sensing system (data not shown), thus confirming our zymographic analysis. We also monitored the activity of the P_estA::luxAB fusion along the growth curve, but no growth phase-dependent regulation was observed. Furthermore, we investigated whether the source of carbon or the availability of phosphate influences transcriptional activity of the estA promoter. However, under none of the investigated growth conditions was a significant effect observed (Table 2 and data not shown), suggesting that estA may be constitutively transcribed.

EstA is not required for swarming motility.

On appropriate surfaces, S. liquefaciens MG1 is capable of swarming motility. This form of surface translocation is characterized by the differentiation of short motile rods at the periphery of a colony into elongated, polyploid, and hyperflagellated swarm cells that migrate coordinately in rafts atop the agar surface (13, 15). The fact that formation of a swarming colony is controlled by the swr quorum-sensing system prompted us to investigate whether EstA plays a role in this multicellular behavior. To this end, the S. liquefaciens wild-type strain MG1, the estA mutant, and, as a control, the swrI mutant MG44 were inoculated on swarming plates containing different carbon sources. In these assays the behavior of the estA mutant was indistinguishable from that of the wild type, suggesting that EstA is not required for the development of a swarming colony (Fig. 7). As production of AHLs is greatly reduced when the estA mutant is growing on a medium containing Tween 20 as the carbon source, it might be expected that the mutant would exhibit at least reduced swarming. However, no difference between the wild type and the mutant was observed.

FIG. 7.

Swarming behavior of the S. liquefaciens wild-type MG1, the estA mutant MG1-estA, and the swrI mutant MG44 on LB and on AB supplemented with 0.25% Casamino Acids and either 0.4% glucose (ABG) or 1% Tween 20 (ABT). Plates were photographed after 12 h of incubation at 30°C.

The swr system is known to control expression of the biosurfactant serrawettin W2, which is required for swarming motility (31). This compound, a cyclic lipodepsipentapeptide carrying a 3-hydroxy-C₁₀ fatty acid side chain, lowers the surface tension of the surrounding medium, a process that is indispensable for the development of an S. liquefaciens MG1 swarming colony. As a consequence, the swrI mutant cannot swarm unless the medium is supplemented either with AHLs to restore serrawettin W2 production or with a surfactant (e.g., serrawettin W1, a surfactant produced by Serratia marcescens; surfactin from Bacillus subtilis; or trace amounts of SDS) (15, 31). Thus, our results can be explained by the fact that Tween 20 is a detergent that lowers the surface tension of the medium. This hypothesis is strongly supported by the fact that the swrI mutant also swarms on Tween 20-containing medium in the absence of AHLs (Fig. 7).

DISCUSSION

Sequence analysis of the DNA regions flanking the swr quorum-sensing locus of S. liquefaciens MG1 revealed the presence of an esterase gene, designated estA, upstream of swrR. EstA belongs to the class II of lipolytic enzymes (2, 45). Members of this enzyme family differ from other lipases in the location and structure of the active-site consensus motif G-X-S-X-G. In the case of class II enzymes, this motif is located very close to the N terminus and consists of G-D-S-L-S, with the terminal glycine of the characteristic lipase consensus replaced in most cases by serine. The active site of these enzymes is believed to consist of a catalytic triad formed by the amino acids serine, histidine, and aspartate, with the serine being embedded in the consensus sequence at the active site. Ester hydrolysis is mediated by a nucleophilic attack of the active serine on the carbonyl group of the substrate in a charge-relay system with the two other amino acid residues (38). Based on amino acid sequence similarities, we predict Ser38, Asp327, and His330 to be the catalytic active residues of the S. liquefaciens MG1 esterase (Fig. 2). Another typical feature of class II esterases is that they are anchored to the outer membrane. Both enzymatic measurements and zymographic analyses of cellular fractions showed that EstA of S. liquefaciens MG1 is located in the outer membrane. Previous work has provided evidence that the C-terminal regions of class II esterases contain an autotransporter domain, which is often found in self-secreted bacterial virulence factors (22, 32, 50). It is believed that this domain is folded into several amphiphatic β-sheets, forming an aqueous pore in the outer membranes through which the catalytic N-terminal domain can transit (Fig. 2).

Esterases and other lipolytic enzymes have attracted considerable interest from industry because of their biotechnological potential. The wide range of properties with respect to substrate specificity and regio- and enantioselectivity has opened a broad spectrum of applications for these hydrolytic enzymes (26). Despite their industrial importance, there is only very little information on the physiological functions of esterases available (37). Some esterases appear to be involved in metabolic pathways that provide access to carbon sources, e.g., the acetyl and cinnamoyl esterases that take part in the degradation of hemicellulose (11, 16). In the case of plant pathogens, such cell wall-degrading esterase activities are believed to represent important virulence factors (34). The inactivation of biocides has been recognized as another function of esterases. For example, B. subtilis has been shown to produce an esterase which hydrolyzes the phytotoxin brefeldin A (49).

To investigate the physiological role of the S. liquefaciens MG1 esterase we constructed a defined estA mutant with the aid of a gene replacement technique. While the mutant and the wild type grew equally well on media containing glucose as a carbon source, the mutant exhibited a marked growth defect on media containing Tween 20 or Tween 80, indicating that the esterase is required for the catabolism of certain fatty acid esters.

As estA is located proximal to the swrI-swrR system, we speculated that expression of estA may be quorum sensing regulated. However, a comparison of the EstA expression patterns of the wild-type MG1 and the swrI mutant MG44 did not support this hypothesis. In agreement with this result, we were also unable to identify a lux box-like sequence, the proposed binding site for LuxR homologs, in the estA promoter region. During the course of these experiments we noticed that the estA mutant produced greatly reduced amounts of AHLs when Tween was used as a carbon source, while no differences were evident on media containing glucose. These data suggest that EstA is required for AHL biosynthesis when cells are grown on certain lipidic substrates. LuxI-type proteins such as SwrI are thought to catalyze the synthesis of AHLs from charged acyl-ACPs and S-adenosylmethionine (23, 27, 36, 39, 44). When cells are grown on lipidic substrates, the enzymatic action of the outer membrane esterase will provide the cell with fatty acids. As a consequence, the cellular pool of charged acyl-ACPs may be replenished, which otherwise may be the bottleneck for AHL synthesis under these growth conditions. Recent work has established a link between fatty acid biosynthesis and quorum sensing in P. aeruginosa (23), and it was shown that modulation of FabG activity of the fatty acid biosynthetic pathway determines the acyl chain lengths of the AHL signal molecules produced by the organism (25). Whether EstA is also involved in AHL biosynthesis in S. liquefaciens under more natural life conditions, e.g., during colonization of plant roots, has yet to be investigated.