Present address: Rubén Avendaño-Herrera, Departamento de ciencias biológicas, Facultad de ciencias biológicas, Universidad de Andrés Bello, Santiago, Chile.
Correspondence: Ana Otero, Departamento de Microbiología, Facultad de Biología-CIBUS, Universidad de Santiago de Compostela, 15782 Santiago, Spain. Tel.: +34 98 156 3100, ext. 16967; fax: +34 98 159 2210; e-mail: email@example.com
Tenacibaculum maritimum (formerly Flexibacter maritimus) is a filamentous, biofilm-forming member of the Cytophaga–Flavobacterium–Bacteroides group (or Bacteroidetes), which causes the widely distributed marine fish disease tenacibaculosis. A search for N-acylhomoserine lactones (AHLs) quorum-sensing (QS) signals in the culture media of nine representative strains of this species using different biosensor strains revealed the presence of short-type AHL activity in all of them. N-butyryl-l-homoserine lactone (C4-HSL) was identified in T. maritimum NCIMB2154T by LC-MS. A degradation activity for long-acyl AHLs (C10-HSL) was subsequently demonstrated in T. maritimum NCIMB2154T. The acidification of the culture medium after degradation did not allow the recovery of C10-HSL, which indicates a possible acylase-type degradation activity. Even though the physiological processes under the control of AHL-mediated QS in T. maritimum need to be further characterized, this discovery extends the paradigm of AHL-mediated QS signalling beyond the Proteobacteria and reinforces its ecological significance.
Many bacterial species coordinate responses to environmental changes using complex cell–cell communication mechanisms in a cell-density-dependent manner. This phenomenon is known as quorum sensing (QS) and relies on the accumulation of signal molecules in the surrounding environment to threshold concentrations at which target structural genes are activated (Williams et al., 2007). Several communication systems exist that use different signal molecules, also known as autoinducers (Waters & Bassler, 2005; Williams et al., 2007). In Gram-negative bacteria, the most intensively studied QS systems rely on the use of N-acylhomoserine lactones (AHLs), a family of signal molecules differing in the length and substituents of the acyl chain. The use of these molecules as QS signals has been well established, and their role in the control of important physiological processes such as bioluminescence, biofilm formation, plasmid conjugation, production of exoenzymes and virulence factors, swarming, etc., has been shown in a number of Proteobacteria, including several important animal and plant pathogens (Williams et al., 2007). The production of AHLs has so far been limited to a few genera within the Proteobacteria (Williams et al., 2007), which has raised questions with regard to the ecological significance of these molecules (Manefield & Turner, 2002). Outside Proteobacteria, the production of AHLs has been recently demonstrated for the colonial cyanobacterium Gloeothece PCC6909 and for several strains of Bacterioidetes isolated from marine biofilms, although the physiological processes under the control of the QS system were not completely elucidated (Sharif et al., 2008; Huang et al., 2009). AHL-like activity was also found in the haloalkalophilic archaeon Natronococcus occultus (Paggi et al., 2003), but the biochemical nature of the signal has not been confirmed. Short-chain AHL-type activity was also found in Flavobacterium sp., a member of the Cytophaga–Flavobacterium–Bacteroides (CFB) cluster, but the presence of AHL could not be confirmed by GC-MS (Wagner-Döbler et al., 2005).
QS seems to be of special significance in the marine environment. AHL signal molecules are produced by more than half of the marine Alphaproteobacteria isolated from various marine habitats (Wagner-Döbler et al., 2005). Moreover, the production of AHLs is common among marine and fish pathogenic Proteobacteria (Bruhn et al., 2005; Wagner-Döbler et al., 2005), controlling the expression of key virulence factors (Defoirdt et al., 2005). Because of the prevalence of QS systems among these pathogens, the inhibition of these processes has been proposed as an alternative to the use of antibiotics in aquaculture (Defoirdt et al., 2004). The inhibition of AHL-mediated QS processes was first described in the marine alga Delisea pulchra (Givskov et al., 1996) and has now been described in several eukaryotes and bacteria of terrestrial origin (reviewed by Dong & Zhang, 2005). The isolation and characterization of marine bacterial strains capable of inhibiting QS, a process known as quorum quenching (QQ), either enzymatically or through the production of inhibitors or antagonists may help to develop new biotechnological tools. Several results indicate the viability of this approach for the treatment of infections in aquaculture (Rasch et al., 2004).
The marine bacteria Tenacibaculum maritimum (formerly Flexibacter maritimus) (Suzuki et al., 2001) is a filamentous member of the CFB group causing the fish ‘gliding bacterial disease’ or tenacibaculosis/flexibacteriosis (Avendaño-Herrera et al., 2004). Tenacibaculum maritimum belongs to the CFB cluster, which is also known as Bacteroidetes (Ludwig & Klenk, 2001), and constitutes one of the dominant heterotrophic bacterial groups in aquatic habitats. The fact that T. maritimum shifts abruptly from a biofilm to a planktonic mode of growth, a characteristic that could be related to a QS-controlled process (Rice et al., 2005; Wagner-Döbler et al., 2005), led us to investigate the possible production and degradation of AHLs by this fish pathogen.
Materials and methods
Strains and culture conditions
The T. maritimum strains NCIMB2154T, NCIMB2153 and NCIMB2158 were obtained from The National Collections of Industrial, Food and Marine Bacteria Ltd (Aberdeen, UK). In addition, six strains isolated in our laboratory from fish farm disease outbreaks from Spain and Portugal were used. These strains belong to the main serotypes and clonal lineages described within this pathogen (Table 1) (Avendaño-Herrera et al., 2004, 2006), and were confirmed as T. maritimum by PCR-based analysis (Toyama et al., 1996). The strains were routinely cultured at 20 °C on F. maritimus medium (FMM) agar or broth (Pazos et al., 1996) and on marine broth (MB, Difco) for some of the experiments. Liquid cultures were inoculated with a 10% volume of a 24-h liquid culture and maintained in a shaker at 100 r.p.m. Cultures were double-checked for purity on Marine Agar (Difco) and FMM before and after each experiment.
Table 1. AHL activity in acidified culture media extracted with dichloromethane (1 : 1 v/v) of nine different isolates of Tenacibaculum maritimum as detected by TLC using the lux-based reporter strain Escherichia coli JM109 pSB536 (specific for short-length AHLs)
Three lux-based Escherichia coli JM109 AHL biosensor strains that respond to AHLs with different side chain lengths were used for the detection of AHL production (Swift et al., 1997; Winson et al., 1998). The biosensor strains were grown at 37 °C in Luria–Bertani (LB) broth or agar supplemented with the adequate antibiotics. Additionally, the AHL biosensor strains Chromobacterium violaceum CV026 (McClean et al., 1997) and C. violaceum VIR07 (Morohoshi et al., 2008) were used for the AHL-degradation assays in solid plates as explained below (McClean et al., 1997). These strains were routinely cultured on LB medium supplemented with kanamycin (50 μg mL−1) at 30 °C.
Sampling and AHL extraction
Samples (100 mL) from cultures of nine different strains of T. maritimum grown in liquid FMM were obtained 24 and 48 h after inoculation, acidified to pH 2 with HCl 1 M in a shaker at 200 r.p.m. for 12 h at 20 °C, to ensure the absence of any AHL lactonolysis products, and extracted with dichloromethane as described previously (Yates et al., 2002). Dried extracts were reconstituted in 1 mL ethyl acetate and stored at −20 °C until further analysis.
Thin-layer chromatography (TLC) detection assays of extracted samples
Two microlitres of the ethyl acetate extracts obtained from the spent culture media were run on reversed-phase hydrocarbon-impregnated C18 silica gel plates using methanol :water 60 : 40 v/v as the mobile phase (Shaw et al., 1997). Two microlitres of synthetic AHLs (Sigma, stock concentration 50 μg mL−1) were run as controls: N-octanoyl-l-homoserine lactone (C8-HSL) for E. coli JM109 pSB401, N-butyryl-l-homoserine lactone (C4-HSL) for E. coli JM109 pSB536 and N-dodecanoyl-l-homoserine lactone (C12-HSL) for E. coli JM109 pSB1075 (Winson et al., 1998). Plates were dried and overlaid with 3 mL of semi-solid LB medium (8% agar) inoculated with 30 μL of an overnight culture of the corresponding sensor strain. Plates were incubated at 37 °C and every hour, radiographic plates were laid over them to detect the emission of bioluminescence.
LC-MS AHL identification
LC-MS analyses were carried out simultaneously in the laboratories in Nottingham and Santiago using different equipment and slightly different conditions to confirm the presence of AHLs unequivocally. In Nottingham, a Shimadzu series 10AD VP equipped with binary pumps, a vacuum degasser and an SIL-HTc autosampler and column oven (Shimadzu, River Drive, MD) was used as the LC system. As column a Phenomenex Gemini C18, 150 × 2 mm (5 μm particle size), at 45 °C was used. The mobile phase was built by 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B). The flow rate was 0.45 mL min−1. The elution conditions were as follows: 1 min 0% B, linear gradient to 50% B for 0.5 min and then a linear gradient from 50% to 90% B over 4 min, then 2.5 min 99% B over 2 min, then ramped back to the starting conditions in 0.2 min. The column was re-equilibrated for a total of 4 min. Samples were redissolved in 50 μL acetonitrile before use and a 10-μL volume was injected onto the column (Ortori et al., 2007). Parallel analyses were carried out using an HPLC 1100 series (Agilent, Santa Clara, CA) equipped with a C8 precolumn (2.1 × 12.5 mm, 5 μm particle size) and a ZORBAX Eclipse XDB-C18 2.1 × 150 mm (5 μm particle size) column. Temperature and mobile phases were the same as above, but the flow rate was set at 0.22 mL min−1. In this equipment, the elution conditions were as follows: 0 min 35% B, linear gradient to 60% B in 10 min and then a linear gradient from 60% to 95% B over 5 min, then 5 min 95% B and then in 1 min, ramped back to the starting conditions in 9 min. The column was re-equilibrated for a total of 5 min. A 2-μL volume was injected onto the column.
The MS experiments shown were conducted in Santiago on an API 4000 triple-quadrupole mass spectrometer (Applied Biosystems, Foster City, CA) equipped with a TurboIon source using positive ion electrospray, multiple reaction monitoring (MRM) mode. The MRM signals were used to generate relative quantification information and to trigger subsequent quality product ion spectra (product ion PI, MS2). The conditions for the generation of the MRM-triggered spectra were as follows: DP ramped from 35 to 57, CE 14-28, CXP 8. Analyses were confirmed on a 4000 QTRAP hybrid triple-quadrupole linear ion trap mass spectrometer (Applied Biosystems) equipped with a TurboIon source used in positive ion electrospray, MRM mode. In this case, the conditions for the generation of the MRM-triggered spectra were as follows: DP ramped from 25 to 50, CE 15-45, CXP 12.
AHL with or without a 3-oxo or a 3-hydroxy substitution and with an acyl side chain length of 4 (C4-HSL, 3-oxo-C4-HSL and 3-hydroxy-C4-HSL), 6 (C6-HSL, 3-oxo-C6-HSL and 3-176 hydroxy-C6- HSL), 7 (C7-HSL), 8 (C8-HSL, 3-oxo-C8-HSL and 3-hydroxy-C8-HSL), 10 (C10-HSL, 3-oxo-C10-HSL and 3-hydroxy-C10-HSL), 12 (C12-HSL, 3-oxo-C12-HSL and 3-hydroxy-C12-HSL), 13 (C13-HSL, 3-oxo-C13-HSL and 3-hydroxy-C13-HSL) or 14 (C14-HSL, 3-oxo-C14-HSL and 3-hydroxy-C14-HSL) were used as standards. Acyl-HSLs were identified and confirmed by comparing both the elution time and the spectra from any peaks obtained with those of the standards.
AHL-degradation activity assay
Chromobacterium violaceum-based solid plate assays (McClean et al., 1997) were carried out to detect AHL degradation activity in T. maritimum NCIMB2154T. Two different sensor strains were used to detect AHL degradation. Chromobacterium violaceum CV026 (McClean et al., 1997) was used to measure the degradation of C6-HSL and C. violaceum VIR07 was used to measure the degradation of C10-HSL (Morohoshi et al., 2008). Twenty microlitres of stock solutions of C6-HSL or C10-HSL were added to 500 μL of an overnight culture of T. maritimum NCIMB2154T in MB (final concentration 2 μg mL−1) and incubated for 24 h at 20 °C. The same amount of AHL was added to 500 μL of spent culture medium obtained from a 24-h-old culture by filtration through 0.22 μm. The amount of remaining AHL in the culture media of T. maritimum was evaluated in LB plates overlaid with 5 mL of semi-solid LB agar seeded with 500 μL of overnight cultures of C. violaceum CV026 for C6-HSL or C. violaceum VIR07 for C10-HSL. Fifty microlitres of culture supernatants were loaded in wells and adjusted to 100 μL with distilled water. Sterile MB and MB plus C4 or C10-SHLs were set as controls. The same experiment was carried out in FMM broth (data not shown). Plates were incubated for 12–24 h and the production of violacein was examined.
To evaluate the possible type of AHL degradation activity, two flasks with 15 mL of FMM were supplemented with C10-HSL to a final concentration of 2 μg mL−1. One of them was inoculated with 1 mL of a 48-h culture of T. maritimum NCIMB2154T and the other was maintained as control. Flasks were incubated in a shaker at 22 °C under soft agitation (110 r.p.m.). After 24 h, 500 μL of normal and acidified culture media were extracted three times with ethyl acetate, dehumidified onto MgSO4, evaporated under nitrogen flux and resuspended in acetonitrile for LC-MS analysis as described above. Before inoculation, 500 μL of FMM+C10-HSL were also extracted and the value of C10-HSL obtained was used to calculate the percentage of degradation.
Results and discussion
The presence of AHLs was investigated in ethyl-acetate extracts of acidified culture media of the type strain T. maritimum NCIMB2154T obtained at 24 and 48 h using the three E. coli JM109 lux-based biosensor strains carrying pSB536, pSB401 or pSB1075 to detect a wide range of AHLs differing in the length of their acyl chain. TLC analysis revealed the presence of short-chain AHLs using the E. coli JM109 pSB536 biosensor (Fig. 1). A search for AHL-type QS signals in extracts obtained from the culture media of another eight representative isolates of T. maritimum using the same technique revealed the presence of short-chain AHL activity in all of them, although differences were recorded in relation to their peak in activity (Table 1). LC-MS analysis confirmed the presence of N-butyryl-l-homoserine lactone (C4-HSL) in the culture media of T. maritimum NCIMB2154T grown in both FMM (Fig. 2) and MB (data not shown). This AHL was unequivocally identified by comparison of its mass spectra with those of pure standards (Fig. 3). As this is the first description of the production of AHLs by a pathogenic member of the CFB cluster, the analyses were carried out simultaneously in both laboratories using different chromatographic conditions. The results confirmed unequivocally the presence of the C4-HSL.
So far, no physiological role other than as QS signals has been assigned to AHLs, except as a chelator, for tetramic acid (a derivative of 3-oxo-C12) or antibiotic activity for both 3-oxo-C12-HSL and tetramic acid (Kaufmann et al., 2005; Schertzer et al., 2009). In addition, a role as biosurfactant has been attributed to long-chain AHLs (Daniels et al., 2006). Therefore, even though the physiological features under the control of these molecules in T. maritimum remain to be investigated, the production of C4-HSL by T. maritimum strains extends the paradigm of AHL-mediated QS beyond the Proteobacteria. As the physiological processes under the control of AHL-mediated QS have so far been described for a limited number of genera of the Alpha-, Beta- and Gammaproteobacteria, many of them human or plant pathogens (Williams et al., 2007), the ecological significance of AHL-mediated QS has been questioned as a key switch controlling gene expression within bacterial populations in nature (Manefield & Turner, 2002). The fact that genera outside the Proteobacteria produce the same signal molecules, and that AHL-degrading activity has been found in Gram-positive, Gram-negative and Cyanobacteria (Dong & Zhang, 2005; Romero et al., 2008) and in mammalian cells (Chun et al., 2004), reinforces the ecological significance of AHL-mediated QS processes. The presence of AHL-mediated QS beyond the Proteobacteria is not surprising, as a phylogenetic study based on the LuxI/LuxR genes suggested that QS mechanisms were established very early in the evolution of bacteria, although horizontal transfer may have also played an important role in the distribution of QS genes, at least within this group (Lerat & Moran, 2004). The cloning of genes responsible for the synthesis and detection of AHLs in T. maritimum will undoubtedly provide new insights into the evolutionary history of QS.
The production of AHLs was demonstrated for all isolates of T. maritimum analysed (Table 1), therefore being a conserved trait within this species, which is not the case in some other marine pathogens such as Aeromonas salmonicida (Bruhn et al., 2005). Some contradictory results have been published previously regarding the production of AHLs by the genus Flavobacterium belonging to the Bacteroidetes group. While AHL-like activity was detected in a planktonic isolate of Flavobacterium sp. using E. coli MT102 carrying the biosensor plasmid pJBA132 based on the luxR gene from Vibrio fischeri, the presence of AHLs could not be demonstrated by GC-MS (Wagner-Döbler et al., 2005). Furthermore, no AHL activity was found in different pathogenic strains of Flavobacterium psychrophilum using the sensor strains C. violaceum CV026 and Agrobacterium tumefaciens NT1 (Bruhn et al., 2005). The differences in the AHL activity described probably depend on the assay conditions and the sensor strain utilized. In our experience, data based on the direct evaluation of culture media of marine bacteria, usually MB, should be interpreted with caution, as the media composition could result in inhibition of growth or bioluminescence production by the sensor strain (unpublished data). On the other hand, due to the ability of different compounds to activate the AHL biosensors (Holden et al., 1999), the results should be viewed with caution unless the presence of AHLs can be confirmed by analytical chemical methods. On the basis of our results and as the detection of the QS activity is strongly dependent on the biosensor strain used and on the culture conditions, it is possible that AHL-based QS systems are more widespread than described so far (Wagner-Döbler et al., 2005).
An in vivo degradation assay was carried out using two biosensor strains of C. violaceum. Chromobacterium violaceum CV026 was used to detect degradation of short AHLs (C6-HSL), and C. violaceum VIR07 was used to detect degradation of long AHLs (C10-HSL). Complete degradation of C10-HSL was observed after 24 h, but no changes in C6-HSL activity were observed (Fig. 4a). The activity should be cell bound, as no significant degradation was obtained when the C10-HSL was added to cell-free spent culture medium (Fig. 2a). HPLC analysis of the degradation of C10-HSL revealed that 90% of the AHL was degraded after 24 h of exposition to T. maritimum cultures, and no recovery of the AHL could be achieved by medium acidification, which may discard a lactonase-type degrading enzyme (Fig. 4b). Further analyses are required to confirm an acylase-type activity. The presence of AHL degradation enzymes has been described in Gram-negative bacteria, possibly as a mechanism to outcompete Gram-positive neighbours (Roche et al., 2004). In Gram negatives, these enzymes can be used for the self-modulation of the level of the signals as already found in A. tumefaciens (Zhang et al., 2002). In the case of the bacteroidete T. maritimum, the presence of a QQ enzyme for long AHLs may represent an exclusion mechanism to interfere with the QS systems of competitors (Dong & Zhang, 2005).
Evidence is beginning to accumulate indicating that QS and QS inhibition processes, including enzymatic degradation of the signal or QQ, are important in the marine environment. Besides the well-characterized phenomenon of the production of furanones by the red alga D. pulchra to avoid surface colonization by Gram-negative biofilm formers (Givskov et al., 1996), QS systems mediated by AHLs have been found in many species of marine pathogenic bacteria (Bruhn et al., 2005). AHLs also seem to play an important role in the eukaryotic–prokaryotic interactions in the marine environment, as demonstrated by the importance of the production of AHLs by marine biofilms for the surface selection and permanent attachment of zoospores of the green alga Ulva (Tait et al., 2005), for spore release of the red alga Acrochaetium sp. (Weinberger et al., 2007), and for some initial larval settlement behaviours in the polychaete Hydroides elegans (Huang et al., 2007). As most of the isolates involved in algal morphogenesis belong to the CFB group (Hanzawa et al., 1998; Matsuo et al., 2003), the discovery of the production and degradation of AHLs by members of this group provides the possibility of new interactions between bacteria and eukaryotes in the marine environment.
For the first time, the production of AHL-type QS signals and QQ activity has been demonstrated simultaneously in a pathogenic member of the CFB group. Because of the ecological significance of the Cytophaga–Flavobacterium cluster, especially in the marine environment, the discovery of AHL-mediated QS processes among their members will advance our understanding of the microbial interactions in complex ecosystems. Moreover, cell-to-cell communication phenomena should be reconsidered in other habitats in which the Bacteroidetes play an important role, such as intestinal flora or dental plaque. As QS controls the expression of important virulence factors in many pathogenic bacteria, the disruption of QS mechanisms in T. maritimum and other fish pathogenic bacteria may represent a new strategy for the treatment of infections in aquaculture.
This work was financed by a grant from Consellería de Innovación e Industria, Xunta de Galicia, Spain (PGIDIT06PXIB200045PR). M.R. is supported by an FPU fellowship from the Spanish Ministry of Science and Education. We would like to thank Noemi Ladra (University of Santiago) and Catherine Ortori (University of Nottingham) for LC-MS analysis. The sensor Chromobacterium violaceum VIR07 was kindly provided by Prof. T. Morohoshi.